Specific features of HIV-1 integrase inhibition by bisphosphonate derivatives

Article abstract of DOI:10.1016/j.ejmech.2013.11.028

The integration of viral DNA into the cell genome is one of the key steps in the replication cycle of human immunodeficiency virus type 1 (HIV-1). Therefore, the viral enzyme integrase (IN) catalyzing this process is of great interest as a target for new

Full text of DOI:10.1016/j.ejmech.2013.11.028

Accepted Manuscript
Specific features of HIV-1 integrase inhibition by bisphosphonate derivatives
Julia Agapkina, Dmitry Yanvarev, Andrey Anisenko, Sergey Korolev, Jouko
Vepsäläinen, Sergey Kochetkov, Marina Gottikh
PII:
S0223-5234(13)00770-8
10.1016/j.ejmech.2013.11.028
EJMECH 6574
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 16 September 2013
Revised Date: 24 November 2013
Accepted Date: 27 November 2013
Please cite this article as: J. Agapkina, D. Yanvarev, A. Anisenko, S. Korolev, J. Vepsäläinen, S.
Kochetkov, M. Gottikh, Specific features of HIV-1 integrase inhibition by bisphosphonate derivatives,
European Journal of Medicinal Chemistry (2014), doi: 10.1016/j.ejmech.2013.11.028.
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

ACCEPTED MANUSCRIPT
1
6
7 bisphosphonates of 5 different groups were synthesized including 3 earlier unknown
compounds inhibited both strand transfer and 3’-processing catalyzed by integrase
SAR study has been reported and explained using liquid NMR experiments
OH / NH in the chelating group switches competitive inhibitor into noncompetitive
2

ACCEPTED MANUSCRIPT
Specific features of HIV-1 integrase inhibition by bisphosphonate derivatives
a,||
*,b,||
a
a
Julia Agapkina , Dmitry Yanvarev , Andrey Anisenko , Sergey Korolev , Jouko
c
b, a
a
Vepsäläinen , Sergey Kochetkov , Marina Gottikh
a
-
Lomonosov Moscow State University, Chemistry Department and Belozersky Institute of
b
Physical and Chemical Biology, Moscow, Russia; -Engelhardt Institute of Molecular Biology,
c
Russian Academy of Sciences, Moscow, Russia; -School of Pharmacy, Biocenter Kuopio,
University of Eastern Finland, Kuopio, Finland.
|
| these authors contributed equally to this work
*
corresponding author: Dmitry V. Yanvarev. E-mail: JDmitry@hotmail.com,
tel.+79267025449
Abstract
The integration of viral DNA into the cell genome is one of the key steps in the
replication cycle of human immunodeficiency virus type 1 (HIV-1). Therefore, the viral enzyme
integrase (IN) catalyzing this process is of great interest as a target for new antiviral agents. We
performed a structural-functional analysis of five different series of methylene bisphosphonates
3 2 3 2
(BPs), PO H -C(R)(X)-PO H , as IN inhibitors with the goal of assessing structural elements
required for the inhibitory activity. We found that IN is inhibited only by BP bearing a
chlorobenzyl substituent R at the bridging carbon of the P-C-P backbone. These BP inhibited
both IN-catalyzed reactions with similar efficacies. They were also active toward some INs with
mutations characteristic for HIV-1 strains resistant to strand transfer inhibitors. The study of the
mechanism of the IN inhibition by various BP showed that it is effected by the nature of the
second substituent (X) at the bridging carbon. Among the tested compounds, only the BP with
the amino group bound directly to the BP bridging carbon was found to be a noncompetitive
inhibitor and, hence, it can be promising for further studies as potential inhibitor of the IN
activity within the preintegration complex.
Keywords
HIV, integrase, bisphosphonates, inhibitors, SAR
1
. Introduction
One of the key steps of the HIV-1 life cycle is the integration of the viral DNA copy into
the host genome. This process is catalyzed by a viral enzyme called integrase (IN). When the
reverse transcription of the viral genome is completed, IN binds to its DNA copy and catalyzes
the 3’-processing resulting in the cleavage of GT nucleotides from both DNA 3’-ends. Then IN
catalyzes the strand transfer reaction, that is, the integration of the viral DNA to the host cell

ACCEPTED MANUSCRIPT
DNA, to form a provirus. Further regulated expression of the proviral genome and assembling
give new virions, which leave the cell and infect other cells.
Figure 1.
Integrase is of great interest as a target for the search of new antiviral agents, since,
unlike other viral enzymes, it has no cell analogues and its inhibitors must not affect normal cell
processes [1],[2]. Several types of integration inhibitors differing in the mechanism of action
have been reported, but only two compounds, raltegravir and elvitegravir, are currently used for
the treatment of HIV-infection as components of highly active antiretroviral therapy (HAART)
[
3],[4]. Both of them inhibit the IN function at the strand transfer stage by a similar mechanism,
which results in the emergence of viral cross-resistance to these drugs [5],[6],[7]. Dolutegravir, a
strand transfer inhibitor (INSTI) of the second generation, which is currently under phase III of
clinical trials, is active toward some raltegravir-resistant HIV-1 strains. However, the partial
cross-resistance with raltegravir and elvitegravir has been already found for this compound
[
7],[8]. Taking into consideration the above facts the optimal way for designing new integration
inhibitors seems to be the search for the compounds with the mechanism of action differing from
that of raltegravir and its analogues. Such inhibitors may be active toward INSTI resistant viral
strains.
Phosphonates being hydrolytically stabile analogs of biogenic phosphates are widely used
in antiviral drug design [9], [10], [11], [12].
Methylenebisphosphonates (BPs) are mimics of inorganic pyrophosphate, a substrate of
HIV-1 reverse transcriptase (RT) in the pyrophosphorolysis reaction. These compounds can
inhibit the cleavage of monophosphates of RT nucleoside inhibitors, which is the key reason for
the HIV-1 resistance to azidothymidine and its analogues [13],[14]. Previously we found that
some BPs can also inhibit HIV-1 IN at micromolar concentrations [15]. This allowed us to
regard them as inhibitors of the new generation capable of inhibiting two viral enzymes
operating at the early stage of viral infection.
In this work we studied the BP structure-activity relationship toward IN inhibition with
the goal of evaluating their mechanism of action and optimizing their structure. Only BPs
bearing a chlorobenzyl residue as a substituent R (Fig. 1) were found to display inhibitory
properties. They inhibited both IN-catalyzed reactions with similar efficacies. These BPs were
also active toward some INs with mutations inherent for drug-resistant HIV-1 strains. For the
evaluation of the inhibition mechanism, kinetic models were designed, which demonstrated that
2
+
2+
the nature of the substituent X as well as the metal-cofactors, Mg and Mn , affected this

ACCEPTED MANUSCRIPT
mechanism. This conclusion was also supported by the results of studies of the BP effect on the
IN-DNA binding and IN catalytic activity within the IN-LEDGF complex.
2
. Results
2
.1 Chemistry
Seventeen bisphosphonates of five different groups were synthesized including earlier
unknown (3b), (3e), and (4c) (Fig. 1). The compounds (1a) [16], (1b) [17], (1d) [18], (2a) [14],
2c) [19], (3a) [20], (3c) [21], (3d) [19], (5) [19], (6) [22] were synthesized as earlier described
(
with minor modifications. General synthetic procedures and physicochemical characteristics of
all the prepared BPs are presented in Materials & Methods and SI.
Four general methods used are known from literature [17],[23],[24]. In brief, involved
either (A)- phosphonylation of a carboxylic acid that resulted (1a, 1b, 1c, 3a, 3c-e) or a
corresponding nitrile that afforded (1d, 2c, 3b). (B)- Michael addition of amine to
vinylidenediphosphonic acid to prepare (4a-c). (C)- Reaction of a acylphosphonate with
dialkylphosphite followed by removal of alkoxy group with TMSBr (2a, 2b). (D)- Reaction of
benzylchlorides with sodium salt of methylenediphosphonic acid tetraester yielded (2d, 5).
Vinylidenediphosphonic acid (VDPA) for synthesis of BPs 4a-c was prepared using the
optimization of the procedures described in the patent [25]. We used different thermal conditions
for pyrolysis of etidronate to reach its 100% conversion and we modified procedure for VDPA
2 4
deionization to avoid using strong ion-exchange resins and H SO which yields to partial
hydration of VDPA.
2
.2. SAR analysis of the BP anti-integrase activity
Recently we demonstrated the capacity of two BPs bearing aromatic substituents to
inhibit the IN-catalyzed 3’-processing [15]. Here we report the results of a structural analysis of
a series of BPs in the attempt to find structural peculiarities essential for the inhibition of HIV-1
IN.
First, we analyzed the dependence of the BP inhibitory activity in 3’-processing reaction
versus the substituent R structure (Fig. 1) and found that only compounds with a chlorobenzyl
radical as the substituent R were active (Fig. 1, compounds 2a-d, and 5). Their inhibitory
activities are presented in Table 1 (column 3); all other compounds were not active up to 250 µM
concentration. We assumed that the reason for this effect could be the electron deficiency of the
aromatic ring due to electron acceptor properties of the chlorine atom. To confirm this
hypothesis, we synthesized compounds 3a-e containing cationic pyridinium residues, which are
more electron deficient than chlorobenzyl groups. However, these compounds did not inhibit IN

ACCEPTED MANUSCRIPT
up to concentrations of 250 µM. It is also noteworthy that compound 2b, although lacking the
chlorine atom at the meta-position and, hence, less electron deficient, was a more effective
inhibitor than the corresponding 3,4-dichlorobenzyl derivative 2a (Table 1, column 3). These
results imply that a chlorine atom at the para-position of the aromatic ring is necessary for BPs
to display inhibitory properties. Compound 5 bearing two 3,4-dichlorobenzyl residues was
somewhat a more effective inhibitor than compounds 2a, 2c, and 2d containing one 3,4-
dichlorobenzyl residue, but its activity was comparable with that of the monochlorobenzyl
derivative 2b. It should be noted that a halogen-containing aromatic residue is a mandatory
structural element of all efficient INSTIs [1],[2],[8],[26] and some of 3’-processing inhibitors
[
27],[28],[29]. In the case of INSTIs, a halogenated benzyl group was shown to be involved in
stacking interactions with the cytosine of the conserved CA dinucleotide forcing the 3′-OH of the
terminal adenosine away from the IN active site [30],[31]. However, the actual role of the
halogen atom in this process has been vague so far.
2+
2+
It is known that BPs exhibit strong affinity to metal ions, in particular, Ca and Mg
32],[33]. We supposed that the BP inhibitory activity toward IN can be explained by their
[
2
+
capacity to coordinate Mg ions, which are bound by amino acids D64, D116, and E152 in the
enzyme active site. The nature of the substituent X is likely to have a substantial impact on BP
metal binding properties and, therefore, their inhibitory activity. Due to the oxygen lone electron
pairs, the hydroxyl group at the P-C-P backbone (compound 2a) can be involved in the
coordination of the metal ion. However, the replacement of the hydroxyl group by the hydrogen
atom (compound 2d) did not change the inhibitory activity (Table 1, column 3). This result has
2
+
thrown into doubt the assumption on the obligatory presence of the Mg -coordinating site in the
inhibitor structure. Therefore, we evaluated two dichlorobenzyl derivatives containing carboxyl
and phosphonate fragments in place of the bisphosphonate backbone (compounds 6 and 7). Both
of them lacked the inhibitory activity up to the concentration of 500 µM. This result allowed us a
conclusion that the bisphosphonate backbone is an obligatory structural element of the inhibitors
of this type.
The most potent IN inhibition was observed for compound 2c, which had an amino group
at the P-C-P backbone (Table 1, column 3). The transfer of the amino group into the linker that
joins the P-C-P backbone and the aromatic fragment (compounds 4a-4c) caused the complete
loss of the BP inhibitory activity without regard to the position of the amino group in the linker
or the presence of a halogen atom (iodine) in the aromatic ring.
Thus, the results of the SAR analysis showed that the BP inhibitory activity is determined
by the presence of two structural moieties: a methylenediphosphonic backbone and a p-
chlorobenzyl substituent.

ACCEPTED MANUSCRIPT
2
.3. The effect of the IN metal cofactor on the BP inhibitory activity
The absolute necessity of the P-C-P backbone for the BP inhibitory activity indicated that
the latter may be associated with the coordination of the metal ion. With the goal of testing this
assumption we studied the inhibition of 3’-processing and strand transfer reactions in the
2
+
2+
presence of two IN metal cofactors, Mg and Mn , using the most active inhibitors 2a, 2b, 2c,
d, and 5.
First, we found that all the compounds inhibited the IN activity in both reactions. In the
2
2
+
presence of Mg ions, the efficiency of the 3’-processing and strand transfer inhibition was
nearly the same for compounds 2d and 5; compound 2c inhibited the strand transfer twice as
effectively; and compounds 2a and 2b bearing a hydroxyl group in the P-C-P backbone
somewhat better inhibited the 3’-processing (Table 1, columns 3 and 5).
2
+
In the presence of Mn ions all the compounds inhibited both reactions with nearly the
same efficiency (Table 1, columns 4 and 6). It is noteworthy that the inhibitory activities of
compounds 2a, 2b, 2d, and 5 considerably grew under these conditions. It was especially
noticeable for inhibitor 5 containing two 3,4-dichlorobenzyl residues. Also, the inhibitory
activities of compounds 2a, 2b, and 2d were comparable with that of compound 2c, whereas
2
+
they markedly differed in the presence of Mg .
In accordance with these results, all the tested BPs could be divided into two groups:
group I, which included only compound 2c with an amino group at the P-C-P backbone, and
group II, with all the rest BPs (compounds 2a, 2b, 2d, and 5). The fact that the metal nature was
essential for the inhibitory activity of only group II inhibitors allowed us to make two
assumptions. First, the mechanism of the IN inhibition by compound 2c was independent of the
metal cofactor nature. Second, in the presence of different metal cofactors compounds of group
II inhibited IN by different mechanisms. Thus, the inhibition mechanism was the same for both
2
+
2+
groups in the presence of Mn and differed in the presence of Mg .
2
.4. NMR-studies of BP interactions with metal ions
The major reason for such a great difference in the behavior of the inhibitors of the two
groups may be differences in electron structures of these compounds. One of these differences is
the capacity of the amino group of compound 2c (group I) to be protonated, whereas the
substituents X in the group II compounds lack this property under the tested conditions.
Evidently, a positive charge in the BP metal-binding site should affect its interaction with metal
ions and, hence, the character of IN inhibition.

ACCEPTED MANUSCRIPT
With the goal of evaluating the impact of X substituents on the BP binding to metal ions,
we studied the distribution of the electron density in inhibitors 2a and 2c using NMR
spectroscopy, which is an accurate and sensitive method characterizing the molecule structure in
solution.
1
It was observed in H NMR spectra that the electron density in the hydrocarbon
backbones of these inhibitors was distributed in a different way. This can be explained by the
+
electron withdrawing effect of the NH
3
-function in compound 2c. As is seen in Fig. S1 (see SI),
compound 2c is characterized by a decreased shielding and a downfield shift of proton
resonances. For the linker -CH - group the reduction of the electron density resulted in a
2
decrease in the bond lengths, and for the aromatic system, in a decrease in the electron density of
the π-system. The synchronous downfield shift of the aromatic protons implies that the cationic
+
function does not cause the redistribution of the π-system electron density. Nevertheless, NH -
3
function can influence cation-π and anion-π interactions (reviewed in [34], which may be formed
between the inhibitor and IN.
The effect of the X substituent on the interaction of compounds 2a and 2c with
3
1
magnesium ions was studied using P NMR spectroscopy. Unlike H- and HO-BPs, the shift of
the phosphorus resonance of NH -BPs at pH close to pK of their amino group is much more
2
a
pronounced [35]. This considerable downfield shift of the phosphorus resonance can be ascribed
+
to deprotonation of the NH
3
-group. This effect was detected for inhibitor 2c (pK
a
of its amino
group is 8.3), whereas the resonance of compounds 2a was nearly constant (Fig. 2A).
2
+
We also studied the interaction of compounds 2a and 2c with Mg ions at pH 1, 7, and
9
.5, i.e., for the amino group of the latter was completely protonated, partially protonated, and
3
1
completely deprotonated, using the P NMR line broadening method. This approach is based on
2
+
the principle that formation of coordination bonds between Mg and a phosphonate group
31
causes the broadening of the phosphorus resonance in P NMR [36] . As is seen in Fig. 2B, the
broadening of the phosphorus resonance of the amino derivative 2c at pH 7 and 9.5 occurs much
more intensely than that of the hydroxyl analogue 2a. Since this effect is based on the formation
2
+
of coordination bonds between BP and Mg , we believe that the P-C-P backbone of compound
2
+
2
a does not form coordination bonds with Mg , whereas in the case of compound 2c such bonds
are formed. Moreover, the greater is the portion of nonprotonated NH -groups, the greater is the
2
efficacy of their formation. These experiments evidenced that the lone electron pair of the
nitrogen atom nonprotonated amino group of compound 2c was involved in the formation of
2
+
coordination bonds with Mg ions.
It is noteworthy that the IN inhibition was studied at рН 7.2, that is, when the amino
group of compound 2c is nearly fully protonated (Fig. 2A) and thus not capable to coordinate

ACCEPTED MANUSCRIPT
2
+
+
Mg ions. However, the NH -group can also strengthen metal binding properties of compound
3
2
a
c. It is known that α-amino BPs have decreased pK of phosphonate groups if compared with
the analogues bearing protons or hydroxyl groups at the P-C-P backbone (1–2 log units [33]).
+
3
Such a decrease is a result of the considerable electron acceptor effect of the NH
cation, which
markedly enhances the coordination capacity of α-amino BPs toward bivalent metal ions.
When summarizing the above facts one can conclude that BP bearing amino groups in the
P-C-P backbone possess some physicochemical and, probably, structural specific features,
particularly, an increased acidity of protons of aromatic substituents and the capacity to
2
+
coordinate Mg ions. It is likely that these specific features lead to different mechanisms of IN
2
+
2+
inhibition by group I and II BPs in the presence of Mg . As for Mn ions, we failed to study the
BP interactions with them because of the paramagnetic nature of manganese.
Figure 2.
2
. 5. Studies of the inhibition mechanism
The kinetic studies for the evaluation of the inhibition type were performed with
compounds 2a (group II) and 2c (group I). We studied the dependence of the initial rate of the
3
’-processing reaction vs the DNA substrate concentration at varied inhibitor concentrations in
2
+
2+
the presence of Mg or Mn ions. The results are given as double-reciprocal plots (Fig. 3).
2
+
First, we studied this dependence for compound 2a in the presence of Mg ions. As is
seen in Fig. 3A, an increase in the compound 2a concentration practically did not affect the
maximal rate of the reaction (Vmax) but caused an increase in the apparent Michaelis constant
m
(K ). This implies that in general the interaction of compound 2a with IN follows the mechanism
of competitive inhibition, i.e., the inhibitor binds to and competes with the substrate in the
enzyme active site. Based on this model, the calculated values of the Michaelis constant K and
m
I
inhibition constant K were determined as 12+/-6 nM and 38+/-8 µM, respectively, for the IN
2+
inhibition by compound 2a in the presence of Mg ions.
Figure 3.
For compound 2c these dependences were shown to have another shape (Fig. 3B). In the
2
+
presence of Mg ions an increase in the inhibitor concentration resulted in the reduction of the
Vmax, whereas K did not change. This mode corresponds to the noncompetitive inhibition, i.e.,
m
the inhibitor binds to IN at the site differing from the substrate binding site. Based on the
m I
calculated K value (see above), the value of the inhibition constant K was found 2.0+/-0.5 µM.

ACCEPTED MANUSCRIPT
2+
In the case of Mn ions the dependences of the 3’-processing initial rate vs the DNA
substrate concentration at varied inhibitor concentrations were similar and supported the
noncompetitive mechanism for both inhibitors (Figs. 3C and D). According to this model the
calculated inhibition constants for compounds 2a and 2c were 4.4+/-0.8 and 1.0+/-0.3 µM,
respectively.
Thus, the study of the dependence of the reaction initial rate vs the DNA substrate
concentration at varied inhibitor concentrations demonstrated that the mechanism of IN
inhibition for group II BPs (compound 2a) changes from competitive to noncompetitive with the
2
+
2+
cofactor change from Mg to Mn , whereas group I BP (compound 2c) always inhibited IN by
the noncompetitive mechanism.
2
.6. The BP effect on the IN-DNA binding
Upon the competitive inhibition the inhibitor competes with the substrate for the binding
in the enzyme active site. Based on this principle we assumed that the group II inhibitors
compounds 2a, 2b, 2d, and 5) would prevent the formation of the IN-DNA complex in the
(
2
+
presence of Mg ions. Upon the noncompetitive inhibition, which was observed for all the tested
2
+
BP in the presence of Mn ions, the inhibitor bound to the enzyme at some other site and,
therefore, is unlikely to have an impact upon the formation of the enzyme-substrate complex.
We studied the influence of the inhibitors of both groups on the complex formation of the
IN with the U5 substrate in the presence of ions of both metals in the direct experiment using the
DRaCALA (Differential Radial Capillary Action of Ligand Assay) approach [37]. Since this
method was not previously used for characterizing the IN-DNA binding, first we evaluated its
d
applicability for this purpose. To this end, we determined the dissociation constant, K , of the
IN-DNA complex, (Fig. 4A). The calculated value of 20+/-6 nM, was close to that earlier found
for the IN-DNA complex by the methods of gel shift[38] and fluorescence anisotropy change
[
39] . Thus, the DRaCALA method was proved to be useful for studying the IN-DNA binding.
As we suggested, none of the tested BP affected the formation of the IN-DNA complex
2
+
2+
in the presence of Mn ions (Table 1, column 2). In the presence of Mg ions compounds 2a,
b, 2d, and 5, but not 2c, inhibited the IN binding to the substrate (Table 1, column 1). These
results confirmed the previous conclusions about different types of inhibition for group I and II
2
2
+
2+
BPs in the presence of Mg ions and the change of the mechanism upon the Mg replacement
2
+
by Mn . However, it is noteworthy that the inhibition of DNA binding was observed at higher
BP concentrations than the inhibition of the IN catalytic activity (Table 1, columns 1, 3, and 5).
This can probably be explained by the fact that the DRaCALA method only allows the
evaluation of overall DNA-IN binding including both the substrate-specific and nonspecific

ACCEPTED MANUSCRIPT
binding and, hence, we can only study the effect of the inhibitor in total. The inhibitor
concentrations, at which they inhibit the formation of the specific IN-substrate complex, can be
lower than those found by us and can be closer to those, at which they inhibit the 3’-processing
and strand transfer reactions.
2
.7. The inhibitor effect on the IN activity within the LEDGF complex
With the goal of further studying the mechanism of the IN inhibition with BP derivatives
we tested the activity of these compounds toward the pre-formed IN-LEDGF complex composed
of four IN and two LEDGF molecules [40]. Previously we demonstrated that the efficiency of
the IN inhibition with a competitive inhibitor derived from a dimeric bisbenzimidazole reduced
in the presence of LEDGF [41]. For the BPs tested, we found that group II compounds (2a, 2b,
2
+
and 2d) did not inhibit IN within the IN-LEDGF complex in the presence of Mg ions up to
2
+
concentrations of 250 µM (Table 1, column 7). In the presence of Mn ions they inhibited the
complex with a similar efficiency as IN alone (Table 1, columns 4 and 8). As for compound 2c,
its inhibitory activity toward IN alone and IN-LEDGF complex was similar in the presence of
both ions. These results supported once again the hypothesis on different mechanisms of IN
inhibition by the tested BP derivatives.
Since competitive inhibitors compete with the substrate for the enzyme binding, the
inability of compounds 2a, 2b, and 2d to inhibit the IN-LEDGF complex can be explained by
kinetic or thermodynamic factors, that is, faster or tighter binding of the DNA substrate to IN
within its complex with LEDGF than to the free IN. The study of the strength of the DNA-IN-
2
+
LEDGF complex in the presence of Mg ions showed that K
which is comparable with the K value for the DNA-IN complex (Fig. 4A). We also calculated
the rate of the DNA binding to IN and the IN-LEDGF complex using the DRaCALA method
Fig. 4B). The initial rate of the DNA binding to the complex (V 3.0 nМ/min) was found to be
considerably higher than that of the binding to IN (V = 0.7 nМ/min).
d
of this complex was 25+/-10 nM,
d
(
o
o
Thus, the lack of activity of inhibitors of group II toward the IN-LEDGF complex can be
explained by kinetic factors: the DNA binds to the complex faster than the inhibitor does. In the
case of free IN the situation is inverse.
Figure 4.
2
.7. Inhibition of IN mutants with increased resistance toward strand transfer inhibitors

ACCEPTED MANUSCRIPT
All the experiments performed imply that compounds of groups I and II followed
different mechanisms while inhibiting IN, that is, they are likely to bind to the enzyme in
different modes. In order to find out their binding sites, we tested their activities toward IN
preparations bearing E92Q, E138K, G140S, and N155H mutations that are known to affect the
binding to IN of such inhibitors as L-chicoric acid, diketoacid L-731,988, raltegravir, elvitegravir,
and dolutegravir, a second generation INSTI [7],[42],[43]. For these experiments we also used
2
+
compounds 2a and 2c and Mg ions, i.e., the conditions under which different mechanisms are
involved and, hence, binding to IN occurs at different sites. However, both compounds inhibited
the 3’-processing catalyzed by all of the mutants and wild type IN with a similar efficacy: in all
the cases IC50 was 35+/-7 µM for compound 2a and 7+/-2 µM for compound 2c. These results
demonstrated that the character of the BP binding to IN differed from that of the known
inhibitors, but unfortunately they were useless for specifying the binding sites.
3
. Discussion and Conclusions
The interest in BPs is stimulated by their capacity to inhibit the pyrophosphorolytic
activity of HIV-1 reverse transcriptase and thus reduce the susceptibility of drug-resistant virus
strains to azidothymidine [44]. We demonstrated that BPs could also inhibit another HIV-1
enzyme, integrase. The SAR analysis of the BP anti-integrase activity showed that these
compounds must contain a chlorobenzyl residue as an R substituent for being effective IN
inhibitors (Fig. 1). Substituents with other halogen atoms were not studied in the present work,
and we cannot rule out their positive effect on the inhibitory activity of BPs. Of note, a halogen-
containing aromatic substituent is a mandatory structural element of many of the IN inhibitors,
but the halogen role is still unclear. A reason for the increased activity of halogen-containing
inhibitors may be the formation of halogen bonding[45],[46]. This noncovalent interaction
formed by the halogen atom and the nucleophile containing a lone electron pair is similar to the
hydrogen bonding in both the properties and the way of formation.
For some enzymes it was found that a significant increase in the activity of their
inhibitors was a result of the formation of halogen-bonded complexes [47],[48],[49]. Chlorine,
bromine, and iodine atoms, but not the fluorine one, which possesses specific properties [50] can
be involved in the formation of halogen bonds [45],[49]. Therefore, it is likely that the inhibitory
activity of BPs with a chlorobenzyl substituent can be explained by the halogen bonding
formation, which stabilizes the IN-inhibitor complexes.
Also, we found that the X substituent has an impact on the IN inhibition mechanism. In
2
+
the presence of Mg ions the inhibition mechanism of BP 2c with X = NH
2
(group I) differs
from that of all other BP (group II). This fact indirectly indicates that the group I and II

ACCEPTED MANUSCRIPT
compounds bind to the enzyme using different modes (probably, at different binding sites). We
assume that the reason for the different behavior of group I and II compounds may be the
1
difference in their electron structures. Indeed, the H NMR spectra indicated different
31
distribution of the electron density in the hydrocarbon backbone of these compounds. The
P
NMR spectra demonstrated that the P-C-P backbone of the group II inhibitors did not coordinate
2
+
Mg ions, whereas in the case of compound 2c such coordination occurred with the efficacy that
is growing with an increase in the amino group deprotonation degree.
Although currently it is difficult to detect directly the differences in the BP-IN binding,
our results support further studies of BPs as IN inhibitors. Among all the compounds studied, α-
amino BPs (analogues of compound 2c) may be of the greatest interest, since, being IN
noncompetitive inhibitors, they may inhibit the IN functioning in the preintegration complex. It
is also essential that BP-based inhibitors are active toward IN mutants resistant to several known
INSTI.
4
. Experimental protocols
Materials and Methods
The structures of 17 bisphosphonates studied are shown in the Figure 1.
All reagents used were purchased from Aldrich or Acros Organics.
The acyl halides used in Method C were prepared by refluxing the corresponding acids
with 2 eq of SOCl in dry dichloromethane for 3-4 hours followed by vacuum distillation to give
2
7
5–88% of the target compounds.
The pyridinium acetic acids or acetonitriles were prepared reacting the corresponding
pyridines with bromoacetic acid or bromoacetonitrile. For details see SI.
Flash chromatography was performed on Kieselgel (40-63 ꢀm, Merck, Germany).
1
The inhibitor concentrations were measured by H NMR using the known concentrations
of tert-butanol-d1 in D O as an internal reference standard (relaxation delay (D1) was set to 60 s
2
1
for H measurements).
High-resolution electrospray mass spectra were obtained on Applied Biosystems/MDS
1
31
Sciex QSTAR XL. NMR spectra were recorded on 400 MHz ( H), 161.9 MHz ( P), and 100.6
1
3
MHz ( C) AMX III -400 Bruker spectrometer. Chemical shifts are reported in parts per million
1
13
31
(
ppm) using tetramethylsilane ( H), tert-butanol-d1 ( C), and 85% H
3
PO
4
( P) as external
13
31
standards. C and P NMR spectra were proton-decoupled unless otherwise specified.

ACCEPTED MANUSCRIPT
Method A.
A mixture of carboxylic acid (2 mmol) or nitrile (2 mmol), phosphorous acid (2 mmol)
and methanesulfonic acid (5 mL) in dry benzene (10 mL) was refluxed for 10-20 min till the
3
mixture became homogenous. On cooling to 60-65 °C a solution of POCl (3 mmol) in dry
benzene (10 mL) was added dropwise for 30 min, maintaining the temperature below 70°C. The
reaction mixture was refluxed with stirring for addition 8-12 h (24 h for (3b)), then the viscous
yellowish mixture was cooled to ~40°C and pooled into the ice-cooled 3 N HCl (50 mL) under
vigorous stirring followed by the reflux for 2 h. The aqueous layer was separated, concentrated
in vacuo to a volume of 3 mL and kept for 12-24 h at +5°C till the bisphosphonates 3a-e, 1d, 2c
were crystallized. For BPs 1a-c, the pH of the water layer was first adjusted to 4 with 50%
NaOH and then kept for 5-12 h at +5°C for crystallization.
Method B.
Vinylidenediphosphonic acid.
VDPA was prepared by thermal dehydration of tetrasodium salt of etidronic acid followed by
partial deionization of tetrasodium VDPA by CO gas. Sodium bicarbonate formed was removed
2
by filtration. For details and NMR data see SI.
Synthesis of 2-amino-1,1-bisphosphonates 4a-4c.
A water solution of disodium VDPA (1 eq) in water and the corresponding amine (1.5 eq)
was concentrated in vacuo to homogenous syrup and heated at 120°C for 3 h. The reaction
mixture was poured into a 50% water/ethanol mixture, acidified with HCl to pH 1 and allowed to
1
13
crystallize at +5°C. The precipitated zwitterionic BPs were of 95% purity according to H, C,
31
and P NMR data.
Method C.
A solution of acyl chloride (3 mmol) in dry benzene (10 mL) was added dropwise at 0°С
to triethylphosphite (3 mmol) in dry benzene (10 mL) under vigorous stirring and followed by
stirring for additional 2 h at +5°С. Diethylphosphite (3 mmol) and diisopropyl amine (0.3 mmol)
were added and stirring was continued for 4–5 h at+5°С. Reaction mixture was concentrated in
3
vacuo and 2a, 2b were purified by column chromatography on silica gel with CHCl -MeOH (0
to 15% MeOH) as an eluent. Solvents were removed in vacuo and ethyl groups were routinely
removed by using 5 equivalents of TMSBr overnight at rt followed by methanolysis.

ACCEPTED MANUSCRIPT
Method D.
To a solution of 1.2 mL (4 mmol) tetraisopropyl methylenebisphosphonate in dry THF (5
mL) 160 mg (4 mmol) of NaH (60% suspension in oil) in of dry THF (5 mL) was added at 0°C
with stirring under the Ar atmosphere and stirring was continued for 1 h at 0°C, and then for 3 h
at rt. A solution of 782 mg (4 mmol) 3,4-dichlorobenzyl chloride in THF (5 mL) was added to
the resulted carbanion solution and stirred overnight at rt. The reaction was quenched with
4 2 4
saturated NH Cl (20 mL), extracted with DCM (3x50 mL), dried (Na SO ) and concentrated in
vacuo. Target BPs 2d or 5 were purified by column chromatography on silica gel eluting with
EtOAc – MeOH gradient (0 to 25% MeOH). For bisphosphonate 5 10 mmol NaH and 10 mmol
3
,4-dichlorobenzylchloride were used. Tetraisopropyl esters were refluxed with 48% HBr (4 mL)
for 2 h followed by co-evaporation with water (3x15 mL) and drying in vacuo over P O /KOH to
2
5
give target 2d and 5 as vaxi oils.
1
-Hydroxy-2-phenylethylidene-1,1-bisphosphonic acid (1a). Compound 1 was prepared from
1
phenylacetic acid following Method A as colorless solid (disodium salt, 490 mg, 75%). H NMR
3
1
(D
2
O, pH 1): δ 7.38-7.28 (m, 5H), 3.32 (t, JPH = 14 Hz, 2H). P (D
2
O, pH 1): δ 19.0 s; lit. 19.0
13
s[16]. C (D O, pH 2, t-BuOD 30.3 ppm): δ 7.48-7.36 (5H, m, Ph), 74.0 (t, J = 147.0 Hz),
2
PC
4
0.6 s
Compound 1b was available from previous work [17]
1
-Hydroxy-5-phenylpentylidene-1,1-bisphosphonic acid (1c). Compound 1c was prepared
from phenylpentanoic acid following General Method A as colorless solid (tetrapotassium salt,
1
3
2
91 mg, 41%). H NMR (D
2
O, pH 2):δ 7.35-7.20 (m, Ph, 5H), 2.63 (t, J = 7.1 Hz, CH
2
-PCP,
3
1
13
H), 2.08-1.96 (m, CH
2
-PCP, 2H), 1.66-1.57 (m,-(CH
2
)
2
-, 4H). P (D
2
O, pH 2):δ 18.5 s. C
(D
2
O, pH 2, t-BuOD 30.3 ppm): δ 144.4 s, 129.6 s, 126.8 s, 74.1 t (JPC = 147 Hz), 35.2 s, 33.9 s,
3
1
2.0 s, 30.1 s.
-Amino-2-phenylethylidene-1,1-bisphosphonic acid (1d). Compound 1d was prepared from
1
phenylacetonitrile following Method A as colorless solid (180 mg, 32%). H NMR (D
2
O, pH 3):
31
δ 7.46-7.32 (5H, m, Ph), 3.43 (2H, t, JPC = 12.5 Hz, CH
2
-PCP). P NMR (D
2
O, pH 6, P-H
31
13
coupled) δ 12.4 (t, JPH = 12 Hz). P NMR (D O, pH 2) δ 12.3 s. C NMR (D O, pH 14, t-BuOD
2
2
3
0.3 ppm): δ 141.9 s, 133.3 s, 128.6 s, 126.8 s, 58.5 (t, JPC = 129.4 Hz), 41.2 s.
1
-Hydroxy-2-(3,4-dichlorophenyl)ethylidene-1,1-bisphosphonic acid (2a). Compound 2a was
prepared from 3,4-dichlorophenylacetic acid following Method C as colorless solid (750 mg,

ACCEPTED MANUSCRIPT
1
7
7
1%). H NMR (D O+25% DMSO-d6, pH 2): δ 7.39 (d, J 1.4 Hz, 1H), 7.28 (d, J 8.3 Hz, 1H),
2
1
3
2
.15 (dd, J = 8.3 Hz, 1.4 Hz, 1H), 3.11 (t, JPH = 13 Hz, 2H). C NMR (D O+25% DMSO-d6,
31
pH 2, t-BuOD 30.3): δ 132.5 s, 131.8 s, 131.1 s, 130.8 s, 74.3 (t, JPC = 147.2 Hz), 38.3 s. P
NMR (D O, pH 2): δ 18.0 s
-Hydroxy-2-(4-chlorophenyl)ethylidene-1,1-bisphosphonic acid (2b). Compound 2b was
prepared from 4-chlorophenylacetic acid following Method C as colorless solid (785 mg, 82%).
2
1
1
13
H NMR (D
2
O, pH 5): δ 7,37 (d, J = 8.4 Hz), 7,28 (d, J = 8.4 Hz), 3,23 (t, J = 12.6 Hz). C NMR
(
D O, pH 12, t-BuOD 30.3 ppm): δ 134.1 s, 132.3 s, 131.5 s, 129.5 s, 74.0 (t, JPC = 147 Hz), 38.9
2
3
1
s. P NMR (D
H]+314.9590).
2 8 7 2
O, pH 5): δ 18.0 s. HREIMS: m/z 314.9594 (calcd for C H10ClO P , [M-
1
-Amino-2-(3,4-dichlorophenyl)ethylidene-1,1-bisphosphonic acid (2c). Compound 2c was
prepared from 3,4-dichlorophenylacetonitrile following Method A as colorless solid (273 mg,
1
3
8
1
9%). H NMR (D
2
O, pH 11): δ 7.55 (d, J = 2 Hz, 1H), 7.39 (d, J = 8.3 Hz, 1H), 7.30 (dd, J =
13
.3, J = 2 Hz, 1H), 3.16 (t, J = 12 Hz, 2H). C NMR (D
2
O, pH 11, DMSO-d6 39.52 ppm): δ
40.4 s, 134.4 s, 132.8 s, 131.9 s, 130.9 s, 130.6 s, 57.5 (t, JPC = 123.6 Hz), 40.3 s
3
1
P NMR (D
47.9360).
2 8 2 6 2
O, pH 11): δ 18.9 s. HREIMS: m/z 347.9366 (calcd for C H10Cl NO P , [M-H]+
3
2
-(3,4-dichlorophenyl)ethylidene-1,1-bisphosphonic acid (2d). Compound 2d was prepared
from 3,4-dichlorobenzylchloride following Method D as colorless solid (disodium salt, 379 mg,
1
5
0%). H NMR (D
2
O, pH 10): δ 7.53 (s, 1H), 7.40 (d, J = 8.2 Hz, 1H), 7.27 (dd, J
1
= 8.2 Hz, J
2
1
3
=
2 Hz 1H), 3.06 (ddd, JHH = 7 Hz, JPH = 15 Hz, 2H), 2.10 (dt, JHH = 7 Hz, JPH = 20 Hz, 1H). C
2
NMR (D O, pH 10, t-BuOD 30.3 ppm): δ 157.3 s, 138.6 s, 126.1 s, 125.2 s, 124.3 s, 36.8 (t, JPC
31
=
111 Hz), 26.4 s. P NMR (D
2
O, pH 10): δ 19.4 s. HREIMS: m/z 332.9265 (calcd for
C
8
H
9
Cl , [M-H]+332.9251).
2 6 2
O P
1
-Hydroxy-2-(pyridine-1-yl)ethylidene-1,1-bisphosphonic acid (3a). Compound 3a was
prepared from 2-(pyridine-1-yl)acetic acid following Method A as colorless solid (449 mg,
1
7
9%). H NMR (D
2
O, pH 1): δ 8.9 (d, J = 6 Hz, 2H), 8.6 (t, J = 8 Hz, 1H), 8.0 (t, J = 7Hz, 2H),
13
5
.0 (t, J = 9.5 Hz). C NMR (D
2
O, pH 1, acetone 30.9 ppm): δ 146.7 s, 127.6 s, 73.4 (t, J = 137
31
Hz, PCP), 64.4 s (CH
2
). P NMR (D O, pH 1): δ 14.4 s; lit. 14.3 s [20]
2
1
-Amino-2-(pyridine-1-yl)ethylidene-1,1-bisphosphonic acid (3b). Bisphosphonate 3b was
prepared by Method A from 2-(pyridine-1-yl)acetonitrile (synthesized as described in SI) as
1
colorless solid (221 mg, 39%). H NMR (D O, pH 1): δ 9.0 (d, J = 5.9 Hz, 2H), 8.7 (t, J = 7.8
2
13
2
Hz, 1H), 8.1 (t, J = 6.7 Hz, 2H), 5.2 (t, JPH = 9.5 Hz). C NMR (D O, pH 1): δ 146.9 (s, Py),
31
1
8
2 2
46.0 (s, Py), 127.7 (s, Py), 61.5 (s, CH ), 57.1 (t, JPC = 123 Hz, PCP). P NMR (D O, pH 1): δ
3
1
.5 s. P NMR (D O, pH 14): δ 17.4 s
2

ACCEPTED MANUSCRIPT
1
3
-Hydroxy-2-(4-phenylpyridine-1-yl)ethylidene-1,1-bisphosphonic acid (3c). Bisphosphonate
c was prepared by Method A from 2-(4-phenylpyridine-1-yl)acetic acid (synthesized as
described in SI) as colorless solid (526 mg, 73%).
1
2
H NMR (D O, pH 1): δ 8.8 (d, J = 6.3 Hz, 2H, Py), 8.2 (d, J = 6.3 Hz, 2H, Py), 7.9 (d, J = 6.6
1
3
Hz, 2H, Ph), 7.6 (m, 3H, Ph), 5.0 (t, JPH = 9.5 Hz). C NMR (D
2
O, pH 1): δ 159.9 (s, Py), 148.9
(s, Py), 137.0 (s, Py), 135.1 (s, Ph), 132.6 (s, Ph), 131.0 (s, Ph), 126.8 (s, Ph), 76.1 (t, JPC = 134
3
1
2 2
Hz, PCP), 66.2 (s, CH ). P NMR (D O, pH 1): δ 14.9 s (lit. 14.7 s [21])
1
-Hydroxy-2-(3-hydroxypyridine-1-yl)ethylidene-1,1-bisphosphonic acid (3d).
Bisphosphonate 3d was prepared by Method A from 2-(3-hydroxypyridine-1-yl)acetic acid
1
(
synthesized as described in SI) as colorless solid (282 mg, 47%). HNMR (D
2
O, pH 14): δ 7.8
1
3
(
d, J = 5.6 Hz, 1H), 7.8 (s, 1H), 7.3 (m, 2H), 4.6 (t, J = 9 Hz, 2H). CNMR (D
2
O, pH 10,
acetone 30.9 ppm): δ 164.6 (s, Py), 136.1 (s, Py), 133.2 (s, Py), 130.6 (s, Py), 126.0 (s, Py), 73.8
31
(
t, JPC = 127 Hz, PCP), 64.5 (s, CH
2
). P NMR (D
-Hydroxy-2-(2-ethylpyridine-1-yl)ethylidene-1,1-bisphosphonic acid (3e). Bisphosphonate
e was prepared by Method A from 2-(2-ethylpyridine-1-yl) acetic acid (synthesized as
2
O, pH 14): δ 15.5 s
1
3
described in SI) as colorless solid (319 mg, 51%).
1
2 1 1 1 2
H NMR (D O, pH 1): δ 8.8 (dd, J = 6.4 Hz, J = 1.6 Hz, 1H, Py), 8.3 (ddd, J = 8.0 Hz, J = 1.7
Hz, 1H, Py), 7.8 (dd, J = 8.0 Hz, J = 1.8 Hz, 1H, Py), 7.7 (ddd, J = 6.4 Hz, J = 1.7 Hz, 1H,
1
2
1
2
2 2
Py), 5.0 (t, JHP = 10 Hz, 2H, N-CH ), 3.3 (q, J = 7.5 Hz, 2H, CH -Et),1.27 (t, J = 7.5 Hz, 3H,
13
CH
3 2
). CNMR (D O, pH 10): δ 161.0 (s, Py), 146.7 (s, Py), 144.3 (s, Py), 126.4 (s, Py), 123.0
31
(
s, Py), 73.8 (t, JPC = 126 Hz, PCP), 59.1 ( s, CH -Py), 26.1 (s, CH -CH ), 11.2 (s, CH ). P
2
2
3
3
3
1
NMR (D
2 2
O, pH 10): δ 15.3 s. P NMR (D O, pH 10, coupled): δ 15.3 (t, JPH = 10 Hz)
(2-Phenylethylamino)-2-ethylidene-1,1-bisphosphonic acid (4a). Compound 4a was prepared
from 3.78 mL (30 mmol) 2-phenylethylamine and VDPA (20 mmol) following Method B as
1
colorless solid (4.3 g, 69%, as calculated for VDPA). H NMR (D
2
O, pH 9): δ 7.39-7.27 (5H, m,
-CH -NH),
-Ph), 2.07 (tt, JHH = 6.5 Hz, JPH = 20 Hz, 1H, H-PCP). P NMR
O, pH 9): δ 16.2 s. C NMR (D O, pH 9, t-BuOD 30.3ppm):δ 137.3 s, 129.6 s, 129.5 s,
27.8 s, 48.6 s, 47.2 s, 37.5 (t, JPC = 110 Hz)
Ph), 3.35 (dt, JHH = 6.5 Hz, JPH = 15 Hz, 2H, CH
2
-PCP), 3.27 (t, J = 7.5 Hz, 2H, CH
2
2
31
2
.98 (t, J = 7.5 Hz, 2H, CH
2
13
(D
2
2
1
Phenylamino-2-ethylidene-1,1-bisphosphonic acid (4b)
Compound 4b was prepared from 2.7 mL (30 mmol) freshly distilled aniline and VDPA (20
mmol) following Method B as colorless solid becoming brownish if kept on air (4.1 g, 73%, as
1
calculated for VDPA). H NMR (D
2
O, pH 2):δ 7.51-7.42 (5H, m, Ph), 3.35 (2H, ddd, JHH 7.5 Hz,
1
3
J
2 2
PH 15Hz, JPH 13Hz, CH -PCP), (1H, tt, JHH = 7.5 Hz, JPH = 21 Hz, H-PCP). C NMR (D O, pH

ACCEPTED MANUSCRIPT
2
1
, t-BuOD 30.3 ppm) : δ 134.9 s, 131.0 s, 130.9 s, 130.6 s, 123.1 s, 50.2 s (CH ), 36.3 (t, J =
2
CP
3
1
2
21 Hz). P NMR (D O, pH 2):δ 15.8 s
((4-Iodophenyl)amino)-2-ethylidene-1,1-bisphosphonic acid (4c). To a solution of 562 mg (2
mmol) bisphosphonate 4b in 1M NaHCO (10 mL) 762 mg (3 mmol) of iodine was addad and
3
resulted mixture was stirred overnight at 5°C followed by the addition of 126 mg (1 mmol)
+
Na
2
SO
3
. The resulted solution was applied on a column of Dowex-50 (H ) resin (50 mL) then
column was washed with 50 mL of water and combined eluents were concentrated in vacuo to 5
1
mL and kept at +5°C for crystallization (635 mg, 78%). H NMR (D
2
O, pH 10): δ 7.5 (d, J = 9
Hz, 2H), 6.6 (d, J = 9 Hz, 2H), 3.5 (ddd, JHH = 7 Hz, JPH = 14 Hz, 2H), 2.1 (dt, JHH = 7 Hz, JPH
=
=
3
1
31
2
2
3
2 1 2
0.5 Hz, 1H). P NMR (D O, pH 10): δ 18.5 s. P NMR (coupled): δ 18.5 (dt, J = 14 Hz, J
1
3
0.5 Hz). C NMR: δ (D
2
O, pH 7, t-BuOD 30.3 ppm): δ 147.9 s, 138.5 s. 117.7 s, 79.8 s, 41.9 s,
9.2 (t, J = 114.2 Hz). HREIMS: m/z 405.9150 (calcd for C H INO P , [M-H]+ 405.9106).
8
11
6 2
Bis-(2-(3,4-dichlorophenyl)ethylidene)-1,1-bisphosphonic acid (5). Compound 5 was prepared
from 3,4-dichlorobenzylchloride following Method D as colorless solid (tetra potassium salt, 1.1
1
g, 92%). H NMR (D
2
O, pH 12): δ 7.64 (s, 2H), 7.48 (d, J = 8.2 Hz, 2H), 7.33 (d, J = 8.2 Hz,
1
3
2
1
2
H), 3.02 (t, J = 15.4 Hz, 4H). C NMR (D O, pH 12, t-BuOD 30.3 ppm): δ 158.1 s, 139.0 s,
31
26.9 s, 125.2 s, 125.1 s, 36.8 (t, J = 112 Hz), 26.0 s. P NMR (D O, pH 12): δ 20.9 s.
PC
2
HREIMS: m/z 490.8942 (calcd for C15
H
13Cl
4
O
6
2
P , [M-H]+490.8942).

ACCEPTED MANUSCRIPT
Biology
The detergent-free recombinant IN protein (wt and mutant forms) was produced in
Escherichia coli and purified as previously described[51]. The plasmids containing the mutant
forms of the IN genes (E92Q, E138K, G140S, and N155H substitutions) were obtained by site-
directed mutagenesis of a plasmid encoding wild-type IN using the QuikChange II Site-Directed
Mutagenesis kit (Agilent Technologies, USA). All procedures were performed in accordance
with the manufacturer's instructions.
IN/LEDGF complex was produced as described in [40].
All oligonucleotides were synthesized by the phosphoramidite method on an automatic
ABI 3400 DNA synthesizer (Applied Biosystems, USA) under conditions recommended by
manufacturer and purified by electrophoresis in a 20% polyacrylamide/7 M urea gel.
3
2
P-labeled IN substrate preparation. DNA duplexes consisting of oligonucleotides U5A (5’-
ACTGCTAGAGATTTTCACAC-3’) and U5B (5’-GTGTGGAAAATCTCTAGCAGT-3’) or
U5B-2 (5’-GTGTGGAAAATCTCTAGCA-3’), and mimicking the end of the HIV-1 U5 LTR
were used as IN substrates. The U5B or U5B-2 oligonucleotide (10 pmol) was labeled with T4
32
polynucleotide kinase (Fermentas, Lithuania) and 50 ꢀCi [γ P]ATP (3000 Ci/mmol). After 1 h
of incubation at 37°C, the T4 polynucleotide kinase was inactivated by the addition of EDTA
followed by heating at 65°C for 5 minutes. The U5B or U5B-2 oligonucleotide was then
annealed with an equimolar amount of complementary oligonucleotide, U5A. The resulting
32
U5B/U5A or U5B-2/U5A duplex was then purified from unincorporated [γ- P]ATP by
centrifugation through MicroSpin G-25 Columns (GE Healthcare, UK).
32
Inhibition of 3'-processing and strand transfer reactions. P-labeled duplex U5B/U5A (3
nM) for 3’-processing or U5B-2/U5A (10 nM) for strand transfer was incubated with IN (100
2 2
nM) in 20 ꢀl of the buffer A (20 mM HEPES pH 7.2, 7.5 mM MgCl or MnCl , 1 mM DTT) in
the presence of increasing concentrations of an inhibitor at 37°C for 2 h. The reaction was
stopped by adding 80 ꢀl of stop solution (7 mM EDTA, 0.3 M sodium acetate, and 10 mM Tris–
HCl, pH 8). IN was extracted with 100 ꢀl of phenol–chloroform–isoamyl alcohol mixture
(25:24:1). DNA was precipitated with 250 ꢀl of ethanol at 0°C and assayed by denaturing PAGE
(20%). Gel was visualized in a STORM 840TM Phosphorimager (Molecular Dynamics, United
States). The reaction efficiency was assessed using the Image QuaNTTM 4.1 program.
Inhibition of 3' processing catalyzed by the complex IN/LEDGF was studied analogously using
1
00 mM IN combined with 50 nM LEDGF.

ACCEPTED MANUSCRIPT
3
2
For the inhibition mechanism studies, IN (100 nM) was incubated with P-labeled substrate
U5B/U5A (2.5–12.5 nM) and compound 2a (0–39 µM) or 2c (0–15 ꢀM), in 20 ꢀl of the buffer A
for 1 h at 37°C.
32
DNA binding assay. The P-labeled substrate U5B/U5A (3 nM) was incubated with HIV-1 IN
at different concentrations (3 – 300 nM) in 10 µl of the buffer A at 25°C for 20 min before
loading on nitrocellulose membrane (Amersham Hybond, GE Healthcare). Molecular Dynamics
STORM 840 Phosphorimager was used for analysis and quantification.
3
2
For kinetic measurements, the P-labeled substrate U5B/U5A (3 nM) was incubated with IN
alone (100 nM) or combined with LEDGF (50 nM) in 10 µl of the buffer A at 25°C, and the
aliquots were taken after different times of incubation (0-60 min).
To study the DNA-IN binding inhibition, compounds 2a-d or 5 (0-500 µM) were added to the
DNA/IN mixture.
Abbreviations:
HAART - highly active antiretroviral therapy, , IN – integrase, BP – bisphosphonate, INSTI –
integrase strand transfer inhibitor, LEDGF - lens epithelium derived growth factor, RT - reverse
transcriptase, VDPA – vinylidenediphosphonic acid. DRaCALA - differential radial capillary
action of ligand assay
ACKNOWLEDGEMENTS
We thank Dr. Janne Weisell from the University of Eastern Finland for providing HREIMS data
and Dr. Alex Khomutov from Engelhardt Institute of Molecular Biology RAS (Moscow, Russia)
for his thoughtful comments on the manuscript. The work was supported by the Russian
Foundation for the Basic Research (projects 13-04-91440, 11-04-01586, 12-04-00958) and by
the Presidium of the Russian Academy of Sciences (Programme “Molecular and Cellular
Biology”) and by the Academy of Finland (Grant No 132070), strategic funding of UEF and
PhoSciNet COST action.
Supporting Information Available:
1
13
31
H, C, P NMR data for all new and most other BPs. Detail synthetic procedures for
preparation some intermediate compounds. Supplementary data associated with this article can
be found in the online version, at http://dx.doi.org/

ACCEPTED MANUSCRIPT
REFERENCES
[
1] P.K. Quashie, R.D. Sloan, M.A. Wainberg, Novel therapeutic strategies targeting HIV
integrase, BMC Med, 10 (2012) 34.
2] C. Marchand, K. Maddali, M. Metifiot, Y. Pommier, HIV-1 IN inhibitors: 2010 update and
perspectives, Curr Top Med Chem, 9 (2009) 1016-1037.
[
[
[
[
3] FDA approves raltegravir tablets, AIDS Patient Care STDS, 21 (2007) 889.
4] FDA approves new 4-drug once-a-day HIV treatment, AIDS Policy Law, 27 (2012) 1.
5] E. Serrao, S. Odde, K. Ramkumar, N. Neamati, Raltegravir, elvitegravir, and metoogravir:
the birth of "me-too" HIV-1 integrase inhibitors, Retrovirology, 6 (2009) 25.
6] M. Metifiot, C. Marchand, K. Maddali, Y. Pommier, Resistance to integrase inhibitors,
Viruses, 2 (2010) 1347-1366.
7] M.A. Wainberg, T. Mesplede, P.K. Quashie, The development of novel HIV integrase
inhibitors and the problem of drug resistance, Curr Opin Virol, 2 (2012) 656-662.
8] S.L. Karmon, M. Markowitz, Next-generation integrase inhibitors : where to after
raltegravir?, Drugs, 73 (2013) 213-228.
9] U. Pradere, V. Roy, A. Montagu, O. Sari, M. Hamada, J. Balzarini, R. Snoeck, G. Andrei,
[
[
[
[
L.A. Agrofoglio, Synthesis and antiviral evaluation of bis(POM) prodrugs of (E)-[4'-phosphono-
but-2'-en-1'-yl]purine nucleosides, Eur J Med Chem, 57 (2012) 126-133.
[
10] D.G. Piotrowska, J. Balzarini, I.E. Glowacka, Design, synthesis, antiviral and cytostatic
evaluation of novel isoxazolidine nucleotide analogues with a 1,2,3-triazole linker, Eur J Med
Chem, 47 (2012) 501-509.
[
11] M. Hamada, V. Roy, T.R. McBrayer, T. Whitaker, C. Urbina-Blanco, S.P. Nolan, J.
Balzarini, R. Snoeck, G. Andrei, R.F. Schinazi, L.A. Agrofoglio, Synthesis and broad spectrum
antiviral evaluation of bis(POM) prodrugs of novel acyclic nucleosides, Eur J Med Chem, 67
(2013) 398-408.
[
12] L. Roux, S. Priet, N. Payrot, C. Weck, M. Fournier, F. Zoulim, J. Balzarini, B. Canard, K.
Alvarez, Ester prodrugs of acyclic nucleoside thiophosphonates compared to phosphonates:
synthesis, antiviral activity and decomposition study, Eur J Med Chem, 63 (2013) 869-881.
[
13] C. Cruchaga, E. Anso, A. Rouzaut, J.J. Martinez-Irujo, Selective excision of chain-
terminating nucleotides by HIV-1 reverse transcriptase with phosphonoformate as substrate, J
Biol Chem, 281 (2006) 27744-27752.
[
14] Y. Song, J.M. Chan, Z. Tovian, A. Secrest, E. Nagy, K. Krysiak, K. Bergan, M.A. Parniak,
E. Oldfield, Bisphosphonate inhibitors of ATP-mediated HIV-1 reverse transcriptase catalyzed
excision of chain-terminating 3'-azido, 3'-deoxythymidine: a QSAR investigation, Bioorg Med
Chem, 16 (2008) 8959-8967.
[
15] A. Anisenko, J. Agapkina, T. Zatsepin, D. Yanvarev, M. Gottikh, A new fluorometric assay
for the study of DNA-binding and 3'-processing activities of retroviral integrases and its use for
screening of HIV-1 integrase inhibitors, Biochimie, 94 (2012) 2382-2390.
[
16] M. Lecouvey, I. Mallard, T. Bailly, R. Burgada, Y. Leroux, A mild and efficient one-pot
synthesis of 1-hydroxymethylene-1,1-bisphosphonic acids. Preparation of new tripod ligands,
Tetrahedron Letters, 42 (2001) 8475-8478.
[
17] D.V. Yanvarev, A.N. Korovina, N.N. Usanov, S.N. Kochetkov, Non-hydrolysable
analogues of inorganic pyrophosphate as inhibitors of hepatitis C virus RNA-dependent RNA-
polymerase, Russ J Bioorg Chem, 38 (2012) 224-229.
[
18] M. Fukuda, Y. Okamoto, H. Sakurai, Nonequivalence between the geminal
dihydroxyphosphinyl groups in some 2-substituted 1-aminoethylidenedophosphonic acids,
Chemistry Letters, 6 (1977) 529-532.
[
19] J.H. No, F. de Macedo Dossin, Y. Zhang, Y.L. Liu, W. Zhu, X. Feng, J.A. Yoo, E. Lee, K.
Wang, R. Hui, L.H. Freitas-Junior, E. Oldfield, Lipophilic analogs of zoledronate and

ACCEPTED MANUSCRIPT
risedronate inhibit Plasmodium geranylgeranyl diphosphate synthase (GGPPS) and exhibit
potent antimalarial activity, Proc Natl Acad Sci U S A, 109 (2012) 4058-4063.
[
20] J.M. Sanders, Y. Song, J.M. Chan, Y. Zhang, S. Jennings, T. Kosztowski, S. Odeh, R.
Flessner, C. Schwerdtfeger, E. Kotsikorou, G.A. Meints, A.O. Gomez, D. Gonzalez-
Pacanowska, A.M. Raker, H. Wang, E.R. van Beek, S.E. Papapoulos, C.T. Morita, E. Oldfield,
Pyridinium-1-yl bisphosphonates are potent inhibitors of farnesyl diphosphate synthase and bone
resorption, J Med Chem, 48 (2005) 2957-2963.
[
21] Y. Zhang, A. Leon, Y. Song, D. Studer, C. Haase, L.A. Koscielski, E. Oldfield, Activity of
nitrogen-containing and non-nitrogen-containing bisphosphonates on tumor cell lines, J Med
Chem, 49 (2006) 5804-5814.
[
22] L. Deng, K. Endo, M. Kato, G. Cheng, S. Yajima, Y. Song, Structures of 1-Deoxy-D-
Xylulose-5-Phosphate Reductoisomerase/Lipophilic Phosphonate Complexes, ACS Med Chem
Lett, 2 (2011) 165-170.
[
23] G.R. Kieczykowski, R.B. Jobson, D.G. Melillo, D.F. Reinhold, V.J. Grenda, I. Shinkai,
Preparation of (4-Amino-1-Hydroxybutylidene)bisphosphonic Acid Sodium Salt, MK-217
Alendronate Sodium). An Improved Procedure for the Preparation of 1-Hydroxy-1,1-
bisphosphonic Acids, The Journal of Organic Chemistry, 60 (1995) 8310-8312.
24] M.S. Marma, Z. Xia, C. Stewart, F. Coxon, J.E. Dunford, R. Baron, B.A. Kashemirov, F.H.
(
[
Ebetino, J.T. Triffitt, R.G. Russell, C.E. McKenna, Synthesis and biological evaluation of alpha-
halogenated bisphosphonate and phosphonocarboxylate analogues of risedronate, J Med Chem,
5
0 (2007) 5967-5975.
[
[
25] M. Company, in, Great Britain, 1970, pp. 1-15.
26] S.P. Korolev, Y.Y. Agapkina, M.B. Gottikh, Clinical Use of Inhibitors of HIV-1
Integration: Problems and Prospects, Acta Naturae, 3 (2011) 12-28.
27] K. Ramkumar, K.V. Tambov, R. Gundla, A.V. Manaev, V. Yarovenko, V.F. Traven, N.
[
Neamati, Discovery of 3-acetyl-4-hydroxy-2-pyranone derivatives and their difluoridoborate
complexes as a novel class of HIV-1 integrase inhibitors, Bioorg Med Chem, 16 (2008) 8988-
8
998.
[
28] H. Sharma, S. Patil, T.W. Sanchez, N. Neamati, R.F. Schinazi, J.K. Buolamwini, Synthesis,
biological evaluation and 3D-QSAR studies of 3-keto salicylic acid chalcones and related amides
as novel HIV-1 integrase inhibitors, Bioorg Med Chem, 19 (2011) 2030-2045.
[
29] J. Agapkina, T. Zatsepin, E. Knyazhanskaya, J.-F. Mouscadet, M. Gottikh, Structure–
Activity Relationship Studies of HIV-1 Integrase Oligonucleotide Inhibitors, ACS Medicinal
Chemistry Letters, 2 (2011) 532-537.
[
30] S. Hare, S.S. Gupta, E. Valkov, A. Engelman, P. Cherepanov, Retroviral intasome assembly
and inhibition of DNA strand transfer, Nature, 464 (2010) 232-236.
31] L. Krishnan, X. Li, H.L. Naraharisetty, S. Hare, P. Cherepanov, A. Engelman, Structure-
based modeling of the functional HIV-1 intasome and its inhibition, Proc Natl Acad Sci U S A,
07 (2010) 15910-15915.
32] E. Matczak-Jon, V. Videnova-Adrabińska, Supramolecular chemistry and complexation
abilities of diphosphonic acids, Coordination Chemistry Reviews, 249 (2005) 2458-2488.
33] E. Matczak-Jon, B. Kurzak, A. Kamecka, P. Kafarski, Interactions of zinc(II),
[
1
[
[
magnesium(II) and calcium(II) with aminomethane-1,1-diphosphonic acids in aqueous solutions,
Polyhedron, 21 (2002) 321-332.
[
34] L.M. Salonen, M. Ellermann, F. Diederich, Aromatic rings in chemical and biological
recognition: energetics and structures, Angew Chem Int Ed Engl, 50 (2011) 4808-4842.
35] T. Appleton, J. Hall, A. Harris, H. Kimlin, I. McMahon, N.M.R. study of acid-base
[
+
-
equilibria of aminoalkylphosphonic acids, NH
cyclization in solution, Australian Journal of Chemistry, 37 (1984) 1833-1840.
36] C.T. Burt, H.M. Cheng, S. Gabel, R.E. London, ATP, ADP, and magnesium mixed
solutions: in vitro 31P NMR characterization and in vivo application to lens, J Biochem, 108
1990) 441-448.
3 2 n 3
(CH ) PO H (n = 1,2,3): Evidence for
[
(

ACCEPTED MANUSCRIPT
[
37] G.P. Donaldson, K.G. Roelofs, Y. Luo, H.O. Sintim, V.T. Lee, A rapid assay for affinity
and kinetics of molecular interactions with nucleic acids, Nucleic Acids Res, 40 (2012) e48.
38] M. Smolov, J. Agapkina, M. Buckle, J.F. Mouscadet, M. Gottikh, Characterization of DNA
binding by HIV-1 integrase using Schiff-base formation, FEBS Journal, 272 (2005) 19-20.
39] J. Agapkina, M. Smolov, S. Barbe, E. Zubin, T. Zatsepin, E. Deprez, M. Le Bret, J.F.
[
[
Mouscadet, M. Gottikh, Probing of HIV-1 integrase/DNA interactions using novel analogs of
viral DNA, J Biol Chem, 281 (2006) 11530-11540.
[
40] F. Michel, C. Crucifix, F. Granger, S. Eiler, J.F. Mouscadet, S. Korolev, J. Agapkina, R.
Ziganshin, M. Gottikh, A. Nazabal, S. Emiliani, R. Benarous, D. Moras, P. Schultz, M. Ruff,
Structural basis for HIV-1 DNA integration in the human genome, role of the LEDGF/P75
cofactor, EMBO J, 28 (2009) 980-991.
[
41] S.P. Korolev, V.N. Tashlitskii, M.A. Smolov, A.V. Gromyko, A.L. Zhuze, I. Agapkina,
M.B. Gottikh, [HIV-1 integrase inhibition by dimeric bisbenzimidazoles having different
spacers], Mol Biol (Mosk), 44 (2010) 718-727.
[
42] I. Malet, O. Delelis, M.A. Valantin, B. Montes, C. Soulie, M. Wirden, L. Tchertanov, G.
Peytavin, J. Reynes, J.F. Mouscadet, C. Katlama, V. Calvez, A.G. Marcelin, Mutations
associated with failure of raltegravir treatment affect integrase sensitivity to the inhibitor in vitro,
Antimicrob Agents Chemother, 52 (2008) 1351-1358.
[
43] F. Ceccherini-Silberstein, I. Malet, R. D'Arrigo, A. Antinori, A.G. Marcelin, C.F. Perno,
Characterization and structural analysis of HIV-1 integrase conservation, AIDS Rev, 11 (2009)
7-29.
44] M.M. Prokofjeva, P.V. Spirin, D.V. Yanvarev, A.V. Ivanov, M.S. Novikov, O.A. Stepanov,
1
[
M.B. Gottikh, S.N. Kochetkov, B. Fehse, C. Stocking, V.S. Prassolov, Screening of Potential
HIV-1 Inhibitors/Replication Blockers Using Secure Lentiviral in Vitro System, Acta Naturae, 3
(2011) 55-65.
[
45] P. Politzer, P. Lane, M.C. Concha, Y. Ma, J.S. Murray, An overview of halogen bonding, J
Mol Model, 13 (2007) 305-311.
46] P. Metrangolo, G. Resnati, H.H.D. Arman, Halogen Bonding: Fundamentals and
Applications, Springer, 2008.
47] L.A. Hardegger, B. Kuhn, B. Spinnler, L. Anselm, R. Ecabert, M. Stihle, B. Gsell, R.
[
[
Thoma, J. Diez, J. Benz, J.M. Plancher, G. Hartmann, D.W. Banner, W. Haap, F. Diederich,
Systematic investigation of halogen bonding in protein-ligand interactions, Angew Chem Int Ed
Engl, 50 (2011) 314-318.
[
48] L.A. Hardegger, B. Kuhn, B. Spinnler, L. Anselm, R. Ecabert, M. Stihle, B. Gsell, R.
Thoma, J. Diez, J. Benz, J.M. Plancher, G. Hartmann, Y. Isshiki, K. Morikami, N. Shimma, W.
Haap, D.W. Banner, F. Diederich, Halogen bonding at the active sites of human cathepsin L and
MEK1 kinase: efficient interactions in different environments, ChemMedChem, 6 (2011) 2048-
2
[
054.
49] R. Wilcken, M.O. Zimmermann, A. Lange, A.C. Joerger, F.M. Boeckler, Principles and
Applications of Halogen Bonding in Medicinal Chemistry and Chemical Biology, J Med Chem,
(
[
(
[
2013).
50] W.K. Hagmann, The many roles for fluorine in medicinal chemistry, J Med Chem, 51
2008) 4359-4369.
51] H. Leh, P. Brodin, J. Bischerour, E. Deprez, P. Tauc, J.C. Brochon, E. LeCam, D. Coulaud,
C. Auclair, J.F. Mouscadet, Determinants of Mg2+-dependent activities of recombinant human
immunodeficiency virus type 1 integrase, Biochemistry, 39 (2000) 9285-9294.

ACCEPTED MANUSCRIPT
Figure legends
Figure 1. The general structure of the methylenebisphosphonates under study (A) and structures
of the compounds synthesized (B). First prepared compounds are in italic
31
Figure 2. P NMR spectra of compounds 2a and 2c. (А) The pH dependence of the chemical
shifts of phosphorus atoms. (В) The variations of the phosphorus atom resonances at various pH
2
+
2+
in the presence of Mg ions. The sample contained both BP at a ratio of 1:2 and Mg at a molar
concentration twice as high as that of both BPs.
Figure 3. The dependence of the 3’-processing initial rate vs the DNA substrate concentration at
varied inhibitor concentrations on double-reciprocal plots. (A) Compound 2a in the presence of
2
+
2+
Mg ions; (B) Compound 2c in the presence of Mg ions; (C) Compound 2a in the presence of
2
+
2+
Mn ions; (D) Compound 2c in the presence of Mn ions.
Figure 4. Binding of the DNA substrate to IN alone and IN-LEDGF complex. А. The
dependency of the binding efficacy vs IN concentration ([INeq]=[IN] for IN and [INeq] =
4
*[IN/LEDGF] for the complex) В. The kinetics of the DNA substrate binding to IN alone and
to the IN-LEDGF complex (IN = 100 nM). The binding kinetics at initial time points are shown
eq
at the insert.

ACCEPTED MANUSCRIPT
Table 1. Inhibition of the DNA-binding and catalytic activities of HIV-1 IN by BPs in the
Inhibitory effect of the tested BPs, IC50 (µM)
IN activity
Activity of the
IN-LEDGF
complex in 3’-
processing
Compound
Binding
3’-processing
ST
2
+
2+
2+
2+
2+
2+
2+
2+
Mg
Mn
Mg
Mn
4
Mg
Mn
6
Mg
Mn
1
2
3
5
7
8
2
2
a
150
70
>500
>500
>500
>500
>500
35
20
7
5
45
42
3
4
>250
>250
7
4.5
8
b
5
6
2
c
>256
130
100
2
2
2
2
d
30
20
5
30
16
5
>250
nd*
6
5
0.8
0.8
nd
presence of different metal cofactors.
*
- not determined

ACCEPTED MANUSCRIPT
Figure 1
HO
O
OH
HO
OH
O
A
X
P
P
R
R'
B
O
O
Cl
P(OH)2
X
n
P(OH)2
X
P(OH)2
O
P(OH)2
O
1
1
1
1
a, n=1, X=OH
b, n=2, X=OH
c, n=4, X=OH
d, n=1, X=NH2
2
a, R’=Cl, X=OH
b, R’=H, X=OH
c, R’=Cl, X=NH2
d, R’=Cl, X=H
2
2
2
O
O
R'
P(OH)2
P(OH)2
X
X
+
N
NH
n
P(OH)2
O
P(OH)2
O
R’
4
4
4
a, n=2, R’=H, X=H
b, n=0, R’=H, X=H
c, n=0, R’=I, X=H
3
3
3
3
3
a, R’=H, X=OH
b, R’=H, X=NH2
c, R’=p-Ph, X=OH
d, R’=m-OH, X=OH
e, R’=o-Et, X=OH
Cl
Cl
O
P(OH)2
Cl
P(OH)2
O
Cl
5
6
7

ACCEPTED MANUSCRIPT
Figure 2.

ACCEPTED MANUSCRIPT
Figure 3

ACCEPTED MANUSCRIPT
Figure 4