Effect of S–Se Bioisosteric Exchange on Affinity and Intrinsic Efficacy of Novel N-acylhydrazone Derivatives at the Adenosine A2A Receptor

In this work, we evaluated the conformational effect promoted by the isosteric exchange of sulfur by selenium in the heteroaromatic ring of new N-acylhydrazone (NAH) derivatives (3–8, 13, 14), analogues of the cardioactive compounds LASSBio-294 (1) and LASSBio-785 (2). NMR spectra analysis demonstrated a chemical shift variation of the iminic Csp2 of NAH S/Se-isosters, suggesting a stronger intramolecular chalcogen interaction for Se-derivatives. To investigate the pharmacological profile of these compounds at the adenosine A2A receptor (A2AR), we performed a previously validated functional binding assay. As expected for bioisosteres, the isosteric-S/Se replacement affected neither the affinity nor the intrinsic efficacy of our NAH derivatives (1–8). However, the N-methylated compounds (2, 6–8) presented a weak partial agonist profile at A2AR, contrary to the non-methylated counterparts (1, 3–5), which appeared as weak inverse agonists. Additionally, retroisosterism between aromatic rings of NAH on S/Se-isosters mimicked the effect of the N-methylation on intrinsic efficacy at A2AR, while meta-substitution in the phenyl ring of the acyl moiety did not. This study showed that the conformational effect of NAH-N-methylation and aromatic rings retroisosterism changed the intrinsic efficacy on A2AR, indicating the S/Se-chalcogen effect to drive the conformational behavior of this series of NAH.

Preliminary data obtained from CEREP (Study Number: 15180, 2008) on a panel of GPCRs indicated that both LASSBio-294 (1) [13] and LASSBio-785 (2) (unpublished data) had a moderate affinity for the adenosine A 2A receptor (A 2A R), a putative target for these drugs. The careful structure-property-activity relationship of LASSBio-29 Bio-785 (2) demonstrated that the N-methylation of the NAH moiety ha fect on this scaffold [14], leading to different electronical profiles (identif NMR studies) and physicochemical properties [15,16]. Follow up stud fraction indicated that the two compounds adopted an unexpected distin [15], with 1 having a hair-pin like conformation ( Figure 1A) and 2 a mation ( Figure 1B) [17]. In both conformations ( Figure 1A,B), it is out that the heteroatom of the ring at the imine double bond adopted the s tation viz-à-viz the Nsp 2 corresponding to a typical chalcogen interact mational behavior was investigated, confirming the steric orientation of ring at the terminus of LASSBio-294 (1) and LASSBio-785 (2), as an accou chalcogen interaction [18][19][20] between S-atom and the electron-pair of t the NAH moiety [21]. This observation suggests that this kind of inte helpful in drug design to address conformational restrictions in bioacti tives [21] (Figure 2). The careful structure-property-activity relationship of LASSBio-294 (1) and LASSBio-785 (2) demonstrated that the N-methylation of the NAH moiety has a profound effect on this scaffold [14], leading to different electronical profiles (identified by 1 H and 13 C NMR studies) and physicochemical properties [15,16]. Follow up studies by X-ray diffraction indicated that the two compounds adopted an unexpected distinct conformation [15], with 1 having a hair-pin like conformation ( Figure 1A) and 2 a U-shape conformation ( Figure 1B) [17]. In both conformations ( Figure 1A,B), it is outstanding to note that the heteroatom of the ring at the imine double bond adopted the same steric orientation viz-à-viz the Nsp 2 corresponding to a typical chalcogen interaction. This conformational behavior was investigated, confirming the steric orientation of the 2-thiophene ring at the terminus of LASSBio-294 (1) and LASSBio-785 (2), as an accountable σ-hole or chalcogen interaction [18][19][20] between S-atom and the electron-pair of the N-sp 2 atom of the NAH moiety [21]. This observation suggests that this kind of interaction could be helpful in drug design to address conformational restrictions in bioactive NAH derivatives [21] (Figure 2). The careful structure-property-activity relationship of LASSBio-294 (1) and LASS-Bio-785 (2) demonstrated that the N-methylation of the NAH moiety has a profound effect on this scaffold [14], leading to different electronical profiles (identified by 1 H and 13 C NMR studies) and physicochemical properties [15,16]. Follow up studies by X-ray diffraction indicated that the two compounds adopted an unexpected distinct conformation [15], with 1 having a hair-pin like conformation ( Figure 1A) and 2 a U-shape conformation ( Figure 1B) [17]. In both conformations ( Figure 1A,B), it is outstanding to note that the heteroatom of the ring at the imine double bond adopted the same steric orientation viz-à-viz the Nsp 2 corresponding to a typical chalcogen interaction. This conformational behavior was investigated, confirming the steric orientation of the 2-thiophene ring at the terminus of LASSBio-294 (1) and LASSBio-785 (2), as an accountable σ-hole or chalcogen interaction [18][19][20] between S-atom and the electron-pair of the N-sp 2 atom of the NAH moiety [21]. This observation suggests that this kind of interaction could be helpful in drug design to address conformational restrictions in bioactive NAH derivatives [21] (Figure 2).  Since selenium, an isoster of sulfur in the chalcogen family, has a comparably higher polarizability, we hypothesized that a classical isosteric exchange of S to Se [22,23] in the five-member ring could intensify the chalcogen interaction. Considering this, we prepared a series of new NAH compounds structurally related to LASSBio-294 (1) and LASSBio-785 (2). The aim was to evaluate how the isosteric change, combined with the new chalcogen-induced conformation, would affect the affinity and/or intrinsic efficacy of these compounds at the adenosine A 2A R [13,24].
Our results showed that N-methylation of Se-containing NAH compounds, changed their intrinsic efficacy, with the N-methylated derivatives (2, 6, 7, and 8) having a weak partial agonist profile in contrast to the non-methylated counterparts (1, 3, 4, and 5), which appeared as weak inverse agonists. It was also observed that isosteric S-Se replacement at the five member ring did not affect affinity or intrinsic efficacy at A 2A R. Moreover, we observed that isosteric change has a marked conformational effect in the selenophenic derivatives (3, 5, 6, 8, and 12) probably due the stronger Se-Nsp 2 intramolecular chalcogen interaction, which could be able to mimic the N-methylation effect observed in the S-isostere LASSBio-785 (2).

Molecular Design
The initial approach was to synthesize selenium isosters of 1 and 2, the new NAH derivatives 3 and 6 ( Figure 3). Additionally, previous studies of structurally similar compounds to 1 and 2 demonstrated that the meta-position with a methoxy group in the phenyl ring leads to increased potency in the adenosine receptors [25]. Hence, we also prepared novel isosteric analogues of meta-substituted compounds (4, 5, 7, and, 8, Figure 3).
Since selenium, an isoster of sulfur in the chalcogen family, has a comparably highe polarizability, we hypothesized that a classical isosteric exchange of S to Se [22,23] in the five-member ring could intensify the chalcogen interaction. Considering this, we pre pared a series of new NAH compounds structurally related to LASSBio-294 (1) and LASSBio-785 (2). The aim was to evaluate how the isosteric change, combined with the new chalcogen-induced conformation, would affect the affinity and/or intrinsic efficacy of these compounds at the adenosine A2AR [13,24].
Our results showed that N-methylation of Se-containing NAH compounds, changed their intrinsic efficacy, with the N-methylated derivatives (2, 6, 7, and 8) having a weak partial agonist profile in contrast to the non-methylated counterparts (1, 3, 4, and 5) which appeared as weak inverse agonists. It was also observed that isosteric S-Se re placement at the five member ring did not affect affinity or intrinsic efficacy at A2AR Moreover, we observed that isosteric change has a marked conformational effect in the selenophenic derivatives (3, 5, 6, 8, and 12) probably due the stronger Se-Nsp 2 intramo lecular chalcogen interaction, which could be able to mimic the N-methylation effect ob served in the S-isostere LASSBio-785 (2).

Molecular Design
The initial approach was to synthesize selenium isosters of 1 and 2, the new NAH derivatives 3 and 6 ( Figure 3). Additionally, previous studies of structurally simila compounds to 1 and 2 demonstrated that the meta-position with a methoxy group in the phenyl ring leads to increased potency in the adenosine receptors [25]. Hence, we also prepared novel isosteric analogues of meta-substituted compounds (4, 5, 7, and, 8, Figure  3).

Chemistry
The synthesis of the new NAH compounds (3)(4)(5) was performed in three steps [4]. First, the esters (10a and 10b) [26,27] were prepared by Yamada oxidation in methanol of the corresponding commercial aldehydes (9a and 9b) [28], and then submitted to hydrazinolysis reaction in ethanol to produce the key hydrazide intermediates 11a and 11b (Scheme 1) [29,30]. Finally, the desired NAH derivatives (3)(4)(5) were obtained in high overall yields by an acid-catalyzed condensation reaction with the corresponding aldehydes (2-formylthiophene or 2-formylselenophene, Scheme 1). Next, the N-methylation reaction was performed using iodomethane and potassium carbonate in acetone, at mild heating [15] (Scheme 2). The N-CH3-NAH compounds 6-8 were obtained after chromatographic column purification.   (Table 1). All signals related to hydrogens of the heterocycle ring have a downfield effect upon substitution o sulfur to selenium, suggesting that its polarizability further deshields the neighboring hydrogens. No shield effect was observed on the -N=CH-iminic proton (Table 1) of the NAH derivatives (1-8), with all signals within the range of δ 8.5-8.6 ppm for NAH de rivatives 1, 3-5 and δ 8.1-8.2 ppm for N-methylated compounds 2, 6-8. These unchanged shifts are, however, a strong indication that the E isomerism of the iminic bond was no affected in the new structures [31].
For the 13 C-NMR spectra, a significant difference was observed in the shift of the iminic carbon atom between the isosters pairs 1 and 3; 4 and 5 (Supplementary Material  Table 1). Displacements in the shifts of carbons can also be correlated to the intramolec ular chalcogen interaction [32,33]. A shift of almost 5 ppm was observed between the isosteric pairs (e.g., compounds 1 and 3 (139.2 ppm to 143.9 pm)). This indicates a weakening of the resonance system between imine and the selenophene ring as a con sequence of the σ-hole interaction between Se-Nsp 2 in isosteric compounds 3, 5, 6, and 8 All of this suggests a stronger intramolecular interaction on the selenophenic derivatives (3, 5, 6, and 8). Next, the N-methylation reaction was performed using iodomethane and potassium carbonate in acetone, at mild heating [15] (Scheme 2). The N-CH 3 -NAH compounds 6-8 were obtained after chromatographic column purification. heme for preparation of N-acylhydrazonic compounds 3, 4, and 5. Reaction conditio unds 3-5.
Next, the N-methylation reaction was performed using iodomethan carbonate in acetone, at mild heating [15] (Scheme 2). The N-CH3-NAH were obtained after chromatographic column purification. Detailed 1 H-NMR analysis of the isosteric pairs (1 and 3, 4 and 5, demonstrated the electronic effect of the chalcogen modification (Tab related to hydrogens of the heterocycle ring have a downfield effect upo sulfur to selenium, suggesting that its polarizability further deshields hydrogens. No shield effect was observed on the -N=CH-iminic proton NAH derivatives (1-8), with all signals within the range of δ 8.5-8.6 p rivatives 1, 3-5 and δ 8.1-8.2 ppm for N-methylated compounds 2, 6-8. T shifts are, however, a strong indication that the E isomerism of the imin affected in the new structures [31].
For the 13 C-NMR spectra, a significant difference was observed in  (Table 1). All signals related to hydrogens of the heterocycle ring have a downfield effect upon substitution of sulfur to selenium, suggesting that its polarizability further deshields the neighboring hydrogens. No shield effect was observed on the -N=CH-iminic proton (Table 1) of the NAH derivatives (1-8), with all signals within the range of δ 8.5-8.6 ppm for NAH derivatives 1, 3-5 and δ 8.1-8.2 ppm for N-methylated compounds 2, 6-8. These unchanged shifts are, however, a strong indication that the E isomerism of the iminic bond was not affected in the new structures [31].
For the 13 C-NMR spectra, a significant difference was observed in the shift of the iminic carbon atom between the isosters pairs 1 and 3; 4 and 5 (Supplementary Material, Table S1). Displacements in the shifts of carbons can also be correlated to the intramolecular chalcogen interaction [32,33]. A shift of almost 5 ppm was observed between the isosteric pairs (e.g., compounds 1 and 3 (139.2 ppm to 143.9 pm)). This indicates a weakening of the resonance system between imine and the selenophene ring as a consequence of the σ-hole

Pharmacological Evaluation
The ability of our compounds to interact with the A2A receptor (A2AR) was evaluated through competition for the binding of [ 3 H]-ZM241385 used as a probe for the A2AR present endogenously in rat striatum [24].
In the basal conditions (medium I, Table 2), the isosteric pairs 1 and 3, 4 and 5, 2 and 6, and 7 and 8 had similar IC50 values (Table 2). Therefore, the isosteric replacement of S by Se did not affect the affinity of our compounds. NAH N-methylation also has no significant effect on affinity for A2AR, albeit there is a tendency for a small decrease in affinity. This result indicates that the change of conformation due to the N-methylation has no significant effect on the binding affinity of these ligands, at least for the A2AR in its high-affinity conformation (medium I, Table 2). Table 2. Affinity and intrinsic efficacy of the N-acylhydrazone derivatives at the A2A receptor (A2AR) present in membrane preparations from rat striatum. Average curves of three-four independent experiments (in triplicate) were analyzed by non-linear regression to estimate the IC50 values with their 95% confidence interval (95% CI). For each compound, the Na + -shifts were calculated by dividing the IC50 value obtained in the medium containing NaCl (II) by the IC50 value obtained in the medium containing MgCl2 (I) and statistically analyzed by paired t test on log IC50 values (P). Na + -shifts lower than one indicates that compounds 1 and 3-5 are weak inverse agonists, whereas Na + -shifts higher indicate that N-methylated compounds 2 and 6-8 are weak partial agonists.

Pharmacological Evaluation
The ability of our compounds to interact with the A2A receptor (A2AR) was evaluated through competition for the binding of [ 3 H]-ZM241385 used as a probe for the A2AR present endogenously in rat striatum [24].
In the basal conditions (medium I, Table 2), the isosteric pairs 1 and 3, 4 and 5, 2 and 6, and 7 and 8 had similar IC50 values (Table 2). Therefore, the isosteric replacement of S by Se did not affect the affinity of our compounds. NAH N-methylation also has no significant effect on affinity for A2AR, albeit there is a tendency for a small decrease in affinity. This result indicates that the change of conformation due to the N-methylation has no significant effect on the binding affinity of these ligands, at least for the A2AR in its high-affinity conformation (medium I, Table 2). Table 2. Affinity and intrinsic efficacy of the N-acylhydrazone derivatives at the A2A receptor (A2AR) present in membrane preparations from rat striatum. Average curves of three-four independent experiments (in triplicate) were analyzed by non-linear regression to estimate the IC50 values with their 95% confidence interval (95% CI). For each compound, the Na + -shifts were calculated by dividing the IC50 value obtained in the medium containing NaCl (II) by the IC50 value obtained in the medium containing MgCl2 (I) and statistically analyzed by paired t test on log IC50 values (P). Na + -shifts lower than one indicates that compounds 1 and 3-5 are weak inverse agonists, whereas Na + -shifts higher indicate that N-methylated compounds 2 and 6-8 are weak partial agonists.

Pharmacological Evaluation
The ability of our compounds to interact with the A2A receptor (A2AR) was evaluated through competition for the binding of [ 3 H]-ZM241385 used as a probe for the A2AR present endogenously in rat striatum [24].
In the basal conditions (medium I, Table 2), the isosteric pairs 1 and 3, 4 and 5, 2 and 6, and 7 and 8 had similar IC50 values (Table 2). Therefore, the isosteric replacement of S by Se did not affect the affinity of our compounds. NAH N-methylation also has no significant effect on affinity for A2AR, albeit there is a tendency for a small decrease in affinity. This result indicates that the change of conformation due to the N-methylation has no significant effect on the binding affinity of these ligands, at least for the A2AR in its high-affinity conformation (medium I, Table 2). Table 2. Affinity and intrinsic efficacy of the N-acylhydrazone derivatives at the A2A receptor (A2AR) present in membrane preparations from rat striatum. Average curves of three-four independent experiments (in triplicate) were analyzed by non-linear regression to estimate the IC50 values with their 95% confidence interval (95% CI). For each compound, the Na + -shifts were calculated by dividing the IC50 value obtained in the medium containing NaCl (II) by the IC50 value obtained in the medium containing MgCl2 (I) and statistically analyzed by paired t test on log IC50 values (P). Na + -shifts lower than one indicates that compounds 1 and 3-5 are weak inverse agonists, whereas Na + -shifts higher indicate that N-methylated compounds 2 and 6-8 are weak partial agonists.

Pharmacological Evaluation
The ability of our compounds to interact with the A2A receptor (A2AR) was evaluated through competition for the binding of [ 3 H]-ZM241385 used as a probe for the A2AR present endogenously in rat striatum [24].
In the basal conditions (medium I, Table 2), the isosteric pairs 1 and 3, 4 and 5, 2 and 6, and 7 and 8 had similar IC50 values (Table 2). Therefore, the isosteric replacement of S by Se did not affect the affinity of our compounds. NAH N-methylation also has no significant effect on affinity for A2AR, albeit there is a tendency for a small decrease in affinity. This result indicates that the change of conformation due to the N-methylation has no significant effect on the binding affinity of these ligands, at least for the A2AR in its high-affinity conformation (medium I, Table 2). Table 2. Affinity and intrinsic efficacy of the N-acylhydrazone derivatives at the A2A receptor (A2AR) present in membrane preparations from rat striatum. Average curves of three-four independent experiments (in triplicate) were analyzed by non-linear regression to estimate the IC50 values with their 95% confidence interval (95% CI). For each compound, the Na + -shifts were calculated by dividing the IC50 value obtained in the medium containing NaCl (II) by the IC50 value obtained in the medium containing MgCl2 (I) and statistically analyzed by paired t test on log IC50 values (P). Na + -shifts lower than one indicates that compounds 1 and 3-5 are weak inverse agonists, whereas Na + -shifts higher indicate that N-methylated compounds 2 and 6-8 are weak partial agonists.

Pharmacological Evaluation
The ability of our compounds to interact with the A2A receptor (A2AR) was evaluated through competition for the binding of [ 3 H]-ZM241385 used as a probe for the A2AR present endogenously in rat striatum [24].
In the basal conditions (medium I, Table 2), the isosteric pairs 1 and 3, 4 and 5, 2 and 6, and 7 and 8 had similar IC50 values (Table 2). Therefore, the isosteric replacement of S by Se did not affect the affinity of our compounds. NAH N-methylation also has no significant effect on affinity for A2AR, albeit there is a tendency for a small decrease in affinity. This result indicates that the change of conformation due to the N-methylation has no significant effect on the binding affinity of these ligands, at least for the A2AR in its high-affinity conformation (medium I, Table 2). Table 2. Affinity and intrinsic efficacy of the N-acylhydrazone derivatives at the A2A receptor (A2AR) present in membrane preparations from rat striatum. Average curves of three-four independent experiments (in triplicate) were analyzed by non-linear regression to estimate the IC50 values with their 95% confidence interval (95% CI). For each compound, the Na + -shifts were calculated by dividing the IC50 value obtained in the medium containing NaCl (II) by the IC50 value obtained in the medium containing MgCl2 (I) and statistically analyzed by paired t test on log IC50 values (P). Na + -shifts lower than one indicates that compounds 1 and 3-5 are weak inverse agonists, whereas Na + -shifts higher indicate that N-methylated compounds 2 and 6-8 are weak partial agonists.

Pharmacological Evaluation
The ability of our compounds to interact with the A 2A receptor (A 2A R) was evaluated through competition for the binding of [ 3 H]-ZM241385 used as a probe for the A 2A R present endogenously in rat striatum [24].
In the basal conditions (medium I, Table 2), the isosteric pairs 1 and 3, 4 and 5, 2 and 6, and 7 and 8 had similar IC 50 values (Table 2). Therefore, the isosteric replacement of S by Se did not affect the affinity of our compounds. NAH N-methylation also has no significant effect on affinity for A 2A R, albeit there is a tendency for a small decrease in affinity. This result indicates that the change of conformation due to the N-methylation has no significant effect on the binding affinity of these ligands, at least for the A 2A R in its high-affinity conformation (medium I, Table 2).
As the pharmacodynamic properties of a drug depend not only on its affinity for a receptor but also on its capacity to activate or block the receptor (i.e., efficacy), we also used the Na+-shift assay previously validated for quantifying the intrinsic efficacy of our compounds [24]. This functional binding assay is based on the extended ternary complex model of G-protein coupled receptor (GPCR). According to this model, the activated conformation of GPCR is the one coupled to a G-protein. Moreover, the affinity of an agonist is much higher for the receptor in this conformation than for the uncoupled receptor [34]. In our assay, the substitution of Mg2+ (a divalent cation that favors the formation of the ternary complex) by Na+ (a negative allosteric modulator of agonist binding and activation in many GPCRs) shifts the competition curve of agonists, but not antagonists, to the right [24,35]. On the other hand, a shift of the curve to the left indicates that the compound is an inverse agonist. This profile was observed with all four non-methylated N-acylhydrazone derivatives (compounds 1 and 3-5), as exemplified in Figure 4A for LASSBio-2093 (5), and shown in Figure S1A,C,E for the other derivatives. Note that Na+-shifts to the left observed with these four compounds were small, but statistically significant, as indicated by a ratio (II/I) lower than one (p < 0.05, Table 2). Table 2. Affinity and intrinsic efficacy of the N-acylhydrazone derivatives at the A 2A receptor (A 2A R) present in membrane preparations from rat striatum. Average curves of three-four independent experiments (in triplicate) were analyzed by non-linear regression to estimate the IC 50 values with their 95% confidence interval (95% CI). For each compound, the Na + -shifts were calculated by dividing the IC 50 value obtained in the medium containing NaCl (II) by the IC 50 value obtained in the medium containing MgCl 2 (I) and statistically analyzed by paired t test on log IC 50 values (P). Na + -shifts lower than one indicates that compounds 1 and 3-5 are weak inverse agonists, whereas Na + -shifts higher indicate that N-methylated compounds 2 and 6-8 are weak partial agonists. 3.1 0.039

Compound
As the pharmacodynamic properties of a drug depend not only on its affinity for a receptor but also on its capacity to activate or block the receptor (i.e., efficacy), we also used the Na+-shift assay previously validated for quantifying the intrinsic efficacy of our compounds [24]. This functional binding assay is based on the extended ternary complex model of G-protein coupled receptor (GPCR). According to this model, the activated conformation of GPCR is the one coupled to a G-protein. Moreover, the affinity of an agonist is much higher for the receptor in this conformation than for the uncoupled receptor [34]. In our assay, the substitution of Mg2+ (a divalent cation that favors the formation of the ternary complex) by Na+ (a negative allosteric modulator of agonist binding and activation in many GPCRs) shifts the competition curve of agonists, but not antagonists, to the right [24,35]. On the other hand, a shift of the curve to the left indicates that the compound is an inverse agonist. This profile was observed with all four non-methylated N-acylhydrazone derivatives (compounds 1 and 3-5), as exemplified in Figure 4A for LASSBio-2093 (5), and shown in Figure S1 (A, C, and E) for the other derivatives. Note that Na+-shifts to the left observed with these four compounds were small, but statistically significant, as indicated by a ratio (II/I) lower than one (p < 0.05, Table 2). On the other hand, the competition curves with the N-methyl-N-acylhydrazone compounds (2, 6-8) shifted to the right in the presence of Na + , as shown in Figure 4B for LASSBio-2199 (8), and in Figure S1B,D,F for the other derivatives. The small but statistically significant (p < 0.05, Table 2) shift to the right (II/I ratio) indicates that these compounds act as weak partial agonists in our conditions. Note that this assay was validated previously [24] by successful discrimination between classical A2AR full agonists (CGS21680, NECA, and adenosine, with IC50 ratios = 13-14) and partial agonists (LUF5834 and regadenoson, with IC50 ratios equal to 3 and 10, respectively).

A B
Since the IC50 ratios (II/I) were similar for the isosteric pairs (1 and 3, 4 and 5, 2 and 6,  and 7 and 8), we can conclude that isosteric replacement of S by Se did not affect the in- On the other hand, the competition curves with the N-methyl-N-acylhydrazone compounds (2, 6-8) shifted to the right in the presence of Na + , as shown in Figure 4B for LASSBio-2199 (8), and in Figure S1B,D,F for the other derivatives. The small but statistically significant (p < 0.05, Table 2) shift to the right (II/I ratio) indicates that these compounds act as weak partial agonists in our conditions. Note that this assay was validated previously [24] by successful discrimination between classical A 2A R full agonists (CGS21680, NECA, and adenosine, with IC 50 ratios = [13][14] and partial agonists (LUF5834 and regadenoson, with IC 50 ratios equal to 3 and 10, respectively).
Since the IC 50 ratios (II/I) were similar for the isosteric pairs (1 and 3, 4 and 5, 2 and  6, and 7 and 8), we can conclude that isosteric replacement of S by Se did not affect the intrinsic efficacy of compounds, probably because the conformational behavior of both is similar. On the other hand, N-methylation of the N-acylhydrazone derivatives changed the intrinsic efficacy while simple meta-substitution in the phenyl ring did not. These results can be explained by the conformational change induced by NAH methylation as previously observed for LASSBio-785 (2) [15,17].

Retroisosteric Derivatives (13 and 14)
Previous results from our working group [36] on NAH derivatives with the thiophene ring on the acyl-terminus have demonstrated that the molecule displays an important intramolecular 1-5 σ-hole-S-Nsp 2 interaction. The geometry of such interaction provides a better orientation for the overlap between the non-bonded electronic pair of the nitrogen and the σ-hole in the heterocycle. This kind of interaction in NAH systems leads to a conformational modification in the amide bond to syn-periplanar, as observed for N-CH 3 -NAH derivatives. The X-ray diffraction analysis of compound LASSBio-1834 (12, Figure 5) confirms this assumption [36].

Retroisosteric Derivatives (13 and 14)
Previous results from our working group [36] on NAH deriva phene ring on the acyl-terminus have demonstrated that the molec portant intramolecular 1-5 σ-hole-S-Nsp 2 interaction. The geometry provides a better orientation for the overlap between the non-bonde the nitrogen and the σ-hole in the heterocycle. This kind of interacti leads to a conformational modification in the amide bond to syn-per for N-CH3-NAH derivatives. The X-ray diffraction analysis of comp (12, Figure 5) confirms this assumption [36].
In this context, we decided to investigate how the herein p structural and electronic effects promoted by the bioisosteric chang would apply to the retroisosteric analogs (13 and 14). These retroiso by the interchange between the five-ring heterocycle and the oxyg substituents of acyl and imine terminus, respectively. Due to the 3-methoxy substituted compounds 4 and 5 in the previous assays (  Table 2), this structural pattern of the phenyl ring in the new retroisos and 14) was preserved.  (12) and new isosteric analogu sentation of the syn-periplanar conformation of the amide bond in the NAH teraction, as observed through X-ray analysis.

Chemistry
For the preparation of the thiophenic hydrazide (18) [37] interm initially the thiophene-2-carboxylic acid (16) was converted to methy description in Supplementary Materials) by standard acid-catalyzed F conditions [38]. The hydrazinolisis reaction leads to intermediate id-catalyzed condensation with the 3-methoxybenzaldehyde [4] to a (Scheme 3). The selenophenic hydrazide intermediate 19 [39] (Sche first by Jones oxidation [40] of 2-formylselenophene (15) to afford the In this context, we decided to investigate how the herein previously described structural and electronic effects promoted by the bioisosteric change between S and Se would apply to the retroisosteric analogs (13 and 14). These retroisosters were designed by the interchange between the five-ring heterocycle and the oxygenated phenyl ring, substituents of acyl and imine terminus, respectively. Due to the higher activity of 3-methoxy substituted compounds 4 and 5 in the previous assays (as demonstrated in Table 2), this structural pattern of the phenyl ring in the new retroisosteric derivatives (13 and 14) was preserved.

Chemistry
For the preparation of the thiophenic hydrazide (18) [37] intermediate (Scheme 3), initially the thiophene-2-carboxylic acid (16) was converted to methyl ester (20, synthetic description in Supplementary Materials) by standard acid-catalyzed Fischer esterification conditions [38]. The hydrazinolisis reaction leads to intermediate 18, followed by acid-catalyzed condensation with the 3-methoxybenzaldehyde [4] to afford compound 13 (Scheme 3). The selenophenic hydrazide intermediate 19 [39] (Scheme 3) was obtained first by Jones oxidation [40]   The retroisosterism approach [23] used in 13 and 14 tends to favor a distinct chal cogen intramolecular interaction, which changes the conformational behavior of the amide bond. Since the syn-periplanar conformation is favored by the σ-hole 1,5-Nsp 2… S and 1,5-Nsp 2… Se interaction, these new retrosisosteric NAH derivatives 13 and 14 could possibly adopt a similar conformational behavior to N-methylated analogs 7 and 8. In fact, it was observed in the 1 H NMR that there is a duplicity of signals related to their conformation ( Figure 6A). This was confirmed by coalescence of the related signals with the higher temperature experiment ( Figure 6B). This indicates the possibility of the chalcogen interaction like LASSBio-1834 (10) in the new NAH compounds 13 and 14.
A Scheme 3. Synthetic scheme for the preparation of retroisosteric derivatives 13 and 14. Reaction conditions and yields for the synthesis of thiophenic and seleonphenic hydrazides and subsequent condensation to obtain retroisosteric NAH 13 and 14.
The retroisosterism approach [23] used in 13 and 14 tends to favor a distinct chalcogen intramolecular interaction, which changes the conformational behavior of the amide bond. Since the syn-periplanar conformation is favored by the σ-hole 1,5-Nsp 2 . . . S, and 1,5-Nsp 2 . . . Se interaction, these new retrosisosteric NAH derivatives 13 and 14 could possibly adopt a similar conformational behavior to N-methylated analogs 7 and 8. In fact, it was observed in the 1 H NMR that there is a duplicity of signals related to their conformation ( Figure 6A). This was confirmed by coalescence of the related signals with the higher temperature experiment ( Figure 6B). This indicates the possibility of the chalcogen interaction like LASSBio-1834 (10) in the new NAH compounds 13 and 14.
and 1,5-Nsp 2… Se interaction, these new retrosisosteric NAH derivatives 13 and 14 could possibly adopt a similar conformational behavior to N-methylated analogs 7 and 8. In fact, it was observed in the 1 H NMR that there is a duplicity of signals related to their conformation ( Figure 6A). This was confirmed by coalescence of the related signals with the higher temperature experiment ( Figure 6B). This indicates the possibility of the chalcogen interaction like LASSBio-1834 (10) in the new NAH compounds 13 and 14.

Pharmacological Evaluation
As for the first set of compounds, we performed the Na + -shift assay with the retroisosteric analogues LASSBio-2278 (13) and LASSBio-2279 (14) to evaluate their pharmacological profile ( Figure S2). Comparing the retroisosteres 13 and 14 with their counterparts (4 and 5), the main difference relates to their intrinsic efficacy. Indeed, the IC 50 ratio (II/I) was higher than one ( Table 3), indicating that compounds 13 and 14 are weak partial agonists. Similarly, to NAH N-methylation, simple retroisosterism changed the intrinsic efficacy of our compounds on A 2A R (Table 3). Table 3. Affinity and intrinsic efficacy of retroisosteric compounds 13 and 14 at the A 2A R present in membrane preparations from rat striatum. Average curves of three independent experiments (in triplicate) were analyzed by non-linear regression to estimate the IC 50 values with their 95% confidence interval (95% CI). For each compound, the Na + -shifts were calculated by dividing the IC 50 value obtained in the medium containing Na + by the IC 50 value obtained in the medium containing Mg 2+ and statistically analyzed by paired t test on log IC 50 values (P).

Materials and Methods
All reagents and (anhydrous) solvents are commercially available and were used without further purification. NMR spectra were obtained at UFRJ with a VARIAN 400-MR American Radiolabeled Chemicals, Inc., Maryland Heights, MI, USA. Stock solutions of our compounds were prepared in DMSO and at the final concentration used (less than 0.5%), DMSO had no effect on our biological assays.
3.1. Synthesis of N-Acylhydrazones (3-8, 13, and 14) The corresponding hydrazide (11a, 11b, 18 or 19) [29,30,37,39] was solubilized in EtOH solution with 1% v/v HCl (0.2-0.3 M). The aldehsyde was added, and the reaction stirred at room temperature. After ca. 30 min to 1 h, a precipitate was formed. The reaction was monitoressd via TLC (20-70% EtOAc/n-hex) until completion. The work up was made by evaporating the solvent on a rotary evaporator, the addition of water, and filtration under vacuum on a Büchner funnel.   N-Acylhydrazones (6, 7 and 8) The corresponding N-acylhydrazone (3, 4, or 5) was dissolved in acetone (0.1-0.2 M) in a round-bottom flask, and K 2 CO 3 (3 eq.) and CH 3 I (4 eq.) were added to the reaction mixture. The flask was attached to a reflux condenser and the reaction stirred and heated to 40 • C. TLC control (20 or 50% of EtOAC/n-hex) was made at the completion of the reaction. Quenching was conducted by evaporating the solvent on a rotary evaporator and the addition of water. The product was either filtrated under vacuum or extracted with EtOAc (3 × 15 mL), dried over Na 2 SO 4 , and the solvent evaporated. Further purification was carried out by column chromatography (5-50% EtOAc/n-hex) and recrystallization in EtOH followed by filtration under vacuum, if possible. tissues were homogenized in a Potter apparatus with a motor-driven Teflon pestle at 4 • C in 20 volumes per gram of tissue of ice-cold Tris-HCl 50 mM buffer (pH 7.4) containing MgCl 2 8 mM and EDTA 5 mM. The resulting suspension was ultracentrifuged at 48.000 g max at 4 • C for 20 min. The pellet was resuspended in the same buffer and incubated at 37 • C for 10 min for endogenous neurotransmitter removal. This suspension was cooled on ice and ultracentrifuged twice at 48.000 g av for 10 min at 4 • C. The final pellet was resuspended in buffer yielding 1.5 mL/g tissue and stored in liquid nitrogen until use. The protein concentration was determined by the method of Lowry et al. [43] using bovine serum albumin as the standard.

N-Methylation of
Rat striatum membrane preparation (150 µg protein) and 0.5 nM [ 3 H]-ZM241385 were incubated at 25 • C for 60 min under yellow light in a buffer solution of Tris-HCl 50 mM (pH 7.4) containing EDTA 1 mM and MgCl 2 50 mM or 100 mM NaCl, in the absence or presence of the compounds. The non-specific binding was estimated using 30 µM NECA.
After incubation, samples were rapidly diluted with 3 × 4 mL Tris-HCl 5 mM (pH 7.4) and immediately filtered under vacuum on glass fiber filters (GMF 3; Filtrak, Germany) previously soaked in 0.5% polyethyleneimine. Filters were then dried, immersed in a scintillation mixture (POPOP 0.1 g/L and POP 4.0 g/L in toluene), and the radioactivity retained in the filters was counted using a Packard Tri-Carb 1600 TR liquid scintillation analyzer (Perkin Elmer Waltham, Massachusetts, USA.) In order to determine the profile of intrinsic efficacy of our compounds at the A 2A R, we used the Na + -shift assay that we previously validated [24]. Paired competition curves for the binding of [ 3 H]-ZM241385 were performed either in the presence of 50 mM MgCl 2 or 100 mM NaCl. The concentration-response curves were analyzed by non-linear regression using the model of "one site competition" for determination of the median inhibitory concentration (IC 50 ). The Na + -shift was calculated by dividing the IC 50 value obtained in the medium containing NaCl by the IC 50 value obtained in the medium containing MgCl 2 and statistically analyzed by a paired t test performed on the log IC 50 values (GraphPad Prism 6.0, GraphPad Software, Inc., San Diego, CA, USA), as previously described [24].

Conclusions
In this report, we explored the isosteric relationship between sulfur and selenium and its implication on the intramolecular interaction between the iminic nitrogen and the antibonding orbital of the C-X bond on heterocyclic residues of N-acylhydrazone (NAH) scaffolds. We showed that the higher polarizability of selenium has a deshielding effect on the neighboring hydrogens and seems to influence their electronical profile. However, this isosteric exchange did not modify the affinity or intrinsic efficacy of the compounds at the A 2A receptor (A 2A R). On the other hand, NAH N-methylation significantly inverted the intrinsic efficacy of these compounds (from weak inverse agonists to partial agonists).
Present results obtained from rational simple structural modification in these series of N-acylhydrazone derivatives illustrated how the ligand conformation has an influence on molecular recognition by the bioreceptor.  10a-b, 11a-b, 18, and 19; Figure S1: Na + -shift assay for the estimation of the intrinsic efficacy of the NAH derivatives 1-4, 6, and 7 at the A 2A receptor (A 2A R) in rat striatum membrane preparation. Figure S2: Na + -shift assay for the estimation of the intrinsic efficacy of the retroisosteric analogues at the A 2A receptor (A 2A R) in rat striatum membrane preparation. Figures S1-S18: NMR and HRMS spectra of compounds 3-8, 11, and 12; Table S1: 13 C NMR shifts of the iminic hydrogen of the NAH in isosteric compounds 1-8.