Efficient and Modular Synthesis of Ibogaine and Related Alkaloids

Abstract

Anecdotal reports and preliminary clinical trials suggest that the psychoactive alkaloid ibogaine and its active metabolite noribogaine have powerful anti-addictive properties, producing long-lasting therapeutic effects across a range of substance use disorders and co-occurring neuropsychiatric diseases such as depression and post-traumatic stress disorder. Here, we report a gram-scale, 7-step synthesis of ibogaine from pyridine. Key features of this strategy enabled the synthesis of three additional iboga alkaloids, as well as the first enantioselective total synthesis of (+)-ibogaine and the construction of four novel analogues. Biological testing revealed that the unnatural enantiomer of ibogaine does not produce ibogaine-like effects on cortical neuron growth, while (–)-10-fluoroibogamine exhibits exceptional psychoplastogenic properties and is a potent modulator of the serotonin transporter. This work provides a platform for accessing iboga alkaloids and congeners for further biological study.

MAIN TEXT

The iboga alkaloids represent a class of monoterpene indole alkaloids with hundreds of members.¹ Ibogaine (1) has garnered the most attention due to its remarkable antiaddictive properties in both humans and rodents, which are both long-lasting and extend to several distinct substance use disorders and related conditions.^1,2,3,4 Currently, ibogaine and related alkaloids are obtained via extraction or semi-synthesis from natural plant sources, and thus, efficient total synthesis could reduce concerns about ecological sustainability and access to iboga alkaloids of low abundance. Despite its promise, the therapeutic potential of ibogaine has been tempered by its suboptimal safety profile. Ibogaine inhibits hERG channels in the heart, which can lead to cardiac arrhythmias.^5,6,7,8 To circumvent this issue, novel ibogaine analogues with comparable efficacy in preclinical models and improved safety profiles have been developed.^9,10 However, the effects of these analogues in humans remain unknown and given that the therapeutic mechanisms of ibogaine have not been fully elucidated, it is unclear if these novel compounds will exhibit similar antiaddictive properties in humans. Thus, recent efforts have focused on mitigating the risks associated with ibogaine therapy through a combination of patient screening, careful cardiac monitoring, and magnesium supplementation.^11,12 As a single ibogaine therapy session can lead to remission lasting several months,^13,14 efforts to improve the safety of ibogaine therapy are extremely important.

Though ibogaine is commonly associated with the shrub Tabernanthe iboga, it is only generated in small quantities via extraction (ibogaine is ~0.3–0.4% of the root bark).^15,16 Thus, ibogaine is primarily obtained via semisynthesis from voacangine (5) following its isolation from Voacanga africana (voacangine is ~1% of the root bark).^16,17 While the majority of the enzymes involved in the biosynthesis of ibogaine are known,^18,19 the entire biosynthetic pathway has never been reconstituted in a simple organism, making chemical synthesis an attractive approach for producing large quantities of ibogaine and related alkaloids. Moreover, chemical synthesis offers the additional benefit of enabling structural diversification and analogue generation.

The iboga alkaloids are characterized by three defining architectural features, which include an indole ring, a 7-membered tetrahydroazepine, and a bicyclic isoquinuclidine (Fig. 1a).¹ While there have been numerous approaches to ibogamine (2), catharanthine, and related alkaloids^{20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37} (Fig. 1b), there have only been three total syntheses of ibogaine published to date with step counts and overall yields ranging from 9–15 and 0.14%–4.61%, respectively.^31,32,33 While pioneering, none of these syntheses were scalable, asymmetric, or allowed for late-stage diversification and the generation of additional alkaloids and analogues. The vast majority of iboga alkaloid syntheses can be grouped into three primary endgame strategies—rearrangements to simultaneously generate the B and C rings, tetrahydroazepine (B ring) formation through C2–C16 bond formation, and isoquinuclidine formation (C ring) through transannular cyclization. Unfortunately, these approaches necessitate pre-functionalization of the indole early in the synthesis, which limits the generation of aryl substituted analogues.

Figure 1. Synthetic strategies towards iboga alkaloids.

(a) Structures of major iboga alkaloids. (b) Common synthetic strategies for accessing the iboga core scaffold. The order of indole (A ring), tetrahydroazepine (B ring) and isoquinuclidine (C ring) construction is indicated. (c) Our retrosynthetic analysis suggested that constructing the indole (A ring) last would enable access to multiple natural products and analogues.

Taking inspiration from White’s 16-step synthesis of ibogamine³⁴ as well as pioneering work from Sallay³⁵ and Pierson,³⁶ we reasoned that a late-stage Fischer indolization of a key tricyclic ketone 6 would enable access to a diverse array of natural products and analogues (Fig. 1c) and would constitute a substantial departure from more common routes for constructing the iboga core scaffold. We envisioned that the tetrahydroazepine of 6 could be accessed by exploiting strained cyclopropyl ketone 7, and that strategic placement of a ketone at the C16 position would allow for endo to exo epimerization following construction of the isoquinuclidine via an intermolecular Diels-Alder reaction between a dihydropyridine derived from 8 and enone 9 (Fig. 1c). Development of an asymmetric Diels-Alder reaction would yield an enantioselective variant of the route, and importantly, our strategy would take advantage of low-cost, readily available starting materials.

Total Synthesis of (±)-Epiibogaine

Our initial efforts focused on developing a regioselective reduction of 3-ethylpyridine 8 to dihydropyridine 10 (Extended Data Fig. 1). While NaBH₄ in MeOH yielded a 9:1 ratio favoring the undesired diene 11, the selectivity could be reversed by using LiBH₄ in THF. However, 10 rapidly converts to the more thermodynamically stable 11 following an aqueous quench and attempts to isolate 10 were unsuccessful. The Fukuyama group had previously demonstrated that 10 could be accessed from 12 via a bromination/dehydrohalogenation sequence, though a lengthy 3-step protocol was required to synthesize 12.³⁷ By employing acidic conditions during the reduction, we were able to convert 8 to tetrahydropyridine 12 in a single step. As expected, the bromination/dehydrohalogenation sequence proceeded in excellent yield to afford dihydropyridine 10 in 3 steps from 3-ethylpyridine. (Fig. 2a).

Figure 2. Total synthesis of epiibogaine.

(a) The completed synthesis of epiibogaine is shown. Step-count is indicated by the numbers over each reaction arrow. Calculations of thermodynamic stability (see supporting information section 3) indicated that isomer 10 was lower in energy than isomer 11. (b) HAT reduction of 13a only yielded the endo product 17a. Calculations of thermodynamic stability (see supporting information section 3) indicated that the C20 exo isomer 14b was lower in energy than endo isomer 14a. (c) Carbamate coordination of a metalloradical intermediate could potentially explain the selectivity observed in HAT reactions.

Next, dihydropyridine 10 was subjected to a Diels-Alder cycloaddition with cyclopropyl enone 9 to produce the expected endo isomer as the predominant product. However, subsequent addition of base allows for epimerization of C16 to a 1:1 mixture of epimers. Hydrogenation of 13 with Pd/C gave intermediate 14 in 91% yield as a single epimer at C20. Subjecting intermediate 14 to HBr opened the cyclopropane to produce the corresponding alkyl bromide, which was used without further purification. A subsequent S_N2 reaction closed the 7-membered ring to yield 15 in 40% yield over two steps (Fig. 2a). We had envisioned that dynamic epimerization of C16 under basic conditions followed by S_N2 ring closure could convert both epimers of 14 to 15; however, competing enolization of the secondary α-carbon followed by halide displacement regenerates cyclopropane 14. To study these transformations more closely, we relied on a model system lacking the C20 ethyl group (Extended Data Fig. 2). We found that the identity of the alkyl halide was crucial, as the alkyl chloride failed to react and the alkyl iodide produced the undesired cyclopropane S4 exclusively. Fortunately, use of the alkyl bromide under anhydrous conditions led to 7-membered ring formation as the predominant pathway (Extended Data Fig. 2). Next, 15 was subjected to a Fischer indolization using 4-methoxyphenylhydrazine to complete one of the most efficient syntheses of epiibogaine (16) to date³⁸ in 8 steps and 8.7% overall yield (Fig. 2a).

Synthesis of (±)-Ibogaine and Analogues

To access iboga alkaloids displaying exo ethyl groups at C20, we envisioned selectively reducing olefin 13 from the face opposite the C–N bridge by exploiting an unfavorable steric clash between endo substituents at both C16 and C20. Computational studies (see supporting information section 3) confirmed that C20 exo isomer 14b is 2.3 kcal/mol lower in energy than 14a (Fig. 2b), which led to the hypothesis that reduction via a radical pathway would favor the more thermodynamically stable C20 exo product. However, hydrogen atom transfer (HAT) hydrogenation³⁹ of 13a yielded intermediate 17a as the sole product (Fig. 2b). Typical metal-catalyzed HAT (MHAT) reactions involve the addition of a metal hydride to an olefin to generate a carbon centered radical and metal complex that is formally reduced by one electron (Fig. 2c).^40,41 The resulting carbon- and metal-centered radicals likely associate through solvent caged pairing,⁴² and we hypothesized that additional stabilization of the exo metalloradical 19 by the carbamate might be the origin of the selectivity we observed (Fig. 2c). Subsequent addition of a second metal hydride would proceed from the same face as the stabilized metalloradical, leading to an endo orientation of the C20 substituent.

We realized that we could exploit this highly selective transformation to access C20 exo substituents from an olefin precursor 20 (Fig. 2c), obviating the need to use 3-substituted pyridines in the synthesis of iboga alkaloids. One-pot acylation and reduction of inexpensive and readily available pyridine yielded dihydropyridine 24 in 98% yield on a 20 g scale (Fig. 3a). Subsequent cycloaddition with cyclopropyl enone 9 afforded isoquinuclidine 20 in excellent yield (90%) on decagram scale. Epimerization of intermediate 20 was achieved in the same reaction vessel by diluting with methanol containing sodium methoxide. This 1:1 mixture of C16 epimers was subjected to an olefin-olefin MHAT coupling⁴³ with methyl acrylate in the presence of iron(III) acetylacetonate and phenylsilane followed by ester hydrolysis to produce 25 with exquisite regio- and stereocontrol (only 1 of 4 possible regio-/stereoisomers was observed) (Fig. 3a). Interestingly, MHAT coupling of pure endo intermediate 20a led to a dramatic decrease in reaction conversion and a significant amount of starting material was isolated (Extended Data Fig. 3). Alternative hydroalkylation methods were also explored,^44,45 but these reactions were suboptimal, producing the reduced compound S6 as the primary byproduct (Extended Data Fig. 4).

Figure 3. Total synthesis of (±)-ibogaine and related compounds.

(a) The completed synthesis of (±)-ibogaine is shown. Step-count is indicated by the numbers over each reaction arrow. (b) ¹H NMR data demonstrates that synthetic and natural ibogaine are indistinguishable. (c) Synthetic efficiency was assessed by comparing overall yield, total step count, and quantity of ibogaine produced following the final step (i.e., scale) for the 4 reported total syntheses of (±)-ibogaine (left). The modularity of the synthesis was evaluated by plotting Bottcher’s complexity index vs. step count (i.e., percent through the synthesis) (right). (d–e) Structures of additional natural products (d) and non-natural analogues (e) that were synthesized using our synthetic strategy. Overall yields for their syntheses are highlighted in blue. Compounds with a hydrogen or ethyl group at C20 were synthesized in 6 or 7 steps overall, respectively.

A photoredox-catalyzed hydrodecarboxylation^46,47 of 25 produced 26 in 71% yield on gram scale. We noticed that the cesium counterion played a significant role in facilitating decarboxylation, as conversion was dramatically diminished when KOH, NaOH, or amine bases were used (Extended Data Fig. 5). Clean conversion was observed using both (Ir[dF(CF₃)ppy]₂)dtbpy))PF₆ and 9-mesityl-10-methylacridinium (Mes-Acr-Me); however, we preferred to use the organic photocatalyst when conducting large scale reactions due to its lower cost. When performing the photoredox-catalyzed reaction on gram scale, factors such as stirring rate and flask size were critical for maintaining reproducibility (see supporting information section 2).

Compound 26 could be readily converted to key intermediate 27 through concurrent opening of the cyclopropyl ketone and Cbz deprotection using HBr in AcOH followed by intramolecular S_N2 cyclization. Fischer idolization completed the synthesis of (±)-ibogaine (1) in 31% over 3 steps on a 500 mg scale, and yields were only slightly diminished when conducting the sequence on >1 gram scale (Fig. 3a). Spectral data for synthetic ibogaine matched that for ibogaine obtained from natural sources (Fig. 3b, Extended Data Fig. 6, and Extended Data Fig. 7).

The efficiency of our synthetic strategy is highlighted when comparing step counts, overall yields, and production scale of ibogaine total syntheses to date (Fig. 3c). Our strategy also compares favorably to a number of other iboga alkaloid total syntheses (Extended Data Fig. 8). To better understand the efficiency and modularity of our route, we analyzed changes in molecular complexity (C_m) over the course of the synthesis (Fig. 3c). Bottcher’s complexity index (C_m)^48,49 formalizes the intrinsic information content of a molecule on a per-atom basis by assigning information content (mcbits) to each non-hydrogen atom. Our analysis revealed two distinct points of modularity amenable to analogue generation—diversification via olefin or carbonyl functionalization. Both points of diversification occur at a synthetic intermediate of lower molecular complexity than the target analogues, indicating that these simpler intermediates are useful nodes for adding molecular complexity (Fig. 3c).

To fully showcase the modularity of our approach, we synthesized 4 iboga alkaloids (Fig. 3d) and 4 non-natural analogues (Fig. 3e) in excellent overall yields (6.2–28.8%). In a single step from key intermediate 27, we were able to complete the total synthesis of not only (±)-ibogaine (1), but also (±)-ibogamine (2), (±)-ibogaline (3), and (±)-tabernanthine (4). By altering the hydrazine reagent used in the Fischer indole cyclization and/or hydrogenating olefin 20, we were also able to access the non-natural compounds 28, 29, 30, and 31.

Asymmetric Synthesis of (+)-Ibogaine

Given that several asymmetric strategies have been reported for constructing isoquinuclidine ring systems,^{50,51,52,53,54,55} we attempted to adapt these to our synthetic approach. Unfortunately, use of chiral protic or Lewis acids did not yield any asymmetric induction during the Diels-Alder reaction (Extended Data Fig. 9a). Next, we conducted the Diels-Alder reaction in the presence of chiral organocatalysts, but never observed the desired iminium intermediate S10. Instead, we only produced the conjugate addition product S11, which is consistent with previous reports demonstrating that α-branched enones do not perform well in organocatalyzed Diels-Alder reactions.⁵⁶ Subsequent reactions with various protected dihydropyridines did not yield any isoquinuclidine products (Extended Data Fig. 9b).

Due to the challenges we encountered with performing catalytic asymmetric Diels-Alder reactions using enone 9, we next explored diastereoselective Diels-Alder reactions using chiral auxiliaries. First, we tested a menthol-derived carbamate at both moderate and low temperatures. Given the potential for producing a complex mixture of diastereomers and rotamers, we opted to convert S13 to desethylibogaine (30) prior to measuring asymmetric induction. Unfortunately, chiral HPLC analysis revealed that we had produced a racemate (Extended Data Fig. 9c).

Appending the chiral auxiliary to the dienophile proved much more successful. Reaction of diene S14 with chiral oxazolidinone-derived enamide S15 in the presence of Ti(O-iPr)₂Cl₂ yielded the desired diastereomeric isoquinuclidines S16 and S17 in near quantitative yield and 4:1 dr, respectively, as single epimers (Extended Data Fig. 9d). The use of a phenyl carbamate on the diene partner was critical as the analogous Cbz-protected carbamate led to lower yield (44%) and diastereoselectivity (3:2 dr). Chromatographic separation of the major diastereomer followed by oxazolidinone removal with LiOOH and Na₂SO₃⁵⁷ yielded isoquinuclidine S18, which was converted to the Weinreb amide and subsequently reacted with cyclopropyl magnesium bromide to furnish cyclopropyl ketone S19 (Fig. 4a). Chiral HPLC analysis indicated that S19 was produced in 99:1 er (Extended Data Fig. 9e). Exchange of the phenyl for a benzyl carbamate enabled us to intercept intermediate 21 from our racemic route, which was progressed to complete the first ever enantioselective synthesis of (+)-ibogaine (1) in 11 steps and 9.3% overall yield (Fig. 4a and b). We were also able to access (+)-10-fluoroibogamine (11 steps, 10.1% overall yield) simply by exchanging 4-methoxyphenylhydrazine for 4-fluorophenylhydrazine in the last step of the sequence (Fig. 4b). Finally, we synthesized (–)-10-fluoroibogamine (11 steps, 9.7% overall yield) using the same general procedures but substituting S15 for its enantiomer (Fig. 4b).

Figure 4. Asymmetric synthesis of (+)-ibogaine, (–)-10-fluoroibogamine, and (+)-10-fluoroibogamine.

(a) The completed synthesis of (+)-ibogaine is shown. Step-count is indicated by the numbers over each reaction arrow. (b) Structures of additional non-natural ibogaine analogues that were synthesized using our asymmetric synthetic strategy. Overall yields for their syntheses are highlighted in blue. PMPH = 4-methoxyphenylhydrazine.

Biological Evaluation of (+)-Ibogaine and (–)-29

The mechanisms by which ibogaine and noribogaine exert their therapeutic effects are poorly understood, though several hypotheses have emerged related to neuroplasticity in addiction-related circuitry^58,59,60 and modulation of serotonin transporter (SERT) function.⁶¹ Our asymmetric route to ibogaine enabled us to compare the psychoplastogenic properties of the two enantiomers for the first time. Given that limited neuritogeneisis occurs in vivo, spinogenesis is considered to be much more relevant to the therapeutic effects of psychoplastogens, and thus, we focused our efforts on spinogenesis assays. Treatment of mature rat embryonic cortical cultures (DIV18) with (–)-ibogaine or its active metabolite noribogaine^62,63,64 led to a large increase in dendritic spine density measured 24 h later (Fig. 5a and b). Similar results were observed when the neurons were treated with the positive control brain-derived neurotrophic factor (BDNF). However, (+)-ibogaine did not produce any increase in spine growth, indicating a clear chiral preference for induction of structural plasticity (Fig. 5a and b). While (–)-ibogaine is known to upregulate BDNF transcription in the prefrontal cortex,⁶⁵ the target that mediates ibogaine-induced cortical neuron growth is still opaque. Thus, the unnatural enantiomer of ibogaine could be an important chemical tool for elucidating ibogaine’s molecular mechanism of action.

Figure 5. Effects of non-natural (+)-ibogaine and (–)-10-fluoroibogamine on cortical spinogenesis.

(a) Dendritic spine density on rat embryonic cortical neurons (DIV19) was assessed after treatment (10 μM) for 24 h. Neurons were visualized using a fluorescent conjugate of phalloidin. Unlike natural (–)-ibogaine, non-natural (+)-ibogaine does not promote spine growth. In contrast, (–)-29 promotes spine growth comparable to (–)-ibogaine. (N = 17–18 neurons per treatment); one-way ANOVA with Dunnett’s post hoc test). *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, as compared to the VEH control. VEH = vehicle; BDNF = brain-derived neurotrophic factor; (–)-IBO = (–)-ibogaine 1; (+)-IBO = (+)-ibogaine; nor-IBO = noribogaine; (–)-F-IBO = (–)-10-fluoroibogamine 29.

Like (–)-ibogaine, noribogaine produces robust increases in cortical neuron dendritic spine density (Fig. 5a and b). Both compounds are also known to stabilize the inward-open conformation of SERT,^66,67 an unusual property that has been hypothesized to contribute to their unique therapeutic effects compared to traditional antidepressants.⁶¹ In vivo microdialysis studies have shown that both ibogaine and noribogaine can elevate serotonin levels in the nucleus accumbens, with noribogaine being more effective.⁶² To compare the effects of ibogaine, noribogaine, and iboga analogues on SERT function, we performed inhibition and efflux assays by heterologously expressing SERT on the plasma membrane of HEK293T cells and measuring the uptake (Fig. 6a) or release (Fig. 6b) of radiolabeled serotonin ([³H]5-HT). Both (–)-ibogaine and noribogaine were partial inhibitors of the transporter (~55%) with the latter being approximately an order of magnitude more potent. Noribogaine, but not (–)-ibogaine, also led to a rapid partial efflux of serotonin (Fig. 6b) with an E_max that was ~30% that of full efflux inducers like PCA and (±)-MDMA. Interestingly, (+)-ibogaine was inactive in both SERT efflux and inhibition assays (Fig. 6a,b).

Figure 6. Effects of non-natural (+)-10-fluoroibogamine and (–)-10-fluoroibogamine on SERT function.

(a) Both (+)-29 and (–)-29 inhibit the uptake of [³H]5-HT into HEK293T cells heterologously expressing SERT. Data were normalized to the vehicle control (0%) and 100 μM cocaine (100%). (b) Noribogaine and (–)-29 promote partial efflux of accumulated [³H]5-HT from HEK293T cells heterologously expressing SERT. Data were normalized to the vehicle control (0%) and 100 μM PCA (100%). (N = 3 wells per treatment from 2 independent cultures). (–)-IBO = (–)-ibogaine 1; nor-IBO = noribogaine; (–)-F-IBO = (–)-10-fluoroibogamine 29; (+)-F-IBO = (+)-10-fluoroibogamine 29, 5-HT = serotonin, PCA = p-chloroamphetamine, Coca = cocaine, (±)-MDMA = (±)-3,4-methylenedioxymethamphetamine.

Using our asymmetric synthetic strategy, we were able to replace the hydroxyl group in noribogaine with a fluorine to produce (–)-10-fluoroibogamine (29). We reasoned that fluorine was an ideal hydroxyl bioisostere given that fluorine and oxygen have similar electronegativities though fluorine is significantly more lipophilic.⁶⁸ Moreover, the fluorinated analogue would avoid metabolism via glucuronidation, which is a known metabolic pathway for noribogaine.⁶⁹

Unlike ibogaine and noribogaine, (+)-29 and (–)-29 both appeared to be full inhibitors of SERT (Fig. 6a). Interestingly, (–)-29 also led to a partial efflux of radiolabeled serotonin that was ~50% that of full efflux inducers like PCA and (±)-MDMA (Fig. 6b). The chirality of (–)-29 seems to be very important for this effect, as (+)-29 was approximately two orders of magnitude less potent. When comparing the ratios of compound potencies in the SERT inhibition and efflux assays, ibogaine was a 5-fold more potent as an inhibitor while (±)-MDMA was a 3-fold more potent as an efflux inducer. In stark contrast, (–)-29 was 20-fold more potent as a serotonin releasing agent than a SERT inhibitor. In addition to its effects on SERT-mediated serotonin efflux, (–)-29 possesses the ability to promote cortical neuron spinogenesis (Fig. 5a and b). Given its improved physicochemical properties relative to noribogaine, (–)-29 should be investigated as a treatment for substance use disorders and related neuropsychiatric diseases.

Discussion

Ibogaine has demonstrated remarkable efficacy for treating patients with a variety of substance use disorders, including those with co-occurring depression and/or anxiety disorders.² However, accessing sufficient quantities of natural ibogaine for clinical trials has proven challenging as it is produced in small quantities and overharvesting of Tabernanthe iboga negatively impacts cultures that use it for spiritual purposes. Semi-synthesis from the more abundant natural product voacangine offers a potential solution, as does efficient total synthesis from readily available building blocks. Here, we report the first scalable, asymmetric total synthesis of ibogaine that also provides access to novel iboga analogues with potential advantages over ibogaine.

While modern methods for patient screening and careful cardiac monitoring have improved outcomes for patients seeking ibogaine treatment, hERG inhibition remains a significant risk that could result in death. Thus, there is an urgent need to better elucidate the mechanism by which ibogaine produces therapeutic effects so that we can develop compounds with improved efficacy and safety profiles. Fortunately, our synthetic strategy provides a robust platform for accessing such chemical tools and novel analogues. Using this strategy, we completed the total syntheses of tabernanthine and ibogaline—two natural products found in Tabernanthe iboga that could contribute to the anti-addictive properties of the root bark. With access to sufficient quantities of these iboga alkaloids as well as ibogaine and ibogamine, head-to-head pharmacological studies can now be conducted.

Nature only produces one enantiomer of ibogaine.⁷⁰ By synthesizing its non-natural optical antipode, we were able to demonstrate that ibogaine’s ability to promote cortical neuron dendritic spine growth is highly dependent on its configuration, implicating a receptor-mediated process rather than nonspecific effects such as changes in membrane fluidity or vesicular pH. While the psychoplastogenic effects of ibogaine are likely to contribute to its antidepressant potential,⁷¹ it is unclear what role, if any, cortical spinogenesis plays in the antiaddictive properties of ibogaine.

Both ibogaine and noribogaine are well known to stabilize the inward-open conformation of SERT^66,67—a feature that distinguishes them from other drugs targeting SERT such as selective-serotonin reuptake inhibitors (SSRIs) and drugs of abuse like cocaine. As compounds capable of inducing serotonin efflux such as 3,4-methylenedioxymethamphetamine (MDMA) produce distinct therapeutic effects compared to SSRIs, we tested the effects of ibogaine and noribogaine on SERT-mediated serotonin efflux in SERT-expressing HEK293T cells and found that noribogaine, but not ibogaine, induced serotonin efflux more potently than it inhibited the transporter. However, serotonin efflux in vivo is a more complex process involving activity at both vesicular monoamine transporters and monoamine oxidases, possibly explaining the relatively modest increases in serotonin levels observed in the nucleus accumbens following noribogaine treatment (~3-fold).^62,64 Alternatively, these modest increases in serotonin in vivo might be explained by the fact that noribogaine seems to only result in partial efflux (~30%) compared to full efflux inducers like PCA and (±)-MDMA.

Given that noribogaine has been proposed to be the active metabolite of ibogaine,^62,63,64 we hypothesize that the large doses of ibogaine required to produce anti-addictive effects in the clinic could be due to the conversion of ibogaine to noribogaine. Using our modular total synthesis, we developed (–)-10-fluoroibogamine, a novel noribogaine analogue demonstrating psychoplastogenic effects comparable to ibogaine and serotonin efflux properties superior to noribogaine in HEK293T cells expressing SERT. Serotonin efflux does not promote cortical spine growth,⁷² which suggests that psychoplastogenic properties and effects on SERT function are distinct, and perhaps produce additive/synergistic therapeutic effects. While the ability of (–)-10-fluoroibogamine to promote cortical neuron spine growth and serotonin efflux might be beneficial, potential antidepressant and anti-addictive effects must be confirmed in vivo using standard behavioral tests. These initial biological results demonstrate the power of chemical synthesis for accessing chemical tools and improved neurotherapeutics related to complex iboga alkaloid natural products.

METHODS

Data Analysis and Statistics.

Treatments were randomized, and data were analyzed by experimenters blinded to treatment conditions. Statistical analyses were performed using GraphPad Prism (version 10.0.3) unless noted otherwise. All comparisons were planned prior to performing each experiment. Data are represented as mean ± SEM, unless otherwise noted, with asterisks indicating *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Drugs.

The NIDA Drug Supply Program provided ibogaine hydrochloride (IBO), noribogaine (NOR).

Animals.

All experimental procedures involving animals were approved by the University of California, Davis Institutional Animal Care and Use Committee (IACUC) and adhered to principles described in the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Power analyses were conducted to ensure appropriate sample size for all experiments involving animals. The University of California, Davis is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC).

Spinogenesis Experiments.

Spinogenesis assays were performed according to previously described protocols using cultured embryonic rat cortical neurons.⁶⁰ Neurons were treated on DIV20 and fixed 24 h later (DIV21). Imaging was performed using a Nikon HCA Confocal microscope with 100x/NA 1.45 oil objective. The vehicle and positive controls were DMSO (0.1% in media) and BDNF (50 ng/mL), respectively.

Serotonin Inhibition Experiments.

Cells (HEK293T) were grown in Dulbecco’s Modified Eagle Media (DMEM) supplemented with 10% fetal bovine serum (FBS). Plastic Costar 96-well plates were seeded at a density of 100,000 cells/well 24 h prior to the experiment and concurrently transfected using Lipofectamine^™ 3000 Transfection Reagent according to the manufacturer’s protocol. Cells were transfected with 0.1 μg of hSERT-N1-pEYFP (Addgene #70105) per well. Wells were washed (1 × 200 μL) with 1x Hank’s Balanced Salt Solution supplemented with 2 mM MgCl₂ and 2 mM CaCl₂ (HBSS) as well as 5 mM HEPES and replenished with 100 μL of supplemented HBSS. The plate was placed in a 37 °C water bath for the reminder of the experiment and allowed to incubate for 30 min prior to beginning the experiment. The assay was initiated by adding a 100 μL mixture of 5-[1,2-³H(N)]-hydroxytryptamine creatinine sulfate ([³H]5-HT; #NET498001MC, Lot: 3261835, Revvity) and respective drug onto a plate at final concentrations of 20 nM [³H]5-HT and various concentrations of drug (31.6 nM – 100 μM). After 10 min, uptake was terminated by aspiration and wells washed with supplemented HBSS (3 × 200 μL). Cells were then lysed for 30 min by the addition of 30 μL of 10 mM NaOH, and 120 μL OptiPhase HiSafe (#1200.437; Revvity) was added to the cell lysates. Counts per minute (CPM) were quantified using MicroBeta2 microplate liquid scintillation counter. In GraphPad Prism, concentration responses were plotted as the difference of the VEH (0.1% DMSO) average for the plate and the count per minute (CPM) for each respective drug concentration (VEHavg – drug). The plots were normalized similarly, with the plate vehicle control (VEHavg – VEHavg) average being set to 0% inhibition and the difference between the VEH average and the positive control average (VEHavg – 100 uM Cocaavg) set to 100%. Outliers were removed by using the ROUT method (Q = 10%) for each column separately.

Serotonin Efflux Experiments.

Cells (HEK293T) were grown in Dulbecco’s Modified Eagle Media (DMEM) supplemented with 10% fetal bovine serum (FBS). Plastic Costar 96-well plates were seeded at a density of 100,000 cells/well 24 h prior to the experiment and concurrently transfected using Lipofectamine^™ 3000 Transfection Reagent according to the manufacturer’s protocol. Cells were transfected with 0.1 μg of hSERT-N1-pEYFP (Addgene #70105) per well. Wells were washed (1 × 200 μL) with 1x Hank’s Balanced Salt Solution supplemented with 2 mM MgCl₂ and 2 mM CaCl₂ (HBSS) as well as 5 mM HEPES and replenished with 100 μL of supplemented HBSS. The plate was placed in a 37 °C water bath for the reminder of the experiment and allowed to incubate for 30 min prior to beginning the experiment. Next, 100 μL of HBSS containing 80 nM 5-[1,2-³H(N)]-hydroxytryptamine creatinine sulfate ([³H]5-HT; #NET498001MC, Lot: 3261835, Revvity) was added to the wells at a final concentration of 40 nM and incubated for 30 min to allow uptake into SERT-expressing cells. Uptake was terminated by aspiration of the wells followed by washing with supplemented HBSS (2 × 200 μL) and replenished with 180 μL of supplemented HBSS. Efflux was then initiated by adding 20 μL of drug (316 nM – 1 mM, 1% DMSO) solution in supplemented HBSS for 20 min at final concentrations ranging from 31.6 nM – 100 μM (0.1% DMSO). Following the incubation, wells were washed with supplemented HBSS (3 × 200 μL). Cells were then lysed for 30 min by the addition of 30 μL of 10 mM NaOH, and 120 μL OptiPhase HiSafe (#1200.437; Revvity) was added to the cell lysates. Counts per minute (CPM) were quantified using MicroBeta2 microplate liquid scintillation counter. In GraphPad Prism, concentration responses were plotted as the difference of the VEH (0.1% DMSO) average for the plate and the count per minute (CPM) for each respective drug concentration (VEHavg – drug). The plots were normalized similarly, with the plate vehicle control (VEHavg – VEHavg) average being set to 0% efflux and the difference between the VEH average and the positive control average (VEHavg – 100 uM PCAavg) set to 100%. Outliers were removed by using the ROUT method (Q = 10%) for each column separately.

Extended Data

Extended Data Fig. 1. Regioselective reduction of 3-ethylpyridine.

Product distribution determined by integration of LC-MS spectra obtained using positive ionization mode. All reactions were run at 0.1 M. While compound 11 was isolable, compound 10 could never be isolated and was assigned based on its mass and the fact that mixtures of 10 and 11 would convert to 11 over time. ^a Both methanol and ethanol were tested.

Extended Data Fig. 2. Optimization of 7-membered ring closure.

Isolated yields are shown. All reactions were conducted on a 1 mmol scale at a concentration of 0.1 M using a 1:1 mixture of S1 and S2 and purified via silica gel chromatography (gradient elution, 20:1→10:1 DCM/MeOH). Product S4 was isolated as a mixture of endo and exo epimers.

Extended Data Fig. 3. Optimization of MHAT coupling.

Isolated yields are shown. All reactions were conducted on a 1 mmol scale and purified via silica gel chromatography (gradient elution, 10:1→7:3 hexanes/EtOAc). Products S5 and S6 were isolated as a mixture of C16 endo and exo epimers.

Extended Data Fig. 4. Alternative hydroethylation strategy.

Isolated yields are shown. All reactions were conducted on a 1 mmol scale and purified via silica gel chromatography (gradient elution, 10:1→7:3 hexanes/EtOAc). Products S6, S7, and S8 were tentatively assigned based on LC-MS analysis and isolated as a mixture of endo and exo epimers. PC = propylene carbonate

Extended Data Fig. 5. Optimization of a photoredox-catalyzed decarboxylation.

Isolated yields are shown. All reactions were conducted on a 0.64 mmol scale and purified via silica gel chromatography (gradient elution, 10:1→7:3 hexanes/EtOAc). DIPEA – N,N-diisopropylethylamine, TRIP thiol = 2,4,6-Triisopropylbenzenethiol.

Extended Data Fig. 6. Comparison of ¹H NMR data obtained from natural and synthetic ibogaine.

(a) ¹H NMR data demonstrates that synthetic and natural ibogaine are indistinguishable. (b) a 1:1 molar ratio of natural ibogaine and CH₂Br₂ was treated with 1 equiv. of synthetic ibogaine. Spiking increased the ibogaine signal integration without impacting that of CH₂Br₂. Natural ibogaine was obtained from the National Institute on Drug Abuse (NIDA) as the hydrochloride salt. Natural Ibogaine • HCl was basified using 1 M NaOH and DCM as the extraction solvent.

Extended Data Fig. 7. Comparison of infrared spectroscopy data obtained from natural and synthetic ibogaine.

Infrared spectroscopy data demonstrate that synthetic and natural ibogaine are indistinguishable. Natural ibogaine was obtained from the National Institute on Drug Abuse (NIDA) as the hydrochloride salt. Natural Ibogaine • HCl was basified using 1 M NaOH and DCM as the extraction solvent.

Extended Data Fig. 8. Total synthesis of iboga alkaloids.

The overall yields and step counts for the total syntheses of various iboga alkaloids from commercially available starting materials are shown. Principal investigators are indicated by color.

Extended Data Fig. 9. Efforts towards an enantioselective synthesis of iboga alkaloids.

(a) Attempts to achieve an enantioselective Diels-Alder reaction using chiral acid catalysts were unsuccessful. (b) Attempts to achieve an enantioselective Diels-Alder reaction using chiral organocatalysts were unsuccessful. (c) Placing a chiral menthol-derived auxiliary on the diene did not lead to any diastereoselectivity in the Diels-Alder reaction. Lack of diastereoselectivity was confirmed by converting S13 to desethylibogaine and then performing chiral HPLC analysis. (d) Placing a chiral auxiliary on the dienophile resulted in a range of diastereoselectivities depending on the reaction conditions. (e) Chiral HPLC analysis revealed that enantiopure S19 could be obtained following a diastereoselective Diels-Alder reaction of S15 and subsequent functional group interconversions.

Data Availability

Data are available at the following link https://doi.org/10.6084/m9.figshare.24531316.