NMR for screening and a biochemical assay: Identification of new FPPS inhibitors exerting anticancer activity

Article abstract of DOI:10.1016/j.bioorg.2019.103449

Farnesyl pyrophosphate synthase (FPPS) is a crucial enzyme for the synthesis of isoprenoids and the key target of nitrogen-containing bisphosphonates (N-BPs). N-BPs are potent and selective FPPS inhibitors that are used in the treatment of bone-related diseases, but have poor pharmacokinetic properties. Given the key role played by FPPS in many cancer-related pathways and the pharmacokinetic limits of N-BPs, hundreds of molecules have been screened to identify new FPPS inhibitors characterized by improved drug-like properties that are useful for broader therapeutic applications in solid, non-skeletal tumours. We have previously shown that N6-isopentenyladenosine (i6A) and its related compound N6-benzyladenosine (2) exert anti-glioma activity by interfering with the mevalonate pathway and inhibiting FPPS. Here, we report the design and synthesis of a panel of N6-benzyladenosine derivatives (compounds 2a-m) incorporating different chemical moieties on the benzyl ring. Compounds 2a-m show in vitro antiproliferative activity in U87MG glioma cells and, analogous to the bisphosphonate FPPS inhibitors, exhibit immunogenic properties in ex vivo γδ T cells from stimulated peripheral blood mononuclear cells (PBMCs). Using saturation transfer difference (STD) and quantitative ¹H nuclear magnetic resonance (NMR) experiments, we found that 2f, the N6-benzyladenosine analogue that includes a tertbutyl moiety in the para position of the benzyl ring, is endowed with increased FPPS binding and inhibition compared to the parent compounds i6A and 2. N6-benzyladenosine derivatives, characterized by structural features that are significantly different from those of N-BPs, have been confirmed to be promising chemical scaffolds for the development of non N-BP FPPS inhibitors, exerting combined cytotoxic and immunostimulatory activities.

Full text of DOI:10.1016/j.bioorg.2019.103449

Journal Pre-proofs
NMR for screening and a biochemical assay: identification of new FPPS in-
hibitors exerting anticancer activity
Manuela Grimaldi, Rosario Randino, Elena Ciaglia, Mario Scrima, Michela
Buonocore, Ilaria Stillitano, Verdiana Covelli, Alessandra Tosco, Patrizia
Gazzerro, Maurizio Bifulco, Manuela Rodriquez, Anna Maria D'Ursi
PII:
S0045-2068(19)31112-5
DOI:
https://doi.org/10.1016/j.bioorg.2019.103449
YBIOO 103449
Reference:
To appear in:
Bioorganic Chemistry
Received Date:
Revised Date:
Accepted Date:
16 July 2019
1 October 2019
14 November 2019
Please cite this article as: M. Grimaldi, R. Randino, E. Ciaglia, M. Scrima, M. Buonocore, I. Stillitano, V.
Covelli, A. Tosco, P. Gazzerro, M. Bifulco, M. Rodriquez, A. Maria D'Ursi, NMR for screening and a
biochemical assay: identification of new FPPS inhibitors exerting anticancer activity, Bioorganic Chemistry
(2019), doi: https://doi.org/10.1016/j.bioorg.2019.103449
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version
will undergo additional copyediting, typesetting and review before it is published in its final form, but we are
providing this version to give early visibility of the article. 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.
© 2019 Published by Elsevier Inc.

NMR for screening and a biochemical assay: identification of
new FPPS inhibitors exerting anticancer activity
Manuela Grimaldi,a,d Rosario Randino,a Elena Ciaglia,c Mario Scrima,a Michela Buonocore,a
Ilaria Stillitano,a Verdiana Covelli,a Alessandra Tosco,a Patrizia Gazzerro,a Maurizio
Bifulco,b,c Manuela Rodriquez,a and Anna Maria D’Ursia*
aDepartment of Pharmacy, University of Salerno, Via Giovanni Paolo II 132, 84084, Fisciano
(Salerno), Italy.
bDepartment of Molecular Medicine and Medical Biotechnology, University of Naples
"Federico II", Via Pansini, 80131, Naples, Italy.
cDepartment of Medicine, Surgery and Dentistry "Scuola Medica Salernitana", University of
Salerno, Via Salvatore Allende, 84081, Baronissi (Salerno), Italy.
dInstitute of Polymers, Composites and Biomaterials, National Research Council of Italy, V.le
J.F. Kennedy 54 - Pad. 20 Mostra d’Oltremare, 80125 Naples, Italy.
To whom correspondence may be addressed: dursi@unisa.it
* Correspondence to:
Anna Maria D’Ursi, Department of Pharmacy, University of Salerno, Via Giovanni Paolo II
132, 84084, Fisciano (Salerno), Italy.
e-mail: dursi@unisa.it, Phone:+39089969748
1

Abstract
Farnesyl pyrophosphate synthase (FPPS) is a crucial enzyme for the synthesis of isoprenoids
and the key target of nitrogen-containing bisphosphonates (N-BPs). N-BPs are potent and
selective FPPS inhibitors that are used in the treatment of bone-related diseases, but have poor
pharmacokinetic properties.
Given the key role played by FPPS in many cancer-related pathways and the pharmacokinetic
limits of N-BP, hundreds of molecules have been screened to identify new FPPS inhibitors
characterized by improved drug-like properties that are useful for broader therapeutic
applications in solid, non-skeletal tumours.
We have previously shown that N6-isopentenyladenosine (i6A) and its related compound N6-
benzyladenosine (2) exert anti-glioma activity by interfering with the mevalonate pathway and
inhibiting FPPS. Here, we report the design and synthesis of a panel of N6-benzyladenosine
derivatives (compounds 2a-m) incorporating different chemical moieties on the benzyl ring.
Compounds 2a-m show in vitro antiproliferative activity in U87MG glioma cells and,
analogous to the bisphosphonate FPPS inhibitors, exhibit immunogenic properties in ex vivo
γδ T cells from stimulated peripheral blood mononuclear cells (PBMCs). Using saturation
transfer difference (STD) and quantitative 1H nuclear magnetic resonance (NMR) experiments,
we found that 2f, the N6-benzyladenosine analogue that includes a tertbutyl moiety in the para
position of the benzyl ring, is endowed with increased FPPS binding and inhibition compared
to the parent compounds i6A and 2.
N6-benzyladenosine derivatives, characterized by structural features that are significantly
different from those of N-BPs, have been confirmed to be promising chemical scaffolds for the
development of N-BP FPPS inhibitors, exerting combined cytotoxic and immunostimulatory
activities.
2

Graphical abstract
Keywords
FPPS, isoprenoids, adenosine derivatives, NMR enzymatic assay
1. Introduction
Farnesyl pyrophosphate synthase (FPPS) is a key enzyme involved in the mevalonate (MVA)
pathway and prenylation of downstream proteins; it catalyses the two-step synthesis of the C15
isoprenoid farnesyl pyrophosphate (FPP), a crucial precursor for the synthesis of several classes
of essential metabolites, such as sterols, ubiquinones, and carotenoids.[1-3]
FPPS has been implicated in many cancer-related pathways[4, 5] and plays a significant role
in the maintenance of the neoplastic malignant phenotype. In particular, we recently
documented the deregulated expression and activity of FPPS in a cohort of stage III-IV glioma
patients; in primary glioblastoma-derived cells, we found FPPS alterations, exhibiting a linear
correlation with canonical oncogenic signalling pathways such as signal transducer and
activator of transcription 3 (STAT3), extracellular signal-regulated kinase (ERK), and protein
kinase AKT.[6]
3

Nitrogen-containing bisphosphonates (N-BPs) are potent and selective inhibitors of FPPS that
are used for the treatment of bone-related tumours and osteoporosis diseases.[7-10] Due to their
ability to induce stimulation of innate δ T cytotoxic antitumoural cells, N-BPs are also used in
immunotherapy-based cancer treatment.[11-14] Despite their potency, N-BPs suffer from poor
pharmacokinetics, exhibiting almost negligible distribution to non-skeletal tissues and rapid
clearance from systemic circulation.[7]
Given the important role played by FPPS in cancer manifestation and progression,[15]
hundreds of molecules have been tested in search of new FPPS inhibitors with improved drug-
like properties that are useful for broader therapeutic applications in solid, non-skeletal
tumours.[16, 17] In particular, powerful uncharged FPPS inhibitors targeting an allosteric
pocket of FPPS were recently identified using fragment-based screening by nuclear magnetic
resonance (NMR) and X-ray crystallographic methods.[18, 19]
In our previous investigation, including an in silico inverse virtual screening, NMR
experiments and in vitro enzymatic assays, FPPS was found to be a valuable target for the
structural binding of N6-isopentenyladenosine (i6A) (1 in Figure 1).[20] Compound i6A is a
modified nucleoside belonging to the cytokine family. It is involved in the control of many
processes in plants[21-24] and exerts in vitro and in vivo antitumoural and immunomodulatory
activities.[25-31] Interestingly, in many experiments, the cytotoxicity and anti-proliferative
activity of i6A was shown to be associated with alterations in FPPS expression and activity.[31-
33]
Encouraged by the preliminary data on FPPS-i6A binding, we decided to synthesize and test
i6A analogues, including those with different substituents at the N6position of adenosine. The
N6-benzyladenosine derivative (2, Figure 1),[30] where a benzyl moiety on the adenosine ring
replaced the isopentenyl, confirmed the ability of 2 to bind and inhibit FPPS and exhibited
greater cytostatic and anti-proliferative activities than those of the parent compound i6A.
4

Figure 1. Chemical structures of i6A (1) and N6-benzyladenosine (2).
Intending to obtain N6-benzyladenosine derivatives endowed with improved FPPS binding, in
this report, we describe the efforts to design, synthesize and screen a panel of compounds
resulting from the introduction of different chemical moieties in the para position of the N6-
benzyl ring of 2 (2a-m, Figure 2).
Figure 2. Chemical structures of N6-benzyladenosine derivatives (2a-m).
We employed saturation transfer difference (STD) NMR experiments to study compounds 2a-
m interacting with the FPPS binding pocket; moreover, we developed an innovative NMR-
based FPPS enzymatic assay based on quantitative 1H-NMR measurements of FPPS substrates
5

vs. products. Finally, we confirmed the effects of compounds 2a-m on FPPS by measuring
their immunostimulatory effects on peripheral T-lymphocytes, analogous to bisphosphonate
FPPS inhibitors.[14]
2. Results
2.1 Design of N6-benzyladenosine analogues (2a-m)
The description of the active site of the FPPS enzyme, according to the numerous crystal
structures
deposited
in
the
Brookhaven
PDB
(https://www.rcsb.org/;
https://www.uniprot.org/uniprot/P14324), is consistent with the existence of i) an allylic sub-
pocket where the pyrophosphate moiety of DMAPP/GPP interacts with the enzyme via Mg2+-
mediated interactions and ii) an IPP sub-pocket where the pyrophosphate moiety of IPP
interacts with basic residues through direct salt bridge (K57, R60, R113) or water-mediated
(R112, R351) interactions.[17, 34, 35] We have previously demonstrated that i6A and its
derivative N6-benzyladenosine (2, Figure 1) bind to the FPPS catalytic site, inducing moderate
inhibition of its enzymatic activity.[20, 30] Based on our molecular docking calculations, in
the most represented binding poses, the isopentenyl moiety of i6A and the benzyl ring of 2
interact with the allylic sub-pocket in a region that is large enough to contain additional bulky
substituents. To obtain more active FPPS ligands, we designed a panel of compounds carrying
substituents with different steric hindrances on the N6-adenosine benzyl ring. The designed
compounds were docked in the FPPS binding site using AutoDock Vina (version 1.1.2).[36]
An FPPS X-ray structure that included three Mg2+ ions essential for enzymatic activity was
selected from the PDB (ID code: 5YGI) following the criteria explained in our previous
works.[20, 30] The binding energies of compounds 2a-m resulting from the molecular docking
calculations are reported in the Supplementary Information (Table S1). Compounds 2f, 2g, and
2i, characterized by the best docking score parameters, show improved FPPS binding compared
6

to the parent compounds i6A and 2. Analysis of the most representative binding poses (Figure
3a-b) indicates that all the molecules have a similar orientation: the ribosyl sugar and adenine
rings are located in the highly charged DMAPP sub-pocket, exploiting metal interactions with
an Mg2+ ion, with additional stabilizing interactions involving Asp residues in the conserved
aspartate-rich motif. The para-substituted benzyl ring occupies a large hydrophobic sub-pocket
and engages in cation-π interactions with F98, F99, Y204, and L100.
Figure 3. a) Two dimensional interaction panel showing interactions between 2f, 2g, and 2i
and the FPPS binding site (PDB ID 5YGI) as derived from molecular docking calculations. b)
The 3D superposition of the best binding poses derived from the docking calculations of 2f
7

(orange), 2g (blue) and 2i (green) with the FPPS (PDB ID: 5YGI) binding site. The FPPS
backbone is shown as a grey ribbon.
2.2 Chemistry
Adenosine represents a very suitable scaffold for chemical modifications; its high polarity
confers a peculiar solubility profile to its derivatives. Several methods for the preparation of
ring-substituted N6-benzyladenosines have been previously described.[37-40] Generally, these
synthetic procedures consist of the direct reaction of 6-chloropurine riboside (3) with the
appropriate primary benzylamine. Conventional heating is used for several hours in an
alcoholic solvent with an excessive amount of triethylamine (TEA)[24, 41-44] (Figure 4). By
following this procedure, we first synthesized the N6-benzyladenosine derivatives by heating
3 with three eq. of benzylamine (BA) in ethanol for 3 h at 80°C; we obtained compound 2 in
55% yield (Table S2, entry 1, see Supplementary Information).
Figure 4. Synthesis of N6-benzyladenosine and derivatives 2a-m.
Microwave-assisted (MW) synthetic protocols have been widely used as an alternative to
conventional heating synthetic methods to shorten the reaction time and with respect to green
chemistry and atom economy principles.[45-48]
8

Several MW-assisted protocols for the synthesis of N6-benzyladenosine derivatives have been
previously described[46, 49]; however, by applying experimental procedures earlier developed
in our laboratory[45, 50] and to optimize the reaction conditions, we decided to set up a new
MW-assisted procedure by systematically modifying the reaction parameters.
First, the temperature, MW irradiation power, reaction time, and solvent were modified. A 99%
conversion of 6-chloropurine riboside into 2 was obtained by heating 3 in ethanol for 5 min
under MW irradiation at 180°C in the presence of three eq. of TEA, leading to 2 in 95% isolated
yield (Table S2, entry 7, see Supplementary Information). In an attempt to reduce the reaction
time, we found that after irradiation for just 1 min at 210°C, the conversion of 3 into 2 was
93% (90% isolated yield, Table S2, entry 8 see Supplementary Information). Prolonged
irradiation time at 210°C led to a significant increase in by-products (Table S3, entries 9-12
see Supplementary Information).
These encouraging results led us to investigate the potential effect exerted by the solvent
(testing both green solvents and solvent-free conditions) modulating the reactant as well as the
TEA stoichiometry. The effect of solvent and BA equivalents for the synthesis of 2 are reported
in Table S2 in the Supplementary Information. Reducing the BA equivalents to 1 eq. in the
presence of just one eq. of TEA and switching from EtOH to an ionic liquid and then to H2O
led to improved yields. Nevertheless, solvent-free conditions (Table S3, entry 16 see
Supplementary Information) led us to reach the best results (98% conversion and 94% isolated
yield), merging the most successful conditions previously reported (Table S2 for entries 7 and
8, see Supplementary Information), irradiating 3 for just 1 min at 300 W and 210°C. These
results suggest that the equimolar ratio of BA, TEA, and 3 in solvent-free conditions represent
the most suitable, versatile settings for the MW-assisted synthesis of N6-benzyladenosines with
respect to green chemistry and atom economy principles.[47]
9

These new, faster and more efficient conditions encouraged us to extend our method to
different BAs (Table 1). Additionally, in these cases, the MW irradiation dramatically reduced
the conversion time, improving the yields and purity of the desired N6-substituted adenosine
analogues 2a-m.
Compound Yield (%)a
2a
2b
2c
2d
2e
2f
2g
2h
2i
92
89
92
88
89
91
87
90
88
88
84
90
89
2j
2k
2l
2m
aYields were evaluated on the isolated product.
Table 1. Synthesis of compounds 2a-m by using the optimized MW-assisted methodology.
2.3 FPPS-ligand interaction: NMR analysis
Compounds 2a-m, designed and synthesized as previously described, were characterized by
1
measuring standard H monodimensional, 2D HSQC (H-1/X correlation via double INEPT
transfer) and 2D HMBC (H-1/X correlation via heteronuclear zero and double quantum) NMR
experiments in the solvent MeOD. The 1H and 13C chemical shift assignments of compounds
2a-m are reported in Figures S1-S14 (Supplementary Information). Then, compounds 2a-m
were screened for their ability to bind to the FPPS enzyme by recording STD NMR
experiments.[51, 52] STD is a powerful NMR technique that is used for the identification of
protein-ligand interactions and the determination of protein-ligand dissociation constants (KD
values).
10

The protein concentration to be used in STD NMR experiments depends on the molecular
weight of the protein and is usually within the 10-4-10-10 M concentration range.[53] STD
experiments at different protein concentrations (8 µM, 15 µM 25 µM) were acquired before to
set 8.0 µM FPPS as the concentration to be used in the full set of STD experiments. Indeed,
the limit of protein concentration was imposed by the solubility of the ligands that at 1.25 mM
precipitate, affecting the reliability of the results.
As a reference, we measured the KD constants for compounds i6A and 2, as previously reported
[20, 30] and corresponding to ∼1.0 mM and ∼0.19 mM, respectively. Subsequently, STD-
NMR experiments were collected for each compound 2a-m (Figures S15-S28, see
Supplementary Information). NMR samples containing 8.0 μM FPPS were titrated with
compounds 2a-m to build-up STD at protein-ligand molar ratios of 1:10, 1:20, 1:30, 1:50, 1:70,
and 1:100. For each titration point, STD experiments were acquired using different saturation
times (0.50, 1.00, 1.50, 2.00, 3.00, 4.00, and 5.00 s).[54] Compound 2f, characterized by the
p-tertbutylphenyl ring in the N6 position of adenosine, showed the most significant STD effect,
with a 30% decrease in the intensity of the signals from the p-tertbutyl protons (Figure 5). For
this compound, we quantitatively evaluated the STD data to calculate the KD for the FPPS/2f
complex. We collected data from experiments at different protein-ligand ratios and saturation
time conditions, according to the methodology developed by Angulo et al.[54] The mean KD
value calculated for the single protons of 2f was ∼0.14 mM (Table 2).
11

Figure 5. STD-NMR spectra of FPPS/2f, 1:100 molar ratio. Red shows the off-resonance
spectrum and blue shows the STD spectrum.
protein-ligand
(2f) system
proton H8
KD (from STD-AF
initial slope) [mM]
0.14 ± 0.02 mM
0.14 ± 0.01 mM
0.14 ± 0.02 mM
0.14 ± 0.02 mM
0.13 ± 0.01 mM
0.21 ± 0.04 mM
0.17 ± 0.03 mM
0.06 ± 0.01 mM
0.14 ± 0.02 mM
proton H2
proton H13/17
proton H14/16
proton H1'
proton H2'
proton H3'
proton H5'
proton H19/20/21
Table 2. Dissociation constants (KD) calculated from the isotherms of the initial growth rates
of the STD-AF values of the protein-ligand systems studied herein.
These values represent an improvement in 2f-FPPS binding compared to 2, indicating that the
introduction of the hindrance of the tertbutyl moiety in the para position of the benzyl ring
12

leads to more active FPPS ligands. To verify that the detected interaction of compound 2f with
FPPS involves the active site instead of an allosteric site, we titrated the FPPS-2f complex with
increasing concentrations of the FPPS allosteric modulator 5,7-dichloro-3-(carboxymethyl)-
1H-indole-2-carboxylic acid (IC50 ∼6 μM)[18, 55] following the procedure previously
described.[20] Notably, 80 μM Zol, a known high-affinity active site FPPS inhibitor, induced
observable variations in the FPPS-2f STD spectrum.
2.4 1H-NMR enzymatic assay
The screening of new FPPS inhibitors is generally based on the measurement of enzymatic
activity via a radio-enzymatic test.[56] To avoid experiments that use 31P radioligands, a new
enzymatic assay was developed by employing NMR quantitative analysis. The proposed
methodology includes the quantification of the FPPS-catalysed reaction rate by measuring the
variation of the signal intensities of FPPS substrates IPP/DMAPP vs. the product GPP. The
real-time enzyme kinetics of the substrates DMAPP and IPP converting to the product GPP by
FPPS is shown in Figure 6a. In a preliminary step, we set up the procedure by determining the
exact concentrations of the substrates (DMAPP and IPP) and FPPS enzyme to be used to obtain
1
the valuable IC50 values for FPPS ligands (vide infra, Methods). Accordingly, H-NMR
experiments were recorded on samples containing 200 µM DMAPP, 400 µM IPP, and 8 µM
FPPS enzyme at 310 K. As seen in Figure 6a, by the end of the experiment, the DMAPP (S1)
(5.53 ppm; 3.58 ppm; 1.80 ppm) and IPP (S2) (4.14 ppm; 2.47 ppm; 1.85 ppm) resonances
converted into that of the GPP (P) molecule (∼30 min). To prove that the ability of the
synthesized compounds to bind FPPS (according to STD NMR effects) would correspond to
their ability to inhibit FPPS enzymatic activity, the conversion of substrates DMAPP (S1) and
IPP (S2) was monitored in the presence of FPPS and compounds 2 (Figure S29, see
13

Supplementary Information) and 2f. Figure 6b shows that at time T=30 min, the GPP signals
were 4% of the corresponding signals recorded in the absence of compound 2f.
1
Figure 6. Real-time H-NMR spectra of the enzyme kinetics on a Bruker 600 MHz
spectrometer. a) The conversion of 400 µM IPP (S2) and 200 µM DMAPP (S1) produces GPP
(P); b) GPP (P) production was inhibited by the addition of 1 mM 2f.
To calculate IC50 values for 2 and 2f, the 1H-NMR spectra of IPP and DMAPP in the presence
of FPPS were recorded at different concentrations of 2 (1 mM, 1.4 mM, 1.8 mM and 2 mM)
and 2f (0.5 mM, 0.65 mM, 0.85 mM and 1 mM). By plotting the reaction rate at the given
inhibitor concentration vs. our inhibitor concentration, IC50 values of ∼ 6.10 mM and ∼1.18
mM were calculated for 2 and 2f, respectively (GraphPad Prism)[57] (Figure 7a-b). An
14

orthogonal colorimetric assay, as previously described, validated the calculated IC50 and KD
values (Figure S31, see Supplementary Information).[58]
Figure 7. Variation of GPP concentration vs. time (min) in the presence of a) 2 and b) 2f. From
the maximum slope, which was achieved within approximately 30 min, the trend lines were
generated with their associated equations.
3. Biological activity
3.1 Cell viability of N6-benzyladenosine analogues
Since compound 2 showed significant activity in inhibiting glioma cell growth,[30] the effects
of 2a-m were investigated by the BrdU incorporation assay. Along with all the compounds
15

tested, i6A and 2 were used as reference compounds in a 0.3-20 µM concentration range on
proliferating U87MG cells. The data reported in Figure 8 show that even though no compound
reached the same cytostatic efficacy of 2, compounds 2c, 2d, 2h, and 2i exert significant
cytotoxic activity. This result is consistent with the presence of halogen atoms (for 2c, 2d, 2h)
and a nitro moiety (for 2i) in the para position of the benzyl ring. Among the compounds
showing significant cytotoxicity, compounds 2a and to a greater extent, 2c, maintained discrete
dose-dependent inhibition of proliferation compared to untreated cells (Figure 8).
Figure 8. Effect of analogues 2 on cellular proliferation of the U87MG human glioma cell line.
U87MG cells were cultured for 48 h in the presence of the indicated concentrations (0–20 μM)
of i6A, 2 and 2(a-m) before the analysis of cell proliferation by the BrdU incorporation assay.
The results are representative of three independent experiments performed in triplicate and
expressed as the mean ± SD (ANOVA, *p<0.05, **p<0.01 and ***p<0.001).
16

3.2 Immunostimulatory activity of N6-benzyladenosine analogues
As the inhibition of FPPS achieved by amino bisphosphonates stimulates Vγ9Vδ2 T-cell
expansion among peripheral blood lymphocytes,[13, 14] we tested the γδ T cell-stimulating
ability of the various compounds in primary PBMC cultures. Freshly isolated PBMCs from
healthy donors were incubated for 14 days with i6A, 2 and 2a-m stimuli in the presence of
suboptimal doses of IL-2. Zol and IPP accumulated in antigen-presenting cells (APCs) and
immunostimulatory phosphoantigens for γδ T cells were used as references. When γδ T cells
were efficiently stimulated, they formed clusters on day 5. As expected, clusters and aggregates
of γδ T cells can be observed when the PBMCs were successfully stimulated with Zol and IPP
(Figure 9a).
In contrast, no clusters or aggregates were observed when the growth of γδ T cells was not
adequate (in medium in the absence of IL-2 or IL-2 alone). Interestingly, 2i, 2h, 2f, and 2a
showed an in vitro stimulatory profile that was similar or improved compared to that achieved
by the bisphosphonate Zol (Figure 9b). Furthermore, flow cytometric determination of the
percent of γδ T cells after 14 days showed that compounds 2a-m generally induced toxicity
comparable to Zol; however, these toxicity values were lower than reference compound 2
(Figure 9c).
17

Figure 9. A selective outgrowth of γδ T cells upon stimulation with analogues 2. a) PBMCs
were stimulated with human IL-2 (IL-2), zoledronic acid (Zol), IPP or compounds (i6A, 2, 2a-
m) to final concentrations of 150 IU/ml for IL-2 and 5 μM for all others. Cells were imaged by
brightfield microscopy. Representative fields are shown. Clusters and aggregates of γδ T cells
can be observed on day 7, when γδ T cells, following treatments, were successfully expanded.
b) Determination of the percentage of γδ T cells as evaluated by flow cytometric analysis on
day 14. Histograms report the percent ± SD of CD3+γδ+ positive cells compared to the total
PBMC population. c) At the end of cell culture, the viability of the stimulated PBMCs was
examined by propidium iodide (PI) staining and flow cytometry analysis. Histograms report
the percent ± SD of PI-negative viable cells in all conditions tested.
4. Discussion
The search for new anticancer FPPS inhibitors that can overcome the pharmacokinetic limits
of N-BPs has currently received increasing interest.[15-17] In the present contribution, we
18

report an NMR screening of new ligands of FPPS designed and synthesized starting from the
N6-benzyladenosine scaffold.
N6-isopentenyladenosine (i6A) and N6-benzyladenosine (2) are modified nucleosides
belonging to the cytokine family.[21, 24, 43] They control many processes in plants and exert
in vitro and in vivo cytostatic and pro-apoptotic activities[4] through a mechanism related to
the inhibition of FPPS expression and activity.[32] Based on in silico inverse virtual screening
and NMR experiments, we showed that the biological activities of i6A and 2 depend, at least
in part, on the direct inhibition of FPPS catalytic activity. Computational data and NMR
experiments provided the structural requirements governing the interaction between the
compounds and FPPS.[20, 30]
Many structural models of FPPS are available in the PDB (https://www.rcsb.org/;
https://www.uniprot.org/uniprot/P14324). All these models agree on the FPPS catalytic site
characterized by i) an allylic sub-pocket (DMAPP/GPP binding site) including two Asp-rich
motifs that bind the pyrophosphate moiety of DMAPP/GPP via Mg2+ cation interactions and
ii) a basic region rich in Arg residues that binds IPP through salt-bridge interactions. By
docking i6A and 2 in the FPPS binding pocket (ID code: 5YGI), we observed that in the most
representative binding poses (Figure 3), i) the riboside sugar and the adenine ring occupy the
DMAPP sub-pocket through electrostatic interactions with Mg2+ ions and conserved Asp
residues and ii) the benzyl ring occupies a large hydrophobic cavity capable of accommodating
new bulky chemical moieties. In agreement with these structural requirements, compounds 2a-
m, including chemical substituents characterized by different degrees of steric hindrance in the
para position of the benzyl ring, were synthesized and tested for their ability to bind FPPS
using STD-NMR experiments. Compound 2f, the compound with a tertbutyl moiety, showed
the lowest KD value of ∼0.14 mM, corresponding to an 8-fold and 2-fold increase in FPPS
binding compared to i6A and 2, respectively.
19

31
The activity of FPPS inhibitors has long been measured using P radioligands.[56] More
recently, LC/MS/MS FPPS inhibition assays have been developed.[18, 19] To measure the IC50
of i6A and 2, (IC50 i6A ∼1 mM and IC50 2 ∼0.25 mM) we used in our previous work a
colorimetric assay[58] that unfortunately suffers from poor accuracy and sensitivity. To
overcome these limits and to avoid the drawbacks related to the handling of the 31P radioligand,
we herein report a newly developed NMR enzymatic assay based on the quantitative evaluation
of the 1H signal intensity of FPPS substrates IPP/DMAPP vs. 1H signal intensity of the product
GPP. By applying the law of Michaelis and Menten and its derivations, we obtained IC50 values
of ∼6.10 mM and ∼1.18 mM for 2 and 2f, respectively, confirming the increased binding
potency observed from the KD values. The calculated IC50 values were cross validated using
the previously described colorimetric assay (see Figure S31 in the Supplementary Information).
N-BPs are the only commercially available FPPS inhibitors; they bind to the allylic sub-pocket
mimicking the pyrophosphate moiety of DMAPP/GPP.[7-9, 32] Because of their highly
charged nature and affinity for skeletal tissue, these drugs are limited to the treatment of bone-
related disorders such as osteoporosis and tumour-induced osteolytic metastases.[10]
The recent discovery of an FPPS allosteric pocket as a potential target of N-BP compounds has
fuelled the search for new FPPS inhibitors to overcome the pharmacokinetic limits of
bisphosphonate and become useful in broader therapeutic applications.[18, 19] As a result,
several non-bisphosphonate allosteric ligands have been identified, some of which show
chameleon-like behaviour, with mixed binding to both the catalytic and allosteric sites.[59-61]
Very recently, the X-ray structure of FPPS co-crystallized with 6-tolylthienopyrimidine (IC50
0.86 µM) highlighted the possibility of efficacious interactions at the FPPS binding site, mainly
based on consistent π-π stacking interactions and excluding the participation of the
phosphonate moiety.[16] NMR data and molecular docking calculations show that compounds
2a-m in the FPPS catalytic pocket have an orientation similar to that observed in the crystal
20

structure of FPPS co-crystallized with 6-tolylthienopyrimidine. The sugar and adenine ring
occupy the DMAPP sub-pocket, interacting with the Mg2+ ion and Asp residues of the
conserved aspartate-rich motif.[17, 34, 35] The p-tertbutyl benzyl ring occupies a large
hydrophobic region that engages in π-π stacking interactions with aromatic residues (F98, F99,
Y204). Although the IC50 and the KD values of 2 and 2f have, at the moment, modest relevance
from a drug-design perspective, the biological activity of 2 and 2f confirms that there are many
unexploited possibilities to identify new potent N-BP FPPS inhibitors, and in this way, N6-
substituted adenosine derivatives can be considered promising chemical scaffolds to be further
investigated.
Data collected on glioma cells indicated that 2a-m exerts cytostatic activity in the concentration
range of 10-20 µM, whereas 2 and 2f bind and inhibit FPPS at mM concentration values. On
the other hand, structure-activity relationship analysis indicates that the structural requirements
governing 2a-m cell cytotoxicity are different from those governing the interaction with the
FPPS enzyme. This discrepancy between the potency of the compounds in binding to and
inhibiting FPPS and the potency exhibited in a more complex cellular model was already
evident in our previous investigation. Here, the high efficiency of 2 to inhibit glioma cell
growth through cholesterol depletion provided a clear indication that 2 affects the MVA
metabolic pathway. However, even in this case, 2 showed greater effectiveness in cultured cells
compared to the in vitro FPPS assay. Thus, the previous and presently reported data support
the hypothesis that FPPS might not be the only target of the benzyladenosine compounds, and
other key enzymes and intermediate metabolites of the mevalonate and isoprenoid pathways
can be affected by their action.[62]
Increasing evidence has shown that inhibition of FPPS using N-BPs induces IPP accumulation,
and this, in turn, activates γδ T cells.[13] Monocytes that accumulate IPP become APC and
stimulate Vγ9Vδ2 T cells in the peripheral blood.[63] Therefore, blocking the mevalonate
21

pathway through the inhibition of FPPS has a final immunostimulatory effect. Therefore,
immunotherapy based on the combinatorial use of both IL-2 and Zol for the expansion of γδ T
cells has become a widely used approach in several cancer treatments.[14] To investigate the
ability of compounds 2a-m to affect the action of enzymes and intermediate metabolites in the
mevalonate and isoprenoid pathways, we measured the ability of compounds 2a-m to induce
expansion of γδ T cell cultures analogous to that achieved by interleukin-2 (IL-2) plus Zol,[11]
with N-BP used as control. As observed in Figure 9, 2a, 2f, 2h, and 2i exhibited an in vitro
stimulatory profile that was similar or improved compared to that achieved by Zol (Figure 9b);
analysis of γδ T cell viability after 14 days showed that these compounds induce toxicity
comparable to Zol; however, their effects were lower than that of 2.
Taken together, our data provide evidence that N6-benzyladenosine derivatives may play a role
similar to N-BPs in interacting with FPPS and thus interfering with the mevalonate pathway.
Therefore, these derivatives represent a good chemical scaffold to be further investigated for
the development of N-BP FPPS inhibitors, allowing the ex vivo expansion of cytotoxic cells to
be used in adoptive immunotherapy with a synergistic, discrete, cytostatic ability.
5. Methods
5.1 Molecular docking
AutoDock Vina (version 1.1.2)[36] was used for all docking calculations. The starting
conformations for docking studies of analogues 2 were built with Maestro (version 9.6).[64]
Three-dimensional models of each compound for the subsequent docking calculations were
preliminary optimized by the conjugate gradient (0.05 Å convergence threshold). After
optimization of the analogues, molecular charges were calculated by the Gasteiger-Marsili
method. Finally, the 3D structures were saved in .pdbqt format for molecular docking studies.
The 3D FPPS protein model was obtained from the Protein Data Bank database (PDB ID:
22

5YGI).[35] Water molecules were removed, and the obtained file was then processed with
AutoDock Tools 1.5.6, merging non-polar hydrogens and adding Gasteiger charges. For all
docked ligands, all bonds were considered rotatable. For an exhaustive exploration of
conformational space, all runs were performed with 300 iterations yielding 20 structures. For
all docking calculations, a grid box size of 26.51 Å (x, y, and z) was used and centred on the
target binding site (spatial coordinates: -16.92 x, 29.55 y, -9.78 z). The resulting data with the
most favourable free energies of binding were analysed. All 3D models were depicted using
Maestro 9.6.[64]
5.2 Chemistry
All reagents were purchased from Sigma-Aldrich (Milan, Italy) in the highest available purity
and were used as received. All reactions involving air or moisture-sensitive reagents were
carried out in a dry nitrogen atmosphere using freshly distilled solvents. DCM, CH3CN, and
methanol were distilled from CaH2. Dry DMF was purchased and used without further
distillation. When necessary, compounds were dried in vacuo over P2O5 or by the azeotropic
removal of water with toluene under reduced pressure. All MW reactions were conducted in a
CEM Explorer apparatus under monomode irradiation. Reaction temperatures were measured
externally; reactions were monitored by thin-layer chromatography (TLC) on Merck silica gel
plates (0.25 mm) and visualized by UV light, KMnO4, p-anisaldehyde, or ninhydrin solutions
and drying. Flash chromatography was performed on Merck silica gel 60 (particle size: 0.040-
0.063 mm), and the solvents employed were of analytical grade. Yields refer to
13
chromatographically and spectroscopically (1H- and C-NMR) pure compounds. Silica gel
(grade 60 PF254) was used for preparative TLC (PTLC). NMR spectra were generally recorded
at room temperature on Bruker Avance series 400 and 600 MHz spectrometers. Chemical shifts
(δ) are reported in ppm relative to the residual solvent peak (CHCl3, δ: 7.26, 13CDCl3, δ: 77.0;
23

CD2HOD, δ: 3.35, 13CD3OD, δ: 49.0), and the multiplicity of each signal is designated by the
following abbreviations: s, singlet; d, doublet; t, triplet; q, quartet; m, multiplet; br, broad; app,
apparent. Coupling constants (J) are quoted in Hz. High-resolution mass spectra (HRMS) were
recorded on a high-resolution mass spectrometer equipped with electrospray (ESI) and
nanospray sources, a quadrupole-time of flight hybrid analyser coupled with capillary UPLC
system (Q-TOF Premier/nanoAquity, Waters) in positive mode, and protonated molecular ions
[M+H]+ were used for empirical formula confirmation. Liquid chromatography was performed
on a Waters system (Milford, Massachusetts, USA) consisting of a Waters 486 tunable
absorbance detector and a Varian 9012 pump. Samples were prepared by dissolving
compounds 2a-m in methanol (0.5 mg/mL) at 0.2 mL/min. The injection volume was 10 μL.
Compounds 2a-m were analysed on a Symmetry® (C18 column (4.6 × 250 mm, 5 μm) under
gradient elution at a flow rate of 0.9 mL/min. The mobile phases consisted of 2% (v/v) acetic
acid in water (solvent A) and 2% (v/v) acetic acid in acetonitrile (solvent B). The following
gradient was used: 0–20 min 10–90% B, 20–24 min 90–10% B, and return to the initial
conditions over 3 min. UV detection was obtained at λ = 254 nm (Figure S30, See
Supplementary Information).
General procedure for the solvent free MW synthesis of N6-benzyladenosine derivatives
2a-m. Synthesis of (2R, 3R, 4S, 5R)-2-(6-(benzylamino)-9H-purin-9-yl)-5-(hydroxymethyl)
tetrahydrofuran-3,4-diol, 2. In a 7 mL MW vessel, 6-chloropurinoriboside 3 (20 mg, 0.07
mmol), benzylamine 4 (7.5 mg, 0.07 mmol, 7.7 µL) and triethylamine (7.08 mg, 0.07 mmol,
9.8 µL) were mixed. The solid mixture was stirred in CEM Explorer®. MW Method: T = 210
°C, Power: 300 W, Hold Time: 1 min, P = 250 PSI, Power Max activated. After cooling, the
solvent was removed in vacuo and the crude was dissolved in methanol and then purified on
PTLC (DCM/MeOH 9:1) to afford compound 2 as white solid (24 mg, 94%). 1H-NMR (600
24

MHz, MeOD) δ 8.30 (s, 1H), 8.28 (s, 1H), 7.43 (d, J = 7.5 Hz, 2H), 7.36 (t, J = 7.6 Hz, 2H),
7.29 (t, J = 7.3 Hz, 1H), 6.01 (d, J = 6.4 Hz, 1H), 4.82 – 4.77 (m, 1H), 4.37 (dd, J = 4.9, 2.5
Hz, 1H), 4.22 (d, J = 2.4 Hz, 1H), 3.93 (dd, J = 12.5, 2.3 Hz, 1H), 3.79 (dd, J = 12.5, 2.5 Hz,
1H) ppm. HRMS (ESI-Q-TOF) m/z [M + H]+ Calcd. for C17H20N5O4: 358.1510; Found
358.1515. Rt: 8.92 min.
Synthesis of (2R,3S,4R,5R)-2-(hydroxymethyl)-5-(6-((3-methylbenzyl)amino)-9H-purin-9-
yl)tetrahydrofuran-3,4-diol, 2a. Compound 2a was obtained as a white solid (24 mg, 92%) by
following general procedure for the solvent free MW synthesis of N6-benzyladenosine
derivatives by using 6-chloropurinoriboside 3 (20 mg, 0.07 mmol) 3-methylbenzylamine 4a
(8.48 mg, 0.07 mmol, 8.8 µL) and triethylamine (7.08 mg, 0.07 mmol, 9.8 µL). 1H-NMR (600
MHz, MeOD) δ 8.29 (d, J = 10.4 Hz, 2H), 7.22 (dd, J = 19.2, 9.3 Hz, 3H), 7.11 (d, J = 7.1 Hz,
1H), 6.01 (d, J = 6.3 Hz, 1H), 4.80 (dd, J = 13.0, 7.1 Hz, 3H), 4.39 – 4.36 (m, 1H), 4.22 (s,
1H), 3.93 (d, J = 12.5 Hz, 1H), 3.79 (d, J = 12.5 Hz, 1H) ppm. HRMS (ESI-Q-TOF) m/z [M +
H]+ Calcd. for C18H22N5O4: 372.1666; Found 372.1659. Rt: 9.59 min.
Synthesis of (2R,3R,4S,5R)-2-(6-(([1,1'-biphenyl]-3-ylmethyl)amino)-9H-purin-9-yl)-5-
(hydroxymethyl)tetrahydrofuran-3,4-diol, 2b. Compound 2b was obtained as a white solid (27
mg, 89%) by following general procedure for the solvent free MW synthesis of N6-
benzyladenosine derivatives by using 6-chloropurinoriboside 3 (20 mg, 0.07 mmol),3-
phenylbenzylamine 4b (14.66 mg, 0.07 mmol) and triethylamine (7.08 mg, 0.07 mmol, 9.8
µL). 1H-NMR (600 MHz, MeOD) δ 8.31 (s, 2H), 7.70 (s, 1H), 7.63 (d, J = 7.5 Hz, 2H), 7.55
(d, J = 7.2 Hz, 1H), 7.44 (dt, J = 18.8, 7.2 Hz, 4H), 7.36 (t, J = 7.2 Hz, 1H), 6.01 (d, J = 6.3
Hz, 1H), 4.80 (t, J = 5.6 Hz, 1H), 4.39 – 4.35 (m, 1H), 4.22 (s, 1H), 3.93 (d, J = 12.5 Hz, 1H),
3.79 (d, J = 12.5 Hz, 1H) ppm. HRMS (ESI-Q-TOF) m/z [M + H]+ Calcd. for C23H24N5O4:
434.1823; Found 434.1829. Rt: 12.19 min.
25

Synthesis
of
(2R,3R,4S,5R)-2-(6-((4-fluorobenzyl)amino)-9H-purin-9-yl)-5-
(hydroxymethyl)tetrahydrofuran-3,4-diol, 2c. Compound 2c was obtained as a white solid (24
mg, 92%) by following general procedure for the solvent free MW synthesis of N6-
benzyladenosine derivatives by using 6-chloropurinoriboside 3 (20 mg, 0.07 mmol), 4-fluoro-
benzylamine 4c (8.76mg, 0.07mmol, 8.0µL) and triethylamine (7.08 mg, 0.07 mmol, 9.8 µL).
1H-NMR (600 MHz, MeOD) δ 8.29 (d, J = 9.8 Hz, 4H), 7.45 (dd, J = 7.9, 5.7 Hz, 4H), 7.08 (t,
J = 8.7 Hz, 4H), 6.00 (d, J = 6.4 Hz, 2H), 4.81 – 4.77 (m, 3H), 4.37 (dd, J = 4.7, 2.3 Hz, 2H),
4.21 (d, J = 2.1 Hz, 2H), 3.93 (dd, J = 12.5, 2.0 Hz, 2H), 3.79 (dd, J = 12.5, 2.2 Hz, 2H) ppm.
HRMS (ESI-Q-TOF) m/z [M + H]+ Calcd. for C17H19N5O4: 376.1416; Found 376.1411. Rt:
9.00 min.
Synthesis
of
(2R,3R,4S,5R)-2-(6-((4-bromobenzyl)amino)-9H-purin-9-yl)-5-
(hydroxymethyl)tetrahydrofuran-3,4-diol, 2d. Compound 2d was obtained as a white solid (27
mg, 88%) by following general procedure for the solvent free MW synthesis of N6-
benzyladenosine derivatives by using 6-chloropurinoriboside 3 (20 mg, 0.07 mmol), 4-bromo-
1
benzylamine 4d (13.0 mg, 0.07 mmol) and triethylamine (7.08 mg, 0.07 mmol, 9.8 µL). H-
NMR (600 MHz, MeOD) δ 8.32 (d, J = 13.8 Hz, 2H), 7.53 (d, J = 7.8 Hz, 2H), 7.38 (d, J = 7.7
Hz, 2H), 6.03 (d, J = 5.9 Hz, 1H), 4.82 (d, J = 5.4 Hz, 2H), 4.39 (s, 1H), 4.24 (s, 1H), 3.96 (d,
J = 12.5 Hz, 1H), 3.82 (d, J = 12.5 Hz, 1H) ppm. HRMS (ESI-Q-TOF) m/z [M + H]+ Calcd.
for C17H19BrN5O4: 436.0615; Found 436.0611. Rt: 10.53 min.
Synthesis of (2R,3S,4R,5R)-2-(hydroxymethyl)-5-(6-((4-methylbenzyl)amino)-9H-purin-9-
yl)tetrahydrofuran-3,4-diol, 2e. Compound 2e was obtained as a white solid (23 mg, 89%) by
following general procedure for the solvent free MW synthesis of N6-benzyladenosine
derivatives by using 6-chloropurinoriboside 3 (20 mg, 0.07 mmol), 4-methyl-benzylamine 4e
(8.5 mg, 0.07 mmol, 9.0 µL) and triethylamine (7.08 mg, 0.07 mmol, 9.8 µL). 1H-NMR (600
MHz, MeOD) δ 8.15 (d, J = 18.7 Hz, 2H), 7.15 (d, J = 7.7 Hz, 2H), 7.03 (d, J = 7.5 Hz, 2H),
26

5.85 (d, J = 6.5 Hz, 1H), 4.64 (t, J = 5.4 Hz, 3H), 4.21 (d, J = 2.8 Hz, 1H), 4.06 (s, 1H), 3.78
(d, J = 11.6 Hz, 1H), 3.64 (d, J = 11.3 Hz, 1H) ppm. HRMS (ESI-Q-TOF) m/z [M + H]+ Calcd.
for C18H22N5O4: 372.1666; Found 372.1661. Rt: 9.73 min.
Synthesis
of
(2R,3R,4S,5R)-2-(6-((4-(tert-butyl)benzyl)amino)-9H-purin-9-yl)-5-
(hydroxymethyl)tetrahydrofuran-3,4-diol, 2f. Compound 2f was obtained as a white solid (26
mg, 91%) by following general procedure for the solvent free MW synthesis of N6-
benzyladenosine derivatives by using 6-chloropurinoriboside 3 (20 mg, 0.07 mmol), 4-tert-
butyl-benzylamine 4f (17.13 mg, 0.07 mmol) and triethylamine (7.08 mg, 0.07 mmol, 9.8 µL).
1H-NMR (600 MHz, MeOD) δ 8.29 (d, J = 7.6 Hz, 2H), 7.41 (d, J = 7.9 Hz, 2H), 7.35 (d, J =
8.0 Hz, 2H), 6.00 (d, J = 6.3 Hz, 1H), 4.81 – 4.77 (m, 2H), 4.39 – 4.34 (m, 1H), 4.21 (s, 1H),
3.93 (d, J = 12.5 Hz, 1H), 3.79 (d, J = 12.5 Hz, 1H) ppm. HRMS (ESI-Q-TOF) m/z [M + H]+
Calcd. for C21H28N5O4: 414.2136; Found 414.2140. Rt: 12.61 min.
Synthesis of (2R,3S,4R,5R)-2-(hydroxymethyl)-5-(6-((quinolin-2-ylmethyl)amino)-9H-purin-
9-yl)tetrahydrofuran-3,4-diol, 2g. Compound 2g was obtained as a white solid (25 mg, 87%)
by following general procedure for the solvent free MW synthesis of N6-benzyladenosine
derivatives by using 6-chloropurinoriboside 3 (20 mg, 0.07 mmol), 2-Quinolinemethanamine
1
4g (11.07 mg, 0.07 mmol) and triethylamine (7.08 mg, 0.07 mmol, 9.8 µL). H-NMR (600
MHz, MeOD) δ 8.33 (dd, J = 22.2, 13.7 Hz, 9H), 8.10 (d, J = 8.5 Hz, 3H), 7.95 (d, J = 8.1 Hz,
3H), 7.80 (t, J = 7.6 Hz, 3H), 7.61 (dd, J = 16.0, 8.1 Hz, 6H), 6.03 (d, J = 6.3 Hz, 3H), 4.81 (t,
J = 5.6 Hz, 4H), 4.38 (d, J = 2.3 Hz, 3H), 4.22 (s, 4H), 3.93 (d, J = 12.4 Hz, 4H), 3.81 (dd, J =
19.7, 12.5 Hz, 4H) ppm. HRMS (ESI-Q-TOF) m/z [M + H]+ Calcd. for C20H21N6O4: 409.1619;
Found 409.1612. Rt: 6.93 min.
Synthesis
of
(2R,3R,4S,5R)-2-(6-((4-chlorobenzyl)amino)-9H-purin-9-yl)-5-
(hydroxymethyl)tetrahydrofuran-3,4-diol, 2h. Compound 2h was obtained as a white solid (25
mg, 90%) by following general procedure for the solvent free MW synthesis of N6-
27

benzyladenosine derivatives by using 6-chloropurinoriboside 3 (20 mg, 0.07 mmol), 4-chloro-
benzylamine 4h (9.9 mg, 0.07 mmol, 8.85 µL) and triethylamine (7.08 mg, 0.07 mmol, 9.8
µL). 1H-NMR (600 MHz, MeOD) δ 8.29 (d, J = 15.4 Hz, 2H), 7.42 (d, J = 8.3 Hz, 2H), 7.36
(d, J = 8.4 Hz, 2H), 6.00 (d, J = 6.4 Hz, 1H), 4.81 – 4.77 (m, 1H), 4.37 (dd, J = 5.0, 2.5 Hz,
1H), 4.21 (d, J = 2.4 Hz, 1H), 3.93 (dd, J = 12.5, 2.3 Hz, 1H), 3.79 (dd, J = 12.5, 2.5 Hz, 1H)
ppm. HRMS (ESI-Q-TOF) m/z [M + H]+ Calcd. for C17H19ClN5O4: 392.1120; Found
392.1127. Rt: 10.34 min.
Synthesis of (2R,3S,4R,5R)-2-(hydroxymethyl)-5-(6-((4-nitrobenzyl)amino)-9H-purin-9-
yl)tetrahydrofuran-3,4-diol, 2i. Compound 2i was obtained as a white solid (27 mg, 88%) by
following general procedure for the solvent free MW synthesis of N6-benzyladenosine
derivatives by using 6-chloropurinoriboside 3 (20 mg, 0.07 mmol), 4-nitro-benzylamine 4i
(13.20 mg, 0.07 mmol) and triethylamine (7.08 mg, 0.07 mmol, 9.8 µL). 1H-NMR (600 MHz,
MeOD) δ 8.33 (s, 1H), 8.28 (s, 1H), 8.25 (s, 1H), 8.23 (s, 1H), 7.66 (d, J = 8.6 Hz, 2H), 6.01
(d, J = 6.4 Hz, 1H), 4.81 – 4.77 (m, 1H), 4.37 (dd, J = 5.0, 2.5 Hz, 1H), 4.22 (d, J = 2.4 Hz,
1H), 3.93 (dd, J = 12.5, 2.4 Hz, 1H), 3.79 (dd, J = 12.5, 2.6 Hz, 1H) ppm. HRMS (ESI-Q-TOF)
m/z [M + H]+ Calcd. for C17H19N6O6: 403.1361; Found 403.1354. Rt: 8.84 min.
Synthesis of (2R,3S,4R,5R)-2-(hydroxymethyl)-5-(6-((4-methoxybenzyl)amino)-9H-purin-9-
yl)tetrahydrofuran-3,4-diol, 2j. Compound 2j was obtained as a white solid (25 mg, 88%) by
following general procedure for the solvent free MW synthesis of N6-benzyladenosine
derivatives by using 6-chloropurinoriboside 3 (20 mg, 0.07 mmol), 4-methoxy-benzylamine 4j
(9.6 mg, 0.07 mmol, 9.3µL) and triethylamine (7.08 mg, 0.07 mmol, 9.8 µL). 1H-NMR (600
MHz, MeOD) δ 8.30 (s, 2H), 7.36 (d, J = 8.5 Hz, 2H), 6.92 (d, J = 8.6 Hz, 2H), 6.00 (d, J =
6.4 Hz, 1H), 4.82 – 4.77 (m, 2H), 4.37 (dd, J = 5.0, 2.5 Hz, 1H), 4.22 (d, J = 2.4 Hz, 1H), 3.93
(dd, J = 12.5, 2.3 Hz, 1H), 3.82 (s, 3H), 3.79 (dd, J = 12.6, 2.6 Hz, 1H) ppm. HRMS (ESI-Q-
TOF) m/z [M + H]+ Calcd. for C18H22N5O5: 388.1615; Found 388.1608. Rt: 8.93 min.
28

Synthesis
of
(2R,3R,4S,5R)-2-(6-((4-ethoxybenzyl)amino)-9H-purin-9-yl)-5-
(hydroxymethyl)tetrahydrofuran-3,4-diol, 2k. Compound 2k was obtained as a white solid (24
mg, 84%) by following general procedure for the solvent free MW synthesis of N6-
benzyladenosine derivatives by using 6-chloropurinoriboside 3 (20 mg, 0.07 mmol), 4-ethoxy-
benzylamine 4k (13.58 mg, 0.07 mmol) and triethylamine (7.08 mg, 0.07 mmol, 9.8 µL). 1H-
NMR (600 MHz, MeOD) δ 8.29 (s, 2H), 7.34 (d, J = 7.9 Hz, 2H), 6.91 (d, J = 7.8 Hz, 2H),
6.00 (d, J = 6.3 Hz, 1H), 4.81 – 4.77 (m, 2H), 4.38 – 4.35 (m, 1H), 4.21 (s, 1H), 4.05 (q, J =
6.8 Hz, 2H), 3.93 (d, J = 12.5 Hz, 1H), 3.79 (d, J = 12.5 Hz, 1H) ppm. HRMS (ESI-Q-TOF)
m/z [M + H]+ Calcd. for C19H24N5O5: 402.1772; Found 402.1779. Rt: 9.66 min.
Synthesis of (2R,3R,4S,5R)-2-(6-((3,5-bis(trifluoromethyl)benzyl)amino)-9H-purin-9-yl)-5-
(hydroxymethyl)tetrahydrofuran-3,4-diol, 2l. Compound 2l was obtained as a white solid (31
mg, 90%) by following general procedure for the solvent free MW synthesis of N6-
benzyladenosine derivatives by using 6-chloropurinoriboside 3 (20 mg, 0.07 mmol), 3,5-
bistrifluoromethyl-benzylamine 4l (17.0 mg, 0.07 mmol) and triethylamine (7.08 mg, 0.07
mmol, 9.8 µL). 1H-NMR (600 MHz, MeOD) δ 8.34 (s, 1H), 8.30 (s, 1H), 8.04 (s, 2H), 7.89 (s,
1H), 6.02 (d, J = 6.3 Hz, 1H), 4.80 (t, J = 5.6 Hz, 1H), 4.38 – 4.36 (m, 1H), 4.21 (s, 1H), 3.93
(d, J = 12.5 Hz, 1H), 3.79 (d, J = 12.5 Hz, 1H) ppm. HRMS (ESI-Q-TOF) m/z [M + H]+ Calcd.
for C19H18F6N5O4: 494.1257; Found 494.1263. Rt: 13.03 min.
Synthesis
of
(2R,3R,4S,5R)-2-(6-((3,5-dimethoxybenzyl)amino)-9H-purin-9-yl)-5-
(hydroxymethyl)tetrahydrofuran-3,4-diol, 2m. Compound 2m was obtained as a white solid
(26 mg, 89%) by following general procedure for the solvent free MW synthesis of N6-
benzyladenosine derivatives by using 6-chloropurinoriboside 3 (20 mg, 0.07 mmol), 3,5-
bismethoxy-benzylamine 4m (11.70 mg, 0.07 mmol) and triethylamine (7.08 mg, 0.07 mmol,
1
9.8 µL). H-NMR (600 MHz, MeOD) δ 8.30 (d, J = 12.4 Hz, 2H), 6.60 (d, J = 1.8 Hz, 2H),
6.42 (s, 1H), 6.01 (d, J = 6.4 Hz, 1H), 4.82 – 4.78 (m, 2H), 4.37 (dd, J = 4.9, 2.4 Hz, 1H), 4.22
29