Synthesis of Pseudellone Analogs and Characterization as Novel T-type Calcium Channel Blockers

T-type calcium channel (CaV3.x) blockers are receiving increasing attention as potential therapeutics for the treatment of pathophysiological disorders and diseases, including absence epilepsy, Parkinson’s disease (PD), hypertension, cardiovascular diseases, cancers, and pain. However, few clinically approved CaV3.x blockers are available, and selective pharmacological tools are needed to further unravel the roles of individual CaV3.x subtypes. In this work, through an efficient synthetic route to the marine fungal product pseudellone C, we obtained bisindole alkaloid analogs of pseudellone C with a modified tryptophan moiety and identified two CaV3.2 (2, IC50 = 18.24 µM; 3, IC50 = 6.59 µM) and CaV3.3 (2, IC50 = 7.71 µM; 3, IC50 = 3.81 µM) selective blockers using a FLIPR cell-based assay measuring CaV3.x window currents. Further characterization by whole-cell patch-clamp revealed a preferential block of CaV3.1 activated current (2, IC50 = 5.60 µM; 3, IC50 = 9.91 µM), suggesting their state-dependent block is subtype specific.


Introduction
Alkaloids produced by marine animals are mostly potent cytotoxins evolved for defense [1]. Marine indole alkaloids have been widely explored for their therapeutic potentials, providing potential new drug leads for the treatment of a wide range of diseases including cancer, neurological disorders, and parasitic infections [2,3]. However, the underlying pharmacological targets and mechanisms of the bioactive indole alkaloids remain largely undefined. Triptans, a family of tryptamine-based drugs used for the abortive treatment of migraine headaches, have been well established and characterized as selective agonists of 5-HT 1 B and 5-HT 1 D serotonin receptors [4]. As a result, pharmacological studies on indole alkaloids mainly focus on the discovery of novel serotonin receptor agonists [5]. Compounds with an indole moiety have been explored for their activities on voltage-gated ion channels previously, and a series of tremorgenic indole alkaloids have been reported to potently block potassium channels in smooth muscle [6]. In a more recent study, a novel aryl indole compound was identified as a potent and selective blocker of N-type Ca V 2.2, and showed robust in vivo efficacy in inflammatory and neuropathic rat pain models [7].
In 2015, Liu et al. isolated a tryptophan-derived bisindole alkaloid pseudellone C (1) from the marine fungus Pseudallescheria ellipsoidea along with two other epipolythiopiperazine alkaloids [8]. Following this discovery, in 2017 Sathieshkumar et al. reported the total synthesis of pseudellone C (44% overall yield) [9], although the biological activity of these natural products remains unstudied. Here, we describe an efficient total synthesis of pseudellone C and several new bisindole alkaloid analogs. To investigate their pharmacological potential, we explored their activity in voltage-gated calcium channels (VGCCs) using FLIPR cell-based assays, and further characterized two potent low voltage-activated (LVA) T-type calcium channel (Ca V 3.x) blockers by whole-cell patch-clamp using an automated electrophysiology platform, QPatch 16 X. For the first time, we identified two "pseudellone" bisindole alkaloids as potent and selective Ca V 3.x blockers, demonstrating their potential as leads for the development of new analgesic and antiepileptic agents.

Evaluation of VGCC Activities of the Synthetic Compounds Using FLIPR Cell-Based Assays
Pseudellone C (1), along with its bisindole alkaloid analogs 2, 3, 3a, and its bisindole substrate 4, were evaluated for activity on VGCCs using FLIPR cell-based assays (see Table 1). The 2,2-bis (3,3 -indolyl) propionic acid 4 and the bisindole alkaloid 3a showed poor inhibition of all the Ca 2+ responses. In contrast, methyl-L-tryptophan-analog 3 showed good potency and selectivity for Ca V 3.3 currents measured in T-type window current assays with an IC 50 value of 3.81 ± 1.08 µM (n = 3), which was >8-fold better than its potency at high voltage-activated (HVA) Ca V s: 3 also potently blocked Ca V 3.2 responses with an IC 50 value of 6.59 ± 0.66 µM (n = 3). Comparatively, deoxy pseudellone C (2), which had tryptophan moiety stabilized with a methyl amide group, had a 2-fold reduced potency for Ca V 3.3 window currents with an IC 50 value of 7.71 ± 0.23 µM (n = 3) and a~3-fold reduced potency for Ca V 3.2 window currents with an IC 50 value of 18.24 ± 0.49 µM (n = 3), compared to 3. The natural product pseudellone C (1), which was generated by oxidation of the methylene group of the tryptophan moiety of deoxy pseudellone C (2), had >3-fold further reduced potency for Ca V 3.3 compared to 2. The fluorescent Ca 2+ responses before and after addition of compounds 2 and 3, and their representative concentration response curves, are presented in Figure 1 (2) and Figure 2 (3), respectively.  Interestingly, as indicated in Figure 2A,B and Figure 3, compound 3 showed potent partial blocking of the Ca V 3.2 window current. Also, pseudellone C is a weak calcium channel blocker that showed partial blocking of Ca V 3.1 and Ca V 2.2. Partial blocking of Ca V 3.x has been indicated to be useful for the treatment of epileptic behaviors by reducing synchronized oscillations of thalamocortical neurons [13,14]. Currently, compound 3 analogs are being designed for potential in rodent models of absence epilepsy and pain. Here, 1 produced 63.56% ± 2.80% (n = 3) inhibition for Ca V 3.1 and 39.26% ± 9.18% (n = 3) inhibition for N-type Ca V 2.2, and 3 produced 75.47% ± 2.29% (n = 3) inhibition for Ca V 3.1 and 56.91% ± 7.74% (n = 3) inhibition for Ca V 3.2.

Electrophysiological Characterization of the Selective Ca V 3.x Blockers in QPatch Assays
We also examined the effects of compounds 2 and 3 on the Ca V 3.x by whole-cell patch-clamp using the automated electrophysiology platform QPatch 16 X (Figures 4-6). Surprisingly, both compounds showed preferential blocking of Ca V 3.1 currents (2: IC 50 = 5.60 ± 0.26 µM (n = 5); 3: IC 50 = 9.91 ± 3.00 µM (n = 6)). Unlike the full inhibition by 2 and 3 of Ca V 3.2 and Ca V 3.3 whole-cell currents (E max value ≥ 95), 2 and 3 only showed partial block of the Ca V 3.1 current, with an E max value of 70.31 ± 2.13 (n = 6) and 82.84 ± 2.55 (n = 5), respectively. Although 2 and 3 did not show comparably good potency for the Ca V 3.3 peak current compared to their effects on the Ca V 3.3 window current, low concentrations (200 nM) of 2 and 3 both showed weak inhibition (<10%) of the Ca V 3.3 peak current.
It is noteworthy that both 2 and 3 had a significant voltage-dependent effect on steady-state inactivation of Ca V 3.2, shifting half-maximal inactivation V 50 to hyperpolarized potentials by~6 mV and 11 mV, respectively ( Figure 7). On average, control V 50 (−65.40 ± 1.91 mV, n = 9) was shifted to −76.36 ± 1.56 mV (n = 5, p < 0.01) in the presence of compound 3, whereas in the presence of 2 the V 50 was shifted to −72.40 ± 1.83 mV (n = 4, p < 0.05). In contrast, the voltage-dependence of steady-state inactivation for Ca V 3.1 and Ca V 3.3 was not significantly shifted in the presence of 2 and 3.

Discussion
In this study, we demonstrated that two pseudellone C-derived bisindole alkaloids, 2 and 3, are potent and selective Ca V 3.x current blockers that showed preferential inhibition of Ca V 3.3 window currents measured in FLIPR assays and Ca V 3.1 whole-cell currents measured in QPatch assays. The FLIPR window current assay applied to Ca V 3.x was permissive of a window current resulting from incomplete inactivation [15]. The specific Ca V 3.x window current made Ca V 3.x a privileged gate for the entry of extracellular Ca 2+ during secretion, neurotransmission, and cell proliferation [16,17]. Therefore, the inhibition of window currents in Ca V 3.x should now be targeted for the development of analgesic, antiepileptic, and anticancer drugs. Novel Ca V 3.x blockers that showed prominent experimental analgesic and antiepileptic efficacy [14,15,18] such as TTA-A2, Z941, and Z944, all showed better potency for T-type window currents than the activated current generated from the channel-closed state [14,15].
As mentioned above, tryptophan is the precursor for a wide range of pharmacologically important indole alkaloids [11,12]. Tryptophan-derived indole alkaloids have been receiving wide attention in drug discovery for their promising therapeutic potentials, represented by the powerful chemotherapy drugs vinblastine and vincristine [19][20][21] and the antipsychotic and antihypertensive drug reserpine [22][23][24]. In short, the marine fungal natural product pseudellone C was produced via oxidation of the intermediate 2, which in turn was obtained from amidation of the carboxylic acid 3a. The acid 3a was generated from hydrolysis of 3, which was in turn obtained from amide coupling of 4 and L-tryptophan methyl ester. The resulting suite of compounds have been evaluated for activity on VGCCs using FLIPR cell-based assays, and 2 and 3 were further explored for effects on Ca V 3.x whole-cell currents in QPatch assays.
Both compounds 2 and 3 inhibited Ca V 3.3 currents in the FLIPR window current assay, but to our surprise, they did not show comparable potency for Ca V 3.3 whole-cell currents in QPatch assays, where affinity is determined by interactions with the resting state of the channel. In contrast, 2 and 3 both showed >7-fold better potency for Ca V 3.1-activated currents than Ca V 3.1 window currents, whereas for Ca V 3.2, these two compounds did not show any preference for Ca V 3.2 window currents or Ca V 3.2-activated currents. Moreover, by measuring the voltage dependence of steady-state inactivation of Ca V 3.x, the addition of both 2 and 3 significantly shifted the half-maximal inactivation V 50 of Ca V 3.2 to more hyperpolarized potentials, whereas they did not significantly alter Ca V 3.1 and Ca V 3.3 channel inactivation. In brief, these two new Ca V 3.x blockers preferentially blocked either the window or activated current depending on the subtype and exclusively modulated channel inactivation kinetics for Ca V 3.2, which may help in differentiating subtype-specific properties in further pharmacological studies on T-type Ca V 3.x.
In summary, through a five-step total synthesis of marine fungal product pseudellone C, we achieved several bisindole alkaloid analogs of pseudellone C, two of which later revealed promising and selective inhibition of T-type Ca V 3.x currents. For the first time, naturally sourced bisindole alkaloids were characterized as Ca V 3.x blockers. Moreover, the exquisite differential inhibition of compounds 2 and 3 for the three T-type subtypes makes them useful pharmacological tools to probe the roles of individual Ca V 3.x subtypes, and they could be subjected to further structural modification for the development of analgesic, antiepileptic, and anticancer drug candidates through targeting Ca V 3.x.

General Procedure for Chemical Synthesis
All reagents were used as purchased from sigma Aldrich without further purification. Anhydrous solvents were used. Flash column chromatography was performed using ethylacetate (EtOAc) and petroleum ether as solvent. Preparative HPLC was used for the purification of the compounds using water and acetonitrile containing 0.01% TFA as solvent. 1 H and 13 C NMR spectra were recorded on Bruker (600 MHz) spectrometers. Data for 1 H NMR spectra are reported as chemical shift (ppm), multiplicity (s = singlet, d = doublet, t = triplet, q = quartet, m = multiplet, dd = doublet of doublets, dt = doublet of triplets, br s = broad singlet), and coupling constant (J in Hz). Electrospray ionization mass spectrometry (ESIMS) experiments were carried out on a LC/MSD (quadrupole) instrument in both positive and negative modes. High-resolution ESIMS spectra were obtained on a micrOTOF mass spectrometer by direct injection in MeCN at 3 µL/min using sodium formate clusters as an internal calibrant.

Synthesis of L-Tryptophan Methyl Ester
To a solution of L-tryptophan (0.204 g, 1 mmol) in anhydrous MeOH (5 mL), we added thionyl chloride (0.236 g, 2 equiv.) at 0 • C in dropwise and stirred for 12 h at room temperature until TLC revealed the disappearance of the starting material. The solvent was removed in vacuo and the residue was purified by flash column chromatography (45% ethyl acetate in petroleum ether) to give 5 (0.163 g, 75%) as a white solid. 1  A solution of 3a (0.15 g, 0.29 mmol) in DMF (3 mL) with DCC (200 mg, 1.5 mmol) and HOBT (135 mg, 1 mmol) was stirred at room temperature. After 15 min methyl amine (3 equiv., ethanolic solution) and TEA (404 mg, 4 mmol) was added, and the reaction mixture was stirred at 60 • C for 1 h. After the completion of the reaction, cold water was used to quench the reaction and the mixture was extracted with EtoAc (3 × 15 mL), and the combined EtOAc concentrated in vacuo with the residue purified by HPLC (phenomenex Luna C 18  A solution of 2 (35 mg, 0.07 mol) in 9:1 MeCN/H 2 O (1 mL) was treated with DDQ (32 mg, 0.14 mmol) in 6 portions over intervals of 10 min, and the resulting mixture was stirred at room temperature for 3 h, after which the residue was dissolved in EtOAc, washed with saturated NaCl (5 mL) and water (10 mL), dried over anhydrous Na 2 SO 4 , and concentrated in vacuo to give residue, which was purified by HPLC (phenomenex Luna C 18

HVA Calcium Channel FLIPR Assays
SH-SY5Y cells were seeded into 384-well black wall clear bottom plates at a density of 15,000 cells per well, resulting in 90%-95% confluence after 24 h. The media were then removed from the wells and replaced with 20 µL of 10% calcium 4 dye (Molecular Devices) in physiological salt solution (PSS) (containing 5.9 mM KCl, 1.4 mM MgCl 2 , 10 mM HEPES, 1.2 mM NaH 2 PO 4 , 5 mM NaHCO 3 , 140 mM NaCl, 11.5 mM glucose, and 1.8 mM CaCl 2 , pH 7.4) with 0.1% BSA. As reported [26], for N-type calcium channel FLIPR assays the cells were pre-incubated with 10 µM nifedipine added in the dye to ensure full inhibition of L-type calcium responses. For L-type calcium channel FLIPR assays, the cells were pre-incubated with 1 µM CVID added in the dye to ensure full inhibition of N-type calcium responses. Positive control on the first reagent plate contained 15 µL of PSS (0.1% BSA), whereas PSS (0.1% BSA) containing 1 µM CVID and 10 µM nifedipine (final concentration) was used as a negative control. The fluorescence readings were recorded and converted as described previously [26], and agonist containing 90 mM KCl + 5 mM CaCl 2 was used in the second addition.

Whole-Cell Patch-Clamp Electrophysiology
Whole-cell patch-clamp experiments were performed on an automated electrophysiology platform QPatch 16 X (Sophion Bioscience A/S, Ballerup, Denmark) in single-hole configuration using 16-channel planar patch chip QPlates (Sophion Bioscience A/S). The extracellular recording solution contained, in mM: TEACl 157, MgCl 2 0.5, CaCl 2 5, and HEPES 10; pH 7.4 adjusted with TEAOH; and osmolarity 320 mOsm. The intracellular pipette solution contained, in mM: CsF 140, EGTA 1, HEPES 10, and NaCl 10; pH 7.2 adjusted with CsOH; and osmolarity 325 mOsm. Compounds were diluted in extracellular recording solution with 0.1% BSA at the concentrations stated (DMSO ≤ 0.1%), and the effects of compounds were compared to the control (extracellular solution with 0.1% BSA) parameters within the same cell. Compounds' incubation time varied from two (for the highest concentration) to five (for the lowest concentration) minutes by applying the voltage protocol 10-30 times at 10 s intervals to ensure steady-state inhibition was achieved. The effects of compounds were obtained using 200 ms voltage steps to peak potential from a holding potential of −90 mV. The steady-state inactivation kinetics of Ca V 3.x currents were examined by applying steps lasting 1 s from −110 mV to the indicated voltages in 5-mV increments, followed by 60-ms test pulses to −20 mV. Data were fitted with a single Boltzmann distribution: I/I max = {1 + exp[V − V 50 ]/k} −1 , where V 50 is the half-availability voltage and k is the slope factor. Off-line data analysis was performed using QPatch Assay Software v5.6 (Sophion Bioscience A/S) and Excel 2013 (Microsoft Corporation, Redmond, WA, USA).

Data Analysis
Data were plotted and analyzed using GraphPad Prism v7.0 (GraphPad Software Inc., San Diego, CA, USA). A four-parameter logistic Hill equation with variable Hill coefficients was fitted to the data for concentration-response curves. Data are means ± SEM of n independent experiments. Statistical analysis was performed with a paired Student's t-test with statistical significance at p < 0.05.