Alkaloids Purified from Aristotelia chilensis Inhibit the Human α3β4 Nicotinic Acetylcholine Receptor with Higher Potencies Compared with the Human α4β2 and α7 Subtypes

Article Cite This: J. Nat. Prod. XXXX, XXX, XXX−XXX pubs.acs.org/jnp Alkaloids Purified from Aristotelia chilensis Inhibit the Human α3β4 Nicotinic Acetylcholine Receptor with Higher Potencies Compared with the Human α4β2 and α7 Subtypes Hugo R. Arias,† Marcelo O. Ortells,‡ Dominik Feuerbach,§ Viviana Burgos,⊥ and Cristian Paz*,⊥ † Department of Pharmaceutical Science, School of Pharmacy, American University of Health Sciences, Signal Hill, California 90755, United States ‡ Facultad de Medicina, Universidad de Morón, and CONICET, 1708 Morón, Buenos Aires, Argentina § Novartis Institutes for Biomedical Research, 4001 Basel, Switzerland ⊥ Laboratory of Natural Products and Drug Discovery, Department of Basic Sciences, Universidad de La Frontera, 4780000 Temuco, Chile Downloaded by UNIV OF TOLEDO at 19:47:38:387 on July 05, 2019 from https://pubs.acs.org/doi/10.1021/acs.jnatprod.9b00314. S Supporting Information * ABSTRACT: The alkaloids aristoteline (1), aristoquinoline (2), and aristone (3) were purified from the leaves of the Maqui tree Aristotelia chilensis and chemically characterized by NMR spectroscopy. The pharmacological activity of these natural compounds was evaluated on human (h) α3β4, α4β2, and α7 nicotinic acetylcholine receptors (AChRs) by Ca2+ influx measurements. The results suggest that these alkaloids do not have agonistic, but inhibitory, activity on each receptor subtype. The obtained IC50 values indicate the following receptor selectivity: hα3β4 > hα4β2 ≫ hα7. In the particular case of hα3β4 AChRs, 1 (0.40 ± 0.20 μM) and 2 (0.96 ± 0.38 μM) show higher potencies compared with 3 (167 ± 3 μM). Molecular docking and structure−activity relationship results indicate that ligand lipophilicity is important for the interaction with the luminal site located close to the cytoplasmic side of the hα3β4 ion channel between positions -2′ and -4′. Compound 1 could be used as a molecular scaffold for the development of more potent noncompetitive inhibitors with higher selectivity for the hα3β4 AChR that could serve for novel addiction and depression therapies. N especially its berries, has antiproliferative, anti-inflammatory, antioxidant, antimicrobial, cardio-protective, and nutritional properties.6−9 The leaves of A. chilensis have shown high content of noniridoid monoterpene indole alkaloids together with polyphenolic compounds. 10,11 To have a more comprehensive idea of the functional interaction of these alkaloids with human (h) AChRs, the alkaloids aristoteline (1), aristoquinoline (2), and aristone (3) were purified from the leaves of the Maqui tree and chemically characterized by NMR spectroscopy (Figure 1A). The pharmacological activity of each compound was subsequently determined by Ca2+ influxinduced fluorescence detections using cell lines each expressing only one of these AChR subtypes: hα3β4, hα4β2, or hα7. To determine the most important molecular interactions underlying the differences in receptor selectivity and inhibitory potencies among the compounds, molecular docking and molecular dynamics were performed on the hα3β4 AChR model using the recently determined X-ray crystal structure of the hα4β2 AChR at 3.9 Å resolution.12 The results show for the first time that alkaloids from A. chilensis, except 3, icotinic acetylcholine receptors (AChRs) are members of the ligand-gated ion channel superfamily that also includes the GABAA, glycine, and serotonin type 3 receptors.1 The α4β2 and α7 AChRs are the most abundant subtypes in the brain, whereas α3β4 AChRs are predominantly expressed in the habenula, a brain area involved in the modulation of the brain reward center in the midbrain.2 AChRs are involved in many physiologically vital functions (e.g., muscle contraction, homeostasis, cognition, learning, memory, pain signaling, neuroprotection, and regulation of the immune responses) and are targets for many natural products with agonistic (e.g., epibatidine, cytisine, anatoxin-a, and 6-bromohypaphorine) and antagonistic (e.g., d-tubocurarine, methyllycaconitine, coronaridine congeners, α-bungarotoxin, and α-conotoxins) properties.2−4 Since the malfunctioning of AChRs, due to low expression, mutations, and/or interaction with endogenous ligands, may evolve in important neurological diseases (e.g., epilepsy, schizophrenia, Alzheimer’s and Parkinson’s diseases, nicotine and drug addictions, and myasthenia and myasthenic syndromes),2,5 the search for natural products with AChR activity deserves special attention. Maqui [Aristotelia chilensis (Molina) Stuntz, Elaecarpaceae] is an evergreen tree found in southern Chile. The Maqui tree, © XXXX American Chemical Society and American Society of Pharmacognosy Received: April 9, 2019 A DOI: 10.1021/acs.jnatprod.9b00314 J. Nat. Prod. XXXX, XXX, XXX−XXX Journal of Natural Products Article Figure 1. (A) Molecular structure of alkaloids purified from the Maqui, A. chilensis, including 1, [(1H)indolo[2,3-g](3,5-ethanoquinoline), 2,2,5trimethyl-1,2,3,4,4a,5,11,11a-octahydro-], 3, [(1S,2R,4R,12S,15S,18R)-14,14,18-trimethyl-5,13-diazahexacyclo[13.3.1.02,13.04,12.04,18.06,11]nonadeca6,8,10-trien-16-one], and 2 [4-((1R,2S,5R)-4,4,8-trimethyl-3-azabicyclo[3.3.1]non-7-en-2-yl)quinoline]. (B) Molecular comparisons between aristoteline (black) and bupropion (green). (C) 1 and (D) bupropion molecular surfaces colored by their atomic electrostatic potentials, from more positive (blue) to more negative (red). Figure 2. Functional activity of alkaloids purified from A. chilensis on HEK293-hα3β4 cells by using Ca2+ influx measurements. (A) (±)-Epibatidine (1 μM) (■), but not aristoteline (100 μM) (○), enhanced intracellular calcium. The same results were found for the other alkaloids. Instead, 1.0 μM 1 inhibited (±)-epibatidine-evoked hα3β4 AChR activity (▲). The same inhibitory activity was found for the other alkaloids. (B) The inhibitory activity of these alkaloids was investigated by pretreating (5 min) the cells with different concentrations of 1 (▼), 2 (⧫), and 3 (▲), respectively, followed by hα3β4 activation with 0.1 μM (±)-epibatidine (■). Ligand response was normalized to the maximal (±)-epibatidine response, which was set as 100%, and expressed as mean ± SD (n = 3). The calculated IC50 and nH values are summarized in Table 3. and 13C NMR (Table 2S, Supporting Information). The results are in excellent agreement with previous data of NMR and Xray studies.13−15 Interestingly, the shape and surface electrostatic potential correspondence between 1 and bupropion are remarkable (Figure 1A−D). The benzene rings, the positive ammonium nitrogens, and the side-chain methyl groups are reasonably overlapped. The most important differences between 1 and bupropion are that the natural compound is a rigid molecule with a nitrogen atom in the indole moiety and bupropion has a ketone as part of the side-chain moiety. Pharmacologic Activity of Aristotelia Alkaloids on Human AChRs. The potency of (±)-epibatidine to activate each AChR subtype was first determined by assessing the fluorescence change in HEK293-hα3β4 (Figure 2B), HEK293hα4β2 (Figure 3A), and GH3-hα7 (Figure 3B) cells after noncompetitively inhibit hα3β4 AChRs with higher selectivity compared with hα4β2 and hα7 AChRs, and its luminal binding site is shared with that for bupropion. Since hα3β4 AChRs directly and indirectly modulate the function of the brain reward system,2 these results provide a foundation for further animal studies to determine the potential antiaddictive and antidepressive activity of 1 and 2. RESULTS AND DISCUSSION Chemical Characterization of Alkaloids Purified from A. chilensis. The alkaloids 1, 2, and 3 (Figure 1) were isolated from extracts of leaves of Aristotelia chilensis. The structures of these natural alkaloids were subsequently characterized by 1H NMR (Table 1S, Supporting Information) ■ B DOI: 10.1021/acs.jnatprod.9b00314 J. Nat. Prod. XXXX, XXX, XXX−XXX Journal of Natural Products Article Figure 3. Inhibitory activity of alkaloids purified from A. chilensis on HEK293-hα4β2 (A) and GH3-hα7 (B) cells by using Ca2+ influx measurements. The inhibitory activity of 1 (▼), 2 (⧫), and 3 (▲) was investigated on the respective hα4β2 (A) and hα7 (B) AChR subtype by pretreating (5 min) the cells with different concentrations of each drug followed by AChR activation with 0.1 μM (A) or 1.0 μM (B) (±)-epibatidine (■). The observed alkaloid-induced inhibition depended on the ligand and AChR subtype. Ligand response was normalized to the maximal (±)-epibatidine response, which was set as 100%, and expressed as mean ± SD (n = 3). The calculated IC50 and nH values are summarized in Table 3. Table 1. Inhibitory Potency (IC50) for Alkaloids Purified from A. chilensis at hα3β4, hα4β2, and hα7 AChRsa AChR subtype inhibitory potency (μM) 1 2 3 hα3β4 IC50 nHb IC50 nHb IC50 nHb IC50(hα4β2)/IC50(hα3β4) IC50(hα7)/IC50(hα3β4) 0.40 ± 0.20 1.43 ± 0.10 3.4 ± 0.9 1.22 ± 0.26 27.0 ± 3.8 1.58 ± 0.21 8.4-fold 67.5-fold 0.96 ± 0.38 1.44 ± 0.14 6.7 ± 0.1 1.11 ± 0.27 45.3 ± 0.3 1.08 ± 0.27 7.0-fold 47.1-fold 167 ± 3 1.19 ± 0.04 145 ± 14 1.10 ± 0.08 no activity ― no difference ― hα4β2 hα7 a These values were obtained from Figures 2B (hα3β4), 3A (hα4β2), and 3B (hα7), respectively. bHill coefficient. (±)-epibatidine stimulation. The observed EC50 values for each AChR subtype were in the same concentration range as those previously determined.16,18−21 The agonistic activity of several alkaloids purified from A. chilensis was subsequently tested on different human AChR subtypes by direct stimulation. The Ca2+ influx traces indicated that 1, in contrast to (±)-epibatidine, does not have agonistic activity on hα3β4 AChRs (Figure 2A). The same results were obtained with the other alkaloids on the studied AChRs (data not shown). The inhibitory activity of these natural compounds was subsequently assessed on hα3β4 (Figure 2B), hα4β2 (Figure 3A), and hα7 (Figure 3B) AChRs, respectively. Interestingly, 1 and 2 fully inhibited agonist-evoked AChR activity, whereas 3 showed slight (hα3β4 and hα4β2) or no (hα7) activity, depending on the receptor subtype. In fact, whereas aristoteline decreased hα3β4 activity by 100% in the 10−100 μM concentration range, 3 only reached ∼40% inhibition at 100 μM. The observed inhibitory potency (IC50) for the alkaloids at the studied AChRs follows the rank order: 1 > 2 ≫ 3 (Table 1). On the basis of the calculated IC50 values, the receptor selectivity for the most active alkaloids clearly follows the sequence: hα3β4 > hα4β2 ≫ hα7 (Table 1). In the case of 1, the difference in potency between the hα3β4 AChR and the other subtypes is 8.4- (hα4β2) and 67-fold (hα7), respectively. The same trend is observed for 2. Although there are other synthetic and hemisynthetic compounds that present higher inhibitory potency at α3β4 AChRs [e.g., AT-1001 (a tropane derivative) shows an IC50 value of 35 nM (determined by Ca2+-induced fluorescence assays) and antiaddictive activity,17 this is the first time that a selective inhibitory activity on human AChRs is shown for Aristotelia alkaloids. The observed nH (Hill coefficient) values for Aristotelia alkaloids at the studied AChRs are in general close to unity, suggesting a noncooperative mechanism, whereas the values for 1 and 2 at the hα3β4 and for 1 at the hα7 are clearly larger than unity (Table 1), indicating a cooperative mechanism. Correlation Between the Pharmacological Activity of Aristotelia Alkaloids and Bupropion and Their Physicochemical Parameters. To determine the structure− activity relationship for Aristotelia alkaloids (Figure 1A), a series of physicochemical parameters were calculated (Table 2) Table 2. Physicochemical Parameters of Aristotelia Alkaloids and Bupropion parameter 1 2 3 bupropion polar surface area (PSA) (Å2) LogP (lipophilicity) molecular volume (Å3) 15.10 2.95 214.4 11.26 2.78 231.9 33.32 0.60 222.6 17.30 1.22 174.6 and subsequently correlated with the IC50 values determined with hα3β4 AChR (Table 1). Considering that bupropion and 1 share the same luminal site (Figure 6D, see below), the physicochemical parameters for the antidepressant were also calculated (Table 2), and subsequently correlated with its IC50 value (2.30 ± 0.40 μM).20 Plot analyses indicated that there is an opposite relationship between LogP (lipophilicity, Figure 4A) and PAS (Polar Surface Area, Figure 4B) values for 1, 2, and bupropion. Nevertheless, when the 3 values are C DOI: 10.1021/acs.jnatprod.9b00314 J. Nat. Prod. XXXX, XXX, XXX−XXX Journal of Natural Products Article Figure 4. Correlation between the inhibitory potency (IC50) of 1 (○), 2 (△), 3 (□) (taken from Table 3), and bupropion (◊)20 against the hα3β4 AChR, and their respective LogP (lipophilicity) (A) and PSA (B) values (taken from Table 4). Plot analyses indicated that there is good correlation for 1, 2, and bupropion, but not for 3, showing cutoff values of LogP < 1.2 (A) and PSA > 17 Å2 (B), when the respective 3 value is included. considered, the plots showed cutoff values of LogP < 1.2 (Figure 4A) and PAS > 17 Å2 (Figure 4B), respectively. Molecular Docking of 1 to the hα3β4, hα4β2, or hα7 AChRs. The DOPE score for the h(α3)3(β4)2, h(α4)3(β2)2, or hα7 models were −245 702, −257 027, and −243 892, respectively, compared to −257 162 for the X-ray hα4β2 AChR structure. The h(α4)3(β2)2 stoichiometry has almost the same score as that for h(α4)2(β2)3, while the other two models had slightly lower scores, indicating better, more stable, structures. Considering that 1 is the most potent ligand at the three studied receptors, molecular docking experiments were subsequently performed using this compound as an example of Aristoteline alkaloids. Several conformers were found at different domains of each AChR subtype. To select those conformers according to the experimental data, linear regression analyses between the calculated theoretical binding energies (TBEs) (Table 3) and the experimental IC50 values (Table 1) were performed on each AChR subtype. Figure 5. Correlation between the inhibitory potency (IC50) of 1 for the respective hα3β4 (□), hα4β2 (○), and hα7 (△) AChR (taken from Table 3) and the corresponding TBE values (taken from Table 5) at the luminal (―) and vestibular (---) sites, respectively. Linear regression analyses indicated excellent correlation with both luminal (R2 = 0.999; P = 0.032) and vestibular (R2 = 1.000; P = 0.002) sites at each receptor subtype. interactions of 1 with the luminal (Figure 6D) and vestibular (Figure 6E) sites are shown for this receptor subtype. In the luminal site (Figure 6D), 1 contacted M2 residues located closer to the cytoplasmic side of the ion channel, including α3C238 and β4-C239 (position -4′; currently considered part of the M1-M2 Loop), α3-G239 and β4-G240 (position -3′), α3E240 and β4-E241 (position -2′, intermediate ring), β4-M243 (position 1′), and α3-T243 and β4-T244 (position 2′; threonine ring) (Table 4). A strong H-bond is formed between the hydrogen from the charged nitrogen (H−N+) of 1 and the oxygen (O) of the α3-E240 side chain (distance = 1.9 Å), which is observed with neither 3 nor 2. 1 and 3 established hydrophobic interactions with residues at positions 2′, 1′, -2′, and -4′. Several residues (i.e., O atoms) at positions -2′ and -3′ formed weak H-bonds with both H−N+ and neutral N (distances 3.2−3.6 Å) from either 1 or 3. In the vestibular site (Figure 6E), 1 is positioned behind the noncanonical β4+/α3- interface. The α3 residues P112 and A113 (β5-β6 loop), Q114 (β6 Sheet), D134, D135, K136, T137 (β6-β7 loop) and K138 (forming cation-π interactions with both aristoteline rings) (β7 Sheet) belong to the Cys-loop characteristic of this receptor superfamily.1 The β4 residues include N21 (N terminal helix), A86, K87, R88, I89, W90, L91, and D93 (β3-β4 loop), V95 (β4 Sheet), Y104 and E105 (forming two H-bonds) (β4-β5 loop), Y109 (forming a Hbond) (β5 Sheet), W122 (β6 Sheet), and W153 and T154 (β7-β8 loop) (Table 4). Table 3. Theoretical Binding Energies (TBE) for the Interaction of 1 with Either the Luminal or Vestibular Site at the Respective hα3β4, hα4β2, and hα7 AChR TBE (kcal/mol) site hα3β4 hα4β2 hα7 luminal (high-affinity) vestibular (low-affinity) −92.01 −57.10 −90.97 −53.01 −74.53 −22.00 The linear regressions showed two sites with significant correlation values (Figure 5): a high-affinity binding site located close to the cytoplasmic side of the ion channel, and a low-affinity site located in the extracellular domain (ECD), on the vestibular side of the noncanonical β2+/α4- interface (Figure 6A). The MD simulations (20 ns) indicated that both luminal (Figure 6B) and vestibular (Figure 6C) interactions are quite stable at the three studied receptor subtypes. The calculated RMSD variance obtained from the last third of the simulation gave values ≤0.1 for both luminal [0.005 (hα3β4), 0.008 (hα4β2), and 0.026 (hα7)] and vestibular [0.016 (hα3β4), 0.007 (hα4β2), and 0.014 (hα7)] sites, consistent with stable interactions. The molecular docking results indicated no interaction with the orthosteric sites, supporting a noncompetitive mechanism of inhibition for the alkaloids. Based on the preference of these alkaloids for the hα3β4 AChR, the detailed molecular D DOI: 10.1021/acs.jnatprod.9b00314 J. Nat. Prod. XXXX, XXX, XXX−XXX Journal of Natural Products Article Figure 6. Docking sites for 1 at the h(α3)3(β4)2 AChR model. (A) View of the h(α3)3(β4)2 AChR showing the high-affinity luminal (yellow) and low-affinity vestibular (red) sites for 1 (surface model). α3 (white) and β4 (dark gray) subunits are represented as solid ribbons. (B, C) Molecular dynamics simulations (20 ns) of 1 interacting with the respective luminal (B) and vestibular (C) site at the h(α3)3(β4)2 (----), h(α4)3(β2)2 (―), and hα7 (···) model, respectively. These results indicated that both luminal and vestibular interactions are stable. (D) In the luminal site, 1 (as sticks, and colored by atoms with carbons in dark gray) interacted with M2 residues located between positions 2′ (threonine ring ring) and -4′ (currently considered Loop M1-M2), forming a strong H-bond with α3-E240 (position -2′; intermediate ring) (distance = 1.9 Å) (thick blue line) and several weak H-bonds (distances = 3.2−3.6 Å) (thin white dotted lines). (E) In the vestibular site, aristoteline interacted with several extracellular substructures. In particular, α3-K138 (β7 Sheet) forms two cation-π interactions (red lines), whereas β4-E105 (β4-β5 loop) and β4Y109 (β5 Sheet) form H-bonds (thin white dotted lines) with 1. The complete list of residues interacting at each site is summarized in Table 4. The interacting residues (as ball and sticks) are labeled by their subunit, residue one-letter code, and amino acid sequence number, and they are colored by atoms, including carbon (green), nitrogen (blue), oxygen (red), and hydrogen (white). The molecular docking results on the hα3β4 model indicated that the inhibitory activity of 1 is mediated by two different noncompetitive binding sites: a high-affinity luminal site and a low-affinity vestibular site. The luminal site is located close to the cytoplasmic side of the ion channel, between positions 2′ and -4′. In this site, only aristoteline forms a strong H-bond with the α3-E240 side chain (intermediate ring). Interestingly, the site for 1 is located apart from those previously found within the hα3β4 AChR ion channel, suggesting singular interactions with Aristotelia alkaloids. In general, ligands with antidepressant activity such as tricyclic antidepressants (e.g., imipramine),16 selective serotonin reuptake inhibitors (e.g., fluoxetine),18 N,6-dimethyltricyclo [5.2.1.02,6]decan-2-amine enantiomers,24 and mecamylamine, 16 as well as antiaddictive compounds such as coronaridine congeners (e.g., 18-methoxycoronaridine),19,20 interact with sites spread out between the middle and upper portions of the ion channel (i.e., between positions 6′ and 20′). On the other hand, the vestibular site is located within the ECD, behind the noncanonical β2+/α4- interface. The cooperative mechanism of inhibition (i.e., nH > 1) for 1 and E DOI: 10.1021/acs.jnatprod.9b00314 J. Nat. Prod. XXXX, XXX, XXX−XXX Journal of Natural Products Article Table 4. Residues Involved in the Docking of 1 to the Respective Luminal and Vestibular Site at the h(α3)3(β4)2 AChRa residues (position) [interaction] site receptor domain α3 subunit β4 subunit luminal (high-affinity) M2 C238 (-4′; currently M1-M2 loop) G239 (-3′) [1 MHB] E240 (-2′) [4 SHB] C239 (-4′; currently M1-M2 loop) G240 (-3′) [1 MHB] E241 (-2′) M243 (1′) T244 (2′) vestibular (low-affinity) ECD β5-β6 loop T243 (2′) P112 A113 Q114 D134 D135 K136 T137 K138 [cation-π] β6 sheet β6-β7 loop (Cys-loop) β7 sheet (Cys-loop) NTH-β1 sheet loop β3-β4 loop W122 N21 A86 K87 R88 I89 W90 L91 D93 V95 Y104 E105 [2 SHB] Y109 [1SHB] W153 T154 β4 sheet β4-β5 loop β5 sheet β7-β8 loop a MHB: Main chain donor hydrogen bond. SHB: Side chain donor hydrogen bond. NTH: N terminal helix. calculated LogP cutoff (<1.2) might suggest that this mechanism is limited for 3, which has a relatively lower LogP value (0.60). Although each alkaloid forms several weak H-bonds, only 3 has a relatively high PAS value, producing the observed PAS cutoff at >17 Å2. This dichotomy might be due to the fact that the aromatic rings present in these alkaloids are not considered for the calculations of the PAS values.26 Thus, the results support the view that ligand lipophilicity and the strong H-bond found only for 1 are the most important structural features for ligand binding. This is the first time showing a noncompetitive inhibitory, but not agonistic, activity of alkaloids purified from A. chilensis on human AChRs, with higher selectivity for hα3β4 AChRs. α3β4 AChR inhibition produces antiaddictive effects to many different drugs of abuse, including nicotine.27 In the particular case of bupropion, which inhibits α3β4 AChRs with relatively higher selectivity,21 this antidepressant is also used to quit smoking.28 In this regard, 1 could be used as a molecular scaffold for the development of novel noncompetitive antagonists with higher potency and higher selectivity for α3β4-containing AChRs, and consequently better therapeutic profile for the treatment of both depression and drug addiction. 2 supports the existence of more than one noncompetitive site in the hα3β4 AChR. Here, 1 could possibly interfere with the subunit rotation and twisting movements known to be elicited upon agonist binding,25 supporting inhibitory mechanisms different to ion channel blockade. Based on the structural and physicochemical similarity between 1 and bupropion (Table 2; Figure 1B−D), additional docking experiments were performed in the hα3β4 AChR model to determine whether the antidepressant overlaps the luminal site for 1. Since our hα3β4 AChR model is based on the crystal structure of hα4β2 AChRs in the desensitized state,12 our docking results likely reflect the interactions in the desensitized hα3β4 AChR. The results indicated that bupropion interacts with the luminal site but with lower affinity than that for 1 (TBEs: −76 and −92 kcal/mol, respectively), in agreement with the higher inhibitory potency for 1 (0.40 ± 0.20 μM) compared to that for the antidepressant (2.30 ± 0.40 μM).21 To determine what structural feature is important for ligand interaction with the luminal site, structure−activity relationship studies were performed. The results indicated that both LogP (lipophilicity) and PAS values correlate, although in opposite fashion, with the inhibitory potency of 1, 2, and bupropion, but not with 3. To explain these results, the following scenario is considered: the ligand interacts first with a hydrophobic environment (e.g., upper portion of the ion channel) before reaching its luminal site (i.e., cytoplasmic side). Previous studies on the hα3β4 AChR showed that bupropion in fact interacts with the upper portion of the ion channel.21 The ■ EXPERIMENTAL SECTION General Experimental Procedures. Optical rotations were recorded on a JASCO P-200 polarimeter (Tokyo, Japan). FTIR spectra were measured on a Nicolet 6700 from Thermo Electron Corporation with the ATR-unit Smart Performer. Melting points were F DOI: 10.1021/acs.jnatprod.9b00314 J. Nat. Prod. XXXX, XXX, XXX−XXX Journal of Natural Products Article determined on a Melting Point SMP10 (Stuart) uncorrected. The 1Hand 13C NMR spectra were recorded in MeOD or CDCl3 solution in 5 mm tubes at RT on a Bruker Avance III spectrometer (Bruker Biospin GmbH, Rheinstetten, Germany) at 600.13 (1H) and 150.61 (13C) MHz, with the deuterium signal of the solvent as the lock and TMS (for 1H) or the solvent (for 13C) as internal standard. All spectra (1H, 13C, gs-H,H−COSY, edited HSQC, and gs-HMBC) were acquired and processed with the standard Bruker software. (±)-Epibatidine hydrochloride was obtained from Tocris Bioscience (Ellisville, MO, U.S.A.). Fetal bovine serum (FBS) and trypsin/EDTA were purchased form Thermo Fisher Scientific (Paisley, U.K.). Ham’s F-12 Nutrient Mixture was obtained from Invitrogen (Paisley, U.K.). Fluo-4 was obtained from Molecular Probes (Eugene, OR, U.S.A.). Probenecid was purchased from SigmaAldrich Chemie GmbH (Buchs, Switzerland). Silica gel (200−300 mesh) was obtained from Merck (Santiago, Chile). Solvents used in this study were distilled prior to use and dried over appropriate drying agents. Salts were of analytical grade. Plant Material. Fresh leaves of A. chilensis (20 kg) were collected on Concepción, Chile (S 36° 50’ 01.51’’ W 73° 01’ 53.75’’), in December 2014. A voucher specimen (code Ach 2014-20) has been deposited at the herbal collection of Department of Chemistry, School of Engineering and Sciences, Universidad de La Frontera, Chile. Extraction and Isolation of Alkaloids from Aristotelia chilensis. The dried and milled plant material from A. chilensis (8.5 kg) was extracted in acid/water (1:3 v/v) (30 L, pH 3, HCl) during 3 days at room temperature (RT) and filtered. The aqueous layer was extracted with ethyl acetate (EtOAc) (3 × 10 L), and the organic layer was evaporated under reduced pressure at 45 °C giving a gummy residue (extract H+: 52 g). The acid water was alkalinized (pH 11, NaHCO3−NaOH diluted) and subsequently extracted with EtOAc (3 × 15 L). The organic layer was evaporated under reduced pressure at 45 °C giving a gummy residue (extract OH−: 87 g). The crude OH− extract (total alkaloid fraction, 87 g) was chromatographed using a silica gel column (200−300 mesh) and increased solvent polarity (from hexane 100% to EtOAc 100%), giving fractions B1−B9. Fractions B1−B4 contained no alkaloids, only fatty acids and chlorophylls. The fraction B6 (5.2 g) was separated by silica gel chromatography using hexane/EtOAc (1:1 v/v), giving 3 (10 mg, colorless crystals, 0.00012% yield). The fraction B7 (6.6 g) was applied to a Sephadex LH-20 column and eluted with EtOAc, giving 1 (210 mg, colorless crystals, 0.0025% yield). The fraction B8 (2.1 g) was also applied to the same column an eluted with EtOAc, but it was further separated by silica gel chromatography using EtOAc, giving 2 (130 mg, yellow oil, 0.0015% yield). Ca2+ Influx Measurements on Cell Lines Each Expressing the hα4β2, hα3β4, or hα7 AChR Subtype. Ca2+ influx measurements were performed on HEK293-hα4β2, HEK293-hα3β4, and GH3-hα7 cells as previously described.16,18−23 Briefly, 5 × 104 cells per well were seeded 72 h prior to the experiment on black 96well plates (Costar, New York, U.S.A.) and incubated at 37 °C in a humidified atmosphere (5% CO2/95% air). Under these conditions, the majority of expressed hα4β2 and hα3β4 AChRs have the (αx)3(βx)2 stoichiometry.21,22 16−24 h before the experiment, the medium was changed to 1% FBS in HEPES-buffered salt solution (HBSS) (130 mM NaCl, 5.4 mM KCl, 2 mM CaCl2, 0.8 mM MgSO4, 0.9 mM NaH2PO4, 25 mM glucose, 20 mM Hepes, pH 7.4). On the day of the experiment, the medium was removed by flicking the plates, and it was replaced with 100 μL of HBSS/1% FBS containing 2 mM Fluo-4 in the presence of 2.5 mM probenecid. The cells were then incubated at 37 °C in a humidified atmosphere (5% CO2/95% air) for 1 h. To determine the antagonistic activity of the Aristotelia alkaloids (Figure 1A), plates were flicked to remove excess of Fluo-4, washed twice with HBSS/1% FBS, refilled with 100 μL of HBSS containing the ligand under study, and incubated for 5 min. Plates were finally placed in the cell plate stage of the fluorimetric imaging plate reader (Molecular Devices, Sunnyvale, CA, U.S.A.), and (±)-epibatidine (0.1 μM for hα4β2 and hα3β4 AChRs or 1.0 μM for hα7 AChRs) was added from the agonist plate to the cell plate using the 96-tip pipettor simultaneously to fluorescence recordings for a total length of 78 s. To determine the agonistic activity of either alkaloid or (±)-epibatidine, each natural compound was added to the cell plate, and the fluorescence was recorded for 78 s. A baseline consisting of 5 measurements of 0.4 s each was previously recorded. The excitation and emission wavelengths are 488 and 510 nm, at 1 W, and a CCD camera opening of 0.4 s. The concentration−response data were curve-fitted by nonlinear least-squares analysis using the Prism software (GraphPad Software, San Diego, CA, U.S.A.). Theoretical Calculations of Physicochemical Parameters for Aristotelia Alkaloids. To determine the physicochemical properties of Aristotelia alkaloids (Figure 1A), several parameters were calculated using Accelrys Discovery Studio 2.5., including LogP (i.e., liophilicity), polar surface Area (PSA), and molecular volume. Molecular Docking of Aristotelia Alkaloids to h(α3)3(β4)2, h(α4)3(β2)2, and hα7 AChRs. The h(α3)3(β4)2, h(α4)3(β2)2, and hα7 AChR models were built using the X-ray structure (PDB ID: 5KXI) of the h(α4)2(β2)3 AChR at 3.9 Å resolution,12 as previously described.21,22 Since the α4 subunit of this X-ray structure has two more amino acids at the beginning of the mature peptide, the sequence numbering is different from that previously used (NP_000735 at www.ncbi.nlm.nih.gov). Since the authors deleted almost entirely the M3-M4 cytoplasmic loop to be able to crystallize the receptor, M4 is also numbered differently as previously. In both cases, we kept the notation of the X-ray structure. Each alkaloid (Figure 1A) was first modeled using VEGA ZZ. Minimization and partial charge calculations were performed using the MOPAC program as implemented in VEGA ZZ and employing the semiempirical AM1 method. Each molecule was subsequently docked into the whole h(α3)3(β4)2, h(α3)3(β4)2, and hα7 AChR model using AutoDock Vina as previously described.21,22 The parameters used were: exhaustiveness = 570 (the maximum value allowed by our computational system), and maximum number of modes = 20. To achieve dockings in few minutes time regime, no flexible residues were allowed in the receptor model. The program gives clusters of superposed conformations from the 20 lowest energy binding poses. Molecular Dynamics Simulations. To determine the stability of each pose within its predicted docking site, 20 ns molecular dynamics (MD) simulations were performed using NAMD and CHARMM force field, and VEGA ZZ as interface as previously described.21,22 To estimate the root-mean-square deviation (RMSD) with respect to the initial structure, the following equation was used:20,21 N RMSD = ∑i = 1 ∂ i2 (1) N where N is the number of atoms from the ligand and ∂ is the distance between the corresponding ligand atoms obtained at each step and the starting conformation. The conformations during the simulation were extracted every 10 ps from the simulation trajectory of 20 ns total time by using VEGA ZZ. Only poses with a variance (VAR) RMSD value below 1 during the last third of the MD were used in this work. Theoretical binding energies (TBE), measured from the individual poses at the end of every MD, were calculated using molecular mechanics as follows:21,22 TBE = CENERGY + L ENTROPY − (L ENERGY + R ENERGY) (2) where C is the complex between the ligand (L) and the receptor (R). The TBE values are estimations used only for comparative purposes among ligands and are not intended to represent absolute binding energies. Larger negative TBE values indicate higher theoretical binding affinities (TBAs). G DOI: 10.1021/acs.jnatprod.9b00314 J. Nat. Prod. XXXX, XXX, XXX−XXX Journal of Natural Products ■ Article (21) Arias, H. R.; Feuerbach, D.; Ortells, M. Neurotransmitter 2018, 5, No. e1631. (22) Arias, H. R.; Feuerbach, D.; Schmidt, B.; Heydenreich, M.; Paz, C.; Ortells, M. O. J. Nat. Prod. 2018, 81, 811−817. (23) Arias, H. R.; Feuerbach, D.; Ortells, M. O. Neurochem. Int. 2016, 100, 67−77. (24) Arias, H. R.; Feuerbach, D.; Targowska-Duda, K. M.; Jozwiak, K. J. Biochem. Mol. Biol. Res. 2015, 1, 19−24. (25) Martin, N. E.; Malik, S.; Calimet, N.; Changeux, J. P.; Cecchini, M. PLoS Comput. Biol. 2017, 13, No. e1005784. (26) Caron, G.; Ermondi, G. Future Med. Chem. 2016, 8, 2013− 2016. (27) Trott, O.; Olson, A. J. J. Comput. Chem. 2009, 31, 455−461. (28) Arias, H. R.; Biała, G.; Słomka, M. K.; Targowska-duda, K.; Biala, G.; Kruk-Slomka, M. Recept. Clin. Investig. 2014, 1, 30−45. ASSOCIATED CONTENT S Supporting Information * The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jnatprod.9b00314. 1 H NMR data for alkaloids 1, 2, and 3 is provided in Table 1S; 13C NMR data for alkaloids 1, 2, and 3 is given in Table 2S (PDF) ■ AUTHOR INFORMATION Corresponding Author *E-mail for C.P.: cristian.paz@ufrontera.cl. Tel: +56-452592825. ORCID Cristian Paz: 0000-0002-4668-4984 Notes The authors declare no competing financial interest. ACKNOWLEDGMENTS This study was supported by the grant REDI170107 and FONDECYT 11181076 from the Chilean government (PI: C.P.; Co-PI: H.R.A.). ■ ■ REFERENCES (1) Ortells, M. O. Neurotransmitter 2016, 3, No. e1273. (2) Dineley, K. T.; Pandya, A. A.; Yakel, J. L. N. Trends Pharmacol. Sci. 2015, 36, 96−108. (3) Kudryavtsev, D.; Shelukhina, I.; Vulfius, C.; Makarieva, T.; Stonik, V.; Zhmak, M.; Ivanov, I.; Kasheverov, I.; Utkin, Y.; Tsetlin, V. Toxins 2015, 7, 1683−1701. (4) Arias, H. R. J. Thermodyn. Catal. 2012, 3, 116. (5) Fuenzalida, M.; Pérez, M. A.; Arias, H. R. Curr. Pharm. Des. 2016, 22, 2004−2014. (6) Bhakuni, D. S.; Bittner, M.; Marticorena, C.; Silva, M.; Weldt, F.; Hoeneisen, M.; Hartwell, J. L. Lloydia 1976, 39, 225−243. (7) Delporte, C. Boletiń Latinoamericano y del Caribe de Plantas Medicinales y Aromáticas 2007, 5, 136. (8) Cespedes, C.; El-Hafidi, M.; Pavon, N.; Alarcon, J. Food Chem. 2008, 107, 820−829. (9) Céspedes, C.; Alarcón, J.; Gavila, G.; Nieto, A. Boletı ́n Latinoamericano y del Caribe de Plantas Medicinales y Aromáticas 2010, 9, 127−135. (10) Paz, C.; Becerra, J.; Silva, M.; Freire, E.; Baggio, R. Acta Crystallogr., Sect. C: Cryst. Struct. Commun. 2013, C69, 1509−1512. (11) Paz, C.; Becerra, J.; Silva, M.; Cabrera-Pardo, J.; Burgos, V.; Heydenreich, M.; Schmidt, B. Rec. Nat. Prod. 2016, 10, 68−73. (12) Morales-Perez, C. L.; Noviello, C. M.; Hibbs, R. E. Nature 2016, 538, 411−415. (13) Stoermer, D.; Heathcock, C. J. Org. Chem. 1993, 58, 564−568. (14) Zabel, V.; Watson, W. H.; Bittner, M.; Silva, M. J. Chem. Soc., Perkin Trans. 1980, 1, 2842−2844. (15) Watson, W. H.; Zabel, V.; Silva, M.; Bittner, M. Cryst. Struct. Commun. 1982, 11, 141−146. (16) Arias, H. R.; Targowska-Duda, K. M.; Feuerbach, D.; Sullivan, C. J.; Maciejewski, R.; Jozwiak, K. Neurochem. Int. 2010, 56, 642−649. (17) Toll, L.; Zaveri, N. T.; Polgar, W. E.; Jiang, F.; Khroyan, T. V.; Zhou, W.; Xie, X.; Stauber, G. B.; Costello, M. R.; Leslie, F. M. Neuropsychopharmacology 2012, 37, 1367−1376. (18) Arias, H. R.; Feuerbach, D.; Targowska-Duda, K. M.; Russell, M. M.; Jozwiak, K. Biochemistry 2010, 49, 5734−5742. (19) Arias, H. R.; Rosenberg, A.; Feuerbach, D.; Targowska-Duda, K. M.; Yuan, X. J.; Jozwiak, K.; Moaddel, R.; Wainer, I. W. Int. J. Biochem. Cell Biol. 2010, 42, 1525−1535. (20) Arias, H. R.; Feuerbach, D.; Targowska-Duda, K. M.; Jozwiak, K. Int. J. Biochem. Cell Biol. 2015, 65, 81−90. H DOI: 10.1021/acs.jnatprod.9b00314 J. Nat. Prod. XXXX, XXX, XXX−XXX