Veraguamide E, a Marine Cyanobacterial Depsipeptide Targeting σ₂R/TMEM97: Chemical and Neurobiological Characterization

Abstract

The human sigma-2 receptor/transmembrane protein 97 (σ₂R/TMEM97) has been identified as a promising target to modulate neuronal excitability in chronic pain and address the unmet need for nonopioid therapeutics. We report the chemical and biological characterization of the cyclic depsipeptide, veraguamide E (Ver E), isolated from a Panamanian marine cyanobacterial collection, as a novel σ₂R/TMEM97 ligand and modulator of calcium in neurons. Ver E’s structure was confirmed using 1D and 2D-NMR, HRMS, and MS/MS molecular networking analyses. NMR titration and computational docking confirmed direct, saturable, and tight binding of Ver E to σ₂R/TMEM97. Functional calcium imaging in primary mouse sensory neurons revealed that Ver E increases intracellular Ca²⁺ levels without modulating store-operated calcium entry (SOCE). Multiwell microelectrode array experiments using human induced pluripotent stem cell (hiPSC) nociceptors showed that Ver E reduced neuronal activity at physiological temperatures, but not under heat-stress. Ver E exhibited no cytotoxicity in HEK293 cells, and immunocytochemistry confirmed it does not alter phosphorylated eIF2α (p-eIF2α) expression, indicating a mechanism distinct from integrated stress response modulators. Collectively, these findings position Ver E as a nontoxic σ₂R/TMEM97 ligand capable of selectively modulating neuronal excitability, creating a starting point for developing novel pain therapeutics.

Graphical Abstract

Millions of individuals worldwide suffer from chronic pain. Whether it presents as neuropathic, inflammatory, or postoperative pain, it is one of the most debilitating and pervasive health issues affecting communities and the leading reason individuals seek medical care.¹ Currently, a variety of pharmacological agents are available, including nonsteroidal anti-inflammatory drugs (NSAIDs), opioids, and adjunctive therapies. However, these therapeutics can cause adverse effects, including substance-use disorders, respiratory depression, and death.^2–4 The limitations of existing treatments underscore the need for new, safer analgesics, including from natural sources.

Marine organisms utilize specialized biosynthetic pathways to produce structurally diverse bioactive secondary metabolites for chemical communication and defense. These compounds have drawn significant scientific interest over the past two decades, leading to the exploration of their anticancer,⁵ antibacterial,⁶ antimalarial,^7,8 anti-inflammatory,⁹ and neuroactive^10,11 properties. Within this spectrum of marine bioactive natural products are the veraguamides, isolated from various marine cyanobacteria.^12,13 Several unique defining features of this class of depsipeptides include a conserved proline residue, multiple N-methylated amino acids, α-hydroxy acid, and 3-hydroxy-2-methyl-7-octynoic acid (HMOYA) a C8-polyketide-derived β-hydroxy acid moiety. The HMOYA tail has characteristic terminal groups that vary, appearing as an alkynyl bromide, alkyne, or vinyl group. Previous reports on the veraguamides isolated from cyanobacteria include 16 named members, which display diverse amino acid incorporations in cyclic and linear forms and have exhibited cytotoxic, antiparasitic, and neuroactive activities.^12,13 Their isolation was reported from collections of Symploca cf. hydnoides and Okeania sp. made in Guam and Panama, underscoring the broad geographic spread of the producing organisms and potentially important biological roles of the veraguamide family.^12–14 In this work, we examine veraguamide E (Ver E) from a collection of Okeania sp (based on phylogenetic analysis) made in Panama.¹⁴ Okeania is a tropical and subtropical marine cyanobacterial genus, abundant in shallow-water benthic habitats.¹⁵ Members of this genus, as well as related cyanobacteria such as Symploca and Oscillatoria, form benthic mats on a variety of underwater substrates and contribute to ecological interactions through chemical defenses.^16–18 While our studies did not directly explore the role of the veraguamides within the context of chemical ecology, their ability to modify calcium, a ubiquitous aspect of cell signaling and function, lead us to believe that there is an ecologically relevant purpose for their existence across geographically disperse species.¹⁴

We evaluated two elements of Ver E’s chemistry and biology. First, we evaluated the potential of Ver E to bind to an emerging new target for analgesic drug development, σ₂R/TMEM97.^19,20 This noncanonical transmembrane protein is associated with the endoplasmic reticulum and has been implicated in a diverse array of cellular functions including cholesterol and calcium homeostasis.^21–26 In the context of pain, modulators of σ₂R/TMEM97 reduce neuropathic pain^27,28 for extended periods of time beyond the predicted half-life of the compounds. σ₂R/TMEM97 is highly expressed in sensory neurons of the dorsal root ganglion (DRG) of mice and humans, suggesting that the analgesic effects are mediated, at least in part, peripherally.^27,28 This pharmacological profile makes σ₂R/TMEM97 an exciting target of exploration for pain therapeutics. We recently reported that a related analog, veraguamide O, showed affinity for σ₂R/TMEM97, suggesting that other members of the veraguamide family may bind to the protein.¹⁴

Second, we evaluated the impact of Ver E on sensory neuron function to understand the impact that it could have through this binding interaction. We explored calcium signaling of primary mouse DRG neurons during Ver E treatment. Calcium signaling is an important part of sensory neuron signaling known to be (1) modulated by σ₂R/TMEM97 and (2) critical to nociceptor function and pain.^29,30 Studies have demonstrated that modulations in calcium influx or SOCE can either potentiate or attenuate pain signals.³¹ Other marine lipopeptides such as barbamide have been reported to act on calcium channels and influence pathways relevant to neuronal excitability.¹⁰ Consequently, testing the impact of Ver E on calcium dynamics in DRG neurons is a promising approach for investigating this compound’s analgesic potential. Next, we evaluated the impact of Ver E on sensory neuron excitability using hiPSC derived nociceptors, an assay with improved translational relevance.³² Finally, we tested the hypothesis that specificity for Ver E’s effects by σ₂R/TMEM97 may lead to modulation of the integrated stress response (ISR), caused by ER stress and known to be associated with chronic pain^33,34 and σ₂R/TMEM97 modulation.²⁷

Overall, we found that Ver E is a structurally distinct compound compared to other identified σ₂R/TMEM97 ligands, while NMR binding and hiPSC studies clearly highlight its potential for functional activity in humans. Ver E provides several points of modification, making it a rich scaffold for both therapeutic development and tool compound generation. Moreover, the lack of cytotoxicity and ability to modulate both mouse and human nociceptors is consistent with potential translational utility. These data suggest that Ver E may mediate its biological effects through σ₂R/TMEM97 but lacks an effect on the ISR providing a unique cellular mechanism beyond those of other recently published σ₂R/TMEM97 modulators.^{19,21–24,27,28} This combined utilitarian chemical structure and biological profile provide impetus for continued exploration of this and related scaffolds for further development.

RESULTS AND DISCUSSION

Collection, Isolation, and Identification of Ver E.

A cyanobacterial sample (DUQ0008) identified as Okeania sp. was collected from the coastal waters of Isla Mina in the Las Perlas Archipelago, Panama (GPS coordinates: N 8°29.717′, W 78°59.947′).¹⁴ From this collection, 2.8 mg of a colorless oily compound was purified from a fraction eluting with 70% MeOH/30% EtOAc using multiple rounds of chromatography (Supplemental Figure S1–S2).

¹H NMR, ¹³C DEPT, and 2D spectra of the purified compound were highly consistent with veraguamide depsipeptides previously isolated from the same collection (Figure 1).¹⁴ HRMS (positive mode) confirmed the molecular formula C₃₉H₆₄N₄O₈, showing peaks at m/z 717.4766 [M + H]⁺ and 739.4581 [M + Na]⁺ (calculated for C₃₉H₆₅N₄O₈, 717.4802) (Figure S11). By comparing the spectra with literature reports, the compound was identified as Ver E (1).^12,13 Diagnostic NMR signals include nine methyl groups between 0.86–1.23 ppm, two sharp N-Me singlets at 2.92 and 2.99 ppm, an acetylene CH (H-8 in Hmoya tail) at 1.94 ppm, proline diastereotopic hydrogens at 3.61 and 3.83 ppm, six deshielded CHs at 4.03, 4.22, 4.74, 4.87, 4.91, and 4.91 ppm, and an NH peak at 6.25 ppm¹³ (Figures S3–S5). Two CHs at 4.91 seem closer to the CH at 4.87 in comparison to the literature spectra, but in the COSY spectrum, we could identify them clearly at separate ppm values, which have distinct COSY correlations (Figure S6). The connectivity and 2D structure of the compound was established through analysis of COSY, HSQC, HMBC, and TOCSY spectra (Figures S6–S10). Utilization of MS/MS classical molecular networking through the GNPS2 platform of the pure compound also provided evidence for it to be a veraguamide and specifically Ver E. The MS/MS fingerprint for the compound yielded a cosine similarity score of 0.4 compared to the Ver E MS/MS fragmentation pattern, with 15 matching peaks identified^35,36 (Figure S12). Additionally, the fragmentation pattern can be justified based on expected fragments and breaks to confidently confirm the planar structure and amino acid connectivity to be the same as Ver E (Figures S13 and S14). The compound networked with two other veraguamide analogs (Figure 1) with minor but distinct modifications of amino acid side chains, in the case of veraguamide D, and in amino acid side chains and HMOYA modification for veraguamide G. Interestingly, there is also connection into a network containing dudawalamide A,³⁷ another marine cyanobacterial cyclic depsipeptide with a 2,2-dimethyl-3-hydroxy-7-octynoic acid (DHOYA) rather than HMOYA lipid tail.

Figure 1. Chemical structures and molecular network.

Chemical structures of veraguamide E (1), veraguamide C (2), veraguamide O (3), veraguamide D (4), veraguamide G (5), dudawalamide A (6), and GNPS2 Molecular Network of Ver E task ID: 0075011f9ad94b779bb735618e11c278.

To compare the 3D structure to the literature report for Ver E, we looked at specific rotation and chemical shift comparison of all proton and carbon signals. The measured specific rotation [α]²⁵_d – 50 (c 0.002, MeOH) for our isolated compound was in close agreement with the reported value of [α]²⁰_d – 56 (c 0.22, MeOH).¹³ Considering that the measured specific rotation is affected by concentration and temperature, these results strongly support the notion that the isolated Ver E is identical to previous report in terms of stereoconfiguration. Careful comparison of the ¹H NMR spectrum revealed only two signals with Δppm values greater than 0.06. One being a shift of 0.09 ppm in a terminal methyl of the HMPA, and the other one being a shift of 0.1 ppm of an alpha proton at the N-Me-Val. (Table S1, Figure S15). All carbon chemical shifts demonstrated <1.4 Δppm values when compared to the reported chemical shifts for Ver E, except for the methylene in the gamma position of the flexible HMOYA tail, which has a 2.7 ppm difference (Table S1, Figure S16). To examine the possibility of inversion of the α-stereocenter at the N-Me-Val residue, we conducted variable temperature NMR experiments (temperature increments from 15 °C to 40 °C) and show that this alpha proton at 4.03 experiences shifts based on temperature up to 0.035 ppm reducing our concern about this proton not exactly matching literature spectra, (Figure S17). Additionally, we obtained the spectra with addition of two droplets of CD₃OD into the CDCl₃ which resulted in significant changes in chemical shifts of the proton signals in the 3–5 ppm range (Figure S18). The variability of these protons to small changes in temperature and solvent reduced our dependence on them and increased our trust in the optical rotation and ¹³C NMR comparison. Thus, based on HR-MS/MS molecular networking analysis, a plethora of 1D and 2D NMR experiments, and specific rotation of our isolated material, we confidently dereplicated and identified our compound as Ver E.

Ver E Directly Binds σ₂R/TMEM97 at a Noncanonical Surface Site.

Due to the unique structure of Ver E along with its relative abundance in our collection and our recently published data showing significant radioligand competitive binding of veraguamides to σ₂R/TMEM97,^14,27,28,38 we wanted to test Ver E’s potential molecular binding partners. In our previous report, both the cyclic and linear depsipeptides veraguamide C and O showed significant binding affinity to σ₂R/TMEM97 as screened by the Psychoactive Drug Screening Program (PDSP).³⁹ While we were unable to test Ver E in this assay due to compound availability, we were able to experimentally determine its binding potential in σ₂R/TMEM97 using protein NMR-detected titration and computational docking experiments.

We tested the potential for direct interactions between Ver E and reconstituted human σ₂R/TMEM97 (Figures S19–21) using NMR-detected σ₂R-ΔER binding studies. ¹H–¹⁵N BEST-TROSY-NMR experiments were recorded at increasing Ver E concentrations. Global evaluation of the binding data by principal component analysis and discrete local chemical shift perturbation binding analysis showed saturable binding isotherms, indicating direct and specific Ver E binding to σ₂R/TMEM97 (Figure 2, Figure S20). The binding data were fit to a single-site binding model, yielding a global K_d of 0.3 ± 0.2 nmol %, confirming that σ₂R/TMEM97 in dodecylphosphocholine (DPC) micelles directly binds Ver E. A control binding experiment was performed, mimicking the vehicle concentrations. This control titration showed no nonspecific binding to σ₂R/TMEM97 (Figure S21), validating a direct interaction between Ver E and σ₂R/TMEM97. As an aside, using mole percent (mol %) units as a binding metric accounts for the preferential partitioning of σ₂R and Ver E into the lipophilic membrane-mimicking micellar environment. Ver E is computationally predicted to have a LogS of −6.8, which gives an estimated aqueous solubility of ~ 0.1 mg/L, and σ₂R is insoluble in an aqueous environment. As such, mol % is the expected unit, where the micelle is the effective solvent for both the receptor and the ligand. In our NMR binding study, the human σ₂R is reconstituted in dodecylphosphocholine (DPC) detergent micelles, meaning that ligand binding does not occur in a simple, homogeneous aqueous phase but instead in the lipophilic DPC micelle environment. As a result, the effective ligand concentration that drives binding depends on the amount of ligand and receptor in the DPC environment and not on the aqueous volume used in molarity units. Using units of molarity in this case does not give a quantitatively meaningful representation of the binding interaction. In quantitative membrane biophysical binding studies of membrane proteins and lipophilic ligands, this has become the standard as it is considered quantitatively accurate.^40–42

Figure 2. Human σ₂R/TMEM97 directly binds Ver E.

A) Superimposed ¹H–¹⁵N BEST-TROSY spectra of σ₂R/TMEM97 titration with Ver E at 37 °C. B) Nonlinear fitting of chemical shift perturbations to a single-site binding model reveals a global K_d of 0.3 ± 0.2 nmol % for Ver E binding to σ₂R/TMEM97.

Next, AlphaFold3 was employed to model the full-length σ₂R/TMEM97 receptor⁴³ (Figure 3). Consistent with previously published modeling exercises, the σ₂R/TMEM97 was predicted to conform to the same architecture as the published bovine σ₂R/TMEM97 receptor.⁴⁴ We employed both generative diffusion-based and energy-based methods to predict the binding pocket (Figure S22). Consistent with the protoypical σ₂R/TMEM97 PB28,⁴⁴ the predicted binding pocket is in the concave region of the σ₂R/TMEM97. Unlike PB28, however, we have found that Ver E is likely to reside outside of the recessed region of the receptor (Figure 3E, Figure 3H). Energy of binding as assessed using AutoDock-GPU predicts similar binding energy to existing ligands of σ₂R/TMEM97.

Figure 3. In silico prediction of binding pocket.

(A-C): 3D representations of the ligand conformers, generated using obgen in a Universal Force Field. (D-F) Most energetically favorable binding conformation against AlphaFold3 generated the σ₂ receptor. (G-I) Protein–Ligand Interaction Profiles of the ligand’s predicted binding pocket.

This is the first example of using NMR binding affinity on reconstituted human σ₂R/TMEM97. The results revealed specific, saturable binding with a global K_d of 0.3 ± 0.2 nmol %. The use of mol % accounts for the partitioning of Ver E between the DPC micelle environment and the bulk solution, providing a more accurate representation of the effective ligand concentration available for binding. The ¹H–¹⁵N BEST-TROSY spectra (Figure 2) also showed broadening and disappearance of some resonances, suggesting conformational exchange. Upon titration with Ver E, selected resonances exhibited dose-dependent chemical shift perturbation (Figure S20), confirming direct binding interaction with σ₂R/TMEM97. The control experiment (Figure S21) demonstrated minimal solvent-induced effects, reinforcing the specificity of Ver E binding. Future studies will aim to identify whether Ver E binds to the σ₂R/TMEM97 orthosteric site. These first-inclass NMR binding data establish that Ver E binds specifically and with high affinity to σ₂R-ΔER, highlighting its promise as a lead compound for developing novel therapeutics targeting σ₂R/TMEM97 for pain management. Both generative diffusion (Figure S22) and energy-based computational docking approaches (Figure 3) predicted binding in the entrance of the PB28⁴⁴ and cholesterol binding pocket. While largely in the same region, the cyclic structure of Ver E precludes its binding in the pocket itself (Figure 3E, Figure 3H), providing an opportunity for side-chain modification to engage the internal pocket for specific and prolonged binding. Similar “anchor-extension” remodeling of macrocycles has successfully deepened pocket engagement in other targets.⁴⁵

Ver E Elevates Intracellular Ca²⁺ without Modifying Store-Operated Entry in DRG Cells.

σ₂R/TMEM97 has been hypothesized to modulate calcium homeostasis and signaling which has obvious impacts on sensory neurons in the context of pain.^31,46 Additionally, marine-derived depsipeptides have been shown to influence neuronal signaling by targeting calcium channels and pathways.⁴⁷ We hypothesized that Ver E, having affinity for σ₂R/TMEM97, would modulate neuronal excitability and calcium dynamics in sensory neurons. To test this, we performed functional calcium imaging recordings on DRG from mice expressing the genetically encoded calcium indicator GCaMP6f. We measured both baseline calcium flux and the ability of Ver E to directly influence DRG excitability under various treatment conditions, including its effects on SOCE, a key pathway for maintaining intracellular calcium homeostasis.^48,49

To assess its impact on calcium homeostasis, 10 μM Ver E was applied directly to DRG neurons, and changes in calcium signaling were monitored during a SOCE activation protocol (Figure 4). In this experiment, the Sarcoendoplasmic Reticulum Calcium ATPase (SERCA) pump inhibitor thapsigargin was applied with or without Ver E in the absence of extracellular Ca²⁺. Under these conditions, thapsigargin induces the release of intracellular Ca²⁺ stores. Upon reintroduction of extracellular Ca²⁺, the SOCE response was measured. This approach enables evaluation of both SOCE buildup during thapsigargin treatment and the subsequent Ca²⁺ influx associated with homeostatic recovery. Ver E significantly increased DRG excitability during its coapplication with thapsigargin compared to thapsigargin alone (P < 0.01), indicating an acute effect on intracellular Ca²⁺ levels. However, no significant change was observed in the subsequent SOCE response, suggesting that Ver E’s mechanism of action does not directly target store-operated calcium pathways.

Figure 4. Effect of Ver E on calcium homeostasis in DRG neurons.

(A) Maximum fluorescence measured in DRGs exposed to vehicle (0.1% DMSO; mean = 0.01323) or Ver E (10 μM; mean = 0.04668). Ver E treatment significantly increased fluorescence compared to vehicle (unpaired two-tailed t test, **P < 0.01). Error bars recorded as standard error of mean (SEM) with n (number of cells) indicated in graph. (B) Maximum fluorescence observed during the SOCE response with vehicle (mean = 2.126) or Ver E (mean = 1.929). No significant difference was detected between groups. (C) Representative real-time calcium imaging traces for vehicle vs Ver E-treated DRGs. Gray bars indicate the period of exposure to each treatment. (D) Representative heat map images of DRG cells illustrating the direct calcium response (middle) and the SOCE response (right) under vehicle or Ver E treatment.

Our results position Ver E as a modulator of neuronal excitability with a distinct mechanism that sets it apart from other marine-derived compounds. The observed increase in intracellular Ca²⁺ levels is consistent with previous studies on cyanobacterial natural products⁵⁰ but occurred independently of SOCE. This suggests that Ver E may act through noncanonical pathways, possibly involving direct modulation of plasma membrane ion channels or stimulation of Ca²⁺ release from intracellular stores unrelated to the STIM1-Orai1 SOCE system.

Ver E Transiently Dampens Spontaneous Nociceptor Firing without Affecting Heat-Evoked TRPV1 Activity.

DRG neurons also play a key role in peripheral nociception and are known to exhibit both spontaneous and evoked firing. To model this activity in vitro, we utilized hiPSC-derived nociceptors cultured on multielectrode array (MEA) plates. By day 28, these neurons show spontaneous firing activity, reflecting baseline hyperexcitability characteristic of neuropathic pain conditions.³² Application of Ver E at 30 μM significantly reduced neuronal activity by approximately 60% (P < 0.05) within the first hour at 37 °C; however, this inhibitory effect was transient, with activity returning to baseline by the 24-h time point (Figure 5). In contrast, during transient heat ramp applications from 37 to 42 °C, Ver E did not produce a reduction in activity compared to vehicle-treated cells. These results indicate that Ver E does not block TRPV1 channels which normally open in response to noxious heat to depolarize neurons. In turn, the initial depolarization recruits voltage-gated sodium (Na_v), potassium (K_v), and calcium (Ca_v) channels, which together drive action potential firing in sensory neurons exposed to high temperatures.

Figure 5. Inhibitory effects of Ver E on sensory neuron activity in vitro.

A) Ver E significantly reduced the activity of hiPSC-derived nociceptors at a concentration of 30 μM at 37 °C. B) A nonsignificant decrease was observed at 42 °C. Data are presented as Mean ± SEM for each concentration point. *p < 0.05 indicates a significant difference compared to DMSO-treated cells at the indicated time point, as determined by two-way ANOVA with Dunnett’s post hoc test.

The transient suppression of spontaneous firing in hiPSC-derived nociceptors at physiological temperature, combined with a lack of effect during thermal stress, suggests that Ver E does not act on thermosensitive channels such as TRPV1. TRPV1 channels normally open in response to noxious heat to depolarize neurons.^51–54 Instead, Ver E’s selective inhibition of baseline excitability points to possible modulation of Na_v, K_v, and Ca_v channels, which are responsible for regulating resting membrane potential and the threshold for action potential generation. This level of selectivity is important in pain therapeutics, where broad inhibition of sensory pathways often results in undesirable side effects.

Ver E Does Not Alter p-eIF2α Levels, Leaving the Integrated Stress Response Intact in DRG Neurons.

The ISR is a conserved cellular pathway that regulates protein synthesis in response to stress and has been implicated in pain sensitization and neurodegenerative disease.^25,27,33 Prior work has shown that certain σ₂R/TMEM97 modulators can disrupt this pathway by reducing phosphorylation of the key ISR regulator eIF2α.²⁷ To assess whether Ver E affects ISR signaling, we performed immunocytochemistry for p-eIF2α in DRG neurons treated with 100 nM or 1000 nM Ver E. No observable differences in p-eIF2α levels were detected compared to vehicle-treated controls (Figure 6). In contrast, treatment with FEM-1689, a known σ₂R/TMEM97 modulator produced a marked reduction in p-eIF2α expression at 30 nM and 100 nM, consistent with published data.²⁷ These results suggest that Ver E does not affect ISR signaling through eIF2α phosphorylation under these conditions.

Figure 6. Impact of Ver E on the Integrated Stress Response.

A) Effect of Ver E (100 nM and 1,000 nM) on p-eIF2α expression in primary mouse DRGs. No significant difference was observed compared to vehicle. Representative immunocytochemistry image on right of primary mouse DRGs treated with Ver E (1000 nM), visualized for p-EIF2α (magenta), and peripherin (green). B) Effect of FEM-1689 (30 nM and 100 nM) on p-eIF2α expression in primary mouse DRGs. FEM-1689 significantly decreased the mean gray intensity compared to vehicle. Representative immunocytochemistry image on right of primary mouse DRGs treated with FEM-1689 (100 nM), visualized for p-EIF2α (magenta), and peripherin (green). Data are presented as mean ± SEM. Statistical significance was determined by one-way ANOVA with Tukey’s multiple comparisons (****P < 0.0001).

In a previous publication, we demonstrated that FEM-1689 drives analgesic responses in mice with neuropathic pain through σ₂R/TMEM97 and that this response is associated with inhibition of the ISR that is absent in σ₂R/TMEM97 knockout mice.²⁷ In the present manuscript, we reproduced this inhibition effect of FEM-1689, but Ver E failed to modulate p-eIF2a. Given that Ver E shows potential to reduce activity in mouse DRG (Figure 4) and hiPSC-derived nociceptors (Figure 5), both of which would be expected to reduce pain, it is possible that there is another antinociceptive mechanism of σ₂R/TMEM97 beyond ISR modulation. Alternatively, Ver E may have binding partners beyond σ₂R/TMEM97.

Ver E Shows No Detectable Cytotoxicity in HEK293 Cells Up to 30 μM.

Previous studies have reported that certain veraguamides exhibit varying degrees of cytotoxicity.¹² However, we recently showed that veraguamide C was nontoxic to HEK293 cells and exhibited only mild cytotoxicity in breast cancer cell lines.¹⁴ To evaluate Ver E, we tested concentrations ranging from 0.03 μM to 30 μM using HEK293 cells, with digitonin included as a positive cytotoxicity control. Ver E showed no evidence of cytotoxicity across the tested concentration range (Figure 7). These findings indicate that the LC₅₀ for Ver E exceeds 30 μM under these conditions, supporting its safety profile and distinguishing it from more toxic veraguamide analogs.

Figure 7. Cytotoxicity profile of Ver E compared to digitonin across a concentration range of 0.03–30 μM.

A) At 30 μM, digitonin caused 100% cell death, whereas Ver E showed no cytotoxicity at the same concentration. B) Based on the logarithmic nonlinear regression curve, the LD₅₀ of digitonin was determined to be 9.68 μM, while Ver E had no detectable LD₅₀, as it did not induce cell death at any tested concentration. A) and B) Each point represents a mean ± SEM ****p < 0.0001 indicates a significant difference compared to Ver E-treated cells, as determined by two-way ANOVA with Sidak’s post hoc test.

Ver E exhibited no cytotoxicity in HEK-293 cells, which is essential for establishing its safety profile and evaluating its potential as a therapeutic agent, particularly given that many marine-derived depsipeptides are known to exhibit neurotoxicity and cytotoxicity.⁵⁵ In summary, Ver E’s ability to modulate calcium dynamics in DRG neurons, its lack of effect on SOCE pathways, and its absence of cytotoxicity position it as a promising candidate for further investigation as a therapeutic agent targeting specific pain pathways.

Ver E Occupies a Distinct Chemical Space among σ₂R/TMEM97 Ligands.

To better understand how Ver E compares to known σ₂R/TMEM97 specific ligands, we performed similarity calculations between Ver E and a select group of previously described σ₂R/TMEM97 ligands^{19,27,56–70} (Figure S24) using an Overlap Analysis tool (InstantJChem, ChemAxon). The selection of σ₂R/TMEM97 ligands was based on manual inspection of compounds within a recently published review article¹⁹ and focused on selecting representative structures of as many different σ₂R/TMEM97 ligand chemotypes (scaffolds). The computational comparison produced Tanimoto scores that quantify the structural similarity of Ver E to each σ₂R ligand selected (Figure 8).⁷¹ Tanimoto scores range from 0 to 1 with 1 being identical and 0 indicating no shared structural elements. Work by Patterson et al.⁷² has been used to justify a Tanimoto value of 0.85 as highly likely to have similar or the same biological activity. Ver E holds the highest similarity with SM-21 (22)⁶² and histatin-1 (23)⁷⁰ (Table S2), with a median Tanimoto score of 0.32 (Max = 0.46, Min = 0.25) across all σ₂R/TMEM97 ligands, suggesting that Ver E is structurally distinct to established σ₂R/TMEM97 ligands. Thus, the predicted and direct binding of Ver E to σ₂R/TMEM97 coupled with the lower Tanimoto scores suggests that Ver E represents a structurally unique scaffold for this target.

Figure 8. Box plot of Tanimoto coefficients for σ₂R ligands compared to Ver E.

Interquartile range (IQR) of all coefficients is represented as a dotted red rectangle and red dashed lines representing Q1 (25th percentile), median, and Q3 (75th percentile).

To further compare Ver E to known σ₂R/TMEM97 ligands, we conducted predictions of physicochemical descriptors of Ver E and σ₂R/TMEM97 ligands using the DataWarrior platform⁷³ (Table S2) and performed Principal Component Analysis (PCA) (Figure 9, Figures S23–S25). Ver E has a peculiar physicochemical profile, characterized by a relatively high predicted drug-likeness score, accompanied by unusually high molecular weight, topological polar surface area, and sp3 atom profile (Table S2). When compared to σ₂R/TMEM97 ligands, Ver E describes its own chemical space. Thus, comparative multivariate analysis of predicted physiochemical descriptors together with Overlap Analysis (Tanimoto) strongly supports the notion that Ver E represents a structurally unique σ₂R/TMEM97 ligand chemotype.

Figure 9.

Principal component analysis (PCA) plot of 13 predicted physicochemical descriptors of Ver E and previously described σ₂R/TMEM97 ligands.

This combination of low Tanimoto score to other known ligands and high drug likeness score drive our enthusiasm for the veraguamide backbone as a scaffold. However, a comprehensive understanding of its interactions with σ₂R/TMEM97 is needed. We will continue to evaluate Ver E and related analogs with computational docking and NMR binding experiments to better understand interactions driving the affinity.

From a medicinal-chemistry standpoint, Ver E’s macrocyclic depsipeptide scaffold is intriguing. Macrocyclic peptides occupy an important chemical space between small molecules and biologics, this “natural products space” provides an interesting area for development as these macrocyclic structures seem to follow different rules for absorption and crossing of membranes from small molecules and biologics.⁷⁴ Ver E provides multiple points for side-chain or ring modifications that could enhance potency, selectivity or pharmacokinetics.

This study provides a biological and chemical characterization of Ver E, a marine-derived depsipeptide from Okeania sp.¹⁴ Ver E exhibited several promising pharmacological properties: (1) direct, saturable binding to σ₂R/TMEM97 with subnanomole percent affinity, (2) modulation of calcium homeostasis without affecting SOCE, (3) reduction of neuronal activity in hiPSC-derived nociceptors at physiological temperature, (4) no alteration of p-eIF2α expression in DRG neurons, and (5) no cytotoxic effects in HEK293 cells up to 30 μM. These findings support Ver E’s potential as a selective and nontoxic analgesic lead compound.

Limitations and Future Directions.

While our study provides a comprehensive initial characterization of Ver E and suggests a favorable safety profile, several important questions remain. First, the precise molecular mechanisms by which Ver E modulates Ca²⁺ signaling have yet to be elucidated, underscoring the need for additional receptor-binding and electrophysiological studies to clarify its target specificity. Second, although we document Ver E’s in vitro efficacy, evaluation of the potential in vivo therapeutic benefits requires validation in animal models of chronic pain (e.g., inflammatory, cancer, and neuropathic). Evaluating its effects on thermal and mechanical sensitivity in vivo will further establish the clinical relevance of the presented findings.

Additionally, the transient nature of Ver E’s inhibitory effect on nociceptors suggests that structural modifications might be necessary to prolong its therapeutic duration, potentially guided by structure–activity relationship (SAR) studies. Finally, it remains critical to assess whether Ver E has off-target activities on other ion channels or signaling pathways to ensure a comprehensive safety profile. By addressing these limitations in future work, we can better define Ver E’s therapeutic potential and optimize it for clinical applications.

EXPERIMENTAL SECTION

General Experimental Procedures.

The optical rotation was measured on a JASCO DIP-370 polarimeter. 1D and 2D NMR spectra were recorded in CDCl₃ (δC 77.2, δH 7.26) on a Bruker 600 (AV4 NEO) MHz spectrometer. The instrument was equipped with a 5 mm ¹H-optimized triple-channel CryoProbe for ¹H observation, with ¹³C/¹⁵N decoupling and ¹³C observation with ¹H decoupling. NMR data were analyzed with MestReNova (version 14.3.0) after processing the reference by solvent. NMR titration analysis was performed on a Bruker Avance III HD 850 MHz spectrometer equipped with a cryogenically cooled 5 mm TCI cryoprobe. High-resolution mass measurements were acquired on a SCIEX Q-TOF 5600 mass spectrometer with a dual-spray electrospray ionization source, connected to a Shimadzu DGU-20A 3R system. For MS/MS spectra, an IonSpray voltage floating 5500 V was used with a collision energy of 3500 V. The source temperature was maintained at 500 °C, with ion source gas 1 and 2 flows of 50 L/h and 55 L/h, respectively. MS/MS survey scans were run as a dependent acquisition, in which the 20 most intense ions in the MS spectrum were selected as MS/MS precursors. The threshold for MS/MS targeting was set at 100, with each targeted ion excluded from further targeting for 15 s. Semipreparative HPLC separations were performed on an Agilent 1260 Infinity II system with UV detection. Silica gel 230–400 mesh (Thermo Fisher) was used for column chromatography.

Cyanobacterial Collection and Isolation.

Sample Collection and Extraction.

A cyanobacterial sample (DUQ0008) was collected from the coastal waters of Isla Mina in the Las Perlas Archipelago, Panama (GPS coordinates: N 8°29.717′, W 78°59.947′) with collection permit (SC-PB-5-12) and exported under permit (SEX-P-44-12) both issued by the Autoridad Nacional de Ambiente (ANAM) of Panama. A description of this collection with phylogenetic analysis of the collected species was previously published.¹⁴ The sample containing Ver E(75 g, dry wt) was extracted using a CH₂Cl₂–MeOH (2:1) solvent mixture, yielding 3.3 g of extract.

The whole extract underwent normal-phase chromatography, using a stepwise gradient solvent system starting from 100% hexane, then gradually increasing polarity by adding EtOAc, and finally adding MeOH. This process yielded nine fractions (A–I). A portion of fraction G (15 mg), eluted with 70% MeOH, was subsequently purified by reversed-phase HPLC (Phenomenex, Kinetex 5 μm, 150 × 10 mm) with a linear gradient of MeCN–0.1% TFA in H₂O (60% MeCN for 5 min, 60–100% MeCN over 12 min, then 100% MeCN for 5 min), affording 2.8 mg of the target compound as a colorless oil (Figure S1, Figure S2).

Molecular Networking.

Molecular networks were generated using the online workflow at the GNPS2 web site (http://gnps2.org). The data were clustered with a parent mass tolerance of 2 Da and an MS/MS fragment ion tolerance of 0.4 Da. Edges in the network were retained if they had a cosine score above 0.4 and more than six matched peaks. The resultant consensus MS/MS spectra were searched against GNPS2’s spectral libraries. A cosine similarity score of 0.4 was obtained when comparing the MS/MS fragmentation pattern of the purified compound with that of Ver E, resulting in 15 matching peaks (Figure S6). In the dependent MS/MS acquisition, the 20 most intense ions in each MS survey scan were selected if they exceeded an intensity threshold of 100, after which they were excluded from reselection for 15 s. Data visualization for the molecular network was also conducted at the GNPS2 web site. A summary of the spectral data is in the SI and publicly available on GNPS2 using the Task ID number 0075011f9ad94b779b-b735618e11c278.

NMR Ligand Binding.

Human σ₂R Expression and Purification.

The gene encoding the human σ₂ receptor (σ₂R) was cloned into the pETSG vector, which includes a C-terminal 8 × His tag linked to thermostable GFP (TGP) followed by a thrombin cleavage site.⁷⁵ The σ₂R gene was obtained from DNASU⁷⁶ (plasmid # HsCD00514886) and pETSG was a gift from Prof. Dianfan Li (Addgene plasmid # 159418). The σ₂R construct was truncated at the ER retention signal (residue Lys168, σ₂R-ΔER) to improve expression. Whole plasmid sequencing (Plasmidsaurus) confirmed the pETSG-σ₂R-ΔER identity.

σ₂R/TMEM97 expression was optimized for the BL21 (DE3) E. coli strain. A 5 mL LB overnight starter culture was prepared from a single colony, supplemented with 0.05 mg/mL kanamycin (Gold Biotechnology), and incubated at 37 °C. The starter culture inoculated 1 L of M9 minimal media (12.8 g Na₂HPO₄·7H₂O, 3.0 g KH₂PO₄, 0.5 g NaCl, 1 g ¹⁵NH₄Cl, 20 mL of 20% w/v d-glucose, 10 mL 100× MEM vitamin solution, 1 mL 1 mM MgSO₄, 1 mL 0.1 mM CaCl₂, 500 μL of 1000× trace metal mixture). The culture was grown at 25 °C, and protein expression was induced at OD_{600 nm} of 0.5–0.6 with 1 mM isopropyl β-D-1-thiogalactopyranoside for 22 h. The resulting cells were harvested by centrifugation at 6000 × g (20 min, 4 °C).

The cell pellet was resuspended in 7 mL of lysis buffer (75 mM Tris-HCl (VWR), 300 mM NaCl (Fisher Scientific), 0.5 mM EDTA (Sigma-Aldrich), pH 7.8) per gram of pellet supplemented with 0.1 mg/mL lysozyme, 0.01 mg/mL DNase, 0.01 m/mL RNase, 10 μL of 0.5 M magnesium acetate (Sigma-Aldrich) and 10 μL of 100 mM PMSF (Sigma-Aldrich) per mL of lysis buffer. The sample was tumbled at room temperature for 30 min and then sonicated at 4 °C (Qsonica S-4000 Ultrasonic Processor) with a 50% duty cycle (5 s on/5 s off) at 55% power. 3% v/v Empigen (N,N-dimethyl-N-dodecylglycine betaine, BOC Sciences) was added to the cell lysate and tumbled at 4 °C for 1 h to extract σ2R-ΔER into micelles. The lysate was then centrifuged (38,500 × g, 25 min, 4 °C) to pellet cellular debris. The supernatant was incubated with 2 mL of preequilibrated Ni(II)-NTA Superflow (Qiagen) resin for 1–1.5 h at 4 °C.

The σ₂R/TMEM97 bound resin was packed in a gravity column and washed with 20 column volumes of Buffer A (40 mM HEPES (Gold Biotechnology), 300 mM NaCl (Fisher Scientific), 2 mM β-ME (Sigma-Aldrich), pH 7.8) with 1% (v/v) Empigen. The resin was washed with 20 column volumes of wash buffer (40 mM HEPES, 300 mM NaCl, 50 mM imidazole (Sigma-Aldrich), 1% (v/v) Empigen, 2 mM β-ME, pH 7.8). Empigen was exchanged for n-dodecylphosphocholine (DPC, Anatrace) with at least 15 column volumes of rinse buffer (20 mM HEPES, 150 mM NaCl, 0.5% (w/v) DPC, 2 mM β-ME, pH 7.8). σ2R-ΔER was eluted with 10 column volumes of elution buffer (20 mM HEPES, 300 mM imidazole, 150 mM NaCl, 0.5% (w/v) DPC, 2 mM β-ME pH 7.8).

The eluted protein from Ni²⁺-NTA column purification was buffer exchanged by ultrafiltration using a 10 kDa cutoff Amicon Ultra-15 centrifugal filter (Millipore) with four 4 mL exchanges with buffer (20 mM HEPES, 150 mM NaCl, pH 7.8) followed by concentration step to 0.5 mL. Three units of thrombin (Novagen) were added to the sample, and the reaction was incubated for 22 h at room temperature. Post-thrombin cleavage, the sample was passed over Ni-NTA resin, and the flow-through containing the thrombin-cleaved σ₂R/TMEM97 was collected. The cleaved protein was concentrated to 0.5 mL and further purified by gel filtration chromatography using a Superdex 200 Prep grade Resin packed in a 16XK column preequilibrated with NMR buffer (20 mM HEPES, 150 mM NaCl, 0.5 mM EDTA, 0.2% (w/v) DPC, pH 6.8). SDS-PAGE was used to assess purity (Figure S19), showing the combined fraction and concentrated to ~ 175 μL for NMR studies. The average σ₂R/TMEM97 yield was ~ 50 μg/L.

NMR-Detected Ver E Binding Studies.

The NMR sample was prepared in a 3 mM NMR tube with a total volume of 180 μL, including 5 μL D₂O and the ¹H–¹⁵N BEST-TROSY NMR experiments (b_trosyf3gpph.2 pulse program)⁷⁷ were recorded. Titrations were diluted from 1 mM stock Ver E prepared in 200 proof EtOH (VWR), and concentrations were measured in mole percent (mol %) and calculated using the equation

$$mol\%VerE=\frac{(molesVerE)}{molesVerE+molesDPC+moles\sigma 2\text{R}-\mathrm{\Delta }\text{ER}}\times 100\%$$

For NMR experiments, a sample containing 53 μM σ₂R/TMEM97 in 4.5% DPC (w/v) was used, and a series of ¹H–¹⁵N BEST-TROSY spectra were recorded as a function of Ver E at 37 °C, with 88 scans per spectrum. The Ver E concentrations used were 0.003, 0.03, 0.3, 0.6, 3, and 7.5 nmol %. NMR data were processed in nmrPipe⁷⁸ and analyzed with the CcpNMR Analysis software.⁷⁹ Chemical shift perturbations induced by Ver E were calculated by

$$\mathrm{\Delta }\delta =\sqrt{{(\mathrm{\Delta }{\delta }_H)}^2+{(0.2(\mathrm{\Delta }{\delta }_N))}^2}$$

where Δδ_H and Δδ_N are the proton and nitrogen chemical shift position differences between the initial titration point (0 nmol % Ver E) and a specific Ver E concentration for a given resonance.⁴¹ For selected resonances, when Δδ values were plotted as a function of Ver E concentration, saturable binding isotherms were observed, indicating specific binding of Ver E to σ₂R/TMEM97. TREND NMR^80,81 was applied, and PCA was used to analyze titration data globally, where PC1 captures the greatest variance and serves as an indicator of ligand binding.⁸² The dissociation constant, K_d, was calculated by fitting (SigmaPlot) the PCA data to a single-site binding model^41,83

$$\mathrm{\Delta }\delta (\text{mol}\%)=\frac{\mathrm{\Delta }{\delta }_{\max }\times \text{mol}\%}{K_{\text{d}}+\text{mol}\%}$$

where Δδ_max is the maximum chemical shift perturbation.

A control binding experiment with the vehicle (EtOH) was performed following a parallel procedure as above to rule out that EtOH binds to σ₂R/TMEM97 (Figure S21).

In Silico Docking Sigma-2 Ligands and Ver E.

Ligand and Receptor Structure Representation.

The sigma-2 receptor sequence was obtained and the structure was prepared using AlphaFold3 on the AlphaFold server.^43,84,85 Two-dimensional representations of Veraguamides C, E, O, and P were rendered in three dimensions using the obgen tool.⁸⁶ The 3-dimensional representations were then assigned hydrogens and charges using “prepligand” python script for Dock6.11.⁸⁷

Autodock-GPU Docking.

To prepare the ligands for AutoDock-GPU, we converted the output.mol2 format to.pdbqt format using OpenBabel version 3.1.0.⁸⁸ After the addition of hydrogens, the.pdbqt format of the receptor was converted to a.maps and.fld file format for docking using AutoGrid4, part of the AutoDockTools 1.5.7 release.⁸⁹ Ligand–receptor interactions were estimated using AutoDock-GPU, with 1000 runs/molecule specified.^90,91 Output scores were saved to.txt formats, which were read using a lab-developed Python script.

Rosetta Ligand Docking and Interaction Analysis.

Conformers of the ligands screened were generated by using obconformer in a scripted loop, then concatenated into a single file using the “cat” Unix shell function.⁸⁸ This conformer file was then processed into a Rosetta-readable.pdb file using the molfile_to_params Python script that is part of the Rosetta suite.⁹² Docking was performed with parsed functions using Rosettascripts with the centerpoint based on the PB28 center, search radius of 20 Å, and 10000 cycles.^93,94 The top-scoring pose was evaluated for interacting residues using the Protein–Ligand Interaction Profiler (PLIP) server.⁹⁵

Diffusion-Based Docking Analysis.

Generative diffusion-based docking was performed using DiffDock.^96,97 We used the Nvidia NIM API for DiffDock (https://build.nvidia.com/mit/diffdock), part of the Nvidia BioNeMo framework.⁹⁸ We generated 100 poses using the AlphaFold3 generated structure of Sigma2 for the Target Protein and the generated 3D representation of Ver E as a ligand. All docking simulations were performed with 20 diffusion steps and 20 time divisions. Poses were scored by their confidence measure, with the top scoring pose evaluated for interacting residues using the Protein–Ligand Interaction Profiler (PLIP) server.⁹⁵

Animals and Ethics.

All animals used for the Ca²⁺ imaging experiments were housed in groups of three to four per cage, maintained under a 12-h light/dark cycle, and provided food and water ad libitum. All protocols and procedures were conducted in compliance with protocol 20–04, approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Texas at Dallas.

Calcium Imaging Experiments.

An 18-week-old male mouse, heterozygous for GCamp6f and positive for Nes-Cre recombinase, was used for the wild-type fluorescent experiments. To generate Nestin-GCaMP6f transgenic mice for Ca²⁺ imaging, C57Bl/6J mice (Jackson Laboratories, Bar Harbor, ME) expressing a Cre-dependent GCaMP6f calcium sensor (B6J.Cg-Gt(ROSA)26Sortm95.1(CAG-GCaMP6f)Hze/MwarJ, JAX #028865) were crossed with C57Bl/6J mice carrying a Nestin promoter-driven Cre recombinase (B6.Cg-Tg(Nes-cre)1Kln/J, JAX #003771). Offspring were then genotyped using primers from Integrated DNA Technologies, Inc.

Isolation of Primary Mouse Sensory Neurons.

Primary DRG neurons were prepared and cultured as follows. Twelve-millimeter glass coverslips (Bellco Glass, cat# 1943–10012A) were spot-coated with 70 μL of 70 μg/mL poly-_d-lysine (PDL; Sigma-Aldrich, cat# P7405) and incubated overnight at 37 °C with 5% O₂. After incubation, coverslips were rinsed three times with sterile water (Sigma-Aldrich, cat# 95284) and dried in a laminar flow hood at room temperature. Next, a mouse was decapitated, and its spinal column was removed and placed in ice-cold Hanks’ Balanced Salt Solution (HBSS; Fisher Scientific, cat# 14–170–161) supplemented with 10 mM HEPES (Sigma-Aldrich, cat# H3375). The spinal cord was bisected, and each half was transferred to ice-cold Hibernate A solution (Fisher Scientific, cat# A1247501). All DRGs were extracted into a 5 mL snap-cap tube containing ice-cold Hibernate A.

To enzymatically dissociate the DRGs, papain was prepared by mixing 2 mL of an “HBSS complete” solution (48.5 mL of HBSS plus 0.5 mL penicillin/streptomycin, 0.5 mL sodium pyruvate [Fisher Scientific, cat# 11–360–070], and 0.5 mL HEPES) with papain (Worthington Biochemical, cat# LS003119) to achieve a final concentration of 20 IU/mL. The papain solution was filtered (0.20 μm) and kept at 37 °C until use. The DRGs were treated with 15 IU/mL of this papain mixture plus 2.25 mg/mL collagenase (Sigma-Aldrich, cat# C6885) at 37 °C for 20 min. The resulting cell suspension was filtered through a 40 μm nylon mesh (Fisher Scientific, cat# 08–771–1), centrifuged at 160 rcf for 4 min, washed once, and pelleted. The pellet was resuspended in Neurobasal A medium (Thermo Scientific, cat# 10888022) supplemented with 5% HI fetal bovine serum (VWR, cat# 10802–772), B-27 (Thermo Scientific, cat# A3582801), 2 mM GlutaMax-1 (Thermo Scientific, cat# 35050061), and 50 IU/mL penicillin/streptomycin (Thermo Scientific, cat# 15070063). Cells were counted (1:1 in trypan blue) and plated at 450,000–500,000 cells/mL. After 4 h of attachment at 37 °C with 5% O₂, 400 μL of warm, complete Neurobasal A medium was added to each coverslip, and the plates were incubated overnight.

Fluorescent Imaging of Mouse Sensory Neurons.

Calcium imaging was performed 24 h postincubation using an upright Olympus BX51WI microscope equipped with a 480 nm interface filter, a 505 nm dichroic mirror, and a 535 nm barrier filter (FITC band-pass). Images were collected at 1 Hz with Cell-Sense software (Olympus) and an Orca Fusion C14440 sCMOS camera (Hamamatsu). Nociceptive cells were identified in the Cell-Sense software, yielding over 100 distinct regions of interest (ROIs). Changes in intracellular Ca²⁺ were tracked over time.

To apply compounds or vehicle controls, a switching valve system (VC-6, Warner Instruments) was employed for bath perfusion into a 358 μL recording chamber (Warner Instruments, RC-21B), regulated by nitrogen gas (VPP-6, Warner Instruments). Coverslips were initially bathed in a baseline recording solution, followed by test compound or vehicle application, and then rinsed again with the baseline solution. At the conclusion of each recording, 60 mM potassium chloride (KCl) was introduced as a positive control for neuronal activation. Neurons responsive to KCl were statistically analyzed across coverslips, focusing on the peak Ca²⁺ signal and area under the curve (AUC) for each treatment group.

Screening for the Analgesic Properties of Ver E in hiPSCs.

To evaluate the analgesic properties of Ver E, a multiwell microelectrode array (MEA) system was employed using hiPSC-derived nociceptors.³² Prior to cell seeding, a 48-well MEA plate (Axion Biosystems, cat# M768-tMEA-48W) underwent a standardized surface-modification protocol: spot-coating the wells with 0.01% Poly-_l-Ornithine (EMD Millipore Sigma, cat# A-004-C) and incubating overnight at room temperature. The plate was then rinsed three times with sterile deionized water. A biocompatible coating—iMatrix-511 silk (Iwai-chem, cat# SKU:N-892022)—was diluted 1:50 in DPBS (Sigma-Aldrich, cat# D8537) and spot-applied to each well, followed by a 3 h incubation at 37 °C.

Chrono sensory neurons (Anatomic, cat# 7009) were thawed, washed in DMEM/F12 (Gibco, cat# 11320–033), and resuspended in Chrono Senso-MM complete growth medium (Anatomic, cat# 1030). Excess iMatrix-511 was removed, and the cells were spot-seeded (5 μL/well) at the density of 35,000 cells/well. After a 20 min attachment period, Chrono Senso-MM medium was gently added to each well (final volume: 400 μL). The medium was partially exchanged (50%) the following day and subsequently every 2–3 days. Cultures were maintained for 4 weeks at 37 °C in a humidified incubator with 5% CO₂.

Electrophysiological recordings were performed on an Axion Maestro Pro system at a 12.5 kHz sampling rate. The data were processed with a single-pole Butterworth bandpass filter (300–5000 Hz) to isolate neuronal signals. Spikes were defined as voltage deflections exceeding ± 5.5 SD of the baseline root-mean-square (RMS), with a minimum firing rate set at 1 spike per minute. At DIV28, baseline recordings were taken at both 37 and 42 °C, followed by administration of Ver E at 1–30 μM or vehicle (0.25% DMSO). Activity was measured for 15 min post-treatment, then at 1, 2, and 24 h, with all time points (except 0–0.25 h) recorded at both temperatures. After 24 h, wells were washed, and a final measurement was taken at least 1 h later at 37 °C.

Immunocytochemistry and p-eIF2α Detection in Primary Mouse DRGs.

Primary mouse DRGs were extracted and plated as described above, then maintained overnight in a complete Neurobasal A medium. After approximately 16 h of treatment with Ver E, FEM 1689, or vehicle (0.01% DMSO), the culture media were removed, and the DRG neurons were fixed by adding 50 μL of 10% formalin in PBS for 10 min. The cells were washed twice with 1× PBS before a blocking/permeabilization step, which involved a 1-h incubation at room temperature with 50 μL of 10% normal goat serum (NGS) in 0.1% Triton X–PBS. Following one additional PBS wash, primary antibodies were introduced in a solution containing 0.1% PBS–Triton X, NGS, and bovine serum albumin.

Specifically, a rabbit anti–p–eIF2α (Cell Signaling, cat# 3398) was used at 1:500, and a chicken antiperipherin (Encor Biotechnology, cat# CPCA-Peri) was used at 1:1000 for primary immunodetection. After overnight incubation at 4 °C, coverslips were washed twice with 0.5% Tween in PBS and once with PBS alone. Cells were then incubated for 1 h at room temperature with secondary antibodies diluted in the same buffer. Secondaries included a goat antirabbit Alexa Fluor 555 (Thermo, cat# A-21428) at 1:1000 and a goat antichicken IgY Alexa Fluor 488 (Thermo, cat# A11039) at 1:2000. After one wash in 0.5% Tween in PBS and two washes in PBS, coverslips were inverted onto microscope slides containing 10 μL of antifade hard-mount DAPI (Fisher, cat# NC9029229) to stain nuclei. Coverslips were stored overnight at 4 °C in the dark. Fluorescence was visualized using an Olympus IX73 microscope with the following filter sets: TRITC (540–570 nm) for p-eIF2α, UV (350–405 nm) for DAPI, and GFP (480–500 nm) for peripherin. Neurons were identified by peripherin staining, and p-eIF2α levels were quantified by measuring the mean gray intensity in regions of interest.

Characterization of the Cytotoxic Profile of Ver E.

HEK-293 cells⁹⁹ (generously provided by Dr. Xintong Dong, University of Texas at Dallas) were cultured in Dulbecco’s Modified Eagle Medium (DMEM; Corning, cat# 10-014-CM) supplemented with 10% fetal bovine serum (VWR, cat# 10802–772) and 100 U/mL penicillin-streptomycin (Gibco, cat# 15–070–063). The cells were maintained at 37 °C in a humidified atmosphere containing 5% CO₂ until they reached approximately 90% confluence. For experimental assays, cells were dissociated with 0.05% trypsin (Gibco, cat# 15–400–054) for 5 min, then neutralized with fresh culture media. The suspension was centrifuged at 300g for 5 min, after which the pellet was resuspended in new culture media.

In 96-well plates, HEK-293 cells were seeded at a density of 6 × 10⁴ cells per well and allowed to adhere for 24 h. Thereafter, cells were treated with Ver E or digitonin, each dissolved in 100% DMSO and then diluted in DMEM, at final concentrations ranging from 0.03 μM to 30 μM. Vehicle-only controls received 0.25% DMSO (VWR, cat# BDH1115–1LP). Following 24 h of compound exposure, cell viability was evaluated using a resazurin reduction assay, wherein living cells convert resazurin to the fluorescent compound resorufin. Resazurin sodium salt (Sigma-Aldrich, cat# 62758–13–8) was freshly prepared at 0.15 mg/mL in DPBS (pH 7.4; Gibco, cat# 14190-144, filter-sterilized (0.2 μm), and protected from light. A 20 μL aliquot of this resazurin solution was added to each well, containing cells in 100 μL of media. After a 4-h incubation at 37 °C, fluorescence was measured at an excitation wavelength of 560 nm and an emission wavelength of 590 nm on a microplate reader. Cell viability was determined by comparing the fluorescence of compound-treated wells to that of vehicle-treated controls.

Computational Analysis of Sigma-2 Ligands and Ver E.

Overlap Analysis to generate Maximum Tanimoto Scores was performed using an academic license of the Instant JChem platform (ver. 23.5.0, 2023). The overlap analysis with Tanimoto Score screening configuration was done using a similarity search mode based on a Chemical Hashed Fingerprint with a 0.01 threshold, where the Ver E was set as the query and selected σ2R ligands (Table S2) as the target. Predicted physicochemical values were determined using DataWarrior (ver. 06.04.02, 2025) and Principal Component Analysis was performed using RStudio (ver. 2024.12.1).

Statistics.

In Vitro Calcium Imaging.

All statistical analyses were performed using GraphPad Prism (version 9.0) and MATLAB. Raw data files, generated after each compound application by Cell-Sense software (Olympus), were transferred into Microsoft Excel and processed via a custom MATLAB script. This script calculated changes in fluorescence over time (ΔF/F₀), the peak fluorescence during compound application (max ΔF/F₀), and the area under the curve (AUC). The first region of interest (ROI), which contained an empty background, was used for subtraction in all other ROIs. The MATLAB script accounted for specific drug application time windows, and the resulting processed data were compiled into new spreadsheets for import into GraphPad Prism, where final visualizations and statistical tests were conducted. In the Ver E assays, unpaired two-tailed t tests were used to compare 10 μM Ver E with 0.1% DMSO vehicle versus vehicle alone in two distinct parts of the recording (direct Ver E and SOCE response).

Screening for the Analgesic Properties of Ver E in Human Induced Pluripotent Stem Cells.

At each experimental time point, the total spike count in each well was normalized to its respective baseline (37 or 42 °C). Statistical comparisons were performed using a two-way ANOVA with Dunnett’s post hoc test to evaluate different Ver E concentrations (1–30 μM) against DMSO vehicle controls. Data are reported as mean ± SEM, with a p-value <0.05 considered statistically significant. At 37 °C, Ver E at 30 μM induced a notable reduction in firing activity (p < 0.05), whereas no significant effect was observed at 42 °C.

Immunocytochemistry and p-eIF2α Detection in Primary Mouse DRGs.

Mean gray intensities corresponding to p-eIF2α staining were calculated for each treatment group and are expressed as mean ± SEM. One-way ANOVA followed by Tukey’s multiple-comparisons test was used to assess statistical differences among Ver E (100 nM, 1000 nM), FEM 1689 (30 nM, 100 nM), and vehicle (0.01% DMSO).

Characterization of the Cytotoxic Profile of Ver E.

A logarithmic nonlinear regression curve [log(inhibitor concentration) vs normalized response] was generated in GraphPad Prism 10 to determine the LC₅₀ of both Ver E and digitonin. The IC₅₀ calculated by this curve-fitting method was taken to represent the LC₅₀ in this experiment. Digitonin yielded an LC₅₀ value of 9.68 μM, whereas Ver E’s LC₅₀ exceeded 30 μM. No additional statistical comparison was required to confirm that Ver E did not induce detectable cytotoxicity within the 0.03–30 μM range.

Supplementary Material

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jnatprod.5c01092.

Full NMR data sets for Ver E including comparison to literature reported values and variable temperature and solvent effect NMR studies, MS and MS/MS data, proposed fragmentation patterns based on the MS/MS, validation of NMR binding studies, diffusion based chemical docking to orthogonally validate modeling, and data used for chemical likeness determination including structures (PDF)

Data Availability Statement

All 1-D and 2-D NMR data (Figures S3–S9) for the following compound have been deposited in the Natural Products Magnetic Resonance Database (NP-MRD; www.np-mrd.org) and can be found at NP0009879 (Veraguamide E). Other raw data related to this study can be found at Open Science Framework (https://osf.io/8vdft).