Novel N-Arylsulfonylindoles Targeted as Ligands of the 5-HT6 Receptor. Insights on the Influence of C-5 Substitution on Ligand Affinity

A new series of twenty-two C-5 substituted N-arylsulfonylindoles was prepared with the aim of exploring the influence of C-5 substitution on 5-HT6 receptor affinity. Eleven compounds showed moderate to high affinity at the receptor (Ki = 58–403 nM), with compound 4d being identified as the most potent ligand. However, regarding C-5 substitution, both methoxy and fluorine were detrimental for receptor affinity compared to our previously published unsubstituted compounds. In order to shed light on these observations, we performed docking and molecular dynamics simulations with the most potent compounds of each series (4d and 4l) and PUC-10, a highly active ligand previously reported by our group. The comparison brings about deeper insight about the influence of the C-5 substitution on the binding mode of the ligands, suggesting that these replacements are detrimental to the affinity due to precluding a ligand from reaching deeper inside the binding site. Additionally, CoMFA/CoMSIA studies were performed to systematize the information of the main structural and physicochemical characteristics of the ligands, which are responsible for their biological activity. The CoMFA and CoMSIA models presented high values of q2 (0.653; 0.692) and r2 (0.879; 0.970), respectively. Although the biological activity of the ligands can be explained in terms of the steric and electronic properties, it depends mainly on the electronic nature.


Introduction
The serotonergic system modulates a diverse variety of physiological functions, such as thermoregulation, sexual and aggressive behavior, learning and memory, endocrine and gastrointestinal functions, and food intake, among others [1]. In order to perform these different tasks, the neurotransmitter 5-hydroxytryptamine (5-HT) interacts with various receptor subtypes among which the 5-HT 6 receptor (5-HT 6 R) is one of the most recently discovered, about a quarter of a century ago [2]. The 5-HT 6 R is involved in several physiological functions, such as learning, memory, and neurodevelopment, and diverse pathological states, such as eating disorders, anxiety, depression, addictive behavior, schizophrenia, epilepsy and Alzheimer's disease [3][4][5][6][7][8]. 5-HT 6 receptor ligands (specially antagonists) have been shown to have procognitive behavioral effects, making them potential cognitive enhancers in conditions associated with cognitive impairment, such as Alzheimer's disease, schizophrenia, and autism spectrum disorder, among others [9][10][11][12][13].
To date, an overwhelming majority of 5-HT 6 ligands, both agonists and antagonists, share the following structural features: a basic ionizable amine group (PI), a sulfonamide moiety as a hydrogen bond acceptor group (HBA) connected to a hydrophobic site (HYD) and a π-electron donor aromatic or heterocyclic ring (AR) [9,14]. However, a few reports claim that the PI pharmacophoric feature is not essential for receptor affinity [15,16]. In this context, and given our interest in exploring the limits of the -HT 6 ligand pharmacophore, we have previously reported the design, synthesis and pharmacological evaluation of a number of extended and weakly basic N-arylsulfonylindole derivatives with moderate to high affinity and antagonistic profile toward the 5-HT 6 receptor [17], where our structural scaffold could be regarded as a masked and extended N-arylsulfonyltryptamine, similar to MS-245 ( Figure 1). In this work, the activity of the ligands was attributed to an additional hydrogen bond between the alcohol group and Asp106 (3.32) and to the interaction of the aromatic ring linked to piperazine with a second hydrophobic pocket, proposed from molecular docking studies.
recently discovered, about a quarter of a century ago [2]. The 5-HT6R is involved in seve physiological functions, such as learning, memory, and neurodevelopment, and diver pathological states, such as eating disorders, anxiety, depression, addictive behavi schizophrenia, epilepsy and Alzheimer's disease [3][4][5][6][7][8]. 5-HT6 receptor ligands (specia antagonists) have been shown to have procognitive behavioral effects, making the potential cognitive enhancers in conditions associated with cognitive impairment, such Alzheimer's disease, schizophrenia, and autism spectrum disorder, among others  To date, an overwhelming majority of 5-HT6 ligands, both agonists and antagonis share the following structural features: a basic ionizable amine group (PI), a sulfonami moiety as a hydrogen bond acceptor group (HBA) connected to a hydrophobic site (HY and a π-electron donor aromatic or heterocyclic ring (AR) [9,14]. However, a few repo claim that the PI pharmacophoric feature is not essential for receptor affinity [15,16]. this context, and given our interest in exploring the limits of the -HT6 liga pharmacophore, we have previously reported the design, synthesis and pharmacologi evaluation of a number of extended and weakly basic N-arylsulfonylindole derivativ with moderate to high affinity and antagonistic profile toward the 5-HT6 receptor [1 where our structural scaffold could be regarded as a masked and extended arylsulfonyltryptamine, similar to MS-245 ( Figure 1). In this work, the activity of t ligands was attributed to an additional hydrogen bond between the alcohol group a Asp106 (3.32) and to the interaction of the aromatic ring linked to piperazine with a seco hydrophobic pocket, proposed from molecular docking studies.  18], MS-245 [19,20], and examples of our previously reported N-arylsulfonylindole antagonists over the 5-HT6 receptor [17].
Considering the excellent affinity presented by several 5-metho arylsulfonyltryptamines derivatives, similar to MS-245 [19,20], we decided to explore t influence of this substitution pattern on our structural scaffold. To have a comparati insight regarding the substitution on C-5 of the indole ring and ligand affinity, we a proposed the synthesis of analogs bearing fluorine as an electron-withdrawing ato Therefore, in this work, we report the synthesis, affinity evaluation over the 5-H receptor, molecular modeling of receptor-ligand interactions through molecular dynam simulations, and 3DQSAR studies of twenty-two new N-arylsulfonylindole compoun with C-5 substitution.

Synthesis
The synthesis of the new N-arylsulfonylindoles targeting the 5-HT6 receptor (4awas performed as is briefly described in Scheme 1 (for details, see the experimen section). First, the synthesis of C-5 substituted 3-bromoacetylindoles 1a-b was perform following a previously reported method for benzenesulfonyltriptamines, which h  18], MS-245 [19,20], and examples of our previously reported N-arylsulfonylindole antagonists over the 5-HT 6 receptor [17].
Considering the excellent affinity presented by several 5-methoxy arylsulfonyltryptamines derivatives, similar to MS-245 [19,20], we decided to explore the influence of this substitution pattern on our structural scaffold. To have a comparative insight regarding the substitution on C-5 of the indole ring and ligand affinity, we also proposed the synthesis of analogs bearing fluorine as an electron-withdrawing atom. Therefore, in this work, we report the synthesis, affinity evaluation over the 5-HT 6 receptor, molecular modeling of receptor-ligand interactions through molecular dynamic simulations, and 3DQSAR studies of twenty-two new N-arylsulfonylindole compounds with C-5 substitution.

Synthesis
The synthesis of the new N-arylsulfonylindoles targeting the 5-HT 6 receptor (4a-v) was performed as is briefly described in Scheme 1 (for details, see the experimental section). First, the synthesis of C-5 substituted 3-bromoacetylindoles 1a-b was performed following a previously reported method for benzenesulfonyltriptamines, which has already been used by our group with slight modifications [17,21,22]. Acylation was performed in a The previously obtained 5-fluoro and 5-methoxy bromoacetylindoles 1a-b were further protected with the appropriate aromatic sulfonyl chlorides under basic conditions to afford the corresponding C-5 substituted N-arylsulfonyl-3-bromoacetylindoles 2a-i with moderate to good yields [17,23]. In the next step, the prepared haloketones were subjected to bromide displacement in a basic medium at room temperature with various arylpiperazines or morpholines to obtain the respective functionalized ethanones 3a-v in moderate to excellent yields [17,24]. Ketones obtained after the N-alkylation reaction described above were subsequently reduced with sodium borohydride in methanol to obtain the corresponding alcohols 4a-v as racemic mixtures. The structures and purity of the new compounds were confirmed through spectral ( 1 H-NMR, 13 C-NMR, IR, HRMS) and chromatographic (TLC) methods. The assignment of the signals from the NMR spectra of the synthesized compounds was carried out by analyzing the multiplicity of the signals and their corresponding coupling constants, and by comparison with the NMR spectra of the unsubstituted C-5 compounds previously reported by us.

Radioligand Binding Assays
Synthesized C-5 substituted N-arylsulfonylindole derivatives were tested in standard radioligand competition binding assays, using HEK-293 cell membranes expressing a recombinant human 5-HT6 receptor. The compounds were assayed as free bases at eight different concentrations in triplicate in order to obtain the dose response curves, determine the half maximal inhibitory concentration (IC50) values, and calculate the inhibitory constant (Ki) values for each one of them (Table 1) The previously obtained 5-fluoro and 5-methoxy bromoacetylindoles 1a-b were further protected with the appropriate aromatic sulfonyl chlorides under basic conditions to afford the corresponding C-5 substituted N-arylsulfonyl-3-bromoacetylindoles 2a-i with moderate to good yields [17,23]. In the next step, the prepared haloketones were subjected to bromide displacement in a basic medium at room temperature with various arylpiperazines or morpholines to obtain the respective functionalized ethanones 3a-v in moderate to excellent yields [17,24]. Ketones obtained after the N-alkylation reaction described above were subsequently reduced with sodium borohydride in methanol to obtain the corresponding alcohols 4a-v as racemic mixtures. The structures and purity of the new compounds were confirmed through spectral ( 1 H-NMR, 13 C-NMR, IR, HRMS) and chromatographic (TLC) methods. The assignment of the signals from the NMR spectra of the synthesized compounds was carried out by analyzing the multiplicity of the signals and their corresponding coupling constants, and by comparison with the NMR spectra of the unsubstituted C-5 compounds previously reported by us.

Radioligand Binding Assays
Synthesized C-5 substituted N-arylsulfonylindole derivatives were tested in standard radioligand competition binding assays, using HEK-293 cell membranes expressing a recombinant human 5-HT 6 receptor. The compounds were assayed as free bases at eight different concentrations in triplicate in order to obtain the dose response curves, determine the half maximal inhibitory concentration (IC 50 ) values, and calculate the inhibitory constant (K i ) values for each one of them (Table 1). Additionally, the structures and K i values of some unsubstituted derivatives previously reported by our research group are shown.   [25]. In this assay, the IC50 value of Clozapine (Clz) was 12.4 nM; Ki = 11.9 nM. b Ki values for unsubstituted derivatives previously reported [17].
All tested compounds exhibited inhibition of [ 125 I]-SB-258585 [26] binding to 5-HT6 receptors (radioligand displacement). Eleven of the twenty-two compounds showed an important binding affinity for the 5-HT6 receptor (equal or better than 400 nM of Ki). However, C-5 substitution, regardless of the substituent, was detrimental for affinity compared to our previous set of ligands within this structural framework [17].
In general, the series with a 5-methoxy substituent exhibited better affinity than the series with 5-fluorine (eight compounds with Ki values below 400 nM including four below 200 nM), even though the best compound of all series was the 5-fluoro derivative 4d with a Ki value of 58 nM. This ligand was the only compound that was better than the endogenous ligand 5-HT and it showed a Ki in the same order of magnitude as the most potent unsubstituted compounds PUC-7, PUC-10, and PUC-32 (See Figure 1 and Table 1). However, this result was not completely surprising given the substitution pattern of 4d, which is like PUC-10. The analogous 5-methoxy derivative 4l also exhibited a good affinity (Ki 160 nM), but lower than that of 4d and PUC-10. Both 4d and 4l ligands have a 2-methoxyphenyl group on Ar2, as do five other compounds (4a, 4i, 4j, 4r and 4u). All these seven ligands are included within the eleven most potent compounds. Thus, the 2methoxyphenyl substitution in the piperazine ring was revealed as the best-performing group. These results are in agreement with our previous report of unsubstituted were determined in triplicate and K i were calculated using Cheng-Prusoff equation [25]. In this assay, the IC 50 value of Clozapine (Clz) was 12.4 nM; K i = 11.9 nM. b K i values for unsubstituted derivatives previously reported [17].
All tested compounds exhibited inhibition of [ 125 I]-SB-258585 [26] binding to 5-HT 6 receptors (radioligand displacement). Eleven of the twenty-two compounds showed an important binding affinity for the 5-HT 6 receptor (equal or better than 400 nM of K i ). However, C-5 substitution, regardless of the substituent, was detrimental for affinity compared to our previous set of ligands within this structural framework [17].
In general, the series with a 5-methoxy substituent exhibited better affinity than the series with 5-fluorine (eight compounds with K i values below 400 nM including four below 200 nM), even though the best compound of all series was the 5-fluoro derivative 4d with a K i value of 58 nM. This ligand was the only compound that was better than the endogenous ligand 5-HT and it showed a K i in the same order of magnitude as the most potent unsubstituted compounds PUC-7, PUC-10, and PUC-32 (See Figure 1 and Table 1). However, this result was not completely surprising given the substitution pattern of 4d, which is like PUC-10. The analogous 5-methoxy derivative 4l also exhibited a good affinity (K i 160 nM), but lower than that of 4d and PUC-10. Both 4d and 4l ligands have a 2-methoxyphenyl group on Ar 2 , as do five other compounds (4a, 4i, 4j, 4r and 4u). All these seven ligands are included within the eleven most potent compounds. Thus, the 2-methoxyphenyl substitution in the piperazine ring was revealed as the bestperforming group. These results are in agreement with our previous report of unsubstituted compounds [17]. A more specific comparison between both series of C-5 substituted ligands indicated that 5-methoxy derivatives were superior to 5-fluoro derivatives as is shown for the following pairs of ligands (except for the mentioned pair of molecules 4d v/s 4l): 4j v/s 4a (K i : 289 nM v/s 390 nM); 4k v/s 4c (K i : 403 nM v/s 12,515 nM); 4n v/s 4e (K i : 153 nM Regarding the Ar 1 substitution pattern, in this work we only employed electron-rich rings, given that our previous results indicated that poor electron density rings were detrimental for affinity [17]. In this way, we confirmed our previous observations that the 1-naphthyl group is the best ring to be used as a hydrophobic region, given that three of the five best compounds bear this substitution in Ar 1 (4d, 4l and 4n). Regarding the substitution in Ar 1 , we also observed that two of the five best compounds bear a 5-bromo-2-thiophenyl group (4r and 4s with K i 167 nM and 130 nM respectively). This group follows the structural guidelines formulated in our previous work on the requirement of electron-rich aromatic or heteroaromatic groups in that position [27]. The affinities showed by derivatives bearing a 5-bromo-2-thiophenyl substitution at Ar 1 are similar to those which possess the p-iodophenyl groups, although the first ones proved to be slightly potent compared to p-iodophenyl derivatives (4i v/s 4a; 4r v/s 4j and 4s v/s 4k). This result could suggest the existence of a halogen bond between these ligands and the active site of the receptor. It is noteworthy that a similar halogen bond interaction was previously described by López-Rodríguez in a series of dimethylaminoethyl-1H-benzimidazol-ylaryl-sulfonamides in an analogous position (i.e., the halogen group is pending from an aryl group of the sulfonamide) [9]. It is worth mentioning that in our previous work on the matter, a derivative containing a p-iodophenyl group presented high affinity (PUC-7), equipotent with PUC-10, the most active compound of the series (see Figure 1) [17].
An interesting observation could be made by comparing these C-5 substituted derivatives with the corresponding unsubstituted derivatives that were previously reported (see Table 1). In addition to the lower activity found for compounds 4d and 4l when compared with PUC-10, both 4a and 4j exhibited decreased affinity at the 5-HT 6 receptor of more than one order of magnitude compared to PUC-7. In addition, the fluorine derivative 4b exhibited ten times less affinity than its unsubstituted derivative PUC-9. Similarly, compound 4g showed half the affinity of its unsubstituted derivative PUC-12. On the other hand, the methoxylated derivative 4t had approximately eight times less affinity than PUC-22, the unsubstituted derivative. The ligand 4p (K i = 2195 nM) was approximately four times less active than the unsubstituted PUC-24 (K i = 479 nM). When comparing derivatives 4o (K i = 972 nM) and PUC-32 (K i = 13.6 nM), the same trend could be observed. Perhaps the only anomalous result in this trend was presented by the trio 4e, 4n and PUC-11. In this case, the fluorine-containing derivative was less potent than the unsubstituted derivative PUC-11 (K i = 801 nM vs. 148 nM), but the methoxylated ligand 4n (K i = 153 nM) was equipotent with PUC-11. Overall, the results seemed to indicate that whatever C-5 substitution is used, this is detrimental for affinity, and in the best case, it is not better than its absence. With the aim of attaining better knowledge about this behavior, we performed molecular docking studies and molecular dynamic simulations over the most potent ligand of this new series 4d, its methoxylated analog 4l and the best overall compound PUC-10.

Docking Studies
Previously, we reported PUC-10 as a high affinity antagonist of the 5-HT 6 receptor (K i = 14.6 nM) [17]. Compounds 4l and 4d correspond to the C-5 substituted methoxylated and fluorinated analogues of PUC-10. Although the affinity of both compounds was not completely abolished, 4d diminished its affinity by 4-fold (K i = 58 nM) compared to PUC-10, whereas 4l diminished its affinity by a factor of 11 (K i = 160 nM).
To gain further molecular insights, these compounds were docked into the 5-HT 6 receptor model, and the R/S stereoisomers of each compound were analyzed separately. The results presented in Figure 2 show that molecules with the hydroxy group in the S configuration allowed the placement of the naphthalene ring into a cavity created between TMH-3 and 5, and the phenylpiperazinyl moiety oriented toward TMH-2 and 7 ( Figure 2A). On the other hand, molecules in R conformation flipped their binding mode by 180 degrees, with Pharmaceuticals 2021, 14, 528 6 of 32 the naphthalene ring extending toward TMH-2 and 7 regions and the phenylpiperazinyl moiety fitting within the cavity between TMH-3 and 5 ( Figure 2B).
The results presented in Figure 2 show that molecules with the hydroxy group in the S configuration allowed the placement of the naphthalene ring into a cavity created between TMH-3 and 5, and the phenylpiperazinyl moiety oriented toward TMH-2 and 7 ( Figure  2A). On the other hand, molecules in R conformation flipped their binding mode by 180 degrees, with the naphthalene ring extending toward TMH-2 and 7 regions and the phenylpiperazinyl moiety fitting within the cavity between TMH-3 and 5 ( Figure 2B). Although the crystal structure of the 5-HT6 receptor was still not available, efforts to describe the orthosteric binding pocket indicated that the key residues included: Asp106 (3.32), which assists in ligand recognition by forming a salt-bridge with charged ligands; Asn288 (6.55), which forms hydrogen bond interactions with HBAs in the ligand; Val107 (3.33), Cys110 (3.36), Ala192 (5.42), Ser193 (5.43) that surround the orthosteric binding pocket; and the hydrophobic cluster on TMH-6 comprised of Trp281 (6.48), Phe284 (6.51) and Phe285 (6.52) [29]. Arylindoles are reported to bind into the described pocket by inserting the indole nucleus into the hydrophobic cluster on TMH-6 with the aryl group extending between TMH-3 and TMH-5 and interacting with the surrounding residues Val107 (3.33) and Ala192 (5.42) [14]. Based on this literature evidence, we reasoned that S stereoisomers better fit the reported binding model of arylindoles in the 5-HT6 receptor. Moreover, considering the hydrophobic nature of the naphthalene ring, it is more reasonable for this moiety to plunge deeper inside the receptor rather than face toward the extracellular region (as in the case of R stereoisomers), where the binding pocket becomes more hydrophilic.
Molecules in the S configuration were further accommodated within the binding pocket by running short MD simulations. The trajectory was then clustered using cpptraj [30], and the final interactions of the ligand-receptor complex were analyzed through the PLIP web server ( Figure 3) [31]. The key salt bridge between Asp106 (3.32) and the piperazinium cation was maintained for all three compounds. In addition, a hydrogen bond was established between Asp106 (3.32) and Tyr310 (7.42), which aided in the stabilization of the ligand. This interaction network between Asp106 (3.32), Tyr310 (7.42) and the ligand is present in many crystal structures of aminergic receptors and is thought to facilitate the interaction of the amine group of the ligand with the receptor [32]. Another common binding feature for all three compounds was the insertion of the naphthalene ring into the hydrophobic cavity between TMH-3 and TMH-5, thereby interacting with Although the crystal structure of the 5-HT 6 receptor was still not available, efforts to describe the orthosteric binding pocket indicated that the key residues included: Asp106 (3.32), which assists in ligand recognition by forming a salt-bridge with charged ligands; Asn288 (6.55), which forms hydrogen bond interactions with HBAs in the ligand; Val107 (3.33), Cys110 (3.36), Ala192 (5.42), Ser193 (5.43) that surround the orthosteric binding pocket; and the hydrophobic cluster on TMH-6 comprised of Trp281 (6.48), Phe284 (6.51) and Phe285 (6.52) [29]. Arylindoles are reported to bind into the described pocket by inserting the indole nucleus into the hydrophobic cluster on TMH-6 with the aryl group extending between TMH-3 and TMH-5 and interacting with the surrounding residues Val107 (3.33) and Ala192 (5.42) [14]. Based on this literature evidence, we reasoned that S stereoisomers better fit the reported binding model of arylindoles in the 5-HT 6 receptor. Moreover, considering the hydrophobic nature of the naphthalene ring, it is more reasonable for this moiety to plunge deeper inside the receptor rather than face toward the extracellular region (as in the case of R stereoisomers), where the binding pocket becomes more hydrophilic.
Molecules in the S configuration were further accommodated within the binding pocket by running short MD simulations. The trajectory was then clustered using cpptraj [30], and the final interactions of the ligand-receptor complex were analyzed through the PLIP web server ( Figure 3) [31]. The key salt bridge between Asp106 (3.32) and the piperazinium cation was maintained for all three compounds. In addition, a hydrogen bond was established between Asp106 (3.32) and Tyr310 (7.42), which aided in the stabilization of the ligand. This interaction network between Asp106 (3.32), Tyr310 (7.42) and the ligand is present in many crystal structures of aminergic receptors and is thought to facilitate the interaction of the amine group of the ligand with the receptor [32]. Another common binding feature for all three compounds was the insertion of the naphthalene ring into the hydrophobic cavity between TMH-3 and TMH-5, thereby interacting with hydrophobic and aromatic side chains that included P161 (4.60), F188 (5.39), V107 (3.33) and A192 (5.42).
hydrogen bond with Asn288 (6.55). However, in 4l ( Figure 3C), this -OH group maintained an interaction with Asp106 (3.32) and did not interact directly with Asn288 (6.55). Regarding the indole ring, it is worth noting the depth that it reaches in each compound within the hydrophobic cluster of TMH-6. PUC-10 plunged the deepest into the pocket, followed by 4d and then by 4l ( Figure 3); the magnitude was in accordance with Ki values. The GPCRs activation mechanism involves the upward movement of TMH-6 toward the extracellular compartment; therefore, the interaction of antagonists with residues in this helix would be pivotal to stabilize the receptor in its inactive form [34]. That said, PUC-10 and 4d lowered the indole ring sufficiently to establish an edge-to-face π interaction with Phe285 (6.52) from TMH-6. This residue has been identified as important for antagonist binding through site-directed mutagenesis [33]. However, in the case of 4l, the methoxy substituent imposed steric hindrance, preventing sufficient lowering of the indole ring to establish a π-stacking with Phe285 (6.52). Instead, 4l interacted with the aromatic residue through the energetically weaker hydrophobic interaction. As for the A difference in the established interactions involved the hydrogen bonding with residue Asn288 (6.55), which has been implicated as an important residue for 5-HT 6 antagonist recognition [33]. In the case of PUC-10 and 4d ( Figure 3A,B), the -OH group, which is in an alpha position to the indole ring, flipped its orientation to establish a hydrogen bond with Asn288 (6.55). However, in 4l ( Figure 3C), this -OH group maintained an interaction with Asp106 (3.32) and did not interact directly with Asn288 (6.55). Regarding the indole ring, it is worth noting the depth that it reaches in each compound within the hydrophobic cluster of TMH-6. PUC-10 plunged the deepest into the pocket, followed by 4d and then by 4l ( Figure 3); the magnitude was in accordance with Ki values.
The GPCRs activation mechanism involves the upward movement of TMH-6 toward the extracellular compartment; therefore, the interaction of antagonists with residues in this helix would be pivotal to stabilize the receptor in its inactive form [34]. That said, PUC-10 and 4d lowered the indole ring sufficiently to establish an edge-to-face π interaction with Phe285 (6.52) from TMH-6. This residue has been identified as important for antagonist binding through site-directed mutagenesis [33]. However, in the case of 4l, the methoxy substituent imposed steric hindrance, preventing sufficient lowering of the indole ring to establish a π-stacking with Phe285 (6.52). Instead, 4l interacted with the aromatic residue through the energetically weaker hydrophobic interaction. As for the case of 4d, electron withdrawing groups, such as -F, decreased the binding energy of the T-shape aromatic interactions when the group was substituted on the facial aromatic system [35]. Therefore, is it suggested that the interaction of 4d with Phe285 (6.52) should be weaker than that established by PUC-10 (Tables 2-4).

CoMFA and CoMSIA Studies
Quantitative structure-activity relationship (QSAR) studies allow systematizing the information of the main structural and physicochemical characteristics for a series of compounds, which are responsible for their biological activity. In this way, these studies minimize the error of a qualitative SAR analysis and allow the design of new compounds based on the information obtained.
In order to find the best models, a sequential search for the best field combinations for comparative molecular field analysis (CoMFA) and comparative molecular similarity indices analysis (CoMSIA) was carried out (34 combinations of steric + electrostatic + hydrophobic + hydrogen-bond donor + hydrogen-bond acceptor; Table S1 in Supplementary Materials). The best models obtained are presented in Table 5. As is shown in Table 5, both final models have a value of q 2 > 0.5 and high r 2 test values (0.786 and 0.726 for CoMFA and CoMSIA respectively). An equilibrium can be seen in terms of the steric and electrostatic contributions to biological affinity. However, in both cases, the electrostatic contribution is higher. Thus, these properties can play a more important role in the affinity. To validate both models, a test set study was carried out for both models. In Table S2 of Supplementary Materials, we report the values of the external validation. In both cases, both models passed validation. Likewise, the experimental affinity values versus predicted values, in logarithmic scale (pK i = −logK i ) for each compound according to the CoMFA and CoMSIA models, are also shown in Table S3 (Supplementary Material).
The CoMFA model presented 24 molecules with a negative residual (pK iexp − pK ipred ) and 22 with a positive residual, while CoMSIA presented 28 predictions with negative residual and 18 with positive residual. Therefore, the CoMFA model presents the best equilibrium in terms of predictive capability, while CoMSIA tends to overestimate prediction values. Figure 4A,B shows the distribution of experimental versus predicted values for CoMFA and CoMSIA. As can be seen in both cases, one compound deviated in more than one logarithmic unit in the predicted value of K i (compound 4s, represented as a triangle). All other training and test set compounds had an adequate distribution along the y = x straight line. The best value fit was obtained for the CoMSIA model.  Unlike a 2D-QSAR, the results of a 3D-QSAR can be represented as contour m around the molecules of the study. The analysis of the different colored polyhedr around a molecule (e.g., the most active compound in the series) allows understand the main characteristics that are favorable or unfavorable for affinity or biological activ Figure 5 shows the results of the steric ( Figure 5A,B) and electrostatic ( Figure 5C,D) m for CoMFA and CoMSIA over the most active compound of our synthetic se (derivative 4d).  Unlike a 2D-QSAR, the results of a 3D-QSAR can be represented as contour maps around the molecules of the study. The analysis of the different colored polyhedrons around a molecule (e.g., the most active compound in the series) allows understanding the main characteristics that are favorable or unfavorable for affinity or biological activity. Figure 5 shows the results of the steric (Figure 5A,B) and electrostatic ( Figure 5C,D) maps for CoMFA and CoMSIA over the most active compound of our synthetic series (derivative 4d). a triangle). All other training and test set compounds had an adequate distribution along the y = x straight line. The best value fit was obtained for the CoMSIA model. Unlike a 2D-QSAR, the results of a 3D-QSAR can be represented as contour maps around the molecules of the study. The analysis of the different colored polyhedrons around a molecule (e.g., the most active compound in the series) allows understanding the main characteristics that are favorable or unfavorable for affinity or biological activity. Figure 5 shows the results of the steric (Figure 5A,B) and electrostatic ( Figure 5C,D) maps for CoMFA and CoMSIA over the most active compound of our synthetic series (derivative 4d).  The CoMFA steric contour map ( Figure 5A) shows a green polyhedron in the phenylpiperazine zone, that is, it is favorable for biological activity to use bulky substituents in that position. The less active compounds present a morpholino group instead of the aryl-piperazine fragment, which may partly explain the lower activity of these derivatives. CoMSIA steric contour map ( Figure 5B) is consistent with this information but shows an additional green polyhedron close to the sulfonyl group region; therefore, the direct connection of bulky rings or alkyl groups to the nitrogen atom of the indole ring would be an interesting option to explore. Close to the hydroxyl group, both the CoMFA and CoMSIA steric contour maps ( Figure 5A,B) show a yellow polyhedron, suggesting that the use of low-volume functional groups in that area is most appropriate. The elimination of the hydroxyl group could be evaluated, although there are additional electronic factors linked to this functional group. Finally, near the naphthyl ring, there is a set of yellow polyhedrons in CoMFA ( Figure 5A) and a large yellow polyhedron in the steric map of CoMSIA ( Figure 5B). Therefore, the insertion of bulky groups at positions 6 and 7 of the naphthyl group should be avoided. On the other hand, the electrostatic contour maps for CoMFA and CoMSIA ( Figure 5C,D) are concordant. In both cases, a red polyhedron can be seen through the piperazine and indole connecting chain. In the case of CoMSIA, the red polyhedron is closer to the piperazine nitrogen atom 1 (N-1) ( Figure 5D). Therefore, the insertion of groups favoring a higher charge density over this nitrogen atom would be favorable. This implies that increasing the basicity of such nitrogen atom would be favorable, which is consistent with our docking studies for the most active derivative 4d, where the nitrogen atom (N-1) of piperazine moiety mediates a key salt bridge with Asp106 (3.32). A red polyhedron over the hydroxyl group on CoMFA map ( Figure 5C) indicates that the presence of an electronegative atom in that position would be beneficial. Another red polyhedron is seen close to the halogen of the indole ring. Complementary to this information, the CoMSIA electrostatic contour map ( Figure 5D) shows a blue polyhedron in the indole ring. Therefore, the presence of electron-attracting groups on the indole scaffold would be beneficial for biological activity. Among them, it would be interesting to explore groups NO 2 , CF 3 , CN and COR among others. In Figure 6, we summarize the structure-activity relationships found in this study. With this information, new active molecules can be designed and synthesized in the future.
The CoMFA steric contour map ( Figure 5A) shows a green polyhedron in the phenylpiperazine zone, that is, it is favorable for biological activity to use bulky substituents in that position. The less active compounds present a morpholino group instead of the arylpiperazine fragment, which may partly explain the lower activity of these derivatives. CoMSIA steric contour map ( Figure 5B) is consistent with this information but shows an additional green polyhedron close to the sulfonyl group region; therefore, the direct connection of bulky rings or alkyl groups to the nitrogen atom of the indole ring would be an interesting option to explore. Close to the hydroxyl group, both the CoMFA and CoMSIA steric contour maps ( Figure 5A,B) show a yellow polyhedron, suggesting that the use of low-volume functional groups in that area is most appropriate. The elimination of the hydroxyl group could be evaluated, although there are additional electronic factors linked to this functional group. Finally, near the naphthyl ring, there is a set of yellow polyhedrons in CoMFA ( Figure 5A) and a large yellow polyhedron in the steric map of CoMSIA ( Figure 5B). Therefore, the insertion of bulky groups at positions 6 and 7 of the naphthyl group should be avoided. On the other hand, the electrostatic contour maps for CoMFA and CoMSIA ( Figure 5C,D) are concordant. In both cases, a red polyhedron can be seen through the piperazine and indole connecting chain. In the case of CoMSIA, the red polyhedron is closer to the piperazine nitrogen atom 1 (N-1) ( Figure 5D). Therefore, the insertion of groups favoring a higher charge density over this nitrogen atom would be favorable. This implies that increasing the basicity of such nitrogen atom would be favorable, which is consistent with our docking studies for the most active derivative 4d, where the nitrogen atom (N-1) of piperazine moiety mediates a key salt bridge with Asp106 (3.32). A red polyhedron over the hydroxyl group on CoMFA map ( Figure 5C) indicates that the presence of an electronegative atom in that position would be beneficial. Another red polyhedron is seen close to the halogen of the indole ring. Complementary to this information, the CoMSIA electrostatic contour map ( Figure 5D) shows a blue polyhedron in the indole ring. Therefore, the presence of electron-attracting groups on the indole scaffold would be beneficial for biological activity. Among them, it would be interesting to explore groups NO2, CF3, CN and COR among others. In Figure 6, we summarize the structure-activity relationships found in this study. With this information, new active molecules can be designed and synthesized in the future.

Materials and Methods
All organic solvents used for the synthesis were of analytical grade. All reagents used were purchased from Sigma-Aldrich (St. Louis, MO, USA), Merck (Kenilworth, NJ, USA) or AK Scientific (Union City, CA, USA) and were used as received. Melting points were determined on a Stuart Scientific SMP30 apparatus (Bibby Scientific Limited, Staffordshire, UK) and are uncorrected. NMR spectra were recorded on a Bruker Avance III HD 400 (Billerica, MA, USA) at 400.1 MHz for 1 H, 100.6 MHz for 13 C-NMR and 376.5 MHz for 19 F-NMR using the solvent signal (CDCl 3 or DMSO-d 6 ) as reference. The chemical shifts are expressed in ppm (δ scale) downfield from tetramethylsilane (TMS). Multiplicity is given as follows: s, singlet; bs, broad singlet; d, doublet; t, triplet; td, triplet of doublets; m, multiplet. Coupling constants values (J) are given in Hertz. Atom numbering and, 1 H and 13 C NMR spectra of the final compounds are available in the ESI. High resolution mass spectra were obtained on mass spectrometer with flight time analyzer (TOF) and Triwave ® system model SYNAPT™ G2 (WATERS, Milford, MA, USA), using atmospheric pressure ionization with electro spray (ESI+/−), Capillarity 3.0, source temperature 100 • C, desolvation temperature 500 • C. The IR spectra were obtained on a Bruker Vector 22 spectrophotometer (Billerica, MA, USA) using KBr discs. Column chromatography was performed on Merck silica gel 60 (70-230 mesh). Thin layer chromatographic separations were performed on Merck silica gel 60 (70-230 mesh) chromatofoils.

2-bromo-1-(5-fluoro-1H-indol-3-yl)ethan-1-one (1a)
To a solution of 5-fluoroindole (1 g, 7.40 mmol) in dry CH 2 Cl 2 (30 mL) was added anhydrous zinc chloride (2.08 g, 15.25 mmol) under N 2 atmosphere; immediately, methylmagnesium bromide 3 M (2.5 mL, 7.50 mmol) was slowly added over a 20 min period and the mixture was vigorously stirred for 2 h at room temperature. After this time, bromoacetyl chloride (0.84 mL, 9.65 mmol) was added in one portion and the mixture was stirred until the starting product disappeared upon checking TLC. The reaction was quenched by adding saturated ammonium chloride solution and extracted with CH 2 Cl 2 . The combined organic layers were dried with anhydrous sodium sulfate and the removal of the solvent afforded a residue, which was further purified by column chromatography on silica gel (CH 2 Cl 2 / EtOAc 5:1) to give 929 mg of (1a) as a light brown solid. Yield: 49% mp:

Selection of Conformers and Molecular Alignment
CoMFA and CoMSIA studies were performed with SYBYL-X v1.2 (Tripos International, St. Louis, MO, USA) software installed in a Windows 10 environment on a PC with an Intel Core i7 CPU. The alignment of the best docking poses was used as a basis for the formulation of the models. For the calculation of the potentials, each compound was assigned MMFF94 loads. The compounds were divided into training and test sets in order to test both the quality of the internal and external predictive capacities of the models. Fifteen compounds were used as test sets (equivalent to 32% of the total) and 32 compounds as training sets (equivalent to 68% of the total). The Ki values were converted to pKi (−logKi).

CoMFA and CoMSIA Field Calculation
To derive the CoMFA and CoMSIA descriptor fields, the aligned training set molecules were placed in a three-dimensional cubic lattice with a grid spacing of 2Å in the x, y, and z directions such that the entire set was included in it. The CoMFA steric and electrostatic field energies were calculated using a sp 3 carbon probe atom with a Van der Waals radius of 1.52 Å and a charge of +1.0. The cut-off values for both steric and electrostatic fields were set to 30.0 kcal/mol. For CoMSIA analysis, the standard settings (probe with charge +1.0, radius 1Å, hydrophobicity +1.0, H-bond donating +1.0, and H-bond accepting +1.0 [54]) were used to calculate five different fields: steric, electrostatic, hydrophobic, donor, and acceptor. Gaussian-type distance dependence was used to measure the relative attenuation of the field position of each atom in the lattice and led to a much smoother sampling of the fields around the molecules when compared to CoMFA. The default value of 0.3 was set for attenuation factor α.

Internal Validation and Partial Least Squares (PLS) Analysis
PLS analysis was used to construct a linear correlation between the CoMFA and CoM-SIA descriptors (independent variables) and the activity values (dependent variables) [55].
To select the best model, cross-validation analysis was performed using the leave-one-out (LOO) method (and sample distance PLS [SAMPLS]), which generated the square of the cross-validation coefficient (q 2 ) and the optimum number of components (N). The non-cross validation was performed with a column filter value of 2.0 to speed up the analysis and reduce the noise. The q 2 , which is a measure of the internal quality of the models, was obtained according to the following Equation (1): where y i , y, and y pred are the observed, mean, and predicted activity in the training set, respectively.

Conclusions
In this work, we reported the synthesis, biological evaluation, and modeling studies of novel N-arylsulfonyl indole derivatives as ligands of the 5-HT 6 receptor. A convenient synthesis was achieved to readily access twenty-two novel diversely substituted derivatives of the extended indole-aryl piperazines. Fifteen of the tested compounds exhibited nanomolar affinity for the 5-HT 6 receptor, with compound 4d being the most potent, having a K i of 58 nM, which is better than the 5-HT endogenous ligand. Protein-ligand molecular dynamics simulations provided us a deeper insight about the influence of the 5-methoxy or 5-fluorine substitutions in the binding mode of the ligands, suggesting that these C-5 substitutions are detrimental to affinity due to them precluding the ligand from plunging inside the binding site. On the other hand, in the 3D-QSAR studies, the CoMFA and CoMSIA models presented high values of q 2 (0.653; 0.692) and r 2 (0.879; 0.970), respectively. The biological activity of the compounds can be explained based on the steric and electronic properties, but mainly on their electronic nature. The information gained in this study will allow the design of novel derivatives with improved activity.