Re-Engineering of Yohimbine’s Biological Activity through Ring Distortion: Identification and Structure–Activity Relationships of a New Class of Antiplasmodial Agents

Abstract

Select natural products are ideal starting points for ring distortion, or the dramatic altering of inherently complex molecules through short synthetic pathways, to generate an array of novel compounds with diverse skeletal architectures. A major goal of our ring distortion approach is to re-engineer the biological activity of indole alkaloids to identify new compounds with diverse biological activities in areas of significance to human health and medicine. In this study, we re-engineered the biological activity of the indole alkaloid yohimbine through ring rearrangement and ring cleavage synthesis pathways to discover new series of antiplasmodial agents. One new compound, Y7j, was found to demonstrate good potency against chloroquine-resistant Plasmodium falciparum Dd2 cells (EC₅₀ = 0.33 μM) without eliciting cytotoxicity against HepG2 cells (EC₅₀ > 40 μM). Y7j demonstrated stage-specific action against parasites at the late ring/trophozoite stage. A series of analogues was synthesized to gain structure—activity relationship insights, and we learned that both benzyl groups of Y7j are required for activity and fine-tuning of antiplasmodial activities could be accomplished by changing substitution patterns on the benzyl moieties. This study demonstrates the potential for ring distortion to drive new discoveries and change paradigms in chemical biology and drug discovery.

Graphical Abstract

Malaria continues to be a major global health concern. Malaria is a mosquito-borne infectious disease caused by a variety of Plasmodium parasites, with Plasmodium falciparum being the most deadly to humans. In 2017, the World Health Organization (WHO) reported 219 million new cases of malaria, an increase of 2 million from 2016.¹ Additionally, malaria-related deaths totaled 435,000, resulting in $12 billion per year with associated healthcare costs. Current clinically used antiplasmodial therapeutics have been developed from drugs that were initially discovered over 30 years ago and primarily belong to one of three drug classes: artemisinins, aminoquinolines, or antifolates.² To make matters worse, each of these drug classes have garnered resistance from the P. falciparum parasite.³ In several parts of the world where malaria has reached epidemic proportions, artemisinin-based combination therapies (ACTs) used for uncomplicated P. falciparum are showing increasing signs of ineffectiveness.^4,5 New antimalarial agents that operate through novel mechanisms are needed to combat widespread resistance to existing drugs.^6,7

Natural products have been a primary source of therapeutic agents due to their outstanding ability to bind and modulate numerous biological targets critical to disease.^8,9 Complex natural products, such as taxol, morphine, and vancomycin, form highly specific binding modes with their respective protein targets due to their complex and chiral molecular architectures that display an array of functional groups critical for such interactions. Despite the many advances and curative natural products that have been introduced into the clinic, there has been a paradigm shift to the high-throughput screening (HTS) of synthetic compound libraries as the primary approach driving current drug discovery efforts.^10–12 It is well established that these compound libraries are composed of structurally simple organic molecules, which have served well for drugging certain biological targets (e.g., kinases); however, these compound libraries lack chemical diversity¹³ and have been implicated in failures to identify viable therapeutic lead compounds for more sophisticated targets (e.g., protein—protein interactions,¹⁴ transcription factors¹⁵).

Innovative strategies to address chemical diversity concerns by increasing the population of complex small molecules within screening libraries have been pursued. The most established approaches include diversity-oriented synthesis (DOS)^16–18 and biology-oriented synthesis (BIOS),^19,20 which have been pioneered by Schreiber and Waldmann, respectively. These strategies aim at generating complex libraries of small molecules by building structural complexity in subsequent synthetic reactions, starting from simple building blocks, and have unveiled a variety of biological discoveries of therapeutic relevance. Interestingly, Schreiber and co-workers have discovered multiple series of new antimalarial agents using DOS.^21–23

Complementary approaches to DOS and BIOS have been developed,^24–27 including Complexity-to-Diversity/ring distortion of available natural products.^28–35 In this approach, natural products are subjected to an array of chemoselective reactions aimed at ring distortion, or the dramatic altering of inherently complex ring systems of select natural products, to generate architecturally unique scaffolds. A major goal of ring distortion is to identify complex small molecules that display biological activity that is distinct from the parent natural product and is of significance in disease and human health. Ring distortion efforts have been reported for a diversity of natural products, including: gibberellic acid,²⁸ abietic acid,²⁹ pleuromutulin,^30,34 quinine,²⁸ sinomenine,³¹ yohimbine,³² lycorine,³³ adrenosterone,²⁸ and other steroids.³⁶

Our laboratory recently reported a tryptoline-enabled ring distortion campaign of yohimbine, a complex indole alkaloid that is commercially available on the decagram scale.^32,35 Utilizing the extended tryptoline substructure of yohimbine, we were able to rapidly generate an array of unique, complex, and diverse small molecules through a series of highly selective ring-cleavage and oxidative indole rearrangement reactions. Using this approach, our ring distortion efforts have culminated in a growing library of small molecules, which are being subjected to various biological screens to identify new active scaffolds relevant to important disease areas.^32,37 Our work presented here details the identification and structure—activity relationship (SAR) studies regarding a new series of antiplasmodial agents derived from the ring distortion of yohimbine.

RESULTS AND DISCUSSION

An initial library of 70 yohimbine-derived ring-distorted compounds was screened for antiplasmodial activity against a chloroquine-resistant P. falciparum strain (Dd2) in an unbiased, cell-based SYBR Green I assay^38–40 (Figure 1). From the initial screen, compounds from the Y1 and Y7 series demonstrated inhibitory activity against Dd2 cells (>40% at 2 μM). Validation of active hits and structurally related analogues were subsequently subjected to dose–response experiments to generate comparative EC₅₀ values to determine overall potency and gain initial insights into SAR. Interestingly, all hit compounds arose from two of the ring-distorted scaffolds accessed from yohimbine, including the Y1 series (6 analogues; synthesized in three steps from yohimbine; EC₅₀ = 1.60–5.18 μM against Dd2) and the Y7 series (8 analogues; synthesized in four steps from yohimbine; EC₅₀ values ranging from 0.33 to 2.57 μM against Dd2) (Figure 2).

Figure 1.

Complex and diverse scaffolds rapidly synthesized from the indole alkaloid yohimbine. Initial hit compounds possessing antiplasmodial activities are highlighted; however, yohimbine was found to demonstrate no antiplasmodial activity. The remainder of the yohimbine-derived scaffolds (Y2–Y6) were found to be inactive against P. falciparum Dd2 cells during initial screens.

Figure 2.

Antiplasmodial activities of Y1 and Y7 series analogues, which were found to be active during initial screens against Dd2 cells. Testing of related analogues provides initial SAR insights and informs follow-up chemical synthesis investigations.

From the initial screen, dibenzylated spiro-oxindole analogue Y7j proved to be the most potent analogue of the library (EC₅₀ = 0.33 μM, Figure 2) and was also found to have no cytotoxicity against HepG2 cells at 40 μM (highest test concentration), demonstrating a clear selectivity for plasmodial cell targeting (selectivity index >121; determined by [Hep G2, EC₅₀]/[Dd2, EC₅₀]). In addition, the parent indole alkaloid yohimbine was completely inactive against Dd2 cells, EC₅₀ of 76.7 μM, demonstrating a re-engineering of its biological activities that occurred through the dramatic altering of its molecular architecture.

To further investigate the antiplasmodial activity of Y7j, we evaluated the developmental stage-specific action of this small molecule by microscopy and flow cytometry.^41,42 Precise delineation of the timing of action of an inhibitor provides valuable insight into the developmental growth and clinical clearance of the parasite. Tightly synchronized parasites were treated at 6 h post-invasion of the merozoites with a 5× EC₅₀ concentration of Y7j (Figure 3A). Microscopic evaluation of Giemsa-stained-thin smears and flow cytometric assessments were done at 12 h intervals. Negative controls represented infected red blood cells exposed to vehicle DMSO (0.1%) only. As can be seen from Figure 3A, compared to untreated cultures, Y7j inhibited the parasite’s late ring/trophozoite stage (18 h post-invasion, HPI).⁴³ The untreated control culture matures as expected through the trophozoite/schizont (30 HPI) and segmenter (42 HPI) and reappears as rings after reinvasion (54 HPI) with a concomitant increase in parasitemia (see flow cytometry data, Figure 3B,C). Although at the trophozoite stage (30 h) the parasite mass increased, there was significant vacuolization noticed in the parasites exposed to Y7j. Interestingly, Y1f also demonstrated stage-specific action at the late ring/trophozoite phase in Dd2 cells and vacuolization, similar to Y7j (see the Supporting Information). These two ring distortion scaffolds, Y1 and Y7, are very different from each other structurally and likely operate through different primary modes of action against plasmodial cells.

Figure 3.

Characterization of stage-specific activity of Y7j in Dd2 cells. Tightly synchronized Dd2 cultures were treated at 6 h time points with Y7j at 5× EC₅₀. Blood smears and Giemsa staining were also taken at each time point to verify (A) phenotype and (B) flow cytometry. Flow cytometry was performed to quantify the stage of action for Y7j at the time of harvesting, following a 6 HPI treatment. For flow cytometric analysis, collected samples were fixed in 0.04% glutaraldehyde in PBS, permeabilized with 0.25% Triton X-100, treated with RNase (50 μg/mL), and stained with YOYO-1. Data acquisition was performed in a CytoFLEX (Beckman) flow cytometer. (C) Parasitemia of Y7j treated and vehicle untreated cultures. Notes: Dihydroartemisinin (DHA), an inhibitor of the late ring/trophozoite stage, was tested as a positive control alongside Y7j (not shown). HPI, hours post-invasion of red blood cells by merozoites.

To determine if Y7j elicits it’s antiplasmodial activity through a parasitocidal or parasitostatic mechanism, we performed a series of kill kinetic experiments in Dd2 cultures.^41,42 We treated asynchronous parasite cultures with 5× EC₅₀ concentration of the compound for different periods of time (6, 12, 24, and 48 h). After washing out the compound at these time points, the growth of the culture was evaluated for 96 h. As is evident from these kill kinetic experiments (Figure 4), the most significant reduction in parasitemia observed following a 24 h treatment with Y7j suggests this compound is a fast-acting parasitocidal agent with rapid clearance of the remaining parasites 48 h post-treatment. Similar to Y7j, significant killing of Dd2 cells was observed following a 24 h treatment of Y1f (see the Supporting Information).

Figure 4.

Rate of killing and parasitocidal/parasitostatic determination of Y7j in Dd2 cultures. To evaluate the rate of killing for Y7j, asynchronous cultures were treated with 5× EC₅₀ (Y7j concentrations) for (A) 6, (B) 12, (C) 24, and (D) 48 h. After each treatment, cultures were washed three times in RPMI, then resuspended in culture media, and monitored daily for parasite growth. Note: DMSO concentration was 0.15% for these experiments. Dihydroartemisinin (DHA) was used as a positive control for rapid parasite killing (test concentration = 50 nM; 5× EC₅₀) in these experiments.

With the interesting antiplasmodial activity we discovered during the initial investigations, we worked to optimize the synthetic route to Y7j and related analogues from yohimbine (Y). Our goals were to explore the structure–activity relationship profiles related to the Y7 series and, ideally, improve antiplasmodial activities. We were able to utilize our previous protocol³² to scale up the synthesis of Y7a (925 mg) from Y via treatment with N-chlorosuccinimide (NCS) and subsequent addition of sodium methoxide (NaOMe), which afforded 925 mg of Y7a in a single run (76% yield over 2 steps, Figure 5). Y7a proved to be an important intermediate as we strategically used the protected amide to selectively monoalkylate the free hydroxyl group of Y7a using sodium hydride and benzyl bromides to synthesize a focused series of compounds for antiplasmodial investigations, following treatment with aqueous trifluoroacetic acid (analogues Y7aa to Y7ad, average yield = 52% over two steps; Figure 5). In addition, Y7a was also used to generate key intermediate Y7b upon treatment with trifluoroacetic acid (TFA; 96% yield, scaled up to 1.63 g of Y7b in a single run). With both Y7a and Y7b in hand, our goals were to synthesize focused subclasses of analogues to gain detailed SAR information for these new antiplasmodial agents. This chemistry provided synthetic routes to diverse mono- and dibenzylated analogues of Y7j for biological evaluation.

Figure 5.

Chemical synthesis of spirocyclic antiplasmodial agents (Y7j analogues), general SAR trends, and the top four most potent analogues from these studies are highlighted.

Symmetrically functionalized derivatives of Y7j were synthesized by treating Y7b with sodium hydride, followed by two equivalents of (hetero)benzyl bromides to afford analogues Y7m–Y7v (average yield = 46%; 10 analogues). This allowed for the incorporation of various heterocyclic bioisosteres⁴⁴ of the phenyl moieties in Y7j, including: pyridines (Y7o, Y7q), pyrazines (Y7t, Y7v), thiophene (Y7p), and furan (Y7r) heterocycles (Figure 5). Complementary diversifications were carried out through mono-N-alkylation of Y7b via treatment with various alkyl bromides and potassium carbonate in N,N-dimethylformamide to afford Y7ae–Y7ai (average yield = 76%; 5 analogues). The free secondary hydroxyl of the resulting amides (Y7ae–Y7ai) was then deprotonated with sodium hydride and alkylated with (hetero)aromatic benzyl bromides to generate mixed functionalized derivatives Y7aj–Y7am (average yield = 71%; 4 analogues).

With new Y7j and Y1f analogues in hand (see the Supporting Information for additional Y1f analogues; Y1i, Y1j), all compounds were subjected to dose–response experiments to determine EC₅₀ values against Dd2 cells. The two new Y1f analogues were found to be inactive (Y1i, Y1j; EC₅₀ > 20 μM), and analogues related to the Y1 series were no longer pursued during these investigations. On the other hand, six new Y7j analogues demonstrated submicromolar antiplasmodial activities with thiophene analogue Y7p (EC₅₀ = 0.32 μM) and 2-chlorobenzyl analogue Y7n (EC₅₀ = 0.52 μM) being the most potent new analogues synthesized against Dd2 cells (Figure 5 and Table 1).

Table 1.

Summary of Antiplasmodial Activities and Cytotoxicity against HepG2 (Liver) Cells for Y and Select Y1 and Y7 Analogues^a

Y code	EC50 Dd2 cells (μM)	EC50 3D7 cells (μM)	EC50 HepG2 cells (μM)	selectivity index (SI)	stage-specific activity
Y	76.7 ± 0.74	-	>100	-	n.a.
Y1c	2.58 ± 0.07	-	-	-	-
Y1e	2.98 ± 0.20	-	-	-	-
Y1f	1.60 ± 0.16	-	18.9	-	A
Y1g	3.18 ± 0.11	-	-	-	-
Y1h	3.47 ± 0.11	-	-	-	-
Y7e	1.50 ± 0.09	-	-	-	-
Y7f	1.41 ± 0.23	-	-	-	-
Y7g	0.79 ± 0.06	0.74 ± 0.07	>40	>51	A
Y7h	1.05 ± 0.08	-	-	-	-
Y7i	0.85 ± 0.05	-	-	-	-
Y7j	0.33 ± 0.03	0.35 ± 0.03	>40	>121	A
Y7k	0.64 ± 0.07	0.68 ± 0.07	-	-	-
Y7l	2.57 ± 0.05	-	-	-	-
Y7m	1.51 ± 0.06	-	-	-	-
Y7n	0.52 ± 0.04	0.56 ± 0.04	>40	>77	-
Y7o	3.51 ± 0.32	-	-	-	-
Y7p	0.32 ± 0.02	0.49 ± 0.05	>40	>125	A
Y7q	0.87 ± 0.04	0.91 ± 0.03	>40	>46	-
Y7r	0.65 ± 0.07	0.68 ± 0.09	>40	>62	-
Y7s	1.32 ± 0.07	1.13 ± 0.09	-	-	-
Y7t	3.08 ± 0.20	4.31 ± 0.16	-	-	-
Y7u	0.89 ± 0.05	0.80 ± 0.09	>40	>45	-
Y7aj	2.38 ± 0.24	-	-	-	-
Y7ak	0.94 ± 0.10	-	-	-	-
Y7am	2.71 ± 0.07	3.00 ± 0.13	-	-	-

^aAll concentrations are reported in micromolar (μM). Notes: “A” = stage-specific activity at the late ring/trophozoite phase of the asexual blood stage of P. falciparum. “n.a.” = yohimbine (Y), which is inactive against P. falciparum. “-” = noted for analogues that were not tested for stage-specific activity. Selectivity index (SI) was generated according to EC₅₀ against HepG2 cells divided by EC₅₀ against Dd2 cells. All values reported in this table resulted from three independent experiments.

From these studies, an interesting structure–activity relationship profile was revealed for the new antiplasmodial agent Y7j and related analogues. The most prominent structural feature regarding antiplasmodial activity of Y7j is the requirement of both 2-chloro-4-methoxybenzyl groups, as removal of either group led to a complete loss in antiplasmodial activity (e.g., Y7ab, EC₅₀ > 20 μM; Figure 6). This SAR information could not have been obtained without the new synthetic efforts described in Figure 4, which allowed for a complementary series of monobenzylated analogues for biological investigations. In total, seven monoalkylated/benzylated analogues of Y7j (containing either a free amide or a free hydroxyl group) were investigated during these studies, and each of these analogues reported an EC₅₀ > 20 μM (e.g., Y7aa and Y7ae). In addition, subtle structural changes were made to Y7j through alkylation of Y7b (Figure 5) with alternative benzyl and heterobenzyl starting materials. Removal of the methoxy groups at the 4-position of both benzyl moieties of Y7j resulted in a 1.6-fold loss in antiplasmodial activity against Dd2 cells (Y7n, EC₅₀ = 0.52 μM), while removal of the 2-chlorine atom of Y7j resulted in a 4.5-fold loss in antiplasmodial activity (Y7m, EC₅₀ = 1.51 μM).

Figure 6.

Detailed structure-activity relationship profiles regarding the antiplasmodial activities of Y7j.

Additional substitutions with six-membered heterocycles containing nitrogen also resulted in losses in activity compared to Y7j (see Figure 6 for detailed SAR profiling). Interestingly, the only new compound to demonstrate equipotent activity against Dd2 cells compared to Y7j was thiophene analogue Y7p (EC₅₀ = 0.32 μM; demonstrated stage-specific activity at the late ring/trophozoite phase of asexual blood stage against Dd2 cells, similar to Y7j; Table 1), as previously mentioned. In addition to SAR, seven Y7 analogues were counter-screened in dose–response experiments against wild-type 3D7 P. falciparum cells, and each analogue demonstrated near equipotent activity when comparing EC₅₀ values to those generated against chloroquine-resistant Dd2 cells (Table 1). Additionally, the same seven Y7 analogues demonstrated no cytotoxicity against HepG2 cells (EC₅₀ > 40 μM; selectivity indexes >45 to >125; Table 1). These combined analyses point to a clear SAR, and we hypothesize that the structural requirements for active analogues related to Y7j result from such analogues interacting with a well-defined biological target. On the basis of our SAR profile, we predict these agents to elicit their activities against a new antiplasmodial target. Additional mechanistic studies are required to substantiate this hypothesis and are currently being pursued.

In conclusion, we identified two diverse and complex scaffolds that possess antiplasmodial activities from ring distortion efforts of the indole alkaloid yohimbine. One new antiplasmodial compound, Y7j, was found to demonstrate good potency against chloroquine-resistant P. falciparum Dd2 cells (EC₅₀ = 0.33 μM) without eliciting cytotoxicity against HepG2 cells (EC₅₀ > 40 μM; selectivity index >121). In addition, Y7j demonstrated stage-specific action against P. falciparum at the late ring/trophozoite stage, preventing an increase in parasitemia as determined by flow cytometry, and rapid parasitocidal activity against Dd2 cultures following 24 h of treatment. A series of focused analogues were also synthesized to gain detailed structure–activity relationship insights, and we learned that both benzyl groups of Y7j are required for activity and that fine-tuning of antiplasmodial activities could be accomplished by changing substitution patterns on the benzyl moieties; however, the only new analogue to demonstrate equipotent activity to Y7j was thiophene bioisostere Y7p. Interestingly, the parent indole alkaloid yohimbine demonstrated no antiplasmodial activities, and our active scaffolds are the result of a re-engineering of yohimbine’s biological activity. This work demonstrates the potential for the ring distortion of indole alkaloids to drive new discoveries in chemical biology and drug discovery.

METHODS

Please see the Supporting Information for the details of the experimental methods and chemical synthesis protocols.

ABBREVIATIONS

3D7

chloroquine-sensitive P. falciuarum strain

ACTs

artemisinin-based combination therapies

BIOS

biology-oriented synthesis

Dd2

chloroquine-resistant P. falciuarum strain

DHA

dihydroartemisinin

DMF

dimethylformamide

DMSO

dimethyl sulfoxide

DOS

diversity-oriented synthesis

EC₅₀

concentration of a compound that gives half-maximal response

equiv

equivalents

HepG2

a human liver cancer cell line

HPI

hour post-invasion

HTS

high-throughput screening

K₂CO₃

potassium carbonate

mg

milligrams

NaH

sodium hydride

NaOMe

sodium methoxide

NCS

N-chlorosuccinimide

nM

nanomolar

SAR

structure–activity relationship

SI

selectivity index

TFA

trifluoroacetic acid

THF

tetrahydrofuran

μM

micromolar

WHO

World Health Organization