Anibamine and its Analogues; Potent Antiplasmodial Agents from Aniba citrifolia

Abstract

In our continuing search for novel natural products with antiplasmodial activity, an extract of Aniba citrifolia was found to have good activity, with an IC₅₀ value less than 1.25 μg/mL. After bioassay-directed fractionation, the known indolizinium alkaloid anibamine (1) and the new indolizinium alkaloid anibamine B (2) were isolated as the major bioactive constituents, with antiplasmodial IC₅₀ values of 0.170 and 0.244 μM against the drug-resistant Dd2 strain of Plasmodium falciparum. The new coumarin anibomarin A (3), the new norneolignan anibignan A (5), and six known neolignans (7 – 12) were also obtained. The structures of all the isolated compounds were determined based on analyses of 1- and 2-D NMR spectroscopic and mass spectrometric data, and the absolute configuration of anibignan A (5) was assigned from its ECD spectrum. Evaluation of a library of 28 anibamine analogues (13 – 40) indicated that quaternary charged analogs had IC₅₀ values as low as 58 nM, while uncharged analogs were inactive or significantly less active. Assessment of the potential effects of anibamine and its analogues on the intraerythrocytic stages and morphological development of P. falciparum revealed substantial activity against ring stages for compounds with two C-10 side chains, while those with only one C-10 side chain exhibited substantial activity against trophozoite stages, suggesting different mechanisms of action.

Graphical Abstract

Malaria remains a serious disease, especially in Africa. A 2018 report from the World Health Organization estimated that it caused 435,000 deaths worldwide, with 61% of these fatalities occurring in children under 5 years of age.¹ The first antimalarial drug was the natural product quinine, but this was largely replaced in the 1940’s by its derivative chloroquine. However, resistance of the malaria parasite Plasmodium falciparum to chloroquine developed as early as 1960,² and the current drug of choice is the natural product artemisinin³ and its analogues such as artemether, artesunate, or dihydroartemisinin, which are used as part of an artemisinin-based combination therapy (ACT).⁴ Unfortunately, the malaria parasite P. falciparum has begun to develop resistance to artemisinin in the Greater Mekong subregion.¹ Thus, there is an urgent need to search for new antimalarial drugs.

In our continuing collaboration to search for novel antiplasmodial agents from plants,^5,6 an extract of Aniba citrifolia (Nees) Mez (Lauraceae) was found to have good antiplasmodial activity against the chloroquine-resistant Dd2 strain of P. falciparum, with an IC₅₀ value less than 1.25 μg/mL. The genus Aniba, which includes at least 48 species, is a neotropical flowering plant genus widely distributed in the Caribbean and Central and South America.⁷ Members of the genus are good sources of flavonoids, alkaloids, and neolignans⁸ with a broad range of bioactivities, including antimicrobial,⁹ chemokine receptor agonist,¹⁰ and cytotoxic activities.^8,9,11,12 However, except for one report on the antimalarial activity of synthetic aniba A-type dimers,¹³ there have been no reports of antimalarial compounds from members of the genus, so this extract was thus selected for investigation of its antiplasmodial constituents.

RESULTS AND DISCUSSION

Isolation and Structure Elucidation.

An extract of the whole plant of Aniba citrifolia was subjected to liquid-liquid partition to afford hexanes, dichloromethane and methanol/H₂O fractions. All three fractions displayed in vitro antiplasmodial activity against the drug-resistant Dd2 strain of P. falciparum, with IC₅₀ values <1.25 μg/mL. Bioassay-directed fractionation with final separation by HPLC on a C₁₈ column with a MeOH-H₂O-TFA solvent system led to the isolation of the indolizinium alkaloids anibamine (1) and anibamine B (2) as the most potently antiplasmodial compounds. The new coumarin anibomarin A (3), related to the known compound licocoumarin A (4), had weak antiplasmodial activity, while the new norneolignan anibignan A (5) was essentially inactive against P. falciparum Dd2 strain (Chart 1).

Chart 1.

Compound 1 was identified as the known alkaloid anibamine by comparison of its spectroscopic data with literature values.^9,10 Compound 2 had the molecular formula C₃₀H₅₂N⁺, based on the molecular ion peak at m/z 426.4094 [M]⁺ in its positive ion HRESIMS. Comparison of the ¹H and ¹³C NMR spectroscopic data of 2 (Table 1) with those of 1 indicated that both compounds had the same skeleton, but compound 2, with a molecular weight 2 Da larger than that of 1, was an analogue of 1 with one of the side chains lacking a double bond.

Table 1.

NMR Spectroscopic Data for Compound 2 in CDCl₃ (500 MHz)^a

	2
position	δH (J in Hz)	δC, type	position	δH (J in Hz)	δC, type
1	3.47 br	31.7, CH	3′	2.28 m	33.2, CH2
2	2.53 br	20.6, CH	4′	1.50 m	28.7, CH2
3	4.86 br	58.0, CH	5′−7′	1.27 m	29.4, CH2
5		147.3, C	8′	1.27 m	31.9, CH2
6		137.2, C	9′	1.29 m	22.6, CH2
7		154.9, C	10′	0.88 t 6.8	14.0 CH3
8		134.9 C	1″	2.67 m	30.1, CH2
9		153.6 C	2″	1.50 m	28.7, CH2
10	2.69 s	18.4 CH3	3″−7″	1.27 m	29.4, CH2
11	2.38 s	17.7 CH3	8″	1.27 m	31.9, CH2
1’	6.22 d 16.0	122.4 CH	9″	1.29 m	22.6, CH2
2’	5.78 dt 16.0, 7.0	141.9 CH	10″	0.88 t 6.8	14.0 CH3

^aA ¹³C NMR spectrum was not obtained because of the limited amount of sample. ¹³C shifts were determined by analysis of HSQC and HMBC spectra.

The side chain with the double bond was assigned to the C-6 position, based on the ²J-HMBC correlation between H-1’ and C-6 and the ³J-HMBC correlations between H-1′ and C-5 and C-7. The ²J-HMBC correlations between H-1″ and C-8 and the ³J-HMBC correlation between H-1″ and C-9 confirmed that the saturated side chain is linked to C-8. These assignments were confirmed by the NOESY correlations between H-1 and H-1″, H-1″ and H-11, H-11 and H-1′, H-1′ and H-10, H-10 and H-3 (Figure 1). The configuration of the C-1′/C-2′ double bond was assigned as E based on the NOESY correlations between H-1′ and H-3′ (Figure 1), and the structure of anibamine B was thus assigned as (E)-6-(dec-1′-en-1′-yl)-8-decyl-5,7-dimethyl-2,3-dihydro-1H-indolizin-4-ium trifluoroacetate.

Figure 1.

HMBC (red) and NOESY (blue) correlations of compound 2

Anibomarin A (3) was isolated as a yellow oil. Its positive-ion HRESIMS revealed a peak for a protonated molecular ion at m/z 571.3412 and for a sodiated molecular ion at m/z 593.3229, both corresponding to the molecular formula of C₃₇H₄₆O₅ for 3. Its UV spectrum, with absorptions at 288 and 355 nm, suggested that it is a 3-arylcoumarin derivative.¹⁴ Its ¹H NMR and HSQC spectra (Table 2) displayed signals for two methoxy groups (δ_H 3.79 and 3.85), six vinyl methyl groups [δ_H 1.75, H-9″ (3H); 1.74, H-9‴ (3H); 1.68, H-8″ and H-8‴ (6H); 1.60, H-10″ and H-10‴ (6H)], six methylene protons [δ_H 3.61, H-1″ (1H); 3.45, H-1‴ (1H); 2.12, H-5″ and H-5‴ (2H); 2.07, H-4″ and H-4‴ (2H)], five aromatic methines (δ_H 7.87, H-4; 7.05, H-7; 6.84, H-4′; 6.83, H-5; 6.67, H-6′), four vinyl methines (5.36, H-2‴; 5.35, H-2″; 5.11, H-6‴; 5.10, H-6″) and a hydroxy group (δ_H 7.15). The presence of a 3-phenylcoumarin skeleton was indicated by ³J-HMBC correlations between H-6′ and C-3 (δ_C 127.7) and H-4 and C-1′ (δ_C 124.1), and supported by the correlations between H-4′ and C-2′ (δ_C 146.9), H-4′ and C-6′ (δ_C 113.1), H-6′ and C-2′ (δ_C 146.9), H-4 and C-2 (δ_C 163.1), H-4 and C-8a (δ_C 146.1), H-4 and C-5 (δ_C 107.1), H-5 and C-8a (δ_C 146.1), H-7 and C-5 (δ_C 107.1), H-7 and C-8a (δ_C 146.1).¹⁵ These data were consistent with the structure proposed for anibomarin A, similar to that of licocoumarin A (4) isolated from Glycyrrhiza glabra (Egyptian licorice).¹⁶

Table 2.

NMR Spectroscopic Data for Compound 3 in CDCl₃ (500 MHz)

	3
position	δH (J in Hz)	δC, type	Position	δH (J in Hz)	δC, type
2		163.1, C	2″	5.35 (m)	120.3, CH
3		127.7, C	3″		138.5, C
4	7.87 (s)	144.3, CH	4″	2.07 (m)	39.9, CH2
4a		119.8, C	5″	2.12 (m)	26.9, CH2
5	6.83 (d 3.0)	107.1, CH	6″	5.10 (m)	124.2, CH
6		156.5, C	7″		131.7, C
7	7.05 (d 3.0)	120.7, CH	8″	1.68 (s)	25.8, CH3
8		131.5, C	9″	1.75 (s)	16.3, CH3
8a		146.1, C	10″	1.60 (s)	17.9, CH3
1’		124.1, C	1‴	3.45 (d 7.3)	29.2, CH2
2’		146.9, C	2‴	5.36 (m)	122.1, CH
3’		133.1, C	3‴		137.2, C
4’	6.84 (d 3.0)	116.9, CH	4‴	2.07 (m)	39.9, CH2
5’		153.8, C	5‴	2.12 (m)	26.7, CH2
6’	6.67 (d 3.0)	113.1, CH	6‴	5.11 (m)	124.4, CH
6-OMe	3.85 (s)	55.9, CH3	7‴		131.8, C
5′-OMe	3.79 (s)	55.9, CH3	8‴	1.68 (s)	25.8, CH3
2′-OH	7.15 (s)		9‴	1.74 (s)	16.4, CH3
1″	3.61 (d 7.3)	27.6, CH2	10‴	1.60 (s)	17.9, CH3

The hydroxy group of 3 was assigned to C-2′ (δ_C 146.9), based on the observed ³J-HMBC correlations between OH and C-1′ (δ_C 124.1) and OH and C-3′ (δ_C 133.1). The two methoxy groups were assigned to C-6 (δ_C 156.5) and C-5′ (δ_C 153.8) according to the ³J-HMBC correlations between OCH₃ (δ_H 3.85) and C-6, and OCH₃ (δ_H 3.79) and C-5′. The remaining 20 carbons and 34 protons composed two geranyl side chains, as confirmed by ³J-HMBC correlations between H-1″ and C-3″, H-2″ and C-9″, H-2″ and C-4″, H-9″ and C-4″, H-4″ and C-6″, H-5″ and C-3″, H-5″ and C-7″, H-6″ and C-8″, H-6″ and C-10″, H-10″ and C-8″; H-1‴ and C-3‴, H-2‴ and C-9‴, H-2‴ and C-4‴, H-9‴ and C-4‴, H-4‴ and C-6‴, H-5‴ and C-3‴, H-5‴ and C-7‴, H-6‴ and C-8‴, H-6‴ and C-10‴, H-10‴ and C-8‴. These two identical side chains were linked to C-8 and C-3′, as indicated by ³J-HMBC correlations between H-1″ and C-7, H-1″ and C-8a, H-1″ and C-2′, H-1″ and C-4′ (Figure 2).

Figure 2.

HMBC (red) and NOESY (blue) correlations of anibomarin A (3)

The configurations of the C-2′ and C-2‴ carbon-carbon double bonds were assigned as E based on the NOESY crosspeaks between H-1″ and H-9″, H-2″ and H-4″, H-1‴ and H-9‴, H-2‴ and H-4‴ (Figure 2). Anibomarin A (3) was thus assigned as 8-((2″E)-3″,7″-dimethylocta-2″,6″-dien-1″-yl)-6-methoxy-3-(3′-((2‴E)-3‴,7‴-dimethylocta-2‴,6‴-dien-1‴-yl-2‱-hydroxyl-5″- methoxyl)-phenylcoumarin.

Anibignan A (5) was isolated as a yellow oil. Its positive-ion HRESIMS revealed a peak for a protonated molecular ion at m/z 331.1543 and for a sodiated molecular ion at m/z 353.1372, both corresponding to a molecular formula of C₁₉H₂₂O₅. The ¹H NMR and HSQC spectra (Table 3) displayed signals for one methoxy group (δ_H 3.67), one methyl group (δ_H 0.87, H-11), one methylenedioxy group (δ_H 5.95), one olefinic methylene group (δ_H 5.18, 5.12, H-10), three alkyl methylene groups (δ_H 2.81, 2.32, H-4; 2.65, 2.49, H-6; 2.38, 2.19, H-8), one olefinic methine (δ_H 5.71, H-9), two alkyl methines (δ_H 3.03, H-3; 2.28, H-2), and three aromatic protons of an ABX spin system (δ_H 6.73 d, J = 1.7 Hz, H-2’; 6.76 d, J = 7.9 Hz, H-5’; 6.69 dd, J = 7.9, 1.7 Hz, H-6’).

Table 3.

NMR Spectroscopic Data for Compound 5 in CDCl₃ (500 MHz)

position	δH (J in Hz)	δC, type	position	δH (J in Hz)	δC, type
1		53.0, C	10	5.18 (dd 10.2, 1.3)	120.0, CH2
2	2.28 (dq 11.7, 7.0)	46.1, CH		5.12 (dd 17.0, 1.3)
3	3.03 (dt 11.7, 7.7)	47.1, CH	11	0.87 (d 7.0)	13.0, CH3
4	2.81 (dd 18.5, 7.7)	39.1, CH2	1′		136.8, C
	2.32 (dd 18.5, 11.7)		2′	6.73 (d 1.7)	107.4, CH
5		219.9, C	3′		146.1, C
6	2.65 (d 16.0)	47.3, CH2	4′		146.5, C
	2.49 (d 16.0)		5′	6.76 (d 7.9)	108.4, CH
7		172.4, C	6′	6.69 (dd 7.9, 1.7)	121.1, CH
8	2.38 (ddt 13.5, 6.5 1.3)	43.0, CH2	OCH2O	5.95 (s)	101.1, CH2
	2.19 (dd, 13.5, 8.5)		OMe	3.67 (s)	52.1, CH3
9	5.71 (dddd, 17, 10.2, 8.5, 6.5)	133.0, CH

The methylenedioxy group in 5 was attached to the benzene ring with the ABX spin system, suggested by ³J-HMBC correlations between OCH₂O (δ_H 5.95) and C-3′ (δ_C 146.1) and C-4′ (δ_C 146.5), H-2′ and C-4′, H-5′ and C-3′, H-6′ and C-4′. The presence of a methyl ester was confirmed by ³J-HMBC correlations between OCH₃ (δ_H 3.67) and C-7 (δ_C 172.4). The presence of an allyl group was suggested by signals for H-8 (δ_H 2.38 ddt, 13.5, 6.5, 1.3; 2.38 dd, J = 13.5, 8.5), H-9 (δ_H 5.71 dddd, J = 17.0, 10.2, 8.5, 6.2) and H-10 (δ_H 5.18 dd, J = 10.2, 1.3; 5.12 dd, J = 17.0, 1.3), and was confirmed by a ³J-HMBC correlation between H-10 and C-8 (δ_C 43.0). The presence of a 4-methylcyclohexanone unit was indicated by signals for the remaining seven carbons and nine protons: H₂-4 (δ_H 2.81 dd, 18.5, 7.7; 2.32 dd, 18.5, 11.7), H₂-6 (δ_H 2.65 d, 16.0; 2.49 d, 16.0), H-3 (δ_H 3.03 dt, 11.7, 7.7), H-2 (δ_H 2.28 dq, 11.7, 7.0), and H₃-11 (δ_H 0.87 d, 7.0). This partial structure was confirmed by ²J-HMBC correlations between H-4 and C-5 (δ_C 219.9), H-6 and C-5 (δ_C 219.9) and ³J-HMBC correlations between H-4 and C-2 (δ_C 46.1), H-6 and C-2, H-3 and C-11 (δ_C 13.0), H-11 and C-1 (δ_C 53.0). The benzene ring was attached to C-3 based on the observed ³J-HMBC correlations between H-3 and C-2′ (δ_C 107.4), H-3 and C-6′ (δ_C 121.1), H-2 and C-1′ (δ_C 136.8). The methyl ester and allyl groups were linked at C-1, indicated by the ³J-HMBC correlations between H-6 and C-7 (δ_C 172.4), H-2 and C-8 (δ_C 43.0), and H-6 and C-8 (Figure 3).

Figure 3.

HMBC (red) and NOESY (blue) correlations of anibignan A

The relative configurations of carbons C-1, C-2 and C-3 were assigned as 1S*,2R*,3S* based on NOESY crosspeaks between H-8 and H-2, H-3 and H-11, as shown in Figure 3. The absolute configuration of 5 was determined by its electronic circular dichroism spectrum, and C-3 was tentatively assigned as S based on its positive n−π* Cotton effect at 298 nm and application of the octant rule, with five heavy atoms in the negative zones and seven in the positive zone.¹⁷ The structure of anibignan A (5) was thus assigned as methyl (1S,2R,3S)-1-allyl-3-(benzo[d][1,3]dioxol-5-yl)-2-methyl-5-oxocyclohexane-1-carboxylate.

Compounds 7-12 (Chart 2) were identified as (+)-burchellin, or (2S,3S,3aR)-3a-allyl-2-(benzo[d][1,3]dioxol-5-yl)-5-methoxy-3-methyl-3,3a-dihydrobenzofuran-6(2H)-one (7),^12,18 (2S,3S,5R)-5-allyl-2-(benzo[d][1,3]dioxol-5-yl)-5-methoxy-3-methyl-3,5-dihydrobenzofuran-6(2H)-one (8),¹⁹ 5-((2S,3S)-6-(allyloxy)-5-methoxy-3-methyl-2,3-dihydrobenzofuran-2-yl)benzo[d][1,3]dioxole (9),²⁰ (2S,3S)-7-allyl-2-(benzo[d][1,3]dioxol-5-yl)-5-methoxy-3-methyl-2,3-dihydrobenzofuran-6-ol (10),^20–22 (1S,5S,6S,7S,8S)-5-allyl-7-(benzo[d][1,3]dioxol-5-yl)-8-hydroxy-3-methoxy-6-methylbicyclo[3.2.1]oct-3-en-2-one (11),²³ and (1S,5S,6R,7R,8R)-5-allyl-7-(benzo[d][1,3]dioxol-5-yl)-8-hydroxy-3-methoxy-6-methylbicyclo[3.2.1]oct-3-en-2-one (12).²³ Since the NMR spectroscopic data reported in the literature for compounds 7 – 12 are not always complete, complete ¹H and ¹³C NMR spectroscopic data are reported for these compounds in the Supporting Information.

Chart 2.

Biological Studies.

All isolated compounds were tested for antiplasmodial activity against P. falciparum Dd2 strain (Table 4). The known indolizinium alkaloids 1 and 2 were the most potent compounds, with IC₅₀ values of 0.170 and 0.244 μM. Of the other isolates, anibomarin A (3) displayed only weak antiplasmodial activity, with an IC₅₀ value of 18.2 μM. However, this modest antiplasmodial activity might possibly be improved by appropriate structural modifications, such as changing the length of the alkyl chain or introducing electron-withdrawing groups on the benzene ring, as was done for selected synthetic coumarins which had submicromolar IC₅₀ values against P. falciparum enoyl-ACP-reductase and micromolar activity against the drug-resistant P. falciparum K1 strain.²⁴

Table 4.

In vitro Biological Activity of Compounds 1 – 12.

compound	P. falciparum Dd2 (μM)	A2780 (μM)
1	0.170 ± 0.004	0.223 ± 0.019
2	0.244 ± 0.005	0.241 ± 0.019
3	18.2 ± 0.4	NT
5	> 100	> 10
7	> 30	> 10
8	NA	> 10
9	~ 30	> 10
10	~ 30	> 10
11	> 100	> 10
12	NA	> 10
artemisinin	0.0065 ± 0.0005	NT
paclitaxel	NT	0.013 ± 0.001

Anibamine (1), anibamine B (2), the new norneolignan anibignan A (5), and neolignans 7-12 were also tested for antiproliferative activity against the A2780 ovarian cancer cells. Only compounds 1 and 2 showed significant antiproliferative activity, as discussed below. Although anibignan A (5) has some structural similarities with the potent antiproliferative compound bifidenone (6),²⁵ neither it nor any of the other compounds 6 – 12 showed any significant antiproliferative activity, with IC₅₀ values of more than 50 μM in most cases. Of these, the most potent compound was 10, with IC₅₀ 35 μM.

The submicromolar antiplasmodial activity of compounds 1 and 2 led us to examine a collection of related compounds prepared by Zhang et al.^26–30 These compounds were of two general types: Type A compounds 1S and 13– 22 (Chart 3), with each having a quaternary nitrogen atom embedded in a bicyclic pyridine derivative with one or two C-10 side chains, while Type B compounds 23 – 40 (Chart 4) consisted of monocyclic uncharged pyridine derivatives with one or two C-10 side chains. Compound 1S is the synthetic version of the isolated natural product 1 with a chloride counterion in place of the trifluoroacetate counterion of 1.

Chart 3.

Chart 4.

The antiplasmodial activities of compounds 1, 1S, 2, and 13 – 40 are presented in Table 5, leading to several observations. The first is that the IC₅₀ values of 1 and 1S against P. falciparum are very similar, with 1S being slightly more potent than 1. This relatively small difference is likely due to the difficulties of precisely weighing small quantities of natural products or to differences in solubility due to different counterions of the parent compound.

Table 5.

In vitro Antiplasmodial Activity of Compounds 1 – 2 and 13 – 40

Cmpd	P. falciparum Dd2 (μM)	Cmpd	P. falciparum Dd2 (μM)
1	0.170 ± 0.004	26	>20
1S	0.110 ± 0.007	27	>20
2	0.244 ± 0.005	28	20
13	0.096 ± 0.006	29	>20
14	0.061 ± 0.001	30	>20
15	0.058 ± 0.001	31	19.3 ± 0.6
16	0.067 ± 0.003	32	>20
17	0.060 ± 0.001	33	>20
18	0.077 ± 0.002	34	6.7 ± 0.4
19	0.060 ± 0.001	35	6.3 ± 0.1
20	0.171 ± 0.010	36	>20
21	0.142 ± 0.005	37	8.8 ± 0.3
22	0.058 ± 0.002	38	>20
23	>20	39	5.3 ± 0.1
24	>20	40	>20
25	>20

An important second observation is that Type A compounds are uniformly more potent than Type B compounds by at least two orders of magnitude, indicating that the quaternary pyridinium ion is crucial for potent antiplasmodial activity. This result is consistent with previous work reporting potent antiplasmodial activity for monomeric pyridinium compounds³¹ and for both symmetrical³² and asymmetrical³³ dimeric pyridinium compounds. The potency of pyridinium compounds as antiplasmodial agents may be due to their ability to inhibit the biosynthesis of phosphatidylcholine, a key ingredient of the plasmodial membrane,³⁴ or to their ability to bind to heme and inhibit the formation of hemozoin.³⁵

The third observation is that the size of the alicyclic ring fused to the pyridinium ring does not appear to have a significant effect on potency, and furthermore that the number of C-10 side chains also does not significantly affect potency. This follows from the observation that the bicyclic pyridinium derivatives 19 and 22, with only one C-10 side chain, are equipotent with compounds 13 – 18 with two C-10 side chains and various alicyclic ring sizes. What does make a difference to the potency is the position of the C-10 side chain in compounds containing only one C-10 side chain. For example, 20 and 21, with C-10 side chains only at the C-6 position, were almost 3-fold less potent than 19 and 22, with side chains at the C-8 position. Finally, the E or Z configuration of the double bond in the C-10 side chains does not appear to be important for activity, since no significant changes in the antiplasmodial potencies were observed with changes in this configuration.

It is worth noting that 1S as well as 13 – 18 have essentially no hemolytic toxicity to red blood cells, with HC₅₀ values all greater than 58 μM, and 1 has a TC₅₀ value of greater than 22 μM against mouse embryonic fibroblast NIH-3T3 cells.³⁶ Anibamine (1S) is thus >129-fold and >98-fold more selective for P. falciparum and A2780 cells, respectively, than for 3T3 cells, suggesting that it has potential for development as an antimalarial agent.

Plasmodium falciparum has a complex intraerythrocytic life cycle that lasts around 48 h, during which the parasite progresses through four morphologically different stages: ring, trophozoite, and schizont stages, ending with the rupture of the erythrocyte and release of merozoites that will invade new erythrocytes. Thus, determination of the timing of action of anibamine (1) and its analogues was performed as the first step in studying their mechanism of action. To this end, assessment of their potential effects on the intraerythrocytic stages and morphological development was performed by Giemsa-stained thin blood smears and light microscopy. This analysis was performed only for Type A synthetic compounds since their IC₅₀ values were below 0.200 μM. Interestingly, these analyses defined two different inhibitory phenotypes. When compounds were added to late schizonts, none of the compounds were able to kill at this stage and prevent reinvasion, as all treated cultures presented normal ring stage parasites and some schizonts were still present after 6 h treatment as compared to untreated controls (Excel file Table S1, Supporting Information). Morphological assessment of these cultures was performed again after 24 and 48 h treatment. After 48 h of treatment, parasites treated with compounds 1S and 13 - 18 did not develop beyond early trophozoites and presented morphological defects as compared to untreated controls. These results were confirmed by starting the treatments in the ring stage and following the development of parasites after 6, 24, 48 and 72 h of treatment where substantial activities against ring stages and no reinvasions (48 and 72 h) were observed. Parasites treated with compounds 19 - 22 were able to develop to trophozoite and early schizont stages after 48 h of treatment but also presented morphological defects. These results suggest that none of the Type A synthetic compounds affects late stage schizonts and erythrocyte reinvasion of P. falciparum and that 1S, and 13 - 18 appear to have substantial activity against ring stage while compounds 19 to 22 acted later during intraerythrocytic development. This difference in their stage of activity may be due to the number of C-10 side chains, since compounds 1S, and 13 - 18 have two C-10 side chains while compounds 19 - 22 have only one.

These studies justify the importance for further understanding the mechanisms of action of this class of compounds as potential antimalarial agents, which apparently is different from that of the postulated molecular target of anibamine (1) in mammalian cells, namely the chemokine receptor CCR5 belonging to the G-protein-coupled receptor (GPCR) family.^10,36–38

EXPERIMENTAL SECTION

General Experimental Procedures.

Optical rotations were recorded on a JASCO P-2000 polarimeter. UV and IR spectroscopic data were measured on a Shimadzu UV-1201 spectrophotometer and a MIDAC M-series FTIR spectrophotometer, respectively. CD spectra were obtained on a JASCO J-815 circular dichroism spectrometer. NMR spectra were recorded in CDCl₃ on Bruker Avance 500 or 600 spectrometers. The chemical shifts are given in δ (ppm), and coupling constants (J) are reported in Hz. Mass spectra were obtained on an Agilent 6220 LC-TOF-MS in the positive-ion mode.

Antiplasmodial Bioassays.

The effect of each fraction and pure compound on parasitic growth of the P. falciparum Dd2 strain was measured in a 72 h growth assay in the presence of the test materials as described previously with minor modifications. Briefly, ring stage parasite cultures (100 μL per well, with 1% hematocrit and 1% parasitemia) were then grown for 72 h in the presence of increasing concentrations of the drug in a 5% CO₂, 5% O₂, and 90% N₂ gas mixture at 37 °C. After 72 h in culture, parasite viability was determined by DNA quantitation using SYBR Green I (25 μL of SYBR Green I in lysis buffer at 0.4 μL of SYBR Green I/mL of lysis buffer). The half-maximum inhibitory concentration (IC₅₀) calculation was performed with GraphPad Prism 6 software (GraphPad Software, Inc.) using nonlinear regression curve fitting. IC₅₀ values are the average of three independent determinations with each determination in triplicate and are expressed ± SEM.⁵

The inhibitory activity of anibamine (1) and its analogs was assessed during intraerythrocytic development by addition of test materials at either the ring stage (2–4 h post-invasion) or the schizont stage (40–42 post-invasion) and then followed by Giemsa-stained blood thin smears after 6, 24, 48 and 72 h of treatment. Results shown on Excel Table S1 (Supporting Information) were obtained from two or more independent experiments.

Antiproliferative Activity.

The A2780 ovarian cancer cell line antiproliferative bioassay was performed at Virginia Tech as previously reported.^39,40 The A2780 cell line is a drug-sensitive ovarian cancer cell line.⁴¹ Paclitaxel was used as the positive control.

Plant Material.

Bark, woody parts, and twigs of a tree identified as Aniba citrifolia (Nees) Mez (Lauraceae) were collected on September 11, 1992 near Groete creek, 27 km from Bartica, Essequibo Islands-West Demerara region, Guyana, at an altitude of 50 m (164 ft). The plant material was collected and identified by Suroojnauth Tiwari, and the identity was confirmed as A. citrifolia by Hans Beck of the New York Botanical Garden. Collection details are available at http://sweetgum.nybg.org/science/vh/specimen_details.php?irn=240345, and a voucher specimen is deposited at the New York Botanical Garden under barcode 01269145.

Extraction and Isolation.

The dried and powdered bark (915 g) was exhaustively extracted with methanol in two 24-hour percolation steps and the resulting solution was evaporated to yield 47.4 g of crude extract. For purposes of fractionation and purification, a 2.55 g portion of this methanol extract, designated 40258–10A, was shipped to Virginia Tech for bioassay-guided isolation. This extract (2.0 g) was dissolved in aqueous MeOH (MeOH-H₂O 9:1, 100 mL), and extracted with hexanes (100 mL) three times. Then the aqueous layer was diluted to 60% MeOH (v/v) with H₂O and extracted with dichloromethane (150 mL) three times. The hexanes fraction (305.0 mg) had an IC₅₀ value of < 1.25 μg/mL. The dichloromethane fraction constituted 1071 mg after being evaporated in a rotavapor, and showed an IC₅₀ value of <<<1.25 μg/mL. The remaining aqueous MeOH fraction (600 mg) had an IC₅₀ value of < 1.25 μg/mL.

The hexanes fraction was loaded on to a Sephadex LH-20 column eluted with CH₂Cl₂/MeOH (1:1) to afford five fractions (1.1–1.5). Fraction 1.2 (38.8 mg), with an IC₅₀ value of ~ 5 μg/mL, was purified by HPLC on a C₁₈ column (Phenomenex 250 × 10 mm) using acetonitrile/H₂O from 45:55 to 100:0 at a flow rate of 2.5 mL/min over 70 min to yield the four known neolignans 7 (0.7 mg), 8 (0.4 mg), 9 (1.0 mg) and 10 (0.9 mg), with retention times of 22.3, 26.5, 45.0 and 47.0 min. Fraction 1.3 (40.0 mg) with an IC₅₀ value of ~ 1.25 μg/mL was separated by HPLC on a C₁₈ column using acetonitrile/H₂O gradient from 80:20 to 100:0 at a flow rate of 2.5 mL/min over 35 min, to yield the new coumarin, anibomarin A (3) (0.7 mg), with a retention time of 12.5 min.

The dichloromethane fraction was also separated by Sephadex LH-20 as above to afford five fractions (2.1 – 2.5). Fraction 2.2 (157 mg), with an IC₅₀ value of 1.25 ~ 2.5 μg/mL was fractionated on a diol open column with chloroform/MeOH (9:1) to generate ten subfractions (2.2.1 – 2.2.10), and the first fraction 2.2.1 (18.4 mg) was further separated by HPLC on a C₁₈ column using an acetonitrile/H₂O gradient from 90:10 to 100:0 over 20 min at a flow rate of 2.5 mL/min and then acetonitrile alone for 15 min, to yield the new norneolignan, anibignan A (6) (1.2 mg), with a retention time of 28.0 min. Fraction 2.3 (108.6 mg) was purified by HPLC on a C₁₈ column using acetonitrile/H₂O from 45:55 to 100:0 over 60 mins to yield the two known neolignans, 11 (0.6 mg) and 12 (0.4 mg), with retention times of 32.8 and 52.0 min.

The aqueous methanol fraction was separated using a 45 × 5 cm Diaion HP-20 column eluted with MeOH:H₂O 40:60, 70:30, and 100:0 followed by acetone to afford the four fractions 3.1 – 3.4. Fraction 3.4 (80.2 mg) with an IC₅₀ value of <<<1.25 μg/mL was separated by HPLC on a C₁₈ column using a gradient of MeOH/0.1% aqueous TFA from 70:30 to 100:0 over 30 mins and then MeOH alone for 10 min, to yield the new compound 2 (1.2 mg), with a retention time of 35.5 min, and the known compound anibamine (1) (6.6 mg), with a retention time of 30.5 min.

Anibamine B (2): yellow oil; UV (c 0.05780 mM, MeOH) λ_max (ε) 221 nm (10554), 285 nm (3922); ¹H NMR (500 MHz, CDCl₃) and ¹³C NMR (125 MHz, CDCl₃), see Table 1; HRESIMS m/z 426.4100 [M]⁺ (calcd for C₃₀H₅₂N⁺, 462.4094).

Anibomarin A (3): yellow oil; UV (c 0.3507 mM, MeOH) λ_max (ε) 208 nm (36480), 288 nm (10655), 355 nm (6195); ¹H NMR (500 MHz, CDCl₃) and ¹³C NMR (125 MHz, CDCl₃), see Table 2. HRESIMS m/z 593.3229 [M+Na]⁺ (calcd for C₃₇H₄₆NaO₅⁺, 593.3237) and 571.3412 [M+H]⁺ (calcd for C₃₇H₄₇O₅⁺, 571.3418).

Anibignan A (5): yellow oil; [α]_D²¹ +40.3 (c 4×10⁻⁴ g/mL, MeOH); UV (c 0.0249 mM, MeOH) λ_max (ε) 204 nm (26278), 233 nm (4683), 285 nm (3408); ECD (c 0.0249 mM, MeOH) Δε₂₉₈ + 3.4; ¹H NMR (500 MHz, CDCl₃) and ¹³C NMR (125 MHz, CDCl₃), see Table 4; HRESIMS m/z 353.1372 [M+Na]⁺ (calcd for C₁₉H₂₂NaO₅⁺, 353.1359) and 331.1543 [M+H]⁺ (calcd for C₁₉H₂₃O₅⁺, 331.1540).

Anibamine Analogues: Anibamine (1S)^27,29 and analogues 13, 21 – 23, 25, 31, 35, 37,^28,30 were synthesized as previously described. The syntheses and characterization data of analogues 24, 26 – 30, 32 – 34, 36, and 38–40 are provided in the Supporting Information.