Conjugable A3 adenosine receptor antagonists for the development of functionalized ligands and their use in fluorescent probes

Article abstract of DOI:10.1016/j.ejmech.2019.111886

Compounds able to simultaneously bind a biological target and be conjugated to a second specific moiety are attractive tools for the development of multi-purpose ligands useful as multi-target ligands, receptor probes or drug delivery systems, with both therapeutic and diagnostic applications. The human A₃ adenosine receptor is a G protein-coupled receptor involved in many physio-pathological conditions, e.g. cancer and inflammation, thus representing a promising research target. In this work, two series of conjugable hA₃AR antagonists, based on the pyrazolo[4,3-e]-1,2,4-triazolo[1,5-c]pyrimidine nucleus, were developed. The introduction of an aromatic ring at the 5 position of the scaffold, before (phenylacetamido moiety) or after (1,2,3-triazole obtained by click chemistry) the conjugation is aimed to increase affinity and selectivity towards the hA₃AR receptor. As expected, conjugable compounds showed good affinity towards the hA₃AR. In order to prove their potential in the development of hA₃AR ligands for different purposes, compounds were also functionalized with fluorescent probes. Unfortunately, conjugation decreased affinity and selectivity for the target as compared to the hA_2AAR. Computational studies identified specific non-conserved residues of the extracellular loops which constitute a structural barrier able to discriminate between ligands, giving insights into the rational development of new highly selective ligands.

Full text of DOI:10.1016/j.ejmech.2019.111886

Journal Pre-proof
Conjugable A adenosine receptor antagonists for the development of functionalized
3
ligands and their use in fluorescent probes
Stephanie Federico, Enrico Margiotta, Stefano Moro, Eszter Kozma, Zhan-Guo Gao,
Kenneth A. Jacobson, Giampiero Spalluto
PII:
S0223-5234(19)31038-4
DOI:
https://doi.org/10.1016/j.ejmech.2019.111886
EJMECH 111886
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 18 September 2019
Revised Date: 12 November 2019
Accepted Date: 12 November 2019
Please cite this article as: S. Federico, E. Margiotta, S. Moro, E. Kozma, Z.-G. Gao, K.A. Jacobson,
G. Spalluto, Conjugable A adenosine receptor antagonists for the development of functionalized
3
ligands and their use in fluorescent probes, European Journal of Medicinal Chemistry (2019), doi: https://
doi.org/10.1016/j.ejmech.2019.111886.
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 Masson SAS.

Conjugable A3 adenosine receptor antagonists for the development of functionalized ligands and
their use in fluorescent probes
Stephanie Federico, *# Enrico Margiotta, § Stefano Moro, § Eszter Kozma, ≠ Zhan-Guo Gao, ≠ Kenneth
A. Jacobson, ≠ and Giampiero Spalluto#
#Dipartimento di Scienze Chimiche e Farmaceutiche, Università degli Studi di Trieste, Via Licio
Giorgeri 1, 34127 Trieste, Italy
§Molecular Modeling Section (MMS), Dipartimento di Scienze del Farmaco, Università di Padova, via
Marzolo 5, 35131 Padova, Italy
≠Laboratory of Bioorganic Chemistry, NIDDK, National Institute of Health, Bethesda, Maryland, USA
To whom correspondence should be addressed.
Tel +39 040 5583671 (S.F.), Fax +39 040 5583903, E-mail: sfederico@units.it
Keywords: G protein-coupled receptor; adenosine receptors; fluorescent ligands; molecular modeling;
molecular probes.
Abbreviations: AR, adenosine receptor; BOC, tert-butyloxycarbonyl; EDCI, 1-ethyl-3-(3-
dimethylaminopropyl)carbodiimide; FITC, fluorescein isothiocyanate; PTP, pyrazolo[4,3-e]-1,2,4-
triazolo[1,5-c]pyrimidine;

Abstract
Compounds able to simultaneously bind a biological target and be conjugated to a second specific
moiety are attractive tools for the development of multi-purpose ligands useful as multi-target ligands,
receptor probes or drug delivery systems, with both therapeutic and diagnostic applications. The human
A3 adenosine receptor is a G protein-coupled receptor involved in many physio-pathological conditions,
e.g. cancer and inflammation, thus representing a promising research target. In this work, two series of
conjugable hA3AR antagonists, based on the pyrazolo[4,3-e]-1,2,4-triazolo[1,5-c]pyrimidine nucleus,
were developed. The introduction of an aromatic ring at the 5 position of the scaffold, before
(phenylacetamido moiety) or after (1,2,3-triazole obtained by click chemistry) the conjugation is aimed
to increase affinity and selectivity towards the hA3AR receptor. As expected, conjugable compounds
showed good affinity towards the hA3AR. In order to prove their potential in the development of
hA3AR ligands for different purposes, compounds were also functionalized with fluorescent probes.
Unfortunately, conjugation decreased affinity and selectivity for the target as compared to the hA2AAR.
Computational studies identified specific non-conserved residues of the extracellular loops which
constitute a structural barrier able to discriminate between ligands, giving insights into the rational
development of new highly selective ligands.

Introduction
Activation or blockade of adenosine receptors (ARs) are responsible for diverse pharmacological
effects in various tissues and organs. Considering the wide distribution of endogenous adenosine, it is
generally accepted that the regulation of ARs has great therapeutic potential [1–6].
ARs are members of the superfamily of G protein-coupled receptors (GPCRs), and four subtypes
are currently known, named A1AR, A2AAR, A2BAR, and A3AR [7]. In recent decades, numerous
medicinal chemistry groups have developed promising ligands, agonists and antagonists for these
receptor subtypes [8]. In particular, the search for selective antagonists held greater appeal than
selective agonists, not only for their potential therapeutic applications but also for their preferential use
as molecular probes for pharmacological characterization of receptors [9].
In this field, our research group and others have extensively investigated the pyrazolo[4,3-e]-
1,2,4-triazolo[1,5-c]pyrimidine nucleus (PTP) that, by appropriate modifications at the N7, N8, and 5
positions, led to very potent and selective human (h) A2A and A3 AR antagonists [10–16]. These
derivatives have previously been used for the study of the functional role of A2A and A3 ARs, in
particular through the preparation of radiolabeled or irreversible derivatives [17–21]. Nevertheless,
another interesting approach for the investigation of AR subtypes could be considered the
functionalization of the PTP scaffold with spacers able to link various groups such as fluorophore or
nanoparticles for drug delivery [22–32].
Using the PTP scaffold we have recently reported a new series of derivatives of general formulae 1, in
which a diamino moiety of different lengths was introduced at the 5 position. Using this approach, a
new series of derivatives bearing a fluorophore group linked to the free amino group was reported as
potential fluorescent probes (Figure 1) [33].
Figure 1. Structures of previously reported AR antagonists and reference compounds.
In this work, we explored the structure activity relationship (SAR) of conjugable AR ligands useful for

the development of probes selective for the hA3AR. In addition to their usefulness to better define the
specific role of the target protein in both physiological and pathological processes, availability of
hA3AR selective probes could be used for diagnostic purposes. In fact, hA3AR was found to be
overexpressed in certain cancers, such as melanoma and breast and colorectal cancer, and the
overexpression level well correlates with the severity of the tumor, thus making it a possible biomarker
of such tumors [34–38]. Unfortunately, the previously synthesized PTP derivatives generally showed a
poor affinity for the hA3AR and low A3AR selectivity versus the other receptor subtypes [33]. A
possible explanation of lack of affinity and selectivity of these compounds could be attributed to the
absence of an aromatic moiety at the 5 position, which previous studies indicated to be important for
both affinity and selectivity at the hA3AR [10].
Taking into account these experimental observations and with the aim of improving selectivity versus
the hA3AR subtype, we decided to design a new series of conjugable PTPs bearing an aromatic portion
at the 5 position (3-18) and anchored to chains of different lengths. These chains act as spacers and at
their terminal position bear a functional group for subsequent functionalization. On the basis of our
modeling of the AR interactions of PTP ligands, the most reliable orientation of the linker is towards
the extracellular loops (ELs) of the receptors [33]. This means that any difference in sequence and
conformation of the ELs of the AR subtypes could affect either selectivity or affinity through direct
ligand interaction. Indeed, our recent work about the selection of the best template for hA3AR
homology modeling highlighted that EL region displays the highest sequence variability between the
hA1, hA2A, and hA3AR subtypes [39], suggesting that the ligand partitioning between the ELs and the
hydrophobic cavity is crucial for binding.
Thus, finally, we decided to conjugate the new functionalized AR ligands with a fluorescent moiety
(e.g. fluorescein) that is able to protrude in the extracellular side, giving new leads for the development
of hA3AR fluorescent probes (Figure 2).
Alternatively, a parallel series of N5 alkynyl functionalized antagonists (19-22) was prepared as
precursors for copper(I)-catalyzed azide-alkyne (click) cycloaddition. Based on previous studies
performed on the triazoloquinazoline nucleus of the nonseletive antagonist CGS15943, which led to a
potent and selective fluorescent hA3AR antagonists (MRS5449, 2, Figure 1) [40], a fluorophore (e.g.
Alexa Fluor-488) was introduced to afford potential A3AR fluorescent antagonist 23 (Figure 2).

Figure 2. Structures of the synthesized PTP derivatives.
Chemistry
All the designed compounds 3-23 have been synthesized as summarized in Schemes 1-3 [41].
Saponification with lithium hydroxide of previously reported PTP ester derivative 24 [42] led to the
corresponding carboxylic acid 3, which after condensation in presence of EDCI with the appropriate
mono BOC-protected diamine 25-29 afforded the desired functionalized compounds 4-8. A common
BOC deprotection with trifluoroacetic acid in dichloromethane provided the free amino compounds 9-
13 as salts. Then, after coupling with fluorescein isothiocyanate (FITC) in methanol under an argon
atmosphere, the final compounds 14-18 were obtained. (Scheme 1)
Alkynyl PTP derivatives 19-22 were instead prepared by reacting the free amino PTP derivative 30
with the appropriate acyl chlorides 35-38 which were obtained by treatment of the corresponding
carboxylic acids 31-34 with oxalyl chloride (Scheme 2).

H
N
NHBoc
OEt
OH
X
O
O
O
HN
HN
HN
i
ii
O
N
HN
O
N
O
N
HN
HN
O
O
O
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
24
3
4-8
iii
H
H
H
N
H
N
+
O
N
N
NH3
X
X
S
O
O
HN
HN
CF3COO-
COO-
O
iv
O
O
N
HN
HN
O
O
N
N
N
N
N
OH
N
N
N
N
N
N
14-18
9-13
X= (CH2)3, (CH2)4, (CH2)5, (CH2CH2O)CH2CH2, (CH2)3O(CH2CH2O)2(CH2)3
Scheme 1. Synthesis of fluorescent adenosine receptor ligands 14-18. Reagents and conditions: i:
LiOH, THF, MeOH, water, rt. overnight; ii: NH2XNHBoc (25-29), EDCI, DMAP, DMF, rt, overnight;
iii: TFA, DCM, rt, 2h; iv: FITC, TEA, MeOH, rt, 72h.
Scheme 2. Synthesis of alkyne derivatives 19-22. Reagents and conditions: i: oxalyl chloride, DMF
cat., CDCl3, 0°C to rt, 3 h; ii: TEA. dioxane, reflux, overnight.

The final derivative 23 was prepared by reacting the alkynyl derivative 22 with Alexa Fluor-488 5-
carboxamido (6-azidohexanyl)bis(triethylammonium salt) in DMF [40]. (Scheme 3)
O
N
N
O
N
NH2
SO3
N
O
N
N
HN
H
-
HN
O
N
N
2(Et3NH)+
i
N
-
CO2
O
N
N
O
N
N
N
-
N
N
SO3
+
NH2
22
23
Scheme 3. Synthesis of Alexa Fluor-488 fluorescent AR ligand 23. Reagents and conditions: i: Alexa
Fluor 488 5-carboxamido-(6-azidohexanyl) bis(triethylammonium salt), aq. sodium ascorbate, aq.
cupric sulfate pentahydrate solution, water, DMF, rt, overnight.
Results and Discussion
Newly synthesized compounds (3-23) were tested at the hA1, hA2A and hA3 ARs expressed in CHO
(hA1AR, hA3AR) and HEK293 (hA2AAR) cells: [3H]RPIA (hA1AR), [3H]CGS21680 (hA2AAR) and
[125I]AB-MECA (hA3AR) were used as radioligands in binding assays (Table 1) [40,43].
Table 1. Binding profile of synthesized compounds 3-23 at the hA1, hA2A and hA3 ARs (see figure 2
for compounds’ structures).
a
b
c
Compd
X
R
hA1 hA2A hA3 hA1/hA3 hA2A/hA3 hA1/hA2A
3
4
5
6
7
8
9
-
(CH2)3
-
4% 1,010 128
> 78
7.8
54.5
88
> 10
> 66
Boc
Boc
Boc
Boc
15%
2%
150 2.75 > 3,636
304 3.46 > 2,890
376 4.47 > 2,237
456 74.8 > 133.6
327 82.4 > 121.3
175 8.06 > 1,240
(CH2)4
> 33
(CH2)5
14%
7%
84.1
6.1
> 26.5
> 22
(CH2CH2O)2CH2CH2
(CH2)3O(CH2CH2O)2(CH2)3 Boc
Hd
11%
9%
3.9
> 30.5
> 57
(CH2)3
21.7

(CH2)4
(CH2)5
Hd
Hd
Hd
4%
6%
172 6.00 > 1,666
203 2.66 > 3,759
28.6
76.3
14
> 58
> 49
> 37.4
> 30
> 96
> 129
> 94
> 94
> 109
> 2.34
-
10
11
12
13
14
15
16
17
18
19
20
21
22
(CH2CH2O)2CH2CH2
17%
1%
267 19.2
328 10.3
110 104
> 520
> 970
> 96
(CH2)3O(CH2CH2O)2(CH2)3 Hd
30.3
1
(CH2)3
(CH2)4
FITC 27%
FITC 28% 60.4 73.6 > 23.7
0.82
1.1
(CH2)5
FITC 21%
106 96.7
> 103
> 73
> 48
> 909
-
(CH2CH2O)2CH2CH2
FITC 29% 95.7 137
0.69
0.44
16.5
48.4
20.7
37.6
(CH2)3O(CH2CH2O)2(CH2)3 FITC 32% 91.5 207
(CH2)2
(CH2)3
(CH2)4
(CH2)5
-
-
-
-
55%
n.d.
133 11.0
199 4.11
151 7.26
482 12.8
891
122.7
> 781
5.9
62%
> 20.7
23e
-
-
n.d.
90
31.8
-
2.83
-
aDisplacement of specific [3H]R-PIA binding at hA1AR expressed in CHO cells, (n=3-6). bDisplacement of specific
[3H]CGS21680 binding at hA2AAR expressed in CHO cells, (n=3-6). cDisplacement of specific [125I]-AB-MECA binding at
hA3AR expressed in CHO cells, (n=3-6). Data are expressed as Ki (nM) or as % of inhibition (in italics) at a 10 ꢀM
d
concentration of radioligand. n.d. Not determined. Data complete of SEM are reported in ESI (Table S1). Compounds as
TFA salts. e MRS5763.
As clearly summarized in Table 1, all the synthesized ureido (3-18) and amido (19-23) compounds
showed affinities at the hA3AR ranging from high nanomolar to nanomolar concentrations, with
different degrees of selectivity versus the other subtypes.
The presence of a carboxylic group at the 5 position of compound 3 led to affinity values of 128 nM at
the hA3AR and 1.01 µM at the hA2AAR, while it was inactive at the hA1AR (4% of radioligand
displacement at 10 µM). Among all the series, compound 3 is the only one to exhibit a micromolar
affinity against the hA2AAR, suggesting that the carboxylic moiety is involved in interactions with

different key residues in the orthosteric binding site of the AR subtypes, which are, thus, responsible
for the observed selectivity.
Conjugation of the acid moiety with the mono-N-BOC protected diamino spacers led to compounds 4-8
which showed high affinity for the hA3AR (Table 1). It should be noted that the length of the diamino
spacer could be a major determinant of the hA3AR interaction. In fact, while shorter spacers (e.g.
compounds 4-6) gave good results in terms of both affinity (Ki hA3 = 2.75-4.47 nM) and selectivity
(hA2A/hA3 = 54.5-88) for the hA3AR (e.g. compound 3, Ki hA3 = 2.75 nM, hA1/hA3 = > 3,636;
hA2A/hA3 = 54.5), longer diamino moieties (e.g compounds 7, 8) led to derivatives showing a
significant loss of affinity at the hA3AR with a consequent reduction of selectivity, especially towards
the hA2AAR subtype (e.g. compound 8, Ki hA3 = 82.4 nM, hA1/hA3 > 133; hA2A/hA3 = 6.1).
However, the corresponding N-BOC deprotected derivatives 9-13 showed good hA3AR affinity
independently of the spacer length, albeit, in general, the selectivity vs the hA2AAR subtype was
significantly reduced with respect to the N-BOC derivatives (4-8) (e.g. compound 10, Ki hA3 = 6 nM,
hA1/hA3 = > 1,666; hA2A/hA3 = 28 vs. compound 5, Ki hA3 = 3.46 nM, hA1/hA3 = > 2,890; hA2A/hA3
= 88). Interestingly, compounds bearing PEG-like spacers (7, 8 and 12, 13) showed a different
selectivity profile, denoting that also the nature of the spacer was important for the AR interactions, in
particular for the hA2A and hA3 ARs. In fact, while in the BOC protected series a very low selectivity
was observed (compound 7, hA2A/hA3 = 6.1 and compound 8, hA2A/hA3 = 3.9), in the free amino series,
compounds gained affinity and selectivity towards the hA3AR subtype (compound 12, hA2A/hA3 = 14
and compound 13, hA2A/hA3 = 30.3). In order to assess if the nature, and in particular the polarity of
the linker could be someway related to the affinity profile displayed by compounds, the logarithm of
the octanol/water partition coefficient (logP(o/w)) was calculated for each compound (Figure 3). As
expected, Boc-free ligands showed lower values than Boc-functionalized ones, with polyethers being
more polar than alkyls. Only for the Boc-free mini-series, the affinity at the hA3AR increased with the
logP(o/w) value of the compound, while in the other homologous series a correlation between polarity
and affinity at the three adenosine receptors was not observed.

Figure 3. Calculated octanol/water partition coefficient logarithm.
Concerning the fluorescent ligands, introduction of a fluorophore such as FITC on amino function gave
compounds 14-18 which showed a significant reduced hA3AR affinity with a drastic reduction of
selectivity vs hA2AAR, independently of the spacer length. (e.g. compound 16, Ki hA3 = 96.7 nM,
hA1/hA3 = > 103; hA2A/hA3 = 1.1). Thus, they can be considered as dual hA2A/hA3AR fluorescent
ligands.
In the amido alkynyl series (19-22), all the synthesized compounds proved to be quite potent at the
hA3AR with affinity ranging from 4 to 12 nM and poor affinities versus the other AR subtypes, even if
the hA2AAR selectivity was not pronounced (e.g. compound 21, Ki hA3 = 7.26 nM, hA1/hA3 = 122;
hA2A/hA3 = 20.7). These results suggest that the triple bond likely formed interactions with
nonconserved residues in the hA3AR binding site.
Finally, Alexa Fluor-488 conjugated derivative 23, a potential fluorescent ligand, retained a good
affinity at the hA3AR (Ki hA3 = 31.8 nM) and selectivity towards the hA2AAR was improved (hA1/hA3
= 2.83) compared to FITC derivatives (19-22), even if further optimization is needed (Table 1).
It is quite clear that, in the present case, it was not possible to obtain a conjugable, selective ligand
which could be differentially functionalized and that retained the selectivity and affinity profile of the
parent compound. Even if the functional moiety (such as the fluorophore), was separated from the
pharmacophore through a long spacer, it influenced the ligand-receptor recognition process.
Molecular modelling studies were performed with a view to rationalize affinity data and SAR of the
tested ligands. Molecular docking simulations were performed on each AR subtype, using inactive
states, in the presence of the sodium ion and its hydration shell [44].

Figure 4. IE map of the electrostatic and hydrophobic interactions between each compound (y-axis)
and each residue (x-axis) of AR subtypes (hA1, A; hA2A, B; hA3, C) The strength of the electrostatic
interaction is represented by a colorimetric scale going from blue to red, from negative to positive
values. The strength of the hydrophobic interaction is represented by a colorimetric scale going from
white to dark green, from low to high values.
Independently of the receptor subtype considered, the ligands retained the classical binding pattern of

AR antagonists: hydrogen bonds with Asn (6.55) and π-π interactions with Phe (EL2) (Videos S1-3).
Moreover, the linker was generally oriented towards the ELs. Despite the overall findings, in the case
of the hA1AR (Video S1), several binding modes result energetically unfavorable or geometrically
unreliable (e.g. pointing the furyl ring towards the ELs), consistent with the high A3AR selectivity
observed vs A1. Interaction energy fingerprint (IEF) maps calculated for each docking pose on three
subtypes, revealed that docking at the hA1AR (Figure 4, A) was less favorable than at hA2A and hA3
ARs (Figure 4, B-C), because of unfavorable electrostatics at the EL2 and TMs 5-7. hA3AR ligand
poses were highly favored at these locations, more than at hA2A and hA1 ARs, also in terms of
hydrophobic contacts. Furthermore, as explained below, in the hA3AR, TM1 contributed significantly
to stabilization by electrostatics, differently from other subtypes.
Per-residue interaction energy profiles (electrostatic, van der Waals and hydrophobic contributions)
were calculated for three representative antagonists: 3 (EA2), 16 (FITC ligand, EA8) and 22 (ethynyl
ligand, EA20). Both conserved and non-conserved residues for each receptor subtype were chosen
from TMs and ELs, in order to give a better understanding of how they affect binding affinity and
selectivity, as shown in Figures 5-7.
Antagonist 3 (EA2) is the parent compound of the series. Nonetheless, it is the only one to show
affinity towards the hA2AAR in the micromolar range (Ki hA2A = 1.01
ꢀM, Ki hA3 = 128 nM).
Interestingly, it is characterized by a negatively charged carboxylic moiety. When docked to the hA1AR,
the ligand pose is highly disfavoured by strong repulsion with Glu172 (EL2) and other polar residues
(Figure 5). Similarly, at the hA2AAR, Glu169 (EL2) disfavors the final state. The mutation of Glu into
Val169 at the EL2, and the mutation of Val in Leu90 (3.32) at the TM3 of the hA3AR are crucial in
improving complex stability by hydrophobic interactions. These results predict that negatively charged
moieties, exposed to the ELs of the hA1 and hA2A ARs, inhibit binding by electrostatic repulsion with
Glu (EL2). Finally, a favourable electrostatic interaction is observed with the conserved residue Glu19
(1.39) at the TM1 of the hA3AR, while the same is absent in other subtypes, giving indications about
the different shape and nature of the A3 binding site. The antagonist 16 (EA8, Figures 6 and 7), bearing
a longer linker and a fluorophore moiety (FITC), shows strong van der Waals repulsion with Glu172
(EL2) and Asn254 (6.55) at the hA1AR, while such interactions are favourable at the hA2AR (Glu169,
Asn253) and accompanied by several hydrophobic contacts (Leu267 (6.70), Val92 (3.32), Ile74 (3.14)).
The hA3AR shows a similar profile, but hydrophobic interactions are not significant, consistently with
the absence of selectivity observed, as compared to A2A. However, nonconserved residues Gly257,
Glu258 and Gln261 (EL3) contribute favourably by electrostatic and van der Waals interactions with
the phenyl-acetamido moiety at the 5 position of the PTP. The FITC moiety is also involved and

interacts with several residues located at the ELs 2 and 3, participating in the overall recognition
process. As shown in Figure 7, the alkyl linker sits between the loops and adapts to their conformation
thanks to its high rotational fredoom. Depending on the latter, the fluorophore is pointed between the
ELs, facing several polar residues like Thr173, Arg174 (EL2) and Asn256 (EL3).
Figure 5. Interaction energy histograms of compound 3 (EA2) for conserved and non-conserved residues of A1, A2A and A3
subtypes.
Ethynyl derivative 22 (EA20) prefers the hA3AR cavity through several hydrophobic contacts with
non-conserved residues and electrostatic interactions, mainly due to the terminal triple bond, which
could explain the selectivity observed (Figure 8). At the hA2AAR, in fact, such interactions are absent,

and even some electrostatic repulsion is observed with Glu169 (EL2) and His250 (6.52).
Figure 6. Interaction energy histograms of compound 16 (EA8) for conserved and non-conserved residues of A1, A2A and A3
subtypes.
Finally, three supervised molecular dynamics (SuMD) simulations were performed with a view to 1)
simulate binding to the hA3AR, 2) inspect which structural domains at the extracellular side or the
hydrophobic cavity may be determinants of the ligand preference. Ligand 3 (EA2) was chosen because
it yields the lowest affinity towards the hA2AAR within the entire series while preserving affinity for
the A3 subtype in the nanomolar range. Moreover, basing on per-residue interaction fingerprints
(Figures 5-6-8), its carboxylic moiety is clearly important for the A2A/A3 selectivity. The ligand
completely reached the binding site only in simulation 1. As shown in Video S4, the ligand intercalates
between the ELs 2-3 and

Figure 7. Selected docking pose of ligand 16 (EA8) at the hA3AR subtype and visualization of the most relevant structural
regions interacting with the linker and the fluorophore (FITC).
finally binds the receptor assuming a pose closely resembling the one predicted by molecular docking
(Video S1). Two residues (conserved and non-conserved, respectively) seem to play an important role
in the recognition process: Glu19 (1.39) and Gln261 (EL3). Glu19, located at the TM1, interacts with
the pyrazole ring favouring binding during the entire process. Gln261 (EL3) (Leu267 in the hA2A AR)
makes hydrogen bonds with the ligand’s carboxylate exposed to the ELs, either in the initial phase of
approach or in the phase of stabilization. Such interaction is detectable in the ligand-protein interaction
energy profile for distances of about 10 Å. Interestingly, in both simulations 2 and 3, the ligand stops
just at 10 Å, between the ELs 2-3, giving the same interaction energy peak observed in simulation 1
(Figure 9). The interaction energy is weaker than the bound state 1, indicating that the ligand can
effectively cross that “barrier”. This result indicates that the ELs’ structural region is crucial for binding,
as ligands can cross it more (or less) efficiently depending on the specific substitution present at the 5
position of PTP. Indeed, the ligand 3 establishes highly favorable contacts even at long distances,
corresponding to the EL3 or EL2 (20-30 Å), where several non-conserved residues are located, in
agreement with per-residue interaction energy calculations. Some contacts are more stabilizing than the
final state 1 (meta-binding sites), however, the great conformational freedom of the EL2 (see Video S4)
could explain the progression of the ligand path.

Figure 8. Interaction energy histograms of compound 22 (EA20) for conserved and non-conserved residues of A1, A2A and
A3 subtypes.

Figure 9. Ligand 3 (EA2)-protein interaction energy vs distance from the binding site, for each SuMD simulation. VdW
surfaces of the respective final states are reported with different colors, along with particular regions of the receptor.
Experimental Section
Chemistry
General. Reactions were routinely monitored by thin-layer chromatography (TLC) on silica gel
(precoated Macherey-Nagel, 60FUV254). Flash chromatography was performed using Macherey-
Nagel, silica 60, 230-400 mesh silica gel. Light petroleum ether refers to the fractions boiling at 40-
1
60 °C. Melting points were determined on a Buchi-Tottoli instrument and are uncorrected. H-NMR
were determined in CDCl3, DMSO-d6 or CD3OD solutions with a Varian Gemini 200, a Bruker 400 or
a Varian 500 spectrometers. Peaks positions are given in parts per million (δ) downfield relative to the
central peak of the solvents, and J values are given in Hz. The following abbreviations were used: s,
singlet; bs, broad singlet; d, doublet; dd, double doublet; bd, broad doublet; t, triplet; m, multiplet.
Electrospray mass spectra were recorded on a ESI Bruker 4000 Esquire spectrometer and compounds
were dissolved in methanol; accurate mass spectra were recorded on a micrOTOF-Q-Bruker and
compounds were dissolved in methanol. Absorption experiments were performed on a Varian Cary

5000 spectrophotometer. Fluorescence measurements were recorded on Varian Cary Eclipse
fluorescence spectrophotometer, with excitation filter at 361 nm, emission filter at 390-600 nm (all
experiments were performed using a λex of 488 nm). Oct-7-ynoic acid was synthesized as reported in
the literature.
Synthesis of 2-(4-(3-(2-(furan-2-yl)-8-methyl-8H-pyrazolo[4,3-e][1,2,4]triazolo[1,5-c]pyrimidin-5-
yl)ureido)phenyl)acetic acid (3). Ethyl ester derivative 24 (1.95 mmol, 900 mg) was dissolved in a
mixture of tetrahydrofuran/methanol/water (4:1:1, 12 mL) and lithium hydroxide monohydrate (5.9
mmol, 245 mg) was added. The suspension was stirred at room temperature overnight. Reaction was
monitored by TLC (ethyl acetate 9.5 / methanol 0.5). Water was added to the mixture, the solution was
cooled to 0°C and 1 M hydrochloric acid was added (pH 2) leading to the precipitation of a pale brown
solid. Solid was filtered and washed with ethyl ether and identified as the desired compound in a
quantitative yield. Mp >300°C; 1H-NMR (200 MHz, DMSO-d6) δ 10.67 (s, 1H) 9.55 (bs, 1H), 8.78 (s,
1H), 7.93 (s, 1H) 7.78-7.36 (m, 3H), 7.38–6.96 (m, 3H), 6.76 (bs, 1H), 4.34 (s, 3H), 3.56 (s, 2H). ES-
MS negative mode (methanol) m/z: 431.0 [M-H]-.
General procedure for the synthesis of amides 4-8. The carboxylic acid derivative 3 (0.58 mmol, 250
mg) was dissolved in DMF (3 mL) and the appropriate mono-BOC-protected diamine (25-29) was
added (0.58 mmol). Reaction was stirred and cooled to 0°C under an argon atmosphere and then 1-
ethyl-3-(3-dimethylaminopropyl)carbodiimmide hydrochloride (EDCI.HCl, 1.16 mmol, 222 mg) and 4-
dimethylaminopyridine (DMAP, 1.44 mmol, 177 mg) were added. Mixture was stirred at room
temperature overnight and monitored by TLC (dichloromethane 9.3 / methanol 0.7). The solvent was
evaporated under reduced pressure. The residue was suspended in water and extracted with ethyl
acetate (3 times). The organic layers were collected and dried over sodium sulfate anhydrous and the
solvent removed under reduced pressure. The crude product was purified on flash silica column
chromatography (dichloromethane 9.7 / methanol 0.3) to afford the desired compound (4-8).
tert-Butyl (3-(2-(4-(3-(2-(furan-2-yl)-8-methyl-8
H-pyrazolo[4,3-
e][1,2,4]triazolo[1,5-c]pyrimidin-
1
5-yl)ureido)phenyl)acetamido)propyl)carbamate (4). Yield 29.4%; white solid; mp 198-205°C; H-
NMR (200 MHz, CDCl3) δ 11.19 (s, 1H), 8.60 (s, 1H), 8.23 (s, 1H), 7.80–7.46 (m, 3H), 7.30–7.26 (m,
3H), 6.62 (dd, J = 3.4, 1.8 Hz, 1H), 6.05 (bs, 1H), 4.94 (bs, 1H), 4.21 (s, 3H), 3.58 (s, 2H), 3.27 (q, J =
6 Hz, 2H), 3.15-3.00 (m, 2H), 1.56 (t, J = 6, 2H), 1.42 (s, 9H); ES-MS (methanol) m/z: 611.3 [M+Na]+.
HRMS (ESI-TOF) m/z: [M+Na]+ Calcd for C28H32N10O5 611.2449; Found 611.2440.
tert-Butyl (4-(2-(4-(3-(2-(furan-2-yl)-8-methyl-8
H-pyrazolo[4,3-e][1,2,4]triazolo[1,5-c]pyrimidin-
1
5-yl)ureido)phenyl)acetamido)butyl)carbamate (5). Yield 34.5%; white solid; mp 186-193°C; H-

NMR (200 MHz, CDCl3) δ 11.19 (s, 1H), 8.60 (s, 1H), 8.22 (s, 1H), 7.90–7.48 (m, 3H), 7.38–7.09 (m,
3H), 6.62 (s, 1H), 5.63 (bs, 1H), 4.60 (bs, 1H), 4.21 (s, 3H), 3.57 (s, 2H), 3.22 (bs, 2H), 3.09 (bs, 2H),
1.75 (bs, 4H), 1.43 (s, 9H); ES-MS (methanol) m/z: 625.4 [M+Na]+, 641.3 [M+K]+. HRMS (ESI-TOF)
m/z: [M+Na]+ Calcd for C29H34N10O5 625.2606; Found 625.2608.
tert-Butyl (5-(2-(4-(3-(2-(furan-2-yl)-8-methyl-8
H-pyrazolo[4,3-
e][1,2,4]triazolo[1,5-c]pyrimidin-
1
5-yl)ureido)phenyl)acetamido)pentyl)carbamate (6). Yield 28%; white solid; mp 78-185°C; H-
NMR (200 MHz, CDCl3) δ 11.20 (s, 1H), 8.61 (s, 1H), 8.23 (s, 1H), 7.66 (bs, 3H), 7.26 (bs, 3H), 6.62
(bs, 1H), 5.47 (bs, 1H), 4.56 (bs, 1H), 4.21 (s, 3H), 3.57 (s, 2H), 3.20 (bs, 2H), 3.07 (bs, 2H), 1.60-1.10
(m, 15H); ESMS (methanol) m/z: 639.4 [M+Na]+, 655.3 [M+K]+. HRMS (ESI-TOF) m/z: [M+Na]+
Calcd for C30H36N10O5 639.2762; Found 639.2764.
tert-Butyl
(2-(2-(2-(2-(4-(3-(2-(furan-2-yl)-8-methyl-8H-pyrazolo[4,3-e][1,2,4]triazolo[1,5-
c]pyrimidin-5-yl)ureido)phenyl)acetamido)ethoxy)ethoxy)ethyl)carbamate (7). Yield 54%; white
solid; mp 157-165°C; 1H-NMR (200 MHz, CDCl3) δ 11.18 (s, 1H), 8.60 (s, 1H), 8.23 (s, 1H), 7.78–
7.56 (m, 3H), 7.42–7.15 (m, 3H), 6.62 (dd, J = 3.2, 1.6 Hz, 1H), 5.97 (bs, 1H), 5.02 (bs, 1H), 4.22 (s,
3H), 3.73 – 3.35 (m, 12H), 3.29 (bs, 2H), 1.42 (s, 9H); ES-MS (methanol) m/z: 685.4 [M+Na]+, 701.4
[M+K]+. HRMS (ESI-TOF) m/z: [M+Na]+ Calcd for C31H38N10O7 685.2817; Found 685.2807.
tert-Butyl (1-(4-(3-(2-(furan-2-yl)-8-methyl-8H-pyrazolo[4,3-e][1,2,4]triazolo[1,5-c]pyrimidin-5-
yl)ureido)phenyl)-2-oxo-7,10,13-trioxa-3-azahexadecan-16-yl)carbamate (8). Yield 21.6%; white
1
solid; mp 151-158°C; H-NMR (270 MHz, CDCl3) δ 11.19 (s, 1H), 8.60 (s, 1H), 8.24 (s, 1H), 7.80–
7.61 (m, 3H), 7.45–7.19 (m, 3H), 6.63 (bs, 1H), 6.20 (bs, 1H), 4.96 (bs, 1H), 4.22 (s, 3H), 3.71–3.49
(m, 14H), 3.35 (bs, 2H), 3.20 (bs, 2H), 1.73 (t, J = 3.2, 4H), 1.49 (s, 9H); ES-MS (methanol) m/z: 757.4
[M+Na]+. HRMS (ESI-TOF) m/z: [M+Na]+ Calcd for C35H46N10O8 757.3392; Found 757.3393.
General synthesis of compounds 9-13 by N-BOC deprotection. The N-BOC protected derivatives
(4-8) were dissolved in a solution of TFA and dichloromethane (1:1) and the mixtures were stirred for 2
hours at room temperature. Reactions were monitored by TLC (dichloromethane 9.3 / methanol 0.7).
The solvent was removed under reduced pressure and the solids were filtered to afford the desired
compounds as trifluoroacetate salts (9-13).
N
-(3-aminopropyl)-2-(4-(3-(2-(furan-2-yl)-8-methyl-8H-pyrazolo[4,3-e][1,2,4]triazolo[1,5-
c]pyrimidin-5-yl)ureido)phenyl)acetamide trifluoroacetate salt (9). Yield 89.8%; white solid; mp
1
142-150°C; H-NMR (200 MHz, CD3OD) δ 11.28 (s, 1H), 8.48 (s, 1H), 8.31 (bs, 1H), 7.76 (s, 1H),
7.62 (d, J = 8.2 Hz, 2H), 7.30 (d, J = 8.2 Hz, 1H), 7.24 (d, J = 3.4 Hz, 1H) 6.66 (bs, 1H), 4.15 (s, 3H),
3.52 (s, 2H), 3.34–3.10 (m, 2H), 2.89 (t, J = 7.0 Hz, 2H), 1.85 (p, J = 7.0 Hz, 2H). HRMS (ESI-TOF)
m/z: [M+H]+ Calcd for C23H24N10O3 489.2106; Found 489.2104.

N
-(4-aminobutyl)-2-(4-(3-(2-(furan-2-yl)-8-methyl-8H-pyrazolo[4,3-e][1,2,4]triazolo[1,5-
c]pyrimidin-5-yl)ureido)phenyl)acetamide trifluoroacetate salt (10). Yield 95.1%; white solid; mp
1
167-173°C; H-NMR (200 MHz, CD3OD) δ 8.48 (s, 1H), 8.19 (bs, 1H), 7.76 (s, 1H), 7.63 (d, J = 8.2
Hz, 2H), 7.39–7.13 (m, 3H), 6.70-6.62 (m, J = 1H), 4.14 (s, 3H), 3.51 (s, 2H), 3.32–3.30 (m, 2H), 2.92
(t, J = 6.7 Hz, 2H), 1.78–1.49 (m, 4H). HRMS (ESI-TOF) m/z: [M+H]+ Calcd for C24H26N10O3
503.2262; Found 503.2260.
N
-(5-aminopentyl)-2-(4-(3-(2-(furan-2-yl)-8-methyl-8
H-pyrazolo[4,3-e][1,2,4]triazolo[1,5-
c]pyrimidin-5-yl)ureido)phenyl)acetamide trifluoroacetate
salt (11). Yield 85.3%; white solid;
1
mp 170-178°C; H-NMR (200 MHz, CD3OD) δ 8.47 (s, 1H), 8.13 (bs, 1H), 7.75 (s, 1H), 7.61 (d, J =
7.7 Hz, 2H), 7.29 (d, J = 7.7 Hz, 2H), 7.25 (s, 1H), 6.66 (s, 1H), 4.14 (s, 3H), 3.49 (s, 2H), 3.33–3.08
(m, 2H), 2.88 (t, J = 7.3 Hz, 2H), 1.77–1.47 (m, 4H), 1.47–1.25 (m, 2H). HRMS (ESI-TOF) m/z:
[M+H]+ Calcd for C25H28N10O3 517.2419; Found 517.2417.
2-(2-(2-(2-(4-(3-(2-(furan-2-yl)-8-methyl-8H-pyrazolo[4,3-e][1,2,4]triazolo[1,5-c]pyrimidin-5-
yl)ureido)phenyl)acetamido)ethoxy)ethoxy)ethanamine trifluoroacetate salt (12). Yield 83.3%;
white solid; mp 152-160°C; 1H-NMR (200 MHz, CD3OD) δ 8.49 (s, 1H), 8.15 (bs, 1H), 7.76 (bs, 1H),
7.62 (d, J = 7.3 Hz, 2H), 7.44–7.13 (m, 3H), 6.66 (bs, 1H), 4.15 (s, 3H), 3.85–3.38 (m, 12H), 3.09 (bs,
2H); ES-MS (methanol) m/z: 563.3 [M+H]+, 585.3 [M+Na]+. HRMS (ESI-TOF) m/z: [M+H]+ Calcd
for C26H30N10O5 563.2473; Found 563.2474.
N
-(3-(2-(2-(3-aminopropoxy)ethoxy)ethoxy)propyl)-2-(4-(3-(2-(furan-2-yl)-8-methyl-8
H-
pyrazolo[4,3- ][1,2,4]triazolo[1,5- ]pyrimidin-5-yl)ureido)phenyl)acetamide trifluoroacetate salt
e
c
1
(13). Yield 94.2%; white solid; mp 133-140°C; H-NMR (270 MHz, CDCl3) δ 11.13 (s, 1H), 8.28 (s,
1H), 7.81 (bs, 2H), 7.66 (s, 1H), 7.60 (d, J = 7.7 Hz, 2H), 7.41–7.13 (m, 3H), 6.62 (s, 1H), 5.40 (bs,
2H), 4.21 (s, 3H), 3.75 (s, 2H), 3.66–3.54 (m, 12H), 3.36 (bs, 2H), 3.24 (bs, 2H), 2.02 (bs, 2H), 1.78
(bs, 2H). HRMS (ESI-TOF) m/z: [M+H]+ Calcd for C38H38N10O6 635.3049; Found 635.3046.
General synthesis of fluorescein conjugates (14-18). 0.075 mmol of amino derivatives 9-13 were
dissolved in 4 mL of dry methanol. 21.6 µL (0.155 mmol) of TEA were added to the solution which
was stirred at room temperature for 30 minutes under an argon atmosphere. Then 30.2 mg (0.075 mmol)
of FITC were added and the reaction was stirred for 72 hours in the dark. The products were purified
by column chromatography starting with dichloromethane:methanol (9:1) as eluent. The obtained
solids were suspended in dichloromethane, filtered and washed several times with ethyl ether yielding
to the desired compounds as orange solids (14-18).
5-(3-(3-(2-(4-(3-(2-(furan-2-yl)-8-methyl-8
H
-pyrazolo[4,3-
e][1,2,4]triazolo[1,5-c]pyrimidin-5-
yl)ureido)phenyl)acetamido)propyl)thioureido)-2-(6-hydroxy-3-oxo-3
H-xanthen-9-yl)benzoate

(14). Yield 54.7%; orange solid; mp 198-204°C; 1H-NMR (200 MHz, DMSO-d6) δ 10.69 (s, 1H), 10.09
(bs, 3H), 8.76 (s, 1H), 8.41–8.05 (m, 3H), 7.98 (s, 1H), 7.76 (d, J = 3.2 Hz, 1H), 7.50 (d, J = 7.4 Hz,
2H), 7.39–7.03 (m, 4H), 6.76 (s, 1H), 6.65–6.47 (m, 6H), 4.12 (s, 3H), 3.50 (bs, 2H), 3.35 (s, 2H), 3.15
(bs, 2H), 1.71 (bs, 2H); 13C-NMR (50 MHz, DMSO-d6) δ 180.29, 170.29, 168.44, 160.06, 159.79,
154.77, 152.93, 151.98, 151.93, 151.91, 149.57, 149.52, 148.96, 148.89, 145.30, 144.99, 141.23,
140.14, 136.60, 131.29, 129.65, 129.59, 129.49, 129.23, 129.20, 129.16, 129.13, 129.01, 126.24,
124.15, 118.95, 112.68, 112.54, 112.18, 109.81, 102.17, 98.68, 45.69, 41.86, 41.44, 36.37, 28.73; UV-
VIS λmax 523 nm; Fluorescence λem 537; ES-MS (methanol) m/z 878.3 [M+H]+, 900.3 [M+Na]+, 916.3
[M+K]+. HRMS (ESI-TOF) m/z: [M+H]+ Calcd for C44H35N11O8S 878.2464; Found 878.2336.
5-(3-(4-(2-(4-(3-(2-(furan-2-yl)-8-methyl-8
H
-pyrazolo[4,3-
e][1,2,4]triazolo[1,5-c]pyrimidin-5-
yl)ureido)phenyl)acetamido)butyl)thioureido)-2-(6-hydroxy-3-oxo-3
H-xanthen-9-yl)benzoate (15).
Yield 53.8%; orange solid; mp 200-207°C; 1H-NMR (200 MHz, DMSO-d6)δ 10.72 (s, 1H), 10.06 (bs,
3H), 8.74 (s, 1H), 8.26 (bs, 2H), 8.07 (bs, 1H), 7.96 (s, 1H), 7.76 (d, J = 6.9 Hz, 1H), 7.50 (d, J = 8.0
Hz, 2H), 7.35–7.08 (m, 4H), 6.75 (s, 1H), 6.70-6.37 (m, 6H), 5.76 (s, 1H), 4.11 (s, 3H) 3.36 (s, 2H),
3.08 (bs, 4H), 1.51 (bs, 4H); 13C-NMR (50 MHz, DMSO-d6) δ 180.21, 169.96, 168.47, 160.33, 160.10,
154.46, 153.01, 152.94, 152.43, 152.19, 149.96, 149.86, 148.88, 148.81, 145.30, 145.03, 141.33,
140.40, 136.61, 131.40, 129.44, 129.10, 129.04, 129.02, 126.20, 124.23, 118.98, 112.94, 112.91,
112.45, 112.16, 109.88, 102.25, 98.62, 45.76, 43.43, 41.85, 38.10, 26.82, 25.84; UV-VIS λmax 524 nm;
Fluorescence λem 535; ES-MS (methanol) m/z: 892.4 [M+H]+, 914.3 [M+Na]+, 930.3 [M+K]+. HRMS
(ESI-TOF) m/z: [M+H]+ Calcd for C45H37N11O8S 914.2439; Found 914.2437.
5-(3-(5-(2-(4-(3-(2-(furan-2-yl)-8-methyl-8
H
-pyrazolo[4,3-
e][1,2,4]triazolo[1,5-c]pyrimidin-5-
yl)ureido)phenyl)acetamido)pentyl)thioureido)-2-(6-hydroxy-3-oxo-3
H-xanthen-9-yl)benzoate
(16). Yield 28%; orange solid; mp 190-200°C; 1H-NMR (200 MHz, DMSO-d6) δ 10.73 (s, 1H), 10.13
(bs, 1H), 8.75 (s, 1H), 8.27 (bs, 2H), 8.11–7.87 (m, 2H), 7.76 (d, J = 5.6 Hz, 1H), 7.50 (d, J = 6.3 Hz,
2H), 7.39–7.02 (m, 4H), 6.75 (s, 1H), 6.71–6.39 (m, 6H), 4.12 (s, 3H), 3.38 (s, 2H), 3.07 (bs, 4H),
1.74–1.10 (m, 4H), 1.08–0.75 (m, 2H); 13C-NMR (50 MHz, DMSO-d6) δ 180.17, 169.90, 168.45,
160.02, 159.91, 154.73, 152.99, 151.96, 149.80, 149.77, 148.92, 148.88, 145.24, 145.03, 141.35,
140.32, 136.63, 131.37, 129.49, 129.44, 129.27, 129.23, 129.02, 129.00, 126.21, 124.08, 119.27,
118.92, 112.94, 112.88, 112.85, 112.46, 112.18, 109.86, 102.17, 98.64, 43.74, 41.84, 40.61, 40.20,
28.79, 28.03, 23.88; UV-VIS λmax 523 nm; Fluorescence λem 537; ES-MS (methanol) m/z: 928.4
[M+Na]+. HRMS (ESI-TOF) m/z: [M+H]+ Calcd for C46H39N11O8S 928.2596; Found 928.2596.
5-(3-(2-(2-(2-(2-(4-(3-(2-(furan-2-yl)-8-methyl-8
H
-pyrazolo[4,3-
e
][1,2,4]triazolo[1,5-
c]pyrimidin-
5-yl)ureido)phenyl)acetamido)ethoxy)ethoxy)ethyl)thioureido)-2-(6-hydroxy-3-oxo-3
H-xanthen-

9-yl)benzoate (17). Yield 54.3%; orange solid; mp 182-185°C; 1H NMR 1H-NMR (200 MHz, DMSO-
d6) δ 10.72 (s, 1H), 10.14 (bs, 3H), 8.75 (s, 1H), 8.30 (s, 1H), 8.13 (d, J = 5.0 Hz, 2H), 7.95 (s, 1H),
7.76 (d, J = 7.7 Hz, 1H), 7.49 (d, J = 7.9 Hz, 2H), 7.37–7.07 (m, 4H), 6.74 (s, 1H), 6.69–6.42 (m, 6H),
4.11 (s, 3H), 3.79–3.48 (m, 8H), 3.51–2.86 (m, 6H); 13C-NMR (50 MHz, DMSO-d6) δ 180.47, 170.26,
168.54, 159.65, 159.61, 154.82, 154.78, 152.92, 151.96, 151.86, 149.68, 148.89, 145.26, 145.00,
141.27, 140.25, 136.60, 131.30, 129.58, 129.47, 128.97, 126.77, 126.27, 124.04, 118.95, 112.70,
112.67, 112.53, 112.19, 109.77, 102.21, 102.19, 98.68, 69.62, 69.52, 69.07, 68.46, 45.70, 43.69, 41.71,
41.03; UV-VIS λmax 524 nm; Fluorescence λem 537; ES-MS (methanol) m/z: 952.4 [M+H]+, 974.4
[M+Na]+, 990.4 [M+K]+. HRMS (ESI-TOF) m/z: [M+H]+ Calcd for C47H41N11O10S 952.2831; Found
952.2833.
5-(3-(1-(4-(3-(2-(furan-2-yl)-8-methyl-8
H-pyrazolo[4,3-e][1,2,4]triazolo[1,5-c]pyrimidin-5-
yl)ureido)phenyl)-2-oxo-7,10,13-trioxa-3-azahexadecan-16-yl)thioureido)-2-(6-hydroxy-3-oxo-3
H-
xanthen-9-yl)benzoate (18). Yield 63%; orange solid; mp 143-149°C; 1H-NMR (200 MHz, DMSO-d6)
δ 10.65 (bs, 1H), 10.36–9.87 (m, 3H), 9.51 (bs, 1H), 8.78 (s, 1H), 8.34–8.11 (m, 2H), 7.99 (s, 2H), 7.74
(d, J = 8.2 Hz, 1H), 7.49 (d, J = 7.8 Hz, 2H), 7.37–7.03 (m, 4H), 6.76 (s, 1H), 6.69–6.45 (m, 6H), 4.13
(s, 3H), 3.65–3.18 (m, 14H), 3.18–2.79 (m, 4H), 1.81 (bs, 2H), 1.63 (bs, 2H); 13C-NMR (50 MHz,
DMSO-d6) δ 180.23, 169.94, 168.39, 159.33, 159.27, 154.82, 151.74, 149.01, 148.96, 148.89, 145.34,
144.91, 141.28, 139.69, 136.40, 131.51, 129.47, 129.38, 129.27, 129.22, 128.92, 128.77, 126.34,
123.87, 118.94, 112.66, 112.49, 112.47, 112.39, 112.21, 112.15, 109.43, 102.16, 98.78, 69.72, 69.53,
69.51, 68.13, 68.01, 41.82, 41.35, 40.62, 35.97, 29.31, 28.58, 9.10; UV-VIS λmax 524 nm; Fluorescence
λem 537; ES-MS (methanol) m/z: 1046.4 [M+Na]+. HRMS (ESI-TOF) m/z: [M+Na]+ Calcd for
C51H49N11O11S 1026.3226; Found 1026.3223.
General synthesis for the N-(2-(furan-2-yl)-8-methyl-8H-pyrazolo[4,3-e][1,2,4]triazolo[1,5-
c]pyrimidin-5-yl)alkynamides (19-22). Alkyne acid (1.175 mmol) (31-24) was dissolved in
deuterated chloroform (2 mL) and a catalytic amount of DMF was added (200 µL). The solution was
cooled to 0°C and oxalyl dichloride was added dropwise (1.175 mmol, 99 µL) and the reaction was
stirred at room temperature for 3 hours. A little amount of reaction was diluted in deuterated
chloroform and monitored by 1H-NMR. The peaks of desired acyl chlorides (35-38) were shifted from
those of the corresponding carboxylic acids. When the conversion was complete, to the reaction were
added 1,4-dioxane (5 mL), triethylamine (1.959 mmol, 273 µL) and the 5-amino-pyrazolo-triazolo-
pyrimidine 30 (0.392 mmol, 100 mg). The reaction was stirred at reflux overnight (TLC ethyl acetate
9.5/ methanol 0.5) then the solvent was removed under reduced pressure. The residue was dissolved in
water and extracted with dichloromethane (3 times), The organic layers were collected and dried over

sodium sulfate anhydrous and the solvent removed under reduced pressure. The crude was purified on
flash silica column chromatography (ethyl acetate 9.8 / methanol 0.2) to afford the desired compound
(19-22).
N
-(2-(furan-2-yl)-8-methyl-8H-pyrazolo[4,3-e][1,2,4]triazolo[1,5-c]pyrimidin-5-yl)pent-4-
ynamide (19). Yield 40%; pale brown solid; mp 264-271 °C d; 1H-NMR (200 MHz, CDCl3) δ 9.11 (s,
1H), 8.22 (s, 1H), 7.65 (s, 1H), 7.23 (s, 1H), 6.61 (s, 1H), 4.20 (s, 3H), 3.50 (t, J = 6.8 Hz, 2H), 2.70 (t,
13
J = 6.8 Hz, 2H), 2.01 (s, 2H); C-NMR (50 MHz, CDCl3) δ 171.47, 155.68, 153.84, 148.92, 145.09,
144.81, 138.67, 125.55, 111.13, 112.24, 100.04, 82.89, 74.39, 69.33, 41.06, 37.05, 13.86. ES-MS
(methanol) m/z: 336.2 [M+H]+, 358.1 [M+Na]+, 374.1 [M+K]+. HRMS (ESI-TOF) m/z: [M+Na]+
Calcd for C16H13N7O2 358.1023; Found 358.1021.
N
-(2-(furan-2-yl)-8-methyl-8H-pyrazolo[4,3-e][1,2,4]triazolo[1,5-c]pyrimidin-5-yl)hex-5-ynamide
(20). Yield 19%; pale yellow solid; mp 210-216 °C d; 1H-NMR (200 MHz, CDCl3) δ 9.05 (s, 1H), 8.21
(s, 1H), 7.65 (s, 1H), 7.26 (s, 1H), 6.62 (s, 1H), 4.21 (s, 3H), 3.26 (t, J = 6.3 Hz, 2H), 2.40 (s, 2H), 2.18
– 1.85 (m, 3H); 13C-NMR (50 MHz, CDCl3) δ 172.67, 156.07, 153.55, 149.05, 145.19, 144.92, 138.68,
124.95, 112.92, 112.16, 100.13, 83.49, 69.33, 40.80, 36.69, 23.34, 18.04. ES-MS (methanol) m/z: 372.2
[M+Na]+. HRMS (ESI-TOF) m/z: [M+Na]+ Calcd for C17H15N7O2 372.1179; Found 372.1182.
N
-(2-(furan-2-yl)-8-methyl-8
H-pyrazolo[4,3-e][1,2,4]triazolo[1,5-c]pyrimidin-5-yl)hept-6-
1
ynamide (21). Yield 15%; white solid; mp 208-215 °C d; H-NMR (200 MHz, CDCl3) δ 9.02 (s, 1H),
8.21 (s, 1H), 7.65 (s, 1H), 7.23 (d, J = 3.2 Hz, 1H), 6.60 (d, J = 3.2 Hz 1H), 4.20 (s, 3H), 3.36 (t, J = 7
Hz, 2H), 2.32-2.24 (m, 2H), 1.97-1.86 (m, 3H), 1.77-1.66 (m, 2H); 13C-NMR (50 MHz, CDCl3) δ
173.09, 155.76, 153.92, 148.96, 144.96, 138.75, 125.38, 112.89, 112.19, 100.05, 84.24, 68.76, 41.03,
37.23, 28.01, 23.52, 18.56. ES-MS (methanol) m/z: 364.2 [M+H]+, 386.2 [M+Na]+, 402.1 [M+K]+.
HRMS (ESI-TOF) m/z: [M+Na]+ Calcd for C18H17N7O2 386.1336; Found 386.1331.
N
-(2-(furan-2-yl)-8-methyl-8
H-pyrazolo[4,3-e][1,2,4]triazolo[1,5-c]pyrimidin-5-yl)oct-7-ynamide
1
(22). Yield 40%; white solid; mp 204-208 °C d; H-NMR (200 MHz, CDCl3) δ 8.99 (s, 1H), 8.21 (s,
1H), 7.65 (d, J = 1.8 Hz, 1H), 7.25 (d, J = 3.2 Hz, 1H), 6.61 (dd, J = 1.8 Hz, J = 3.2 Hz 1H), 4.20 (s,
13
3H), 3.24 (t, J = 7.2 Hz, 2H), 2.23-2.04 (m, 2H), 1.95-1.73 (m, 3H), 1.61-1.56 (m, 4H); C-NMR (50
MHz, CDCl3) δ 173.28, 155.74, 153.95, 148.95, 144.96, 138.76, 125.37, 112.88, 112.18, 100.06, 84.57,
68.4, 41.02, 37.62, 28.47, 28.38, 23.91, 18.48. ES-MS (methanol) m/z: 378.2 [M+H]+, 400.2 [M+Na]+,
416.1 [M+K]+. HRMS (ESI-TOF) m/z: [M+Na]+ Calcd for C19H19N7O2 400.1492; Found 400.1488.
Synthesis of 2-(6-amino-3-iminio-4,5-disulfonato-3
H-zanthen-9-yl)-5-((6-(4-((8-methyl-2-(furan-2-
yl)- 8 -pyrazolo[4,3- ][1,2,4]triazolo[1,5- ]pyrimidin-5-yl)amino)-4-oxobutyl)-1H
H
e
c
-1,2,3-triazol-1-

yl)hexyl)carbamoyl)benzoate (23)
A solution of Alexa Fluor-488 5-carboxamido-(6-azidohexanyl)bis(triethyammonium salt) (0.5 mg,
0.58
of alkyne derivative 20 (0.38 mg, 1.1
sodium ascorbate (0.1 M, 8 l, 5.86
of aqueous copper sulfate pentahydrate solution (0.1 M, 2.94
µ
mol, Invitrogen-Life Technologies, Grand Island, N.Y.) in water (200
mol) in DMF (100 l). A freshly prepared solution of aqueous
mol) was added to the reaction mixture followed by the addition
l, 0.29 mol): The resulting mixture was
µl) was added to a solution
µ
µ
µ
µ
µ
µ
stirred overnight at room temperature; 10 mM aqueous solution of triethylammonium acetate buffer (2
ml) was added with constant mixing. The resulting mixture was lyophilized and purified by HPLC with
a Luna 5µ RP.C18 semipreparative column (250 x 10 mm: Phenomenex, Torrance, Calif.) under the
following conditions: flow rate of 2 ml/min; H2O-MeCN from 100:0 (v/v) to 0:100 (v/v) in 30 min and
isolated at 20.07 min to get the Alexa Fluor-488 conjugate 23 (0.16 mg, 28%) . Analytical purity > 99%
by HPLC (retention time 10.13 min).
Biology
Radioligand Binding to hA1, hA2A and hA3 ARs
[3H]R-PIA and [125I]AB-MECA were utilized in radioligand binding assays to membranes prepared
from CHO cells expressing recombinant hA1and hA3 ARs, respectively, as previously described. [40]
While, [3H]CGS21680 binding was used at hA2AAR expressed in HEK293 cells. ADA (3 units/mL)
was present during the preparation of the membranes, in a preincubation of 30 min at 30 °C, and during
the incubation with the radioligands. All non-radioactive compounds were initially dissolved in DMSO
and diluted with buffer to the final concentration, where the amount of DMSO never exceeded 2%.
Incubations were terminated by rapid filtration over Whatman GF/B filters, using a Brandell cell
harvester (Brandell, Gaithersburg, MD). The tubes were rinsed three times with 3 mL of buffer each. At
least six different concentrations of competitor, spanning 3 orders of magnitude adjusted appropriately
for the IC50 of each compound, were used. IC50 values, calculated with the nonlinear regression method
implemented in Graph-Pad (Prism, San Diego, CA), were converted to Ki values as described. [40,43]
Computational studies
The MOE suite [45] was used for general molecular modeling operations, including logP(o/w)
calculations as well. Preliminary computations were carried out on a 12 CPU (Intel® Xeon® CPU E5-
1650 3.80 GHz) Linux workstation. SuMD simulations were performed with the ACEMD engine [46]
on NVIDIA drivers: NVIDIA GTX 980Ti and NVIDIA GTX 980. The CHARMM36/CHARMM36

general force field (CGenFF) combination was adopted. The GOLD [47] suite (goldscore scoring
function) was used as docking program. Ligands’ 3D structures were constructed by the MOE-builder
tool. Ionization states were predicted using the MOE-protonate 3D tool. Tautomerization and atom
hybridization were checked. Structures were minimized by the MMFF94x until the root mean square
(RMS) gradient was below 0.05 kcal mol−1 A−1. For each ligand, 20 docking simulation runs were
performed on each receptor subtype, searching on a sphere of 20 Å radius, centered on the backbone
nitrogen of the conserved Asn (6.55). Ligand and protein partial charges were calculated by means of
the MMFF94 and AMBER14:EHT [48] force fields, respectively. The hA2A and hA1AR coordinates
were retrieved from the Protein Data Bank [49] using the crystal structures 3PWH [50] and 5UEN [51],
respectively. For the hA3AR a homology model was used, basing on our docking/structural based
model assessment [39]. All molecular docking studies have been carried out with a sodium ion and its
first hydration shell in all adenosine receptor structures, according to our previous studies [33]. The
Ballesteros–Weinstein [52] numbering system was used sometimes to indicate conserved residues. IE
electrostatic and Van der Waals energy contributions to the binding energy were calculated by MOE,
along with per residue electrostatic and hydrophobic interactions. Per-residue information was reported
in the “Interaction Energy Fingerprints”: they are heat maps reporting the strength of the interaction of
each residue (x-axis) and each ligand (y-axis) according to a colorimetric scale going from blue to red
for negative to positive values in the case of electrostatic contributions and from white to dark green for
low to high values for hydrophobic contributions. Interactions of most relevant residues were also
reported by histograms, whose height is proportional to the strength of the interaction. All the plots
were generated using Gnuplot 4.6 [53], except Figure 3. Molecular graphics were performed with the
UCSF Chimera package [54], except Fig. 8, which was obtained using the VMD program (version
1.9.3) [55]. The in-house MMsDocking video maker tool was exploited to produce videos showing the
docking poses, per residue hydrophobic and electrostatic contributions for selected residues,
experimental binding data, and scoring values. Representations of docking poses were produced using
the UCSF Chimera package, 2D depictions were constructed by the cheminformatics toolkit RDKit [56]
and the heat maps were obtained by Gnuplot 4.6; in the end, videos were mounted using MEncoder.
Ligand 3 force field parameters for MD simulations were initially retrieved from the Paramchem web
service and then deeply optimized in concordance with CGenFF, at the MP2/6-31G [57] level of the
theory by using Gaussian 09 [58] and RESP partial charges. Systems were embedded in a 1-palmitoyl-
2oleyl-sn-glycerol-3-phospho-choline (POPC) lipid bilayer, according to the pre-orientation provided
by the Orientations Proteins in Membrane (OPM) database [59] and by using the VMD membrane

builder plugin. Lipids within 0.4 Å from the protein were removed and TIP3P [60] model water
molecules were added to solvate the system by means of Solvate1.0 [61]. Charge neutrality of the
system was obtained by adding Na+/Cl− counterions to a final concentration of 0.154 M. System was
equilibrated through a three-step procedure. In the first step, 1500 conjugate-gradient minimization
steps were applied in order to reduce the clashes between protein and lipids. Then, a 5 ns long MD
simulation was performed in the NPT ensemble, with a positional constraint of 1 kcal mol−1 Å−2 on
ligand, protein, and lipid phosphorus atoms. During the second stage, 10 ns MD simulation in the NPT
ensemble was performed constraining all the protein and ligand atoms but leaving the POPC residues
free to diffuse in the bilayer. In the last equilibration stage, positional constraints were applied only to
the ligand and protein backbone alpha carbons for further 5 ns MD simulation. All the MD simulations
were performed using: (1) an integration time step of 2 fs; (2) the Berendsen barostat [62] to maintain
the system pressure at 1 atm; (3) the Langevin thermostat [63] to maintain the temperature at 310 K
with a low dumping of 1 ps− 1; (4) the M-SHAKE algorithm [64] to constrain the bond lengths
involving hydrogen atoms; and (5) a long-range cutoff of 10 Å. According to the SuMD approach
[65,66], the timescale needed to reproduce binding is in the range of nanoseconds, instead of hundreds
of nanoseconds or microseconds usually necessary with unsupervised MD. Sampling is performed by a
tabu-like algorithm to monitor the distance between the centers of mass of the ligand and the binding
site during short unbiased MD simulations. SuMD considers the ligand atoms and the atoms of user-
defined protein residues to monitor the distance between the centers of mass of the binder and the
binding site. A series of 600 ps unbiased MD simulations are performed and after each simulation, the
distance points collected at regular intervals are fitted into a linear function. If the resulting slope is
negative the next simulation step starts from the last set of coordinates and velocities produced,
otherwise the simulation is restarted by randomly assigning the atomic velocities. Short simulations are
perpetuated under the supervision until the distance between the ligand and the binding site goes below
5 Å, then, the supervision is disabled and a classical MD simulation is performed. For the orthosteric
center of mass, we considered hA3AR residues Asn250, Phe168, His272, and Ser247. NAMD energy
Plugin 1.4 [67] was used to calculate ligand-protein interaction energies.
Conclusions
In this work we reported a new series of hA3AR conjugable derivatives that exhibited a very good
selectivity towards hA1AR and a discrete selectivity versus the hA2AAR. We have developed both
amino compounds, which could be functionalized by reactions with carboxylic acids or isothiocyanates,
and alkyne derivatives, which are useful precursors for the click reaction. Unfortunately, attempts to

use these compounds to obtain selective functionalized hA3AR ligands led mainly to dual hA2A/hA3 AR
ligands, revealing that the functional moiety had an active role in the establishment of ligand-receptor
interactions. Moreover, molecular docking and SuMD simulations highlighted that non-conservative
residues, located at the ELs 2 and 3 (Gln167, Val169, Gln261, Glu258 of the hA3AR), constitute a
specific structural region able to discriminate ligands depending on their chemical and partition
properties. These data can, thus, pave the way for a new understanding of selectivity towards ARs,
leading finally to the rational design of highly selective ligands. In any case, among the functionalized
compounds, Alexa Fluor 488 derivative MRS5763 (23) showed the best hA3A affinity and selectivity
profile, which, by further computer-guided optimization, could lead to a selective hA3AR probe useful
as fluorescent tool for this receptor.
Acknowledgments
KAJ thanks the NIDDK Intramural Research Program (ZIADK31117) for support.
Declaration of interest: none.
References
[1]
[2]
[3]
[4]
J. Sawynok, Adenosine receptor targets for pain, Neuroscience. 338 (2016) 1–18.
doi:10.1016/j.neuroscience.2015.10.031.
A. Guerrero, A2A adenosine receptor agonists and their potential therapeutic applications. An
update, Curr. Med. Chem. 25 (2018) 3597–3612. doi:10.2174/0929867325666180313110254.
B. Hocher, Adenosine A1 receptor antagonists in clinical research and development, Kidney Int.
78 (2010) 438–445. doi:10.1038/ki.2010.204.
B. Cacciari, G. Spalluto, S. Federico, A2A adenosine receptor antagonists as therapeutic
candidates: are they still an interesting challenge?, Mini-Reviews Med. Chem. 18 (2018) 1168–
1174. doi:10.2174/1389557518666180423113051.
[5]
[6]
G. Haskó, B. Csóka, Z.H. Németh, E.S. Vizi, P. Pacher, A2B adenosine receptors in immunity and
inflammation., Trends Immunol. 30 (2009) 263–70. doi:10.1016/j.it.2009.04.001.
S.L. Cheong, S. Federico, G. Venkatesan, A.L. Mandel, Y.-M. Shao, S. Moro, G. Spalluto, G.
Pastorin, The A3 adenosine receptor as multifaceted therapeutic target: pharmacology, medicinal
chemistry, and in silico approaches, Med. Res. Rev. 33 (2013) 235–335. doi:10.1002/med.20254.
B.B. Fredholm, G. Arslan, L. Halldner, B. Kull, G. Schulte, W. Wasserman, Structure and
function of adenosine receptors and their genes, Naunyn. Schmiedebergs. Arch. Pharmacol. 362
[7]

(2000) 364–374. doi:10.1007/s002100000313.
[8]
[9]
P.A. Borea, K. Varani, F. Vincenzi, P.G. Baraldi, M.A. Tabrizi, S. Merighi, S. Gessi, The A3
adenosine receptor: history and perspectives, Pharmacol. Rev. 67 (2014) 74–102.
doi:10.1124/pr.113.008540.
S. Moro, Z. Gao, K.A. Jacobson, G. Spalluto, Progress in the pursuit of therapeutic adenosine
receptor antagonists, Med. Res. Rev. 26 (2006) 131–159. doi:10.1002/med.20048.
[10] S. Redenti, A. Ciancetta, G. Pastorin, B. Cacciari, S. Moro, G. Spalluto, S. Federico,
Pyrazolo[4,3- ][1,2,4]triazolo[1,5- ]pyrimidines and structurally simplified analogs. Chemistry
e
c
and sar profile as adenosine receptor antagonists, Curr. Top. Med. Chem. 16 (2016) 3224–3257.
doi:10.2174/1568026616666160506145831.
[11] P.G. Baraldi, B. Cacciari, R. Romagnoli, G. Spalluto, K. Klotz, E. Leung, K. Varani, S. Gessi, S.
Merighi, P.A. Borea, Pyrazolo[4,3-e]-1,2,4-triazolo[1,5-c]pyrimidine Derivatives as highly
potent and selective human A3 adenosine receptor antagonists, J. Med. Chem. 42 (1999) 4473–
4478. doi:10.1021/jm991114s.
[12] P.G. Baraldi, B. Cacciari, R. Romagnoli, G. Spalluto, S. Moro, K. Klotz, E. Leung, K. Varani, S.
Gessi, S. Merighi, P.A. Borea, Pyrazolo[4,3-e]1,2,4-triazolo[1,5-c]pyrimidine derivatives as
highly potent and selective human A3 adenosine receptor antagonists: influence of the chain at
the N8 pyrazole nitrogen, J. Med. Chem. 43 (2000) 4768–4780. doi:10.1021/jm001047y.
[13] P.G. Baraldi, B. Cacciari, S. Moro, G. Spalluto, G. Pastorin, T. Da Ros, K. Klotz, K. Varani, S.
Gessi, P.A. Borea, Synthesis, biological activity, and molecular modeling investigation of new
pyrazolo[4,3-e]-1,2,4-triazolo[1,5-c]pyrimidine derivatives as human A3 adenosine receptor
antagonists, J. Med. Chem. 45 (2002) 770–780. doi:10.1021/jm0109614.
[14] A. Maconi, S. Moro, G. Pastorin, T. Da Ros, G. Spalluto, Z.-G. Gao, K.A. Jacobson, P.G.
Baraldi, B. Cacciari, K. Varani, P. Andrea Borea, Synthesis, biological properties, and molecular
modeling investigation of the first potent, selective, and water-soluble human A3 adenosine
receptor antagonist, J. Med. Chem. 45 (2002) 3579–3582. doi:10.1021/jm020974x.
[15] S. Federico, A. Ciancetta, D. Sabbadin, S. Paoletta, G. Pastorin, B. Cacciari, K.N. Klotz, S.
Moro, G. Spalluto, Exploring the directionality of 5-substitutions in a new series of 5-
alkylaminopyrazolo[4,3-e]1,2,4-triazolo[1,5-c]pyrimidine as a strategy to design novel human
A3 adenosine receptor antagonists, J. Med. Chem. 55 (2012). doi:10.1021/jm300899q.
[16] G. Pastorin, T. Da Ros, C. Bolcato, C. Montopoli, S. Moro, B. Cacciari, P.G. Baraldi, K. Varani,
P.A. Borea, G. Spalluto, Synthesis and biological studies of a new series of 5-
heteroarylcarbamoylaminopyrazolo[4,3-e]1,2,4-triazolo[1,5-c]pyrimidines as human A3