Computer-Aided Studies for Novel Arylhydantoin 1,3,5-Triazine Derivatives as 5-HT6 Serotonin Receptor Ligands with Antidepressive-Like, Anxiolytic and Antiobesity Action In Vivo

This study focuses on the design, synthesis, biological evaluation, and computer-aided structure-activity relationship (SAR) analysis for a novel group of aromatic triazine-methylpiperazines, with an hydantoin spacer between 1,3,5-traizine and the aromatic fragment. New compounds were synthesized and their affinities for serotonin 5-HT6, 5-HT1A, 5-HT2A, 5-HT7, and dopamine D2 receptors were evaluated. The induced-fit docking (IFD) procedure was performed to explore the 5-HT6 receptor conformation space employing two lead structures. It resulted in a consistent binding mode with the activity data. For the most active compounds found in each modification line, anti-obesity and anti-depressive-like activity in vivo, as well as “druglikeness” in vitro, were examined. Two 2-naphthyl compounds (18 and 26) were identified as the most active 5-HT6R agents within each lead modification line, respectively. The 5-(2-naphthyl)hydantoin derivative 26, the most active one in the series (5-HT6R: Ki = 87 nM), displayed also significant selectivity towards competitive G-protein coupled receptors (6–197-fold). Docking studies indicated that the hydantoin ring is stabilized by hydrogen bonding, but due to its different orientation, the hydrogen bonds form with S5.44 and N6.55 or Q6.58 for 18 and 26, respectively. Compound 26 exerted anxiolytic-like and antidepressant-like activities. Importantly, it demonstrated anti-obesity properties in animals fed palatable feed, and did not show toxic effects in vitro.


Introduction
The serotonin receptor 5-HT6 is one of the recently discovered members of the 5-HTRs family [1]. It is quite unique, since it includes a short third cytoplasmatic loop in parallel, with a long C-terminal tail, and one introne located in the middle of the third cytoplasmatic loop [2,3]. The localization of 5-HT6R is limited to the central nervous system, especially as it is located in the brain areas involved in learning and memory processes. Relatively limited, but intensive, research efforts in this area led to several families of potent 5-HT6R ligands [4,5], which have been proposed for potential treatment in the cognitive dysfunction associated with Alzheimer's disease, anxiety or depression [5][6][7]. Over 20 compounds have qualified for clinical trials, most of which in studies following a cure for Alzheimer's disease [5]. Nonetheless, none has reached the pharmaceutical market yet. Recent lines of evidence also point towards promising properties of some 5-HT6R ligands in the battle against obesity [8,9]. Therefore, the search for new ligands for the 5-HT6R, in particular, those with an action directed against obesity, seems to be a challenge for current medicinal chemistry.
Lines of evidence have identified several groups of chemical compounds, which displayed significant affinities for the serotonin receptors 5-HT6 [4,10,11]. The structure-activity relationship (SAR) analysis performed for the 5-HT6 ligands obtained previously enabled to identify pharmacophore features, including a triangle topology with tops of: A bulky hydrophobic area (HYD), positive ionizable nitrogen (PI), and hydrogen bond acceptor (HBA), as well as a central aromatic fragment (AR) (Figure 1) [12]. Taking into consideration the pharmacophore features, our previous studies have successfully found a new group of potent 5-HT6R agents among methylpiperazine derivatives of benzyl Taking into consideration the pharmacophore features, our previous studies have successfully found a new group of potent 5-HT 6 R agents among methylpiperazine derivatives of benzyl 2-amino-1,3,5-triazines which fit 3 out of 4 pharmacophore features. Two closed HBAs were the missing feature [13]. The computer-aided SAR analysis allowed us to modify the previous pharmacophore features and it has underlined the notable importance of aromatic substituents separated by a short linker from the 1,3,5-triazine. Thus, a bulky aromatic moiety, represented by the naphthyl (1) or m-chloro- (2) or m-methyl-(3) substituted phenyl ring (Figure 1) was beneficial for the kind of 5-HT 6 R affinity evaluated experimentally [13].
Taking into account both, the need to search for the 5-HT 6 R agents with new pharmacological possibilities and the results of previous SAR studies, we decided to explore a novel group of triazine-methylpiperazine derivatives (Figure 2), in which the spacer between 1,3,5-traizine and aromatic fragment is represented by a differently positioned hydantoin system containing both, two closed HBAs and HYD/AR features. In the first step, five pilot compounds (4)(5)(6)(7)(8) were synthesized and evaluated with regard to their affinity for the 5-HT 6 R. These initial studies resulted in two parallel lead structures 5 and 7 ( Figure 2). Chemical modifications of the lead structures subsequently provided a series of compounds 9-27 (Table 1). Therefore, these studies are focused on the design, synthesis and biological evaluation, including affinity and selectivity for the 5-HT 6 R, as well as computer-aided SAR discussion for an entire series of hydantoin-triazines . For representative compounds, anti-obesity and anti-depressive activity in vivo, as well as their "drugability" properties, were also examined.
Molecules 2018, 23, x FOR PEER REVIEW 3 of 26 2-amino-1,3,5-triazines which fit 3 out of 4 pharmacophore features. Two closed HBAs were the missing feature [13]. The computer-aided SAR analysis allowed us to modify the previous pharmacophore features and it has underlined the notable importance of aromatic substituents separated by a short linker from the 1,3,5-triazine. Thus, a bulky aromatic moiety, represented by the naphthyl (1) or m-chloro- (2) or m-methyl-(3) substituted phenyl ring ( Figure 1) was beneficial for the kind of 5-HT6R affinity evaluated experimentally [13]. Taking into account both, the need to search for the 5-HT6R agents with new pharmacological possibilities and the results of previous SAR studies, we decided to explore a novel group of triazine-methylpiperazine derivatives (Figure 2), in which the spacer between 1,3,5-traizine and aromatic fragment is represented by a differently positioned hydantoin system containing both, two closed HBAs and HYD/AR features. In the first step, five pilot compounds (4)(5)(6)(7)(8) were synthesized and evaluated with regard to their affinity for the 5-HT6R. These initial studies resulted in two parallel lead structures 5 and 7 ( Figure 2). Chemical modifications of the lead structures subsequently provided a series of compounds 9-27 (Table 1). Therefore, these studies are focused on the design, synthesis and biological evaluation, including affinity and selectivity for the 5-HT6R, as well as computer-aided SAR discussion for an entire series of hydantoin-triazines . For representative compounds, anti-obesity and anti-depressive activity in vivo, as well as their "drugability" properties, were also examined.

Chemical Synthesis of Compounds 5-26
The final compounds 5-26 were obtained using four different synthesis pathways (Scheme 1a-d).

Molecular Modeling
Molecular docking was performed to gain an insight into the binding mode of the library of compounds synthetized  to the recently developed 5-HT6R homology models [13] built on the β2 adrenergic receptor template and optimized for the structures of Lead 1 and 2. The molecular docking indicated that newly synthesized compounds, generally, exhibited a very consistent binding mode with recently reported 5-HT6 ligands [16][17][18]. An influence of the topology of aromatic substituents and the position of the 1,3,5-triazine substitution at the hydantoin ring on the binding Scheme 1.

Molecular Modeling
Molecular docking was performed to gain an insight into the binding mode of the library of compounds synthetized  to the recently developed 5-HT 6 R homology models [13] built on the β 2 adrenergic receptor template and optimized for the structures of Lead 1 and 2. The molecular docking indicated that newly synthesized compounds, generally, exhibited a very consistent binding mode with recently reported 5-HT 6 ligands [16][17][18]. An influence of the topology of aromatic substituents and the position of the 1,3,5-triazine substitution at the hydantoin ring on the binding has been observed ( Figure 2) and was in good agreement with results of the radioligand binding assay (see below).

Radioligand Binding Assay
Compounds 5-26 were evaluated with respect to their affinity and selectivity for the target 5-HT 6 R in radioligand binding assays based on the human serotonin receptors 5-HT 6 , 5-HT 1A , 5-HT 2A , 5-HT 7b R and dopaminergic D 2L R, all of whom were expressed stably in HEK-293 cells. Five selective radioligands were employed (Table 1). Results expressed as K i indicate that all derivatives of both leads displayed submicromolar affinities for the target 5-HT 6 R. More active agents (6, 7, 18, 23, 25 and 26) exerted K i (5-HT 6 R) values lower than 200 nM. Compound 26 was the most active one (K i = 87 nM) and also highly selective over 5-HT 1A R, 5-HT 2A R, and D 2L R, with some selectivity over 5-HT 7 R (6-fold). Furthermore, a slightly less active 5-HT 6 R agent, compound 25, was highly selective with respect to all competitive GPCRs tested ( Table 1). Most of the compounds displayed much weaker affinities for 5-HT 1A , 5-HT 2A and D 2L R receptors than for 5-HT 6 R. Notably, compounds 8-11, 14, 17 and 24 were more potent for 5-HT 7 R than for the target 5-HT 6 R, in particular, compound 14 (5-HT 7 R; K i = 88 nM), which was also selective toward 5-HT 7 R (Table 1).

Behavioral Tests In Vivo
The most potent 5-HT 6 R agent identified in the radioligand binding assays, i.e., compound 26, was selected for behavioral assays to determine its antidepressant-and anxiolytic-like properties in vivo in male Wistar rats.

.2. Behavioral Tests In Vivo
The most potent 5-HT6R agent identified in the radioligand binding assays, i.e., compound 2 s selected for behavioral assays to determine its antidepressant-and anxiolytic-like properties o in male Wistar rats.

tidepressant-Like Activity of Compound 26
In the forced swim test (FST), compound 26 significantly decreased immobility time about 20 NOVA: F(3,28) = 3.7674, p < 0.05) vs. vehicle treated group. A U-shaped dose-response in the FS t for compound 26 was observed ( Figure 3). The data are presented as the mean ± SEM of 6-8 rats. The data were statistically evaluated by one-way ANOVA followed by Bonferroni's post-hoc test, * p < 0.05 vs. vehicle group.
This effect seems to be typical for some antidepressants with different mechanisms of actio o producing U-shaped dose-response curves in some animal models and evaluated fo idepressant-like activity [19,20]. Moreover, this effect was also observed for other 5-HT6 ligand ., SB258585) [21].
xiolytic-Like Activity of Compound 26 The data are presented as the mean ± SEM of 6-8 rats. The data were statistically evaluated by one-way ANOVA followed by Bonferroni's post-hoc test, * p < 0.05 vs. vehicle group.
This effect seems to be typical for some antidepressants with different mechanisms of action also producing U-shaped dose-response curves in some animal models and evaluated for antidepressant-like activity [19,20]. Moreover, this effect was also observed for other 5-HT 6 ligands (i.e., SB258585) [21].

Anxiolytic-Like Activity of Compound 26
Anxiolytic-like activity was assessed using a Vogel conflict drinking test. As demonstrated in Figure 4, compound 26, administered at a dose of just 3 mg/kg, has significantly increased (by 76%) the number of accepted shocks (ANOVA: F(3,22) = 3.4967, p < 0.05) during the 5 min experimental session. Thus, compound 26 (at a dose of 3 mg/kg) exerted anxiolytic-like activity in this behavioral test.  The data are presented as the mean ± SEM of 6-8 rats. The data were statistically evaluated by one-way ANOVA followed by Bonferroni's post-hoc test, * p < 0.05 vs. vehicle group. .

Metabolic Assays In Vivo
Compounds 6 and 26, representing the most potent 5-HT6R agents of both groups, A (6) and ), were evaluated with regard to their influence on the body weight of male Wistar rats withi tabolic assays in vivo ( Figure 5). Animals that were fed palatable feed, and treated with ibited significantly less weight gain when compared to animals in the control group consuming ndard feed. From day 16 of the experiment, there was a significant difference between the group e body weight of rats treated with 6 having access to preferential feed differed significantly from weight of control animals fed standard feed. Results are shown in Figure 3A,B.
Similarly, animals fed palatable feed and treated with 26 showed significantly less weight gai n animals in the control group consuming a preferential feed ( Figure 5C,D), but there was nificant difference between the groups from day 12 of the experiment,. Additionally, the bod ight of rats treated with 26 and having access to preferential feed did not differ significantly from weight of control animals fed standard feed. No effects on body weight were noted in anima ated either with 6 or with 26 and consuming standard feed. The data are presented as the mean ± SEM of 6-8 rats. The data were statistically evaluated by one-way ANOVA followed by Bonferroni's post-hoc test, * p < 0.05 vs. vehicle group.

Metabolic Assays In Vivo
Compounds 6 and 26, representing the most potent 5-HT 6 R agents of both groups, A (6) and B (26), were evaluated with regard to their influence on the body weight of male Wistar rats within metabolic assays in vivo ( Figure 5). Animals that were fed palatable feed, and treated with 6, exhibited significantly less weight gain when compared to animals in the control group consuming a standard feed. From day 16 of the experiment, there was a significant difference between the groups. The body weight of rats treated with 6 having access to preferential feed differed significantly from the weight of control animals fed standard feed. Results are shown in Figure 3A,B.
Similarly, animals fed palatable feed and treated with 26 showed significantly less weight gain than animals in the control group consuming a preferential feed ( Figure 5C,D), but there was a significant difference between the groups from day 12 of the experiment,. Additionally, the body weight of rats treated with 26 and having access to preferential feed did not differ significantly from the weight of control animals fed standard feed. No effects on body weight were noted in animals treated either with 6 or with 26 and consuming standard feed.  The change in body weight in Wistar rats fed palatable diet or standard diet and in Wistar rats fed palatable diet or standard diet treated for 21 days with 6 (A,B) or 26 (C,D). Results are means ± SEM, n = 6. Multiple comparisons were performed by two-way ANOVA, Sidak post-hoc (A,C) or by one-way ANOVA, Sidak post-hoc (B,D); * p < 0.05, ** p < 0.01, *** p < 0.001 significant vs. control rats fed standard diet;ˆp < 0.05,ˆˆp < 0.01 significant vs. control rats fed palatable diet.

Lipophilicity
The lipophilicity for compounds considered in the search for lead structures (5)(6)(7)(8) in particular for the most active compounds found as derivatives of both lead modifications (18, 23, 25 and 26) was determined with the standard RP-TLC method [22]. For comparison, the active arylmethyl-triazine compounds previously found (1-3) were also examined. Stationary phase RP-18 and mixtures of water/methanol (1:9, v/v) as a mobile phase were employed. R M values were calculated from the R F data obtained. It was found that the R M parameters decreased linearly with increasing amounts of the organic modifier (methanol) in the mobile phases tested. The regression coefficients (r 2 ) determined for all compounds were higher than 0.95. On the basis of the linear relationship between R M values and the volume fraction of methanol, values of R M0 in the systems analyzed corresponding to 100% of water were obtained by extrapolation. R M0 values received for the compounds tested ranged from 3.14 to 4.49 ( Figure 6). The lipophilicity for compounds considered in the search for lead structures (5)(6)(7)(8) in particular for the most active compounds found as derivatives of both lead modifications (18, 23, 25 and 26) was determined with the standard RP-TLC method [22]. For comparison, the active arylmethyl-triazine compounds previously found (1-3) were also examined. Stationary phase RP-18 and mixtures of water/methanol (1:9, v/v) as a mobile phase were employed. RM values were calculated from the RF data obtained. It was found that the RM parameters decreased linearly with increasing amounts of the organic modifier (methanol) in the mobile phases tested. The regression coefficients (r 2 ) determined for all compounds were higher than 0.95. On the basis of the linear relationship between RM values and the volume fraction of methanol, values of RM0 in the systems analyzed corresponding to 100% of water were obtained by extrapolation. RM0 values received for the compounds tested ranged from 3.14 to 4.49 ( Figure 6). The most lipophilic character is associated with compound 23 (RM0 = 4.49), while the least lipophilic hydantoin compound is a weak 5-HT6R agent 5 (RM0 = 3.22). The arylmethyl derivatives previously investigated (1-3) displayed slightly less lipophilic properties than those of the most active hydantoin agents (18)(19)(20)(21)(22)(23)(24)(25)(26). Overall, these results indicate a good "drug-like" lipophilicity for the entire series of compounds investigated, taking into consideration either classical rules (Rules of Five, Ghose or Pfizer) or the latest CNS MPO approach [23].

Toxicity In Vitro
The preliminary evaluation of the safety profile for the most promising 5-HT6R ligands 6 and 26 was performed in vitro with the HEK-293 eukaryotic cell line. The compounds examined were dissolved in culture growth media and incubated with cells for 72 h. Next, the MTS test was applied to determine the viability of the cells. As shown in Figure 7, a statistically significant decrease of viability (*** p < 0.001) was observed only for compound 6 and only at the highest concentration used, i.e., 100 μM. The reference cytostatic drug doxorubicin (DX) at the very low concentration 1 μM was also cytotoxic. Thus, the results obtained confirm a low or no significant toxicity for compounds 6 and 26, respectively. The most lipophilic character is associated with compound 23 (R M0 = 4.49), while the least lipophilic hydantoin compound is a weak 5-HT 6 R agent 5 (R M0 = 3.22). The arylmethyl derivatives previously investigated (1-3) displayed slightly less lipophilic properties than those of the most active hydantoin agents (18)(19)(20)(21)(22)(23)(24)(25)(26). Overall, these results indicate a good "drug-like" lipophilicity for the entire series of compounds investigated, taking into consideration either classical rules (Rules of Five, Ghose or Pfizer) or the latest CNS MPO approach [23].

Toxicity In Vitro
The preliminary evaluation of the safety profile for the most promising 5-HT 6 R ligands 6 and 26 was performed in vitro with the HEK-293 eukaryotic cell line. The compounds examined were dissolved in culture growth media and incubated with cells for 72 h. Next, the MTS test was applied to determine the viability of the cells. As shown in Figure 7, a statistically significant decrease of viability (*** p < 0.001) was observed only for compound 6 and only at the highest concentration used, i.e., 100 µM. The reference cytostatic drug doxorubicin (DX) at the very low concentration 1 µM was also cytotoxic. Thus, the results obtained confirm a low or no significant toxicity for compounds 6 and 26, respectively.

Blood-Brain Barrier Permeability
In order to estimate an ability of the library synthetized (4-26) to penetrate the blood-brain barrier (BBB), the permeability QPlogBB parameter was calculated ( Table 2) using QikProp from Schrodinger Suite 2017 [24]. QikProp predictions are for orally delivered drugs, and thus some natural neurotransmitters, e.g., dopamine and serotonin, are CNS negative, because they are extremely polar to cross the blood-brain barrier. Recommended values range for compounds that penetrate BBB is from −3.0 to 1.2. All the hydantoin-triazines investigated  are predicted as BBB-permeable (Table 2) with QPlogBB values from −0.98 (25) to −0.33 (12).

Computer-Aided SAR Analysis
In order to expand the current knowledge about structural properties responsible for 5-HT6R activity in the group of recently discovered aryl derivatives of 1,3,5-triazine, this study took into account the crucial role of the linker, which was found by comparative SAR analysis for the previously active arylmethyl-and inactive aryl-triazine molecules [13]. The new series of hydantoin-triazines (5-26) were designed as an extension of the methylene linker of the most active 2-arylmethyl-1,3,5-triaines, via an introduction of the inflexible, cyclic hydantoin insert, containing also hydrogen bond acceptors, and occurring in two topological variants found to be profitable in the preliminary lead identification (4)(5)(6)(7)(8). Both leads (6 and 7) and their derivatives (9-18 and 19-26, respectively) displayed significant submicromolar affinities for the target 5-HT6 receptors, much more potent than those of previously investigated linker-free aryl-triazines [13], but slightly less potent if compared to the arylmethyl ones (1-3, Figure 1). Interestingly, the influence of the topology of the aromatic moieties in the group of 1,3,5-triazine derivatives seems to be distinctly

Blood-Brain Barrier Permeability
In order to estimate an ability of the library synthetized (4-26) to penetrate the blood-brain barrier (BBB), the permeability QPlogBB parameter was calculated ( Table 2) using QikProp from Schrodinger Suite 2017 [24]. QikProp predictions are for orally delivered drugs, and thus some natural neurotransmitters, e.g., dopamine and serotonin, are CNS negative, because they are extremely polar to cross the blood-brain barrier. Recommended values range for compounds that penetrate BBB is from −3.0 to 1.2. All the hydantoin-triazines investigated (4-26) are predicted as BBB-permeable (Table 2) with QPlogBB values from −0.98 (25) to −0.33 (12).

Computer-Aided SAR Analysis
In order to expand the current knowledge about structural properties responsible for 5-HT 6 R activity in the group of recently discovered aryl derivatives of 1,3,5-triazine, this study took into account the crucial role of the linker, which was found by comparative SAR analysis for the previously active arylmethyl-and inactive aryl-triazine molecules [13]. The new series of hydantoin-triazines  were designed as an extension of the methylene linker of the most active 2-arylmethyl-1,3,5-triaines, via an introduction of the inflexible, cyclic hydantoin insert, containing also hydrogen bond acceptors, and occurring in two topological variants found to be profitable in the preliminary lead identification (4)(5)(6)(7)(8). Both leads (6 and 7) and their derivatives (9-18 and 19-26, respectively) displayed significant submicromolar affinities for the target 5-HT 6 receptors, much more potent than those of previously investigated linker-free aryl-triazines [13], but slightly less potent if compared to the arylmethyl ones (1-3, Figure 1). Interestingly, the influence of the topology of the aromatic moieties in the group of 1,3,5-triazine derivatives seems to be distinctly linker-dependent. Although the bulky aromatic area of the naphthalene was beneficial for all the groups considered, it was distinctly predominant in the case of the 5-arylhydantoin (25, 26 vs. 7 and 19-24; Table 1), comparable with the 3-methylphenyl and 3-chlorophenyl compounds in the group arylmethyl triazines (1-3, Figure 1), and slightly less favourable than the 4-chlorophenyl substituent within the 5,5-dimethylhydantoin group (18 vs. 6, Table 1). Furthermore, the preferable m-substitution within the phenyl ring, found for the benzyl 1,3,5-triazine 5-HT 6 R agents earlier [13], has not been confirmed in this study, either for the 5,5-dimethyl-1-benzylhydantoin-(6 vs. 10) or for the 5-phenyl-5-methylhydantoin triazine compounds (7 vs. 20). Consequently, an influence of both variants of the hydantoin linkers (1,3,5-substitution, group A or 3,5-substitution, group B) on the 5-HT 6 R activity was dependent on the properties of the aromatic moiety. Thus, both variants appear to be comparable, although the most active compound (26) is a member of the 5-aryl-5-methylhydantoin derivatives (group B).
Results of our docking studies have provided data useful to explain SAR on the molecular level ( Figure 8).
Molecules 2018, 23, x FOR PEER REVIEW 12 of 26 linker-dependent. Although the bulky aromatic area of the naphthalene was beneficial for all the groups considered, it was distinctly predominant in the case of the 5-arylhydantoin (25, 26 vs. 7 and 19-24; Table 1), comparable with the 3-methylphenyl and 3-chlorophenyl compounds in the group arylmethyl triazines (1-3, Figure 1), and slightly less favourable than the 4-chlorophenyl substituent within the 5,5-dimethylhydantoin group (18 vs. 6, Table 1). Furthermore, the preferable m-substitution within the phenyl ring, found for the benzyl 1,3,5-triazine 5-HT6R agents earlier [13], has not been confirmed in this study, either for the 5,5-dimethyl-1-benzylhydantoin-(6 vs. 10) or for the 5-phenyl-5-methylhydantoin triazine compounds (7 vs. 20). Consequently, an influence of both variants of the hydantoin linkers (1,3,5-substitution, group A or 3,5-substitution, group B) on the 5-HT6R activity was dependent on the properties of the aromatic moiety. Thus, both variants appear to be comparable, although the most active compound (26) is a member of the 5-aryl-5-methylhydantoin derivatives (group B). Results of our docking studies have provided data useful to explain SAR on the molecular level ( Figure 8). In general, all of the newly synthesized derivatives of Lead 1 and Lead 2 exhibited a binding mode to 5-HT6R, which was very consistent with derivatives of 1,3,5-tiazines studied previously [13]. In addition to the crucial salt bridge interaction with D3.32, all of the docked ligands formed CH-π aromatic interactions with F6.52, and partially with F6.51. Additionally, the -NH2 group of the 1,3,5-triazine fragment was, usually, hydrogen-bonded with the carbonyl oxygen of A5.42 and V3.33 ( Figure 8A). The methylene linker connecting 1,3,5-triazine and hydantoin rings, due to its tetrahedral conformation, forced the terminal aromatic group into a hydrophobic cavity formed by transmembrane helices 3-5 and extracellular loop 2. Thus, the structure-activity relationship in this series is determined by the different substituents at the terminal aromatic ring, and by the different orientation of hydantoin ring induced by its different positioning within the compound's structure. A comparison of the analogues with different positions of the 1,3,5-triazine substitution at the hydantoin ring indicated that linking by the methylene group at position 3 of hydantoin is profitable (6, Ki = 127 nM), whereas a significant decrease of activity was noted for the derivative where the 1,3,5-triazine was linked with hydantoin via position 1 (5, Ki = 4275 nM). The analysis of the binding mode showed that a differently positioned hydantoin ring induces its different spatial orientation in the binding site ( Figure 8A), and only the complex of 6 with the receptor is stabilized additionally by a hydrogen bond with a Q6.58 side chain.
Interestingly, within the group A and B, the best activity showed 2-naphthyl derivatives (18 and 26, respectively; Figure 8B). It should be noted, that apart from the length of the linker and the positioning of the hydantoin ring, the difference of activity between 18 and 26 is negligible-a In general, all of the newly synthesized derivatives of Lead 1 and Lead 2 exhibited a binding mode to 5-HT 6 R, which was very consistent with derivatives of 1,3,5-tiazines studied previously [13]. In addition to the crucial salt bridge interaction with D3.32, all of the docked ligands formed CH-π aromatic interactions with F6.52, and partially with F6.51. Additionally, the -NH 2 group of the 1,3,5-triazine fragment was, usually, hydrogen-bonded with the carbonyl oxygen of A5.42 and V3.33 ( Figure 8A). The methylene linker connecting 1,3,5-triazine and hydantoin rings, due to its tetrahedral conformation, forced the terminal aromatic group into a hydrophobic cavity formed by transmembrane helices 3-5 and extracellular loop 2. Thus, the structure-activity relationship in this series is determined by the different substituents at the terminal aromatic ring, and by the different orientation of hydantoin ring induced by its different positioning within the compound's structure. A comparison of the analogues with different positions of the 1,3,5-triazine substitution at the hydantoin ring indicated that linking by the methylene group at position 3 of hydantoin is profitable (6, K i = 127 nM), whereas a significant decrease of activity was noted for the derivative where the 1,3,5-triazine was linked with hydantoin via position 1 (5, K i = 4275 nM). The analysis of the binding mode showed that a differently positioned hydantoin ring induces its different spatial orientation in the binding site ( Figure 8A), and only the complex of 6 with the receptor is stabilized additionally by a hydrogen bond with a Q6.58 side chain.
Interestingly, within the group A and B, the best activity showed 2-naphthyl derivatives (18 and 26, respectively; Figure 8B). It should be noted, that apart from the length of the linker and the positioning of the hydantoin ring, the difference of activity between 18 and 26 is negligible-a comparison of its binding modes indicated that the most significant difference is noted for the hydantoin ring interaction.
In both derivatives, the hydantoin ring is stabilized by hydrogen bonding, but due to its different orientation, the hydrogen bonds form with S5.44 and N6.55 or Q6.58 for 18 and 26, respectively. It is worth to note that the shift of the chlorine atom from 3-to 4-position resulted in a 4-(for modification B) to 5-fold (for modification A) increase of affinity for 5-HT 6 R compared to its unsubstituted analogues. This relationship was confirmed by molecular docking ( Figure 8C) revealing that only the presence of chlorine atom in 4-position stabilized the L-R complex through the formation of halogen bond with the carbonyl oxygen of P4.60 residue (Cl···O distance = 3.31 Å, σ-hole angle = 151.47 • ). As the halogen bond appeared to possess a highly directional nature, to explain the decrease in activity by shifting the chlorine atom from position 3 to 4 in the phenyl ring, the interaction sphere [25,26] was plotted onto the relevant backbone carbonyl oxygens ( Figure 8C). The 4-chloro substituent was positioned within the energetically favorable areas of the sphere, whereas the 3-chloro substituent pointed outside of the sphere, indicating no ability to form halogen bond.

Chemistry
Reagents were purchased from Alfa Aesar (Karlsruhe, Germany) or Sigma Aldrich (Darmstadt, Germany). Methanol was dried over calcium oxide Reaction progress was verified using thin layer chromatography (TLC), which was carried out on 0.2 mm Merck silica gel 60 F254 plates. Spots were visualized by UV light or treatment with Dragendorff reagent. Melting points (m.p.) were determined using MEL-TEMP II apparatus (LD Inc., Long Beach, CA, USA) and are uncorrected. The 1 H-NMR and 13 C-NMR spectra were obtained on a Mercury-VX 300 Mz spectrometer (Varian, Palo Alto, CA, USA) in DMSO-d 6 or CDCl 3 . Chemical shifts in 1 H-NMR spectra were reported in parts per million (ppm) on the δ scale using the solvent signal as an internal standard. Data are reported as follows: Chemical shift, multiplicity (s, singlet; br.s., broad singlet; d, doublet; t, triplet; m, multiplet), coupling constant J in Hertz (Hz), number of protons, proton's position (Ind-indole, Naph-naphthalene, Ph-phenyl, Pp-piperazine). LC-MS were carried out on a system Water TQ Detector (Water Corporation, Milford, CT, USA) consisting of a Waters Acquity UPLC, coupled to a Waters TQD mass spectrometer. Retention times (t R ) are given in minutes. The UPLC/MS purity of all final compounds was determined (%).

Molecular Docking
The 5-HT 6 R models selected in IFD procedure were used to study the binding mode of the library synthesized. 3-dimensional structures of the ligands were prepared using LigPrep v3.6 (Schrödinger, New York, NY, USA) [36], and the appropriate ionization states at pH = 7.4 ± 1.0 were assigned using Epik v3.4 (Schrödinger, New York, NY, USA) [37]. Compounds with unknown absolute configuration were docked in all configurations. The Protein Preparation Wizard was used to assign the bond orders, appropriate amino acid ionization states and to check for steric clashes. The receptor grid was generated (OPLS3 force field [38]) by centering the grid box with a size of 12 Å on the D3.32 side chain. Automated flexible docking was performed using Glide at SP level [39].

Plotting Interaction Spheres for Halogen Bonding
The halogen bonding web server was used (accessed on 22 February 2018, http://www. halogenbonding.com/) to visualize (plotting interaction spheres) the potential contribution of halogen bonding to ligand-receptor complexes.

Thin-Layer Chromatography
The mobile phases were prepared by mixing the respective amounts of water and organic modifier (methanol) in a range from 40 to 90% (v/v) in 5% increments. TLC was performed on Silica gel 60RP-18 F 254 plates (7 × 10 cm) plates (Merck, Darmstadt, Germany). Methanol was used to prepare the solutions of the substances. Solutions (10 µL) of the analyzed compounds were applied to the plates as 5 mm bands, 10 mm apart and 10 mm from the lower edge and sides of the plates, by using Linomat V applicator (Camag, Basel, Switzerland). The vertical chamber (Sigma-Aldrich, St. Louis, MI, USA), 20 × 10 × 18 cm in size, was saturated with mobile phase for 20 min. Development was carried out over 9 cm from the starting line at a temperature of 20 • C. Next, the plates were dried at room temperature, and the spots were observed under ultraviolet light at 254 nm and/or 366 nm (UV lamp, Camag, Basel, Switzerland). In each case, sharp and symmetric spots without a tendency for tailing were obtained. Each experiment was run in triplicate and mean R F (retardation factor) values were calculated.
Starting from the R F values, the R M parameters were computed as described in the formula: The linear correlations between the R M values of the substances and the concentration of methanol in mobile phases were calculated for each compound with the Soczewiński-Wachtmeister equation [22]:

Drugs
Compounds were suspended in 1% Tween 80 immediately before administration in a volume of 2 mL/kg-behavioral tests or 1 mL/kg-body weight measures. Compounds were administered intraperitoneally (i.p.) 60 min before testing-behavioral tests. Control animals received vehicle (1% Tween 80) according to the same schedule.

Behavioral Procedures in Rats
Forced Swim Test (FST Test) The experiment was carried out according to the method of Porsolt et al. [41]. On the first day of an experiment, the animals were gently individually placed in Plexiglas cylinders (40 cm high, 18 cm in diameter) containing 15 cm of water maintained at 23-25 • C for 15 min. On removal from water, the rats were placed for 30 min in a Plexiglas box under 60-W bulb to dry. On the following day (24 h later), the rats were replaced in the cylinder and the total duration of immobility was recorded during the whole 5-min test period. The immobility was assigned when no additional activity was observed other than that necessary to keep the rat's head above the water. Fresh water was used for each animal.

Vogel Conflict Drinking Test
The testing procedure was based on a method of Vogel et al. [42] and used the Anxiety Monitoring System "Vogel test" produced by TSE Systems (Germany). It was consisted of polycarbonate cages (dimensions 26.5 × 15 × 42 cm), equipped with a grid floor made from stainless steel bars and drinking bottles containing tap water. Experimental chambers were connected to PC software by control chassis and on electric shocks' generator. On the first day of the experiment, the rats were adapted to the test chambers and drink water from the bottle spout for 10 min. Afterward, the rats were returned to their home cages and were given 30 min free access to water followed by a 24 h water deprivation period. The adaptation session and water deprivation protocols were repeated on the second day of the experiment. On the third day, the rats were placed again in the test chambers 60 min after compound 26 and diazepam administration and were given free access to the drinking tube. Recording data started immediately after the first lick and rats were punished with an electric shock (0.5 mA, lasting 1 s) delivered to the metal drinking tube every 20 licks. The number of licks and the number of shocks received during a 5 min experimental session were recorded automatically.

Metabolic Assays In Vivo
The Effect of compound 6 or 26 on Body Weight by Non-Obese Rats Fed Palatable Diet (Model of Excessive Eating) In order to determine the anorectic activity of 6 or 26, the model of excessive eating was used [43,44]. Male Wistar rats (170-190 g) were housed in a pair in plastic cages in constant temperature facilities exposed to a light-dark cycle. Water and food were available ad libitum. Two groups of 6 rats were fed with diets consisting of milk chocolate with nuts, cheese, salted peanuts, and 7% condensed milk and also had access to standard feed (Labofeed B, Morawski Manufacturer Feed, Kcynia, Poland) and water ad libitum for 3 weeks. Palatable control group (palatable diet + vehicle) received vehicle (1% Tween 80, intraperitoneally), while palatable test group (palatable diet + 6 or palatable + 26) was injected (intraperitoneally) with 6 or 26 at the dose of 5 mg/kg b.w. dissolved in 1% Tween 80 (1 mL/kg). Body weights were measured daily, immediately prior to administration of drugs.