Synthesis and Evaluation of Asymmetric Acyclic Nucleoside Bisphosphonates as Inhibitors of Plasmodium falciparum and Human Hypoxanthine-Guanine-(Xanthine) Phosphoribosyltransferase

Article abstract of DOI:10.1021/acs.jmedchem.7b00926

Acyclic nucleoside bisphosphonates (ANbPs) have previously been shown to be good inhibitors of human hypoxanthine-guanine phosphoribosyltransferase (HGPRT) and Plasmodium falciparum (Pf) hypoxanthine-guanine-xanthine phosphoribosyltransferase (PfHGXPRT). On the basis of this scaffold, a new series of ANbPs was synthesized. One of these new ANbPs, [3-(guanine-9-yl)-2-((2-phosphonoethoxy)methyl)propoxy]methylphosphonic acid, exhibited K_i values of 6 and 70 nM for human HGPRT and Pf HGXPRT, respectively. These low K_i values were achieved by inserting an extra carbon atom in the linker connecting the N⁹ atom of guanine to one of the phosphonate groups. The crystal structure of this ANbP in complex with human HGPRT was determined at 2.0 ? resolution and shows that it fills three key pockets in the active site. The most potent phosphoramidate prodrugs of these compounds have IC₅₀ values in the low micromolar range in Pf lines and low toxicity in human A549 cells, demonstrating that these ANbPs are excellent antimalarial drug leads.

Full text of DOI:10.1021/acs.jmedchem.7b00926

Subscriber access provided by UNIVERSITY OF ADELAIDE LIBRARIES
Article
Synthesis and Evaluation of Asymmetric Acyclic Nucleoside
Bisphosphonates as Inhibitors of Plasmodium falciparum and
Human Hypoxanthine-Guanine-(Xanthine) Phosphoribosyltransferase
Petr Špa#ek, Dianne T. Keough, Marina Chavchich, Martin Dracinsky, Zlatko
Janeba, Lieve Naesens, Michael D. Edstein, Luke W Guddat, and Dana Hocková
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.7b00926 • Publication Date (Web): 16 Aug 2017
Downloaded from http://pubs.acs.org on August 16, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 57
Journal of Medicinal Chemistry
1
1
2
3
4
5
6
7
8
Synthesis and Evaluation of Asymmetric Acyclic Nucleoside Bisphosphonates as
Inhibitors of Plasmodium falciparum and Human Hypoxanthine-Guanine-(Xanthine)
Phosphoribosyltransferase
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Petr Špaček,a Dianne T. Keough,b Marina Chavchich,c Martin Dračínský,a Zlatko Janeba,a
Lieve Naesens,d Michael D. Edstein,c Luke W. Guddat,b* and Dana Hocková,a*
aThe Institute of Organic Chemistry and Biochemistry of the Czech Academy of Sciences,
Flemingovo nám. 2, CZꢀ16610 Prague 6, Czech Republic
bSchool of Chemistry and Molecular Biosciences, The University of Queensland, Brisbane,
Queensland, 4068, Australia
cDepartment of Drug Evaluation, Australian Army Malaria Institute, Enoggera, Brisbane,
Queensland 4051, Australia
dKU Leuven ꢀ Rega Institute for Medical Research, Laboratory of Virology and chemotherapy
Herestraat 49, Bꢀ3000 Leuven, Belgium
Keywords: acyclic nucleoside phosphonates; phosphoramidate prodrugs; hypoxanthineꢀ
guanineꢀ(xanthine) phosphoribosyltransferase, purine salvage pathway, malaria
Abstract
Acyclic nucleoside bisphosphonates (ANbPs) have previously been shown to be good
inhibitors of human hypoxanthineꢀguanine phosphoribosyltransferase (HGPRT) and
Plasmodium falciparum (Pf) hypoxanthineꢀguanineꢀxanthine phosphoribosyltransferase
(PfHGXPRT). Based on this scaffold, a new series of ANbPs was synthesized. One of these
new ANbPs, [3ꢀ(guanineꢀ9ꢀyl)ꢀ2ꢀ((2ꢀphosphonoethoxy)methyl)propoxy]methylphosphonic
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 57
2
1
2
3
4
5
6
7
8
9
acid, exhibited Ki values of 6 and 70 nM for human HGPRT and Pf HGXPRT, respectively.
These low Ki values were achieved by inserting an extra carbon atom in the linker connecting
the N9 atom of guanine to one of the phosphonate groups. The crystal structure of this ANbP
in complex with human HGPRT was determined at 2.0 Å resolution and shows that it fills
three key pockets in the active site. The most potent phosphoramidate prodrugs of these
compounds have IC50 values in low micromolar range in Pf lines and low toxicity in human
A549 cells demonstrating that these ANbPs are excellent antimalarial drug leads.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
INTRODUCTION
Malaria remains as one of the most lethal infectious diseases known to afflict mankind with
429,000 deaths and 212 million infections in 2015.1 The causative agent of malaria is a
protozoan parasite that belongs to Plasmodium genus and is spread by the bite of female
Anopheles mosquito. Of the five Plasmodium species that can infect humans, Plasmodium
falciparum (Pf) and Plasmodium vivax (Pv) are the most lethal and widespread.2 Artemisinin
based combination therapies (ACTs) are used worldwide for the treatment of malaria, but
alarmingly ACT resistance has emerged in Southeast Asia with failure rates of 50% and 26%
in Cambodia3 and Vietnam4, with the use of dihydroartemisininꢀpiperaquine, respectively. On
the basis that malarial infections remain as major cause of morbidity and mortality, the need
to discover new antimalarial drug targets and new therapeutic agents with novel mechanisms
is of the utmost importance.
Hypoxanthineꢀguanineꢀxanthine phosphoribosyltransferase (HGXPRT) is an attractive drug
target for antimalarial drug therapy, as its activity is essential for both the survival and
reproduction of malarial parasites.5 This is because it provides the only pathway for the
synthesis of the 6ꢀoxopurine nucleoside monophosphates required for DNA and RNA
production (salvage pathway). HGXPRT is also indirectly responsible for the synthesis of the
ACS Paragon Plus Environment

Page 3 of 57
Journal of Medicinal Chemistry
3
1
2
3
4
5
6
7
8
9
6ꢀamino nucleoside monophosphate, AMP, as this is produced in these parasites starting from
IMP. The nucleotides, IMP, GMP and XMP, are therefore synthesized from hypoxanthine,
guanine and xanthine which are actively transported from their host cell. Though humans
possess both a de novo pathway starting from simple precursors as well as the salvage
pathway, the malarial parasites rely solely on the salvage pathway. Independent studies have
shown that both purine starvation by inhibition of purine nucleoside phosphorylase5 or
hypoxanthine depletion6 have been effective routes to preventing the growth of malarial
parasites, either in cell culture or in Aotus monkeys. These data provide further confirmation
that purine metabolism is a major flaw in the parasitic armory that can be exploited for
effective antimalarial drug discovery.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
HG(X)PRTs catalyse the formation of the 6ꢀoxopurine nucleoside monophosphates
from a nucleobase and from 5′ꢀphosphoꢀαꢀDꢀribosylꢀ1ꢀpyrophosphate (Fig. 1). The human
and Pv enzymes use only hypoxanthine and guanine as substrates, whereas the Pf enzyme
uses all three bases as substrates, hence the assignment of the enzymes as human HGPRT,
PvHGPRT and PfHGXPRT. The presence of divalent metal ions, presumably magnesium in
vivo, is essential for the reaction to proceed. Despite the fact that there is some redundancy in
the salvage pathway (e.g. IMP dehydrogenase and GMP synthetase are present in the Pf
genome), direct inhibition of HGXPRT7,8 has been shown to be an efficient strategy for the
development of novel lead antiparasitic and antibacterial agents.9ꢀ15
O
O
N
N
HN
HN
N
R
O
N
N
H
R
N
Mg2+
-O P O
O-
O
+
2-O3PO
HO OH
+
O
O
O
-O
O-
O-
3-
OPO3PO3
P O P
-O
HO OH
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 57
4
1
2
3
Figure 1. Reaction catalyzed by HGPRT/HGXPRT. R = H (hypoxanthine), R = NH2
4
5
6
(guanine), R = OH (xanthine).
7
8
9
The chemical structures of bisphosphonate HGPRT/HGXPRT inhibitors16 previously
characterized in a SARꢀstudy are given in Figs. 2A and 2B. The basic structure of the newly
synthesized ANbPs is shown in Fig. 2C.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. General structures of ANbPꢀbased HGPRT/HGXPRT inhibitors: A) ꢀ B)
previously published bisphosphonates16; C) new asymmetric inhibitors.
The design of new inhibitors presented in this paper (Fig. 2C) is based on alterations
to the scaffold of published acyclic nucleoside bisphosphonates (ANbPs, Fig. 2AB). These
ANbPs have a 6ꢀoxopurine base attached either to a symmetric moiety bearing two
phosphonomethyloxy groups (Fig. 2A) or to an asymmetric moiety with two
phosphonoethyloxy groups (Fig. 2B).16 The rationale for the design of new asymmetric
ANbPs (Fig. 2C) is based on the crystal structure of human HGPRT in complex with 1a (Fig.
2A and Fig. 3).16 In new series of compounds, five atoms17 are contained in the linker
connecting the N9 atom of the purine base with one of the phosphonate groups. The crystal
structure of human HGPRT in complex with 1a16 shows that this particular phosphonate
group is located in the 5′ꢀphosphate binding pocket (Fig. 3). The second phosphonate group
ACS Paragon Plus Environment

Page 5 of 57
Journal of Medicinal Chemistry
5
1
2
3
4
5
6
7
8
9
points down towards the PPi binding pocket but this moiety is not located in this pocket.
Instead, a sulphate ion, from the crystallization buffer, was found to occupy this site which is
also the location where the second product of the reaction, PPi, binds. The crystal structure
shows that a magnesium ion is coordinated to one of the sulphate oxygen atoms and one of
the phosphonate oxygen atoms of 1a (Fig. 3).16
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Purine binding pocket
5′‑phosphate
binding
pocket
Sulfate
Figure 3. [(2ꢀ[(guaninꢀ9ꢀyl)methyl]propaneꢀ1,3ꢀdiyl)bis(oxy)]bis(methylene)diphosphonic
acid (1a) bound in the active site of human HGPRT (Protein Data Bank coordinates 4IJQ).16
A sulphate ion from the crystallization media is present in the complex along with two
magnesium ions (purple spheres).
Molecular docking suggested that lengthening one of the linkers by one atom may
allow the second phosphonate group to be located in the PPi binding pocket leading to
decreased Ki values (Fig. 4 yellow structure). In addition, the chemical nature of the purine
base was altered to examine the effects of such changes on the inhibition constants for human
HGPRT and PfHGXPRT.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 57
6
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. Comparison of the crystal structure of the human HGPRT. 1a complex16 with the
docked structure of the new guanine derivative, 20 (Scheme 2), overlaid.
To test the antimalarial activity of the new inhibitors in erythrocyte cell culture infected
by Pf, a prodrug approach was employed.18 The highly polar bisphosphonates are unable to
cross the cell membrane unless the phosphonate groups are masked by moieties which are
cleaved enzymatically once the intact prodrug enters the cells. Phosphoramidate prodrugs
bearing four nonꢀtoxic ethyl ester of Lꢀphenylalanine moieties14,16,19 were prepared, their IC50
values measured against Pf in cell culture and their cytotoxicity in mammalian cells in tissue
culture determined.
RESULTS AND DISCUSSION
2.1. Chemistry
The synthesis of this series of asymmetric ANbPs inhibitors can be described in three
consecutive parts. The first part was to prepare the key hydroxy derivative 8. This precursor
contains two different functional groups – phosphonomethyl and phosphonoethyl – attached
ACS Paragon Plus Environment

Page 7 of 57
Journal of Medicinal Chemistry
7
1
2
3
4
5
6
7
8
9
via ether bonds to the main scaffold (Scheme 1). The second part of the synthesis was the
Mitsunobu reaction affording the connection of this bisphosphonate moiety to Nꢀ9 position of
a purine base. The last part was a sequence of reactions transforming the preformed
derivatives to the final inhibitors (Scheme 2).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Starting material 3 (5ꢀhydroxymethylꢀ2,2ꢀdimethylꢀ1,3ꢀdioxane) was prepared by
simultaneous protection of two hydroxy groups of commercially available 2ꢀhydroxymethylꢀ
1,3ꢀpropanediol using reaction with dry acetone and acid catalyst.20 Obtained alcohol 3 was
transformed to a phosphonomethyloxy derivative 4 by the reaction with diisopropyl
(bromomethyl)phosphonate and sodium hydride in DMF. Intermediate 4 was treated with
DOWEX (50WX8 H+) in methanol to obtain dihydroxy derivative 5, which was a key
intermediate for the subsequent synthetic pathways (Scheme 1).
Scheme 1: a) BrCH2P(O)(OiPr)2, NaH, DMF, 0°C; b) DOWEX (50WX8 H+), 90% MeOH,
50°C; c) TBDMSCl, imidazole, DMF, r.t.; d) Diethyl vinylphosphonate, Cs2CO3, tꢀBuOH,
r.t.; e) DOWEX (50WX8 H+), 90% MeOH, 50°C.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 57
8
1
2
3
4
5
6
7
8
9
In the course of the first attempts to attach the second phosphonate group, dihydroxy
derivative 5 was treated with various molar equivalents of diethyl vinylphosphonate under
oxaꢀMichael reaction conditions.21 However, required alcohol 8 was not obtained in a
sufficient quantity. Depending on the reaction conditions (using various bases and differing
amounts of diethyl vinylphosphonate), we observed either product 9 with two attached
phosphonoethoxy groups or the starting material 5 with a very small amount of desired
product 8. Therefore, we tried to protect one of the hydroxy groups of intermediate 5 using
standard protecting groups. This would subsequently enable a selective introduction of only
one phosphonoethoxy group. Experiments with several alkylating agents (benzyl halogenides
or trityl chloride) under various reaction conditions (low temperature, various dry solvents,
different bases) were carried out. Unfortunately, none of these methods to synthesize
derivatives 6b and 6c were successful on a preparative scale.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Another approach was to prepare a fully protected cyclic carbonate diester derivative
from dihydroxy phosphonate 5 using ethyl chloroformate in THF with the addition of
triethylamine, and subsequently cleave this compound by an acidꢀcatalysed reaction with
silica gel and corresponding alcohol to form carbonate esters 6d and 6e. Isopropyl alcohol and
benzyl alcohol were chosen to diversify possible methods for later deprotection. Attempts
were then made to attach a phosphonoethoxy functional group to intermediates 6d or 6e.
Despite multiple attempts with reaction conditions of oxaꢀMichael addition, phosphonoethoxy
derivatives 7b and 7c could not be obtained.
ACS Paragon Plus Environment

Page 9 of 57
Journal of Medicinal Chemistry
9
1
2
3
4
5
6
7
8
9
Cl
N
Cl
N
N
N
O
N
N
O
O
P OiPr
a)
a)
N
R
R
O
P
O
P
OiPr
O
P
OiPr
O
P
OiPr
O
P
OiPr
OiPr
OH
O
O
OEt
(69%)
(82%)
(74%)
(29%)
12 R = NH2
13 R = H
10 R = NH2
11 R = H
OEt
8
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
OEt
O
OEt
OEt
OEt
b)
c)
O
O
N
O
N
N
N
HN
HN
d)
HN
Br
N
N
R
N
R
R
N
O
P
O
P
O
P
O
P
OiPr
O
P
OiPr
O
P
OiPr
OiPr
OiPr
OiPr
O
O
O
O
(23%)
(43%)
(45%)
(27%)
(99%)
(65%)
18
R = NH2
16 R = NH2
17 R = H
14
R = NH2
19 R = H
15 R = H
O
OEt
OEt
O
OEt
OEt
OEt
OEt
e)
e)
e)
O
O
O
N
N
N
N
HN
HN
HN
R
Br
N
N
R
N
R
N
O
P
O
P
O
P
O- Na+
O- Na+
O
O- Na+
O
O- Na+
O
P
O- Na+
O- Na+
O
O
O
O- Na+
O- Na+
O- Na+
22 R = NH2 (10%)
20 R = NH2 (30%)
O
24 R = NH2 (35%)
P
O- Na+
O- Na+
O
O
P
O- Na+
25
23
21
R = H (39%)
R = H (31%)
R = H (6%)
Scheme 2: a) DIAD, PPh3, 6ꢀchloropurine, 2ꢀaminoꢀ6ꢀchloropurine or corresponding 7ꢀ
deazapurines, THF, ꢀ30°C; b) CF3COOH, H2O, r.t.; c) K2CO3, DABCO, H2O, 120°C; d) Br2,
CCl4, DMF, r.t.; e) BrSi(CH3)3, CH3CN, r.t.
Ultimately, a successful synthetic pathway was developed (Scheme 1). This was via a
tertꢀbutyldimethylsilyl protected intermediate 6a, which was prepared by the reaction of a 1.1
molar equivalent of tertꢀbutyldimethylsilyl chloride (TBDMSCl) and imidazole as a base in
THF.22 Protected alcohol 6a further reacted with diethyl vinylphosphonate under standard
reaction conditions of oxaꢀMichael addition to form bisphosphonate 7a. Intermediate 7a was
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 57
10
1
2
3
4
5
treated with DOWEX (50WX8 H+) to obtain the final hydroxy derivative 8, a key precursor
for the synthesis of the inhibitors and their prodrugs.
6
7
8
9
Further process leading to the synthesis of asymmetrical ANbPs was based on parallel
reaction procedures (Scheme 2). First, bisphosphonate moiety of alcohol 8 was introduced
under Mitsunobu reaction conditions to the Nꢀ9 position of 2ꢀaminoꢀ6ꢀchloropurine, 6ꢀ
chloropurine, 2ꢀaminoꢀ6ꢀchloroꢀ7ꢀdeazapurine and 6ꢀchloroꢀ7ꢀdeazapurine to form
derivatives 10, 11, 12, 13. Transformation to the corresponding 6ꢀoxopurine and 6ꢀoxoꢀ7ꢀ
deazapurine analogues followed. 6ꢀChloropurine derivatives 10 and 11 were treated using
standard conditions with a mixture of trifluoroacetic acid and water to obtain intermediates 14
and 15.16 For intermediates 12 and 13 with 7ꢀdeazaꢀmodified heterocyclic bases a more
vigorous method had to be applied. Thus, 7ꢀdeazapurine compounds 12 and 13 were
hydrolysed by the reaction with potassium carbonate and DABCO in water17,23 to form 6ꢀoxoꢀ
7ꢀdeazapurine derivatives 18 and 19.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Previously, an activity of several 8ꢀbromoguanine derivatives was observed.24
Therefore we also decided to prepare bromo derivatives for testing. Thus, intermediates 14
and 15 were used for preparation of bromo derivatives 16 and 17 by bromination with
elementary bromine in CCl4 and DMF. Synthesis of bromo derivative 27 started from a
previously published16 diester of bisphosphonate 2c
, which was transformed to the
corresponding 8ꢀbromoguanine derivative 26 by the above mentioned bromination method
(Scheme 3).
ACS Paragon Plus Environment

Page 11 of 57
Journal of Medicinal Chemistry
11
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 3: a) Br2, CCl4, DMF, r.t.; b) BrSi(CH3)3, CH3CN, r.t.
The last step for the preparation of the target ANbPs was the cleavage of phosphonate
ester groups.25 Esters 14, 15, 16, 17, 18, 19 and 26 were treated with bromotrimethylsilane in
acetonitrile followed by hydrolysis to obtain free phosphonic acids 20, 21, 22, 23, 24, 25 and
27.
To improve cell penetration properties of the ANbP inhibitors, phosphoramidate
prodrugs were prepared from corresponding esters (Scheme 4). Phosphonate ester precursors
14, 15, 18 and 19 were transformed to the corresponding prodrugs 28, 29, 30 and 31 by
treatment with bromotrimethylsilane in acetonitrile, followed by the reaction of thus
preformed silyl esters with Lꢀphenylalanine ethyl ester, 2,2′ꢀdipyridyl disulfide and
triphenylphosphine in dry pyridine.16,26
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 57
12
1
2
3
4
5
6
7
8
9
O
O
N
O
N
Ph
Ph
Z
N
N
Z
a)
b)
HN
HN
HN
Br
N
N
R
N
R
R
COOEt
Ph
COOEt
O
P
O
P
O
P
O
P
NH
O
P
NH
O
P
OiPr
HN
HN
OiPr
O
O
O
Ph
COOEt
COOEt
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
OEt
HN
O
HN
O
Ph
Ph
OEt
HN
COOEt
HN
COOEt
COOEt
COOEt
Ph
Ph
R = NH2 ; Z = N
R = H ; Z = N
14
15
(62%)
(25%)
(23%)
(12%)
R = NH2
R = NH2 ; Z = N
28
32
33
29
30
31
(15%)
R = H ; Z = N
R = H
18 R = NH2 ; Z = CH
R = NH2 ; Z = CH
19
R = H ; Z = CH
(11%)
R = H ; Z = CH
O
N
O
O
a)
b)
Ph
N
N
Ph
N
N
HN
HN
HN
Br
N
N
N
H2N
H2N
H2N
N
COOEt
Ph
O
O
COOEt
Ph
O
O
O
O
O
P
NH
P
OEt
P
NH
HN
OEt
HN
O
O
COOEt
COOEt
HN
P
O
O
P
OEt
HN
P
O
Ph
Ph
HN
COOEt
OEt
HN
COOEt
COOEt
(13%)
COOEt
35
2c
34
Ph
Ph
Scheme 4: a) 1. BrSi(CH3)3, CH3CN, r.t, 2. PPh3, Aldrithiol, (L)ꢀNH2CH(Bn)COOEt.HCl,
Et3N, pyridine, 70°C; b) NBS, 1,4ꢀdioxane, MW, 100°C.
Direct preparation of phosphoramidate prodrugs of 8ꢀbromo ANbPs from the
corresponding esters is not possible.20 Therefore direct bromination of the final prodrugs with
NBS and 1,4ꢀdioxane in a microwave reactor was used. Using this approach, prodrugs 28, 29
and previously published16 prodrug 34 were transformed to corresponding 8ꢀbromopurine
prodrugs 32, 33 and 35 (Scheme 4).
2.2. Inhibition of human HGPRT and Pf HGXPRT
ACS Paragon Plus Environment

Page 13 of 57
Journal of Medicinal Chemistry
13
1
2
3
4
5
6
7
8
The Ki values of the newly synthesized ANbPs for human HGPRT and PfHGXPRT are given
in Table 1 together for the data obtained with the previously published compounds (1a, 1b
and 2a; Figs. 2A and 2B)16 to allow a direct comparison.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 1: Summary of Ki values for ANbPs with human HGPRT and Pf HGXPRT.
human HGPRT
PfHGXPRT
Base
G
Acyclic moiety
No.
1a
Ki [ꢀM]
Ki [ꢀM]
Series 1
0.03 ± 0.002a
0.07 ± 0.01a
Hx
1 ± 0.1a
5 ± 1a
1b
G
0.6 ± 0.02a
0.7 ± 0.1
0.5 ± 0.01a
NIb
2a
2b
Series 2
Hx
8ꢀBrꢀG
0.4±0.2
0.9±0.3
27
G
0.006 ± 0.001
1.8±0.6
0.07±0.05
3±1
20
21
22
23
24
25
Series 3
Hx
8ꢀBrꢀG
0.1 ± 0.03
2.5 ± 1
2±1
8ꢀBrꢀHx
7ꢀdeazaꢀG
7ꢀdeazaꢀHx
NIc
0.1±0.5
4±2
4.6 ± 2
6±3
aData from Ref. 16.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 57
14
1
2
3
b,cNo inhibition for 2b and 23 was observed when these compounds were tested at a
4
5
6
concentration of 33 ꢀM and 50 ꢀM, respectively.
7
8
9
Compound 1a is a potent inhibitor of both the human and Pf enzymes (Table 1).16
Surprisingly, changing the attached purine base from guanine to hypoxanthine resulted in a
large increase in Ki values (33ꢀfold for human HGPRT and 71ꢀfold for PfHGXPRT). This is
partly understandable for the human enzyme as guanine has a twoꢀfold lower Km for guanine
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
compared with hypoxanthine (1.9 ꢀM for guanine compared with 3.4 ꢀM for hypoxanthine).
The Pf enzyme, on the other hand, has a lower Km value for hypoxanthine compared with
M for hypoxanthine).24 The significant
guanine (0.8 ꢀM for guanine compared with 0.07 ꢀ
increase in the Ki value due to the change in the base from guanine to hypoxanthine is evident
when the Ki values of 20 and 21 are compared resulting in a 300ꢀfold increase for human
HGPRT and a 43ꢀfold increase for PfHGXPRT. Thus, for these two ANbP scaffolds, the
nature of the purine base, whether hypoxanthine or guanine, has a significant influence on the
Ki values.
In Series 2, substitution of guanine by hypoxanthine had little effect on the Ki
values for the human enzyme (0.4ꢀ0.7 ꢀM) (Table 1) though, once again, a large increase was
observed for PfHGXPRT. In this series, there was no change in the Ki value when guanine
was replaced by 8ꢀbromoguanine for either enzyme. The only difference in the structures in
the ANbPs in Series 2 compared with Series 3 is the position of the oxygen atom in the linker
(Table 1). Thus, the higher Ki values for the guanine containing compounds in Series 2
compared with Series 3 is attributed to the change in position of this oxygen atom. This atom
is located in the linker connecting the N9 atom of the base with the phosphonate group
predicted to bind in the 5′ꢀphosphate binding pocket. This prediction is based on the crystal
structure of the human HGPRT.1a complex (Fig. 3).16 The location of the oxygen atom can
ACS Paragon Plus Environment

Page 15 of 57
Journal of Medicinal Chemistry
15
1
2
3
4
5
6
7
8
influence the conformation of the linker and, thus, the location of the phosphonate group and
the purine base in the active site. Therefore, it is clear that subtle chemical modifications can
have a large effect on the orientation of the ANbPs in the active site.
9
In Series 3, the attachment of 8ꢀbromoꢀ or 7ꢀdeaza bases, either guanine or
hypoxanthine derivatives, resulted in weak inhibition of both enzymes. Both these bases are
very weak substrates and this, obviously, influences the affinity of their ANbP. The crystal
structure of 20 (Series 3) in complex with human HGPRT was obtained and compared with
that of the human HGPRT.1a structure to explain the decrease in the Ki values (cf. Series 3
and Series 1). This structure also provides an entry into the design of next generation
compounds based on the structure and activity of these ANbPs.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2.3. The crystal structure of the human HGPRT.20 complex.
To explain the potent binding of 20, the crystal structure of the human HGPRT.20 was
determined to 2.03 Å resolution. The refinement statistics of this structure are given in Table
2. The crystals contains a tetramer in the asymmetric unit (Table 2) and 20 is bound in the
active site of each subunit (Figure 5). Figure 5 shows that the large mobile loop of ~20
amino acid residues is disordered. This compares with its structure when ImmGP27/ImmHP8
/acyclic immucillin phosphonate (AIP)10, PPi and Mg2+ are bound to human HGPRT or to
PfHGXPRT where this region becomes ordered and closes over the active site. The
immucillins are compounds that are inert mimics for the predicted transition state of the
reaction (Fig. 1). When AIP binds, the large mobile loop becomes ordered and covers the
active site of PfHGXPRT. The ordering is due to the formation of hydrogen bonds between
(i) the backbone amide S115 (Pf numbering) and a pyrophosphate oxygen, (ii) the backbone
amide of Y116 (Pf numbering) and the same pyrophosphate oxygen and (iii) the hydroxyl of
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 57
16
1
2
3
Y116 and a phosphonate oxygen of AIP. Similar sets of stabilizing interactions have also been
4
5
6
observed when ImmGP or ImmHP bind to human HGPRT or PfHGXPRT.8,27
7
8
9
Subunit A
Subunit B
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Subunit C
Subunit D
Figure 5. Crystal structure of the tetrameric human HGPRT.20 complex. Magenta highlights
the location of the ordered regions adjacent to the large mobile loop, residues R100ꢀD120,
which are not visible in this complex.
Table 2. Data collection and refinement statistics for the human HGPRT.20 complex.
Crystal parameters
a, b, c (Å)
56.67, 128.40, 65.33
α, β, γ (°)
90.0, 103.0, 90.0
P 21
Space group
Crystal size (mm)
Resolution range (Å)
Observations
0.3 x 0.1 x 0.05
47.30 ꢀ 2.03
438 819 (31 172)
58 747 (4 242)
99.7 (98.1)
Unique reflections
Completeness (%)
bRmerge
0.081 (0.654)
ACS Paragon Plus Environment

Page 17 of 57
Journal of Medicinal Chemistry
17
1
2
3
4
5
6
7
8
9
cRp.i.m.
0.038 (0.277)
13.9 (2.2)
47.30 – 2.35
0.1962
<I>/<σ(I)>
Resolution limits (Å)
Rwork
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Rfree
0.2487
dClashscore
RMSD
11.0
Bond lengths (Å)
Bond angles (o)
Tetramers
Inhibitors
Waters
0.0036
0.713
1
4
521
4
Mg2+
Favoured
Outliers
96.91
2.33
aValues in parentheses are for the outer resolution shell (2.08ꢀ2.03
at 100 K.
Å). All data were collected
ꢋ
ꢎ
ꢋ
ꢎ
ꢋ
ꢎ
ꢀꢁꢂꢃꢄꢂ = ∑ꢅꢆꢇ∑ꢈꢉꢊꢈ ℎꢌꢍ − ꢏꢊ ℎꢌꢍ ꢐꢉ/∑ꢅꢆꢇ∑ꢈꢊꢈ ℎꢌꢍ
ꢗ/ꢘ
1
ꢋ
ꢎ
ꢋ
ꢎ
ꢋ
ꢎ
ꢀꢑ.ꢈ.ꢁ. = ∑ꢅꢆꢇ
ꢒ
ꢖ
∑ꢈꢉꢊꢈ ℎꢌꢍ − ꢏꢊ ℎꢌꢍ ꢐꢉ/∑ꢅꢆꢇ∑ꢈꢊꢈ ℎꢌꢍ
ꢓ ꢋ ꢕ
ꢎ
ꢔ ℎꢌꢍ − 1
ꢋ
ꢎ
where Ii(hkl) is the observed intensity and ꢊ ℎꢌꢍ is the average intensity obtained from
multiple observations of symmetry related reflections. dClashscore is defined as the number of
bad overlaps ≥ 0.4 Å per thousand atoms.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 57
18
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(A)
(B)
Figure
6.
(A)
Electron
density
map
of
[3ꢀ(guanineꢀ9ꢀyl)ꢀ2ꢀ((2ꢀ
phosphonoethoxy)methyl)propoxy]methylphosphonic acid (20) bound in subunit C. The pale
blue sphere shows the location of the single magnesium ion in the active site. (B)
Superimposition of subunits, A, B, C and D showing the location of 20 in each of the four
active sites.
Figure 6A shows the electron density for 20 in the active site and Figure 6B, the
superimposition of the structure of the inhibitor bound in each of the four subunits. The
location of the three critical structural elements of 20 i.e. the purine ring and the two
phosphonate groups are the same in each of the four subunits, though there are slight
differences in the conformation of the linker (Fig. 6B). The only significant difference occurs
in subunit B. In this subunit, the oxygen atom is pushed further back into the active site than
in the other three subunits. In this position, it is located within 2.5 Å of a water molecule that
forms a hydrogen bond to the N3 of the purine ring (Fig. 7). This water molecule also forms a
hydrogen bond to one of the phosphonyl oxygen atoms. There are no interactions between
either of the two linkers and active site residues or bridging waters.
ACS Paragon Plus Environment

Page 19 of 57
Journal of Medicinal Chemistry
19
1
2
3
4
5
6
7
D193
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
[3-(guanine-9-yl)-2-((2-
phosphonoethoxy)methyl)propoxy]
methylphosphonic acid (20)
Figure 7. The interaction between the oxygen atom in one of the linkers and the bridging
water molecule when 20 is bound in the active site of subunit B of human HGPRT. The
hydrogen bond between the exocyclic amino group and the carbonyl of D193 occurs in all
subunits.
Comparison of the crystal structures of the human HGPRT.1a and HGPRT.20
complexes. Superimposition of these two structures reveals that the location of the purine
base and the phosphonate group bound in the 5′ꢀphosphate binding pocket are identical (Fig.
8) and that the structure of the flexible loops that surround the active site are unchanged (Fig.
8). The differences between the two structures is in the location of the second phosphonate
group (Fig. 8).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 57
20
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 8. Comparison of the binding mode of 20 and 1a (PDB code: 4IJQ) in the active site
of human HGPRT. Compound 1a with the associated sulfate ion (dark blue); Compound 20
(colored by element).
Thus, the differences in the Ki values between 20 and 1a for human HGPRT are
attributed to how the second phosphonate group which points down into the PPi pocket binds
in the active site. In the human HGPRT.20 complex, this group points back into the active
site towards the PPi binding pocket while, in 1a, it points out of the active site leaving room
for a sulfate and magnesium ion to bind (Fig. 8). It is, therefore, speculated that the higher Ki
value for PfHGXPRT compared to human HGPRT (Table 1) may be due to differences in the
mode of binding of this phosphonate group.
The specific interactions that the phosphonyl oxygens of the second phosphonate
group form at the active site in the human HGPRT.20 complex. A network of hydrogen
bonds holds this phosphonate group in place when 20 is bound in the active site of human
HGPRT (Fig. 8). One of the phosphonyl oxygen atoms forms hydrogen bonds to three water
molecules. These form hydrogen bonds with each other and with the highly conserved
residues, D193 (OD), R199 (NH2) and G69 (backbone nitrogen). A second hydrogen bond is
also evident between one of the other phosphonyl oxygen atoms and a water molecule that
ACS Paragon Plus Environment

Page 21 of 57
Journal of Medicinal Chemistry
21
1
2
3
4
5
6
7
8
forms a hydrogen bond with the OD atom of D134 and is also coordinated to a magnesium
ion (Fig. 9). Thus, this network of bonds via bridging waters makes a strong contribution to
anchoring 20 in position.
9
In the human HGPRT.20 complex, the phosphonate group (Fig. 9) is located in the same
position as one of phosphate groups of pyrophosphate when this second product of the
reaction is bound in the active site of human HGPRT together with ImmGP27 and magnesium
ions (Fig. 10). The space in the active site occupied by the second phosphate group of PPi is,
instead, occupied by water molecules when 20 binds (cf. Figs. 9 and 10). In the human
HGPRT.1a complex), sulfate ions replaces one of the phosphate group of PPi (Fig. 10). As
there are no negatively charged ions in this vicinity in the human.20 complex i.e. phosphate or
sulfate, the magnesium ion found in the HGPRT.ImmGP.PPi and HGPRT.1a complexes is
not present.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 9. The network of hydrogen bonds that the phosphonyl oxygen atoms of the
phosphonate groups which points into the PPi binding pocket forms at the active site of the
human HGPRT.20 complex. The green sphere is magnesium ion and the red spheres are
waters.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 57
22
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 10. Superimposition of the human HGPRT.20 complex with the human enzyme in
complex with (2ꢀ[(guaninꢀ9ꢀyl)methyl]propaneꢀ1,3ꢀdiyl)bis(oxy)]bis(methylene)diphosphonic
acid (1a) and sulfate (yellow)16 (PDB code: 4IJQ) and ImmGP and pyrophosphate (dark
blue)27 (PDB code: 1BZY).
The network of hydrogen bonds is a strongly contributing factor leading to the low Ki
value of 20 for human HGPRT. It is unclear though why the hypoxanthine derivative binds
much more weakly to both the human and Pf enzymes (Table 1). The absence of a single
hydrogen bond between the backbone carbonyl of D193 and the exocyclic amino group seems
unlikely to cause such a dramatic decrease if the ANbP was bound in exactly the same place
as its guanine counterpart. Thus, the most logical conclusion is that this is not the case and,
when the base is hypoxanthine, this change influences the whole orientation of the ANbP in
the active site and, thus, has a much higher Ki value.
2.4. In vitro antimalarial study in erythrocyte cell cultures infected by Plasmodium
falciparum. Phosphoramidate prodrugs (Scheme 4) for each inhibitor in Table 1 were
synthesized and tested for their ability to restrict the growth of Pf in cell culture. The results
are given in Table 3.
ACS Paragon Plus Environment

Page 23 of 57
Journal of Medicinal Chemistry
23
1
2
3
4
5
6
Table 3: In vitro antimalarial activity and cytotoxicity of ANbP prodrugs in cell culture
7
8
assays.
9
CC50 (ꢀM)
SI
Compound
Y = CH(CH2Ph)COOEt
IC50 (µM)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Base
Acyclic moiety
D6a
W2b
A549c
D6/A549
No.
36
Series 1
G
9.7 ± 1.6d
3.8 ± 0.5d
7.1 ± 2.1d
>300d
>30
27
Hx
4.0 ± 0.9d
101 ± 17d
37
G
6.6 ± 2.3d
3.6
7.7 ± 1.4d
1.9
107 ± 55d
90 ± 29
16
25
Series 2
34
38
Hx
8ꢀBrꢀG
ND
ND
188 ± 43
ꢀ
35
G
5.7 ± 0.3
1.4 ± 0.2
5.6 ± 0.1
ND
7.4 ± 0.7
3.2 ± 0.5
8.3 ± 1.2
ND
212 ± 16
82 ± 11
151 ± 27
36 ± 5
37
59
27
ꢀ
28
29
32
33
30
31
Series 3
Hx
8ꢀBrꢀG
8ꢀBrꢀHx
7ꢀdeazaꢀG
7ꢀdeazaꢀHx
1.8 ± 0.0
ND
6.5 ± 2.6
ND
106 ± 23
59
ꢀ
20 ± 4
B = purine base; ND = not determined a Pf strain sensitive to most drugs; Chloroquineꢀ and
b
c
d
pyrimethamineꢀresistant Pf strain; Human lung carcinoma cells; Data from Ref. 16. SI =
selectivity index = CC50/IC50. IC50 and CC50 represent the concentrations of the drug where
cell growth is reduced by 50% in Pf and human cells, respectively
The IC50 values range between 1.4ꢀ9.7
ꢀM. The precise reason for the variability is
unknown as it could be attributed to a number of factors. These include differences in
transport across the plasmodial membrane and/or hydrolysis of the prodrug to its active
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 57
24
1
2
3
4
5
6
7
8
9
component. The phosphoramidate prodrugs have several advantages: They have a long
intracellular halfꢀlife that allows infrequent dosing, the nontoxic amino acid moieties serve as
an attachment and their properties can be fineꢀtuned, and there is no stereochemistry issue (in
contrast to ProꢀTide approach). Thus, the phosphoramidates can serve as good model
compounds to show the in vitro activity of inhibitors and to compare individual types of
inhibitors with each other. An encouraging result is the fact that 28, whose parent compound
20 is the most potent inhibitor of human HGPRT (Ki = 6 nM), has an extremely low
cytotoxicity in human A549 cell line. This is also true for compound 36. It confirms the
assumption that HGXPRT is an excellent drug target based on the fact that the purine salvage
pathway is much more important for the parasite than for its host. Further studies are in
progress to investigate if modifications to these current prodrugs can improve antimalarial
activity.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
CONCLUSIONS
The new ANbPs exhibit a range of Ki values for human HGPRT (0.006ꢀ4.6
ꢀM) and for
PfHGXPRT (0.07ꢀ6.0 M). These values strongly depend on the nature of the purine base.
ꢀ
The crystal structure of the most potent of these inhibitors in complex with human HGPRT
has been obtained. This suggests that one of the factors that could contribute to the potency
and/or the selectivity for human HGPRT and PfHGXPRT could be the binding of one of the
phosphonate groups in the binding pocket of the second product of the reaction, PPi. Variable
micromolar antiparasitic activity of phosophoramidate prodrugs of these inhibitors in Pf
infected erythrocyte cultures was observed. The toxicity data in human cell line show that
even a prodrug (28) of potent inhibitor of human HGPRT (20) does not affect host cells, but
at the same time can be toxic to parasites.
ACS Paragon Plus Environment

Page 25 of 57
Journal of Medicinal Chemistry
25
1
2
3
EXPERIMENTAL SECTION
4
5
6
7
8
Synthesis and Analytical Chemistry. Unless stated otherwise, solvents were
evaporated at 40°C/2 kPa, and compounds were dried in vacuum oven at 2 kPa.
Tetrahydrofuran (THF) was distilled with lithium aluminium hydride pellets under argon.
TertꢀButanol was stored over annealed molecular sieves (4 Å).
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1H and 13C NMR spectra were measured on a Avance 500 spectrometer (500 MHz and
126 MHz) in DMSOꢀd6, D2O or D2Oꢀ(NaOD additive) and referred to TMS, residual solvent
signal or added solvent (tBuOH or 1,4ꢀdioxan). Complete assignment is based on
heteronuclear correlation experiments HSQC and H, CꢀHMBC. Chemical shifts (δ) are in
ppm and coupling constants (J) in Hz. The numbering system for the assignment of NMR
signals is described in Fig. 11. GC/MS spectra were measured on Agilent 5975B MSD
spectrometer coupled to 6890N gas chromatograph. Determination of lowꢀresolution mass
spectra is described below (procedure 1 a procedure 2). High resolution mass spectra were
measured: on QꢀTof micro (Waters) using ESI technique for lowꢀresolution mass spectra,
followed by LTQ Orbitrap XL (Thermo Fisher Scientific) spectrometer using ESI technique
for highꢀresolution mass spectra.
For flash chromatography, Teledyne ISCO CombiFlash Rf200 with dual λ absorbance
detector and various types of columns were used: a) original Teledyne ISCO columns
RediSepRf® variation “HP Silica GOLD” in sizes 12 g, 40 g, 80 g and 120 g; b) original
Teledyne ISCO columns RediSepRf® variation “HP C18 Aq GOLD” in sizes 50 g and 100 g;
c) column CHROMABOND Flash DL 40, DL 80, DL 120 and DL 200, filled by FLUKA
silica gel 60. Separation of UV nonꢀabsorbing substances were made: a) with internal UV
absorption standard (like caffeine) with similar Rf as required substance; b) by capturing all
fractions followed by analysis.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 57
26
1
2
3
4
5
Purity of target compounds was ≥95%. Purity was determined by the combination of UPLCꢀ
PDAꢀMS, NMR and HRꢀMS.
6
7
8
UPLC-MS:
9
Procedure 1: Analytical sample was prepared typically so that 50 µl reaction mixture or
chromatographic fraction was diluted in suitable solvent (MeOH or acetonitrile) to
concentration about 1 mg of sample in 1 mL of solution. Mixture was diluted to 1ꢀ2 mL
volume and filtered over syringe filter (0.22 µM, membrane material = PVDF).
Samples were tested by universal LC method (eluent H2O/CH3CN, gradient 0ꢀ100 %, run
length 7 min) and timeꢀtested MS method (ESI+, cone voltage = 30V, mass detector range ~
100ꢀ1000 Da in standard case; 500ꢀ1600 Da for measurement of prodrugs).
Analytical arrangement contains: Waters UPLC HꢀClass Core System, (column Waters
Acquity UPLC BEH C18 1.7 µm, 2.1x100 mm), Waters Acquity UPLC PDA detector, Mass
spectrometer Waters SQD2.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Procedure 2: Procedure is almost identical as described above in Procedure 1. There are
differences in solvent for dilution of samples (water) and MS method (ESIꢀ, detector range ~
100ꢀ1000 Da). This procedure is suitable for measurement of free phosphonic acids or their
sodium salts.
ACS Paragon Plus Environment

Page 27 of 57
Journal of Medicinal Chemistry
27
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 11. Structures and a numbering system for the assignment of NMR signals of
compounds in experimental section.
Synthetic procedures
General procedure for Mitsunobu reaction (Method A): To a solution of
triphenylphosphine (3.5 g, 13.4 mmol) in dry THF (50 mL) cooled to –30°C under argon
atmosphere diisopropyl azadicarboxylate (DIAD, 2.4 mL, 12.2 mmol) was added. The
mixture was stirred for 30 min to form a complex. The solution of the corresponding alcohol
(4 mmol) in dry THF (30 mL) and 6ꢀchloropurine (1 g, 6.4 mmol = 1.6 eq.) or 6ꢀchloroꢀ7ꢀ
deazapurine (1 g, 6.4 mmol = 1.6 eq.) was added to the preformed complex and the reaction
flask was filled with argon again. The resulting mixture was slowly warmed up to room
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 57
28
1
2
3
4
5
6
7
8
9
temperature and stirred for two days. MeOH (30 mL) was added to dissolve all possible
heterogeneous particles. Solvent was evaporated, and the crude mixture was purified by flash
chromatography (silica gel; eluent CHCl3/MeOH, gradient 1ꢀ10%). The pure product was
obtained as a yellowish solid or a foam.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
General procedure for the Mitsunobu reaction (Method B): Starting from 2ꢀaminoꢀ6ꢀ
chloropurine or 2ꢀaminoꢀ6ꢀchloroꢀ7ꢀdeazapurine and the corresponding alcohol procedure
was identical as described above – Method A. Then after the stirring of the reaction mixture
for two days, water (30 mL) and MeOH (30 mL) were added and the mixture was heated at
80°C for 30 h. Solvent was evaporated, the residue was codistilled with toluene, and the crude
mixture purified by flash chromatography (silica gel; eluent CHCl3/MeOH, gradient 2ꢀ15%).
General procedure for hydrolysis of 6-chloropurine and 2-amino-6-chloropurine
derivatives to corresponding 6ꢀoxopurine and 2ꢀaminoꢀ6ꢀoxopurine derivatives (Method C):
The 6ꢀchloropurine or 2ꢀaminoꢀ6ꢀchloropurine derivative (4.13 mmol) was dissolved in
aqueous trifluoroacetic acid (75%, 60 mL) and stirred overnight at room temperature. The
solvent was evaporated and the residue was codistilled with water (3×) and toluene (2×). The
crude product was purified by flash chromatography (silica gel; eluent CHCl3/MeOH,
gradient 2ꢀ15%). Pure product was obtained as a solidified oil.
General procedure for hydrolysis of 6-chloro-7-deazapurine and 2-amino-6-chloro-
7-deazapurine derivatives to corresponding 6ꢀoxoꢀ7ꢀdeazapurine and 2ꢀaminoꢀ6ꢀoxoꢀ7ꢀ
deazapurine derivatives (Method D): The appropriate 6ꢀchloropurine derivative (3 mmol)
was dissolved or suspended in water (40 mL) containing K2CO3 (1.69 g, 12.2 mmol) and
DABCO (0.69 g, 6.2 mmol). The reaction mixture was stirred and heated to 120 °C for 1ꢀ4 h.
The resultant solution was cooled down, neutralized by 1 M HCl and a product extracted to
CHCl3 (3x 50 mL). Organic phase was washed with water (2x 100 mL), followed by brine (1x
ACS Paragon Plus Environment

Page 29 of 57
Journal of Medicinal Chemistry
29
1
2
3
4
5
6
7
8
9
100 mL) and dried over MgSO4. Compounds were purified by flash chromatography (silica
gel; CHCl3/MeOH, gradient 2ꢀ10 %). Reaction was monitored by LCꢀMS (Procedure 1).
General procedure for synthesis of sodium salts of free bisphosphonates (Method
E): A mixture of bisphosphonic acid diester (diisopropyl or diethyl ester) (1 mmol),
acetonitrile (20 mL) and BrSiMe3 (2 mL) was stirred overnight at room temperature. After
evaporation, the residue was stirred with MeOH:water (1:1, 100 mL) for 2 h at room
temperature. The solvent was evaporated in vacuo. The residue was dissolved in solution of
TEAB (5 mL 2M aqueous TEAB), evaporated in vacuo and codistilled with water (3x). That
TEABꢀsalt was purified by flash chromatography (C18 silica gel, 0.125 M aqueous
TEAB/MeOH, gradient 0ꢀ40%) and evaporated in vacuo and codistilled with water (3x). The
residue was dissolved in small volume of water (5 mL) and percolated through DOWEX
(50WX8 Na+, 5 g) to obtain sodium salt of the bisphosphonic acid. Filtrate was partially
evaporated and prepared for following manipulation by lyophilisation from water.
General procedure for bromination of 6-oxopurine derivatives (Method F): To a
solution of corresponding 6ꢀoxopurine derivative (2 mmol) in DMF (10 mL), 0.3 M solution
of bromine in CCl4 (13 mL, 4 mmol of Br2, 2 eq.) was added. Reaction mixture was stirred for
2ꢀ12 h at room temperature. The mixture was evaporated in vacuo, codistilled with toluene
(3x) and purified by flash chromatography (silica gel; CHCl3/MeOH, gradient 1ꢀ10%).
Reaction was monitored by LCꢀMS (Procedure 1). The product was obtained as
colourless/yellowish solidified oil.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
General procedure for synthesis of tetra-( -phenylalanine ethylester) prodrug of
ȯ
bisphosphonates (Method G): Bisphosphonic acid tetraester (diisopropyl or diethyl ester) (1
mmol) was codistilled with dry acetonitrile (2x). Dry acetonitrile (20 mL) and
bromotrimethylsilane (2 mL) were mixed with bisphosphonic acid tetraester and stirred
overnight at room temperature under argon. After evaporation in vacuo and codistillation with
ACS Paragon Plus Environment