Exploring Mannosylpurines as Copper Chelators and Cholinesterase Inhibitors with Potential for Alzheimer’s Disease

Alzheimer’s Disease (AD) is characterized by a progressive cholinergic neurotransmission imbalance, with a decrease of acetylcholinesterase (AChE) activity followed by a significant increase of butyrylcholinesterase (BChE) in the later AD stages. BChE activity is also crucial for the development of Aβ plaques, the main hallmarks of this pathology. Moreover, systemic copper dyshomeostasis alters neurotransmission leading to AD. In the search for structures targeting both events, a set of novel 6-benzamide purine nucleosides was synthesized, differing in glycone configuration and N7/N9 linkage to the purine. Their AChE/BChE inhibitory activity and metal ion chelating properties were evaluated. Selectivity for human BChE inhibition required N9-linked 6-deoxy-α-d-mannosylpurine structure, while all three tested β-d-derivatives appeared as non-selective inhibitors. The N9-linked l-nucleosides were cholinesterase inhibitors except the one embodying either the acetylated sugar or the N-benzyl-protected nucleobase. These findings highlight that sugar-enriched molecular entities can tune bioactivity and selectivity against cholinesterases. In addition, selective copper chelating properties over zinc, aluminum, and iron were found for the benzyl and acetyl-protected 6-deoxy-α-l-mannosyl N9-linked purine nucleosides. Computational studies highlight molecular conformations and the chelating molecular site. The first dual target compounds were disclosed with the perspective of generating drug candidates by improving water solubility.


Introduction
Alzheimer's Disease (AD) is a devastating disease with a multifactorial pathogenesis. The main histopathological hallmarks of AD are the generation and deposition of senile plaques containing extracellular aggregates of neurotoxic amyloid-β (Aβ) peptide, along with intracellular neurofibrillary tangles of hyperphosphorylated τ-protein and the impairment of the cholinergic neurotransmission, due to a decrease of the level of the neurotransmitter acetylcholine (ACh), leading to a degeneration of cholinergic neurons and neuronal loss [1,2]. In physiological conditions, ACh is predominantly hydrolyzed by acetylcholinesterase (AChE), while butyrylcholinesterase (BChE) has a marginal role. Inspired by these first results, we designed a new series of purine nucleosides, aiming to insert a copper chelating molecular moiety for the development of multitarget agents against AD acting as cholinesterase inhibitors and/or as copper chelators. Starting from the general structure in Figure 1, the new compounds were synthesized and assayed as cholinesterase inhibitors aiming to evaluate the role on compound bioactivity of anomeric stereochemistry, sugar D, L configuration, sugar 6-deoxygenation, and glycosyl group regioligation to the purine base at positions N 7 or N 9 . For preliminary information about protecting group requirements, the peracetylated analogue of the most potent perbenzylated nucleoside was synthesized. New compounds were also tested for their copper chelating efficiency. The results, here presented and discussed, highlight the potent activity of this purine nucleoside family of compounds and demonstrate how copper chelation and anticholinesterase activity/selectivity can be tuned by the regiochemistry of glycone ligation, glycone conformation, and also by the anomeric absolute configuration.

Chemistry
Synthesis of compounds 5-10, 12, 15 and16 (Table 1, Schemes 1-3) was accomplished via TMSOTf catalyzed reaction of fully protected methyl glycosides with the purine base activated by reaction with BSA, a procedure adapted from a methodology previously described by us for other purine nucleosides [19]. The glycosyl donors were the perbenzylated methyl α-D-mannopyranoside (3), its 6-deoxy analogue (4), the peracetylated 6-deoxy-L-mannose (L-rhamnose) (11), and the perbenzylated methyl 6-deoxy-α-L-mannopyranoside (14). Reaction conditions optimization was required resulting from the very low yields obtained with the previously reported procedure. Several temperatures and reaction times were tried, under conventional and microwave-assisted syntheses, aiming to improve the overall nucleoside yield. Microwave irradiation at 150 W and 65 °C for 60 min gave the best results, and nucleoside isolated yields are given in Table 1. The four isomers differing in the anomeric configuration and linkage of the glycosyl group to the  [19]; (b) General structure of the hypothesized purine nucleosides' copper complex.

Chemistry
Synthesis of compounds 5-10, 12, 15 and 16 (Table 1, Schemes 1-3) was accomplished via TMSOTf catalyzed reaction of fully protected methyl glycosides with the purine base activated by reaction with BSA, a procedure adapted from a methodology previously described by us for other purine nucleosides [19]. The glycosyl donors were the perbenzylated methyl α-D-mannopyranoside (3), its 6-deoxy analogue (4), the peracetylated 6-deoxy-Lmannose (L-rhamnose) (11), and the perbenzylated methyl 6-deoxy-α-L-mannopyranoside (14). Reaction conditions optimization was required resulting from the very low yields obtained with the previously reported procedure. Several temperatures and reaction times were tried, under conventional and microwave-assisted syntheses, aiming to improve the overall nucleoside yield. Microwave irradiation at 150 W and 65 • C for 60 min gave the best results, and nucleoside isolated yields are given in Table 1. The four isomers differing in the anomeric configuration and linkage of the glycosyl group to the purine, either at the N 7 or the N 9 position, were formed by reacting the perbenzylated methyl α-D-mannopyranoside (3), but both the N 9 anomers were obtained in higher yield than the corresponding N 7 isomers. However, by reacting the perbenzylated methyl 6-deoxy-D-mannopyranoside donor (4), only the N 9 isomers were obtained in good yield, while the N 7 isomers were detected as traces (≤1%) by 1 H NMR. Replacing the benzyl by the acetyl protecting group, the peracetylated L-rhamnose gave the N 9 -linked α-nucleoside (12) as the single product, isolated in 63% yield (Table 1, Scheme 2). Compound 13 was then obtained in 15% isolated yield by reacting 12 with sodium hydroxide, tetrabutylammonium iodide, and benzyl bromide. Finally, with the perbenzylated methyl 6-deoxy-α-L-mannoside (14) as the starting material, only the N 9 isomers were obtained, as also observed for the reaction with compound 4 (Table 1, Scheme 3). deoxy-α-L-mannoside (14) as the starting material, only the N 9 isomers were obtained, as also observed for the reaction with compound 4 (Table 1, Scheme 3). Both the anomer configuration and the ligation at N 7 or N 9 were determined based on 1 H and 13 C NMR data and COSY, HMQC, and HMBC experiments. The chemical shift of purine carbon 5 is characteristic and permits to distinguish between the kinetic regioisomer N 7 and the N 9 thermodynamic one [19]. These values were between δ 118 and 112 ppm for compounds 7 and 8 (N 7 isomers), while ligation at N 9 resulted in chemical shifts of C-5 between ca δ 129 and 122 ppm (as in compounds 5, 6, 9, 10, 12, 13, and 15, 16).
By the NMR spectra analysis, the structures of ten nucleosides were proposed, six of them being unexpected. Usually, the most stable glycosyl conformation presents its voluminous protecting groups in equatorial positions, which does not happen in nucleosides 5, 7, 9, 12, and 15. Moreover, compound 6 seems to present a non-usual β-conformation, and 12 (α-anomer) was obtained as a single product. These unexpected results led us to perform computational studies, which unambiguously confirmed the proposed structures deduced from NMR data. Figure 2 and Tables S1-S7 in Supplementary Materials show the comparison of experimental and calculated NMR data, and the formation of 12 as a single product ( Table 2) was unambiguously confirmed. The tables comparing Both the anomer configuration and the ligation at N 7 or N 9 were determined based on 1 H and 13 C NMR data and COSY, HMQC, and HMBC experiments. The chemical shift of purine carbon 5 is characteristic and permits to distinguish between the kinetic regioisomer N 7 and the N 9 thermodynamic one [19]. These values were between δ 118 and 112 ppm for compounds 7 and 8 (N 7 isomers), while ligation at N 9 resulted in chemical shifts of C-5 between ca δ 129 and 122 ppm (as in compounds 5, 6, 9, 10, 12, 13, and 15, 16).
The anomeric configuration was assigned from the coupling constant 3 J 1 ,2 and confirmed by COSY and NOESY experiments. The coupling constants 3 J 1 ,2 for compounds 5, 7, 12, and 15 were 8.0 Hz, 9.3 Hz, 6.3 Hz, and 7.1 Hz, respectively, thus suggesting the presence of trans-diaxial like protons as in 1 C 4 conformation for the D series and 4 C 1 for the L series. The coupling constant 3 J 1 ,2 of 7.4 Hz for the β-anomer (6) was unexpected, indicating a conformation possibly bearing H-1 and H-2 approaching a parallel orientation, tentatively depicted in Scheme 1. 1 H NMR spectra of the β-anomers 8, 10 and 16 show coupling constants 3 J 1 ,2 lower than 1 Hz, as expected.
By the NMR spectra analysis, the structures of ten nucleosides were proposed, six of them being unexpected. Usually, the most stable glycosyl conformation presents its voluminous protecting groups in equatorial positions, which does not happen in nucleosides 5, 7, 9, 12, and 15. Moreover, compound 6 seems to present a non-usual β-conformation, and 12 (α-anomer) was obtained as a single product. These unexpected results led us to perform computational studies, which unambiguously confirmed the proposed structures deduced from NMR data. Figure 2 and Tables S1-S7 in Supplementary Materials show the comparison of experimental and calculated NMR data, and the formation of 12 as a single product ( Table 2) was unambiguously confirmed. The tables comparing experimental and calculated NMR data can be found in the Supplementary Materials (Tables S1-S7), the latter in line with the experimental results. Interestingly, according to the acquired NMR data and the knowledge to date, all the studied nucleosides exhibited a preferred equatorial orientation of the purine base, which is in full agreement with the exoanomeric effect known for N-nucleosides [20]. Table 2. Energy difference between the α and the β anomers for compound 12.

Anticholinesterase Activity
The in vitro inhibitory activities of novel purine nucleosides 5-10, 12, 13 and 15, and 16 were evaluated against both heterologous (electric eel eeAChE, horse eqBChE) and human (recombinant) cholinesterases isoforms, by applying the Ellmann colorimetric assay, with slight modifications [21,22]. The results are summarized in Table 3, thus reporting the half maximal inhibitory concentration (IC50) values, or the percent of inhibition at the final tested concentration of 20 μM, with an exception for compounds 15 and 16, which were tested at 5 μM due to their lower water solubility, but they were not able to produce more than 50% of inhibition for eeAChE and eqBChE. Compounds 6, 8, 9, 10, and 13 were

Anticholinesterase Activity
The in vitro inhibitory activities of novel purine nucleosides 5-10, 12, 13 and 15, and 16 were evaluated against both heterologous (electric eel eeAChE, horse eqBChE) and human (recombinant) cholinesterases isoforms, by applying the Ellmann colorimetric assay, with slight modifications [21,22]. The results are summarized in Table 3, thus reporting the half maximal inhibitory concentration (IC 50 ) values, or the percent of inhibition at the final tested concentration of 20 µM, with an exception for compounds 15 and 16, which were tested at 5 µM due to their lower water solubility, but they were not able to produce more than 50% of inhibition for eeAChE and eqBChE. Compounds 6, 8, 9, 10, and 13 were the most interesting ones. Nucleosides 8 and 9 show eeAChE IC 50 values of 2.6 and 4.7 µM, 6 and 13 show eqBChE IC 50 values of 4.3 and 5.1 µM, respectively, whilst compound 10 proved moderate inhibition of BChE. Furthermore, compound 9 was an eeAChE selective inhibitor, while compounds 7 (IC 50 = 10 µM) and 13 were selective for eqBChE. The most potent compound was the dual inhibitor compound 8.
The reported activities highlight that sugar, the geometry of the glycosidic bond, and the regioisomerism of substitution pattern on the nucleoside affected the anticholinesterase activity. It has been, indeed, reported that depending on sugars stereochemistry, the acceptor/donor hydrogen bond properties of glycosylated compounds can significantly change and deeply affect binding properties to the target [23].
The relationship between glycone D-configuration, N-ligation, and selectivity is in agreement with our previous findings for related nucleosides with a 6-chloropurine in their structure, while those for the L-configuration are herein reported for the first time [19,24]. With the exception of the 2,3,4-tri-O-benzyl-6-deoxy-α-D-mannosyl purine derivative 9, exhibiting a glycone 1 C 4 conformation and reported as a selective AChE inhibitor, compounds bearing a 2,3,4,6-tetra-O-benzyl-β-D-mannosyl fragment (6 and 8) showed from moderate to good inhibition of heterologous AChE and BChE and resulted most active than the quite inactive corresponding α anomers. Moreover, regarding N 9 or N 7 ligation of sugar, the N 7linked β-glycosyl regioisomer 8 proved a better activity than the N 9 regioisomers against both cholinesterases, thus resulting in compound 8 as a dual inhibitor against human cholinesterase isoforms as well.
Amongst the ten synthesized nucleosides, the peracetylated N 9 -linked 6-deoxy-α-L-manno nucleoside 12 was not active towards any of the cholinesterases, in agreement with previous findings, which revealed that aromatic rings are preferred for the anticholinesterase activity of the studied purine nucleosides [24], and all the benzylated compounds, except the N 9 -linked α-D-manno nucleoside 5, were active at some extent. However, benzylation of amide nitrogen in N 9 -linked N-benzyl 6-deoxy-α-L-manno nucleoside 13, exhibiting a glycosyl 4 C 1 conformation, deeply modified the activity profile, thus inducing a selective inhibition of the equine BChE, which disappeared in human BChE.
Substitution of D-mannosyl with L-mannosyl sugar moiety resulted in a dramatic reduction of aqueous solubility in compounds 15 and 16, which were tested at a lower concentration (5 µM) without showing any significant inhibition.
In order to verify compound behavior against the human cholinesterases, the most potent inhibitors (6, 8, and 9), dual inhibitor 10, and nucleosides 15 and 16 were tested against hAChE and hBChE. The previously known dual inhibitors 6, 8, and 10 seemed to keep this property, however, with a slight IC 50 value variation. Compound 6 IC 50 value for hAChE increased from 10.6 to 17.0 µM, while the value for hBChE did not undergo such a variation (from 4.3 to 4.6 µM), suggesting a higher affinity of this compound to BChE. Nucleoside 8, with IC 50 values of 4.8 µM and 4.6 µM for hAChE and hBChE, respectively, seemed to keep the dual inhibition with no greater affinity for one of the enzymes. Furthermore, for compound 10, a higher IC 50 value for hAChE (IC 50 = 17.8 µM) was found when compared to eeAChE. However, the IC 50 value found for hBChE was lower than the one found for eqBChE. Nucleoside 9, a selective eeAChE inhibitor, in the human enzymes inhibitory assays turned into a hBChE selective inhibitor, while compound 13 showed no activity against the human enzymes. For compounds 15 and 16, the lower concentration which was tested was not able to produce 50% of enzyme inhibition. Finally, also in Table 3, the inhibition results for the latter compounds at 5 µM final concentration are disclosed. Nucleosides 15 seemed to have a slightly higher affinity to AChE, producing 31 and 28% inhibition of eeAChE and hAChE, respectively, while 16 seemed to present a higher inhibition towards BChE, with a higher inhibition value for the hBChE (30%). However, both seem to show potential as non-selective inhibitors. These results then suggest that by improving these compounds' water solubility, the inhibition rates could increase, and the nucleosides could become interesting candidates as cholinesterase inhibitors.

Chelating Activity
The chelating activity of all purine nucleoside probes was evaluated by the UV-Visible method in order to determine compounds' absorption spectra in the absence and in the presence of Cu 2+ at concentrations ranging from 1 µM to 10 µM. Amongst nucleosides 5-10 and 12, the latter displayed the most interesting profile (Figure 3). The resulting wavelength shift from λ = 280 to λ = 325 nm indicates a conformational rearrangement induced by interaction with Cu 2+ . Furthermore, selectivity towards the biologically relevant ions Zn 2+ , Fe 2+, and Al 3+ was also investigated, thus demonstrating for most of them a negligible quenching. Hence, although 12, the only acetylated nucleoside, did not show promising cholinesterase inhibition, it became the first nucleoside-based molecule with metal chelating properties and, therefore, became a starting point for the investigation of structural mod- Hence, although 12, the only acetylated nucleoside, did not show promising cholinesterase inhibition, it became the first nucleoside-based molecule with metal chelating properties and, therefore, became a starting point for the investigation of structural modifications that could lead to the desired dual-target compounds against AD.
In order to understand which structural alterations could be carried out to achieve our goal, computational studies were performed. Molecular electrostatic potential maps, in which the red and blue hues represent negatively and positively charged regions, respectively, and green tones depict neutral regions, were generated for compounds 5-7, 9, and 12. In all cases, a significant electron density in the amide group of the purine base was found, and the benzyl units were found to be almost neutral. Moreover, compound 7, exhibiting N 7 purine ligation, also exhibits high electron density between atoms N3 and N9, and compound 12 shows some negative density in the acetyl groups. Hereinafter, the generated maps for compounds 7 and 12 are disclosed (Figure 4), the remaining being found in the Supplementary Materials (Figures S48-S50).  Based on the molecular electrostatic potential map generated for compound 12, the geometry for the complex compound 12-Cu 2+ was optimized, starting with the copper ion placed near the amide group of the purine. The electronic transitions in water for the two possible conformers, "b" and "c," of compound 12 in the absence and in the presence of Cu 2+ ions are presented in Figure 5. In comparison with the experimental results, the calculated transitions are displaced to shorter wavelengths; however, the bathochromic shift Based on the molecular electrostatic potential map generated for compound 12, the geometry for the complex compound 12-Cu 2+ was optimized, starting with the copper ion placed near the amide group of the purine. The electronic transitions in water for the two possible conformers, "b" and "c," of compound 12 in the absence and in the presence of Cu 2+ ions are presented in Figure 5. In comparison with the experimental results, the calculated transitions are displaced to shorter wavelengths; however, the bathochromic shift observed experimentally with the addition of Cu 2+ is also noticed in the calculated results. The larger red shift in λ max from 254 to 284 nm is observed for conformer "b." Also, the energy difference between both complexes listed in Table 4 shows that complex 12a-Cu 2+ is more stable. Finally, the optimized molecular geometry of complexes 12-Cu 2+ is shown in Figure 6, and the electrostatic potential map for the most stable complex, the 12a-Cu 2+ complex, is presented in Figure 7.           Furthermore, having noticed that the glycosyl groups have a small contribution to the molecule electrostatic potential, the interaction with the metal ions of the purine in Furthermore, having noticed that the glycosyl groups have a small contribution to the molecule electrostatic potential, the interaction with the metal ions of the purine in conformer "b" was studied after replacing the glycosyl unit of compound 12 with a hydrogen atom at N9. In Figure 8, the simulated absorption spectra based on electronic transitions calculated for 6-benzamidopurine in the absence and in the presence of the metal ions in water are presented. The spectra were simulated based on the assumption that all the solutions have the same concentration (10 µM) and that the metal ions are totally binded to the purine. As observed experimentally, the calculated absorption spectra indicate that the binding constant between the purine and Cu 2+ is larger than the one with the other metal ions. Also, it is important to stress that the spin multiplicity of Cu 2+ is doublet while all other metal ions' spin multiplicity is singlet. conformer "b" was studied after replacing the glycosyl unit of compound 12 with a hydrogen atom at N9. In Figure 8, the simulated absorption spectra based on electronic transitions calculated for 6-benzamidopurine in the absence and in the presence of the metal ions in water are presented. The spectra were simulated based on the assumption that all the solutions have the same concentration (10 μM) and that the metal ions are totally binded to the purine. As observed experimentally, the calculated absorption spectra indicate that the binding constant between the purine and Cu 2+ is larger than the one with the other metal ions. Also, it is important to stress that the spin multiplicity of Cu 2+ is doublet while all other metal ions' spin multiplicity is singlet. Having the computational studies performed, three new nucleosides, 13, 15, and 16, were synthesized aiming at confirming the latter findings, 13 being a benzylated analogue of compound 12 with an additional N-benzyl group, 15 a benzylated analogue of 12 and 16 the β-anomer of 15. Amongst these new compounds, 15 and 16 displayed the most interesting profiles since 13 did not show chelating activity towards any of the biologically relevant ions. Compound 15 was found to be a selective copper chelator, as it is shown in Figure 9. In the complex 15-Cu 2+ spectrum, there is a perceptible wavelength shift from λ Having the computational studies performed, three new nucleosides, 13, 15, and 16, were synthesized aiming at confirming the latter findings, 13 being a benzylated analogue of compound 12 with an additional N-benzyl group, 15 a benzylated analogue of 12 and 16 the β-anomer of 15. Amongst these new compounds, 15 and 16 displayed the most interesting profiles since 13 did not show chelating activity towards any of the biologically relevant ions. Compound 15 was found to be a selective copper chelator, as it is shown in Figure 9. In the complex 15-Cu 2+ spectrum, there is a perceptible wavelength shift from λ = 284 to λ = 334 nm, indicating a conformational rearrangement due to the interaction with Cu(II). Through the analysis of the remaining spectra, it is possible to verify that there is no relevant interaction of this nucleoside with the other ions. Compound 16 was shown to be a non-selective chelator, as the resulting spectra showed conformational rearrangements induced by interaction with all the tested ions, as can be verified in Figure 10. Interactions with copper, zinc, iron and aluminum ions result in wavelength shifts from λ = 290 nm to λ = 345, 346, 339, and 351 nm, respectively.  To conclude, the computational studies revealed that for metal chelation, the glycosyl units had a small contribution, being the amide group of the purine the most relevant. The results obtained with compounds 13 and 15 confirmed these findings. Nucleoside 15, the benzylated analogue of 12, also showed selective copper chelation over the remaining metals, proving that the glycosyl-protecting groups had no influence on this property. Moreover, compound 13, being similar to 15 with an extra N-benzyl group in the purine moiety, showed no activity, showing that the metal chelation site should be found in the purine. Finally, the computational studies indicated that this purine base should have a higher affinity to Cu 2+ ; however, experimentally, compound 16 showed no selectivity over the four metal ions.

Discussion
A series of novel nucleosides were synthesized and tested, leading to new insights into the structural features required for BChE or AChE selective inhibition and Cu 2+ binding. Starting from lead compounds 1 and 2, the replacement of 6-chloropurine by 6-benzamidopurine led to two novel and selective Cu 2+ chelating agents 12 (with the peracetylated α-L-rhamnosyl group) and its analogue perbenzylated compound 15 (with the perbenzylated α-L-rhamnosyl group). Moreover, the latter modifications led to a novel metal Figure 10. Absorbance spectra of compound 16 in the absence and in the presence of copper ion, zinc ion, iron ion, and aluminum ion (10 µM) and spectra of copper, zinc, iron, and aluminum ions, determined by UV-Visible spectroscopy with water/DMSO (90:10) as the solvent.
To conclude, the computational studies revealed that for metal chelation, the glycosyl units had a small contribution, being the amide group of the purine the most relevant. The results obtained with compounds 13 and 15 confirmed these findings. Nucleoside 15, the benzylated analogue of 12, also showed selective copper chelation over the remaining metals, proving that the glycosyl-protecting groups had no influence on this property. Moreover, compound 13, being similar to 15 with an extra N-benzyl group in the purine moiety, showed no activity, showing that the metal chelation site should be found in the purine. Finally, the computational studies indicated that this purine base should have a higher affinity to Cu 2+ ; however, experimentally, compound 16 showed no selectivity over the four metal ions.

Discussion
A series of novel nucleosides were synthesized and tested, leading to new insights into the structural features required for BChE or AChE selective inhibition and Cu 2+ binding. Starting from lead compounds 1 and 2, the replacement of 6-chloropurine by 6-benzamidopurine led to two novel and selective Cu 2+ chelating agents 12 (with the peracetylated α-L-rhamnosyl group) and its analogue perbenzylated compound 15 (with the perbenzylated α-L-rhamnosyl group). Moreover, the latter modifications led to a novel metal chelator 16 ( 1 C 4 conformer), obtained with the same substrate as 15, with an affinity for Cu 2+ , Zn 2+ , Fe 2+ , and Al 3+ . Importantly, all synthesized nucleosides showed the equatorial orientation of the base, most probably resulting from the exoanomeric and steric effects. Amongst the new nucleosides tested, 12, the only acetylated compound, did not show cholinesterase inhibition, while all the benzylated compounds, except 5 (the N 9linked α-D-mannosyl nucleoside), were active to some extent, highlighting the importance of aromatic rings for this property. With regard to eeAChE and eqBChE, the selective inhibitors exhibit a glycosyl 1 C 4 conformation regardless of the ligation position to the purine base, while the dual inhibitors show a glycone 4 C 1 conformation or derived thereof. Moreover, N 9 ligation is shown for AChE and N 7 ligation for BChE, in agreement with our previous results shown by the 6-chloropurine series [19]. However, for human cholinesterases, only compound 9 (N 9 -linked 6-deoxy-α-D-manno) was selective for BChE inhibition, and no nucleoside showed AChE selective inhibition, while the known eeAChE and eqBChE dual inhibitors kept this property, with 6 (N 9linked β-D-manno) and 10 (N 9linked 6-deoxy-β-D-manno) having a higher affinity for BChE and 8 (N 7 -linked β-D-manno) showing no greater affinity for any of the enzymes. Finally, compounds 15 and 16, with the glycone L configuration, N 9 purine ligation, and aromatic rings in their structure, showed metal chelation and some inhibition of the cholinesterases, both the animal and the human ones, at the tested concentration, becoming the first nucleoside-based molecules with potential to act as dual-target compounds against AD. This work illustrates the role of sugars by tuning bioactivity and selectivity, encouraging further investigation of this family of purine nucleosides towards glycone structure optimization to access anticholinesterase compounds and copper chelators as multitarget agents against AD.

Materials and Methods
Reagents were purchased from commercial suppliers (Sigma Aldrich, St. Louis, MI, USA) and used without further purification. The starting materials 3, 4, and 11 were available from Sigma Aldrich, Glentham Life Sciences (Corsham, UK), Toronto Research Chemicals (North York, ON, Canada). Compound 14 was obtained by benzylation of methyl α-Drhamnoside purchased from Alfa Aesar. The physical and spectroscopic data of 14 was in full agreement with those previously reported [26]. Microwave-assisted synthesis was performed with a CEM Discover and Explorer SP. NMR spectra were recorded on a BRUKER Advance 400 spectrometer at 298 K or on a BRUKER Advance III 500 spectrometer at 298 K. Spectra were referenced internally to residual proton-solvent ( 1 H) or carbon-solvent ( 13 C) resonances, and reported relative to tetramethylsilane (0 ppm). Chemical shifts (δ) are given in ppm and coupling constants (J) in Hz. UV spectra were recorded on a SHIMADZU UV-1800 spectrometer with quartz cells. Optical rotations were determined on a Perkin-Elmer 341 polarimeter and on an Anton Paar MCP 100 polarimeter. HRMS analyses were performed on an Agilent 6530 Accurate Mass Q-TOF, only significant m/z peaks, with their percentage of relative intensity in parentheses, are reported. TLC was performed on silica gel (Merck 5554), and column chromatography for compound isolation was carried out with silica gel. Solvents were dried before use according to the conventional procedures.

UV-Visible Chelating Studies
UV-Visible spectroscopy offers the advantage of high sensitivity towards small changes that affect the electronic properties of ligand receptors [27], so absorption spectra of all reported compounds have been evaluated by recording UV-visible properties in subsequent conditions: 1.5 mL 1 × 10 −4 M of the probe are added to 1.5 mL 2 × 10 −5 M of each ion chloride salt's solution (Copper Chloride (CuCl 2 ), Zinc Chloride (ZnCl 2 ), Iron (II) Chloride (FeCl 2 ) and Aluminum Chloride (AlCl 3 )). A mixture of distilled water/DMSO 95:5 has been selected as a solvent for compounds 5-10 and 12. For compounds 14-16, the selected solvent was a mixture of distilled water/DMSO 90:10 due to poor water solubility. Furthermore, a titration assay of each compound has been carried out by adding increasing concentration of selected ion (final ion concentration = 1 µM, 2 µM, 3 µM, 5 µM, and 10 µM).

Computational Studies
The initial geometries were optimized in Gaussian16 with CAM-B3LYP/6-31G(d) and integrated equation formalism polarized continuum model (IEFPCM) for water [28][29][30]. Final geometries were optimized with CAM-B3LYP/def2TZVP and IEFPCM for water. The vibrational frequencies were checked for true minima in all cases (no imaginary frequencies were found). To accelerate the search for the best configurations, the molecules were split into two parts: the sugar unit and the purine moiety. With respect to the purine moiety, the geometry optimization started from some initial purine geometries with small differences in the dihedral angle between the rings, and the optimized geometries always ended up in two conformers. The geometry optimization for the glycosyl groups followed the same procedure. The optimization was started from a series of different initial geometries for each sugar unit, and the most stable conformer was selected. For the final geometry optimizations, the best glycosyl conformer found was coupled with the two purine conformers, and a full optimization of the whole molecule was performed.
NMR parameters were evaluated from single-point calculations with CAM-B3LYP/def2TZVP and IEFPCM for water. The computed NMR shielding tensors were converted to chemical shifts with empirical scaling factors determined from a test set of 80 molecules at the same level of theory [31].