Antiprotozoal Nor-Triterpene Alkaloids from Buxus sempervirens L.

Malaria and human African trypanosomiasis (HAT; sleeping sickness) are life-threatening tropical diseases caused by protozoan parasites. Due to limited therapeutic options, there is a compelling need for new antiprotozoal agents. In a previous study, O-tigloylcyclovirobuxeine-B was recovered from a B. sempervirens L. (common box; Buxaceae) leaf extract by bioactivity-guided isolation. This nor-cycloartane alkaloid was identified as possessing strong and selective in vitro activity against the causative agent of malaria tropica, Plasmodium falciparum (Pf). The purpose of this study is the isolation of additional alkaloids from B. sempervirens L. to search for further related compounds with strong antiprotozoal activity. In conclusion, 25 alkaloids were obtained from B. sempervirens L., including eight new natural products and one compound first described for this plant. The structure elucidation was accomplished by UHPLC/+ESI-QqTOF-MS/MS and NMR spectroscopy. The isolated alkaloids were tested against Pf and Trypanosoma brucei rhodesiense (Tbr), the causative agent of East African sleeping sickness. To assess their selectivity, cytotoxicity against mammalian cells (L6 cell line) was tested as well. Several of the compounds displayed promising in vitro activity against the pathogens in a sub-micromolar range with concurrent high selectivity indices (SI). Consequently, various alkaloids from B. sempervirens L. have the potential to serve as a novel antiprotozoal lead structure.


Introduction
Buxus sempervirens L. (common box; Buxaceae) is an evergreen shrub or small tree, with nor-triterpene alkaloids of the nor-cycloartane type as its main chemical constituents. Decoctions of leaves were ethnopharmacologically used against a variety of indications, including malaria [1,2].
Malaria is a life-threatening infectious disease caused by protozoans of the genus Plasmodium. The vector of these parasites is the infected female Anopheles mosquito. The World Health Organization (WHO) estimated that 229 million malaria cases and 409,000 malaria deaths occurred worldwide in 2019. The incidence has remained virtually unchanged in recent years and, because of the coronavirus pandemic, the WHO estimated 100,000 additional deaths in 2020. The progress in the fight against malaria has stalled. There is a high incidence of treatment failure due to resistance and, thus, the development of new antimalarial drugs is indispensable.
In a previous study by our group [6], O-tigloylcyclovirobuxeine-B, after bioactivityguided isolation from the dichloromethane extract of B. sempervirens leaves, was identified as the constituent mainly responsible for the plant's antiplasmodial activity. It appeared likely that further congeneric alkaloids also yield contributions to the overall antiplasmodial effect of the total leaf extract. Moreover, a lupane triterpene of B. sempervirens L. showed prominent bioactivity even against drug-resistant malaria parasites [7]. Consequently, the isolation, as well as the antiplasmodial testing of further Buxus-alkaloids, is of great interest.
In a recent study [8], it was determined that the alkaloid-enriched fraction of Pachysandra terminalis (Buxaceae) possessed promising activity against another protozoan parasite, Trypanosoma brucei rhodesiense (Tbr), the causative agent of East African sleeping sickness, a poverty-related neglected tropical disease. This plant contains aminosteroids structurally related to the nor-triterpene alkaloids of B. sempervirens L. Furthermore, an isolated compound of B. sempervirens L., cyclovirobuxeine-B, was found to be highly active and selective against Tbr in vitro [9]. These findings suggested that additional Buxus-alkaloids could likewise represent strong trypanocides. The chemotherapeutic agents currently in use against African sleeping sickness (human African trypanosomiasis (HAT)) are toxic and have many other associated disadvantages, such as long hospitalization. Accordingly, there is a compelling need for new treatments.
The aim of this study was the identification and characterization of potent and selective antiprotozoal compounds as lead structures against tropical diseases. For this purpose, we report on the systematic isolation of 25 alkaloids from a B. sempervirens L. leaf extract, including eight new natural products and one compound first described for this plant. The compounds were obtained by centrifugal partition chromatography (CPC), preparative high performance liquid chromatography (prep-HPLC), and column chromatography (CC). Additionally, we present the results of the in vitro testing of the isolated alkaloids against the pathogens of malaria and HAT.
For compound 11, the molecular formula was determined as C 26 H 43 NO 3 by LC/MS ( Figures S71-S73, Supplementary Materials). In the 1 H, 13 C, and HSQC-NMR spectra for 11 ( Table 2;  In the 13 C-NMR spectrum, a signal at δc = 217.33 could be detected in the low field in the typical shift range of a ketone [29]. The 2 J and 3 J couplings in the 1 H/ 13 C-HMBC spectrum of the ketone carbon with positions 1, 2, 29 and 30 clearly assigned it to C-3. This position also appears plausible from a biosynthetic point of view [28] for the elucidated new structure ( Figure 2).
In the +ESI-QqTOF MS/MS spectrum ( Figure S121, Supplementary Materials), compound 23 exhibited identical fragmentation (the fragmentation pathway is reported for the first time in Figure S122 Figure S114, Supplementary Materials), was not detectable. The signals of the NMR spectra (Table 3; Figures S124-S133, Supplementary Materials) of 22 and 23 were very similar for the most part. Deviations occurred in the chemical shifts at positions C-11 and C-19 in the A-ring. In the 1 H/ 13 C-HSQC spectrum of compound 23, two signals could be detected at δ H = 5.42 (br s)/δc = 121.95, as well as δ H = 5.31 (br m)/δc = 119.37, which indicated olefinic structural elements. In the 1 H/ 13 C-HMBC spectrum, these signals did not show any interaction with each other, as would be expected in a conjugated ∆ 9(11),10(19) system as found in 22. Instead, the chemical shifts agree with a (9-(10→19))abeo-pregnane with the presence of two isolated double bonds, i.e., ∆ 1(10) and ∆ 9 (11) [28,30]. Consequently, compound 23 could be identified as a constitutional isomer of 22 with a double bond between C-1 and C-10, instead of C-10 and C-19 ( Figure 2).    (Table 3; Figures S137-S145, Supplementary Materials), compound 24, in analogy with 23, could clearly be assigned to the Buxus-alkaloids with a (9-(10→19))abeo-pregnane backbone and isolated double bonds between C-1 and 10 and between C-9 and 11. The acetamide group, already suspected by the fragmentation, could be located at the nitrogen atom at C-20 by the cross signal of the carbonyl carbon (δ C = 170.18) with the proton at position 20 (δ H = 3.99) in the 1 H/ 13 C-HMBC spectrum. This amide group has a significantly reduced basicity in contrast to the secondary amine group in position 3, which provides an explanation for the low intensity of the [M + 2H] 2+ quasimolecular ion in the +ESI-QqTOF mass spectrum. The resulting structure is presented in Figure 2.
The  (Table 3; Figures S149-S158, Supplementary Materials), compound 25 was clearly identified as a Buxus-alkaloid with a (9-(10→9))abeopregnane skeleton and isolated double bonds between C-1 and C-10 and between C-9 and C-11. A methine proton signal at δ H = 3.84 (δ C = 73.7 according to the 1 H/ 13 C-HSQC spectrum) resonated as a quartet (J = 6.1 Hz) and should therefore be in geminal position with a methyl group and must represent position 20. In the 1 H/ 13 C-HMBC spectrum, this proton showed cross peaks with carbons at δ C = 18.

In Vitro Antiprotozoal Activity of Isolated Compounds
After isolation, all alkaloids were tested in vitro for antiplasmodial and antitrypanosomal activity against Plasmodium falciparum (Pf) and Trypanosoma brucei rhodesiense (Tbr). Furthermore, to assess their selectivity against the parasites, cytotoxicity against L6 rat skeletal myoblasts, as mammalian control cells, was tested (Table 4). All IC 50 values are expressed in µg/mL and values in µM are given in parentheses. While * n = 3 was reported as the mean value from three independent measurements with the standard deviation, all other values were determined with n = 2 as the mean value from two independent measurements with the fluctuation range. Note that the purity of compound 10 was <90%, so that activity was not determined.  1, 2, 5, 9, 11-16) and the steroidal alkaloid (19) displayed moderate activities with IC 50 values in the range of 1.05 to 4.3 µM. The (9-(10→19))abeo-5α-pregnanes (20)(21)(22)(23)(24)(25) were less active with IC 50 values of >4.0 µM.

Plant Material
The same plant material was used as described previously [11].

Extraction and Isolation of Alkaloids from the B. sempervirens Leaf Extract
The extraction of plant material, the first part of the isolation procedure, and the isolation of O-tigloylcyclovirobuxeine-B (1) was equal to our previous study (Scheme 1) [11].
UHPLC/+ESI-QqTOF-MS/MS measurements were performed as described previously [11]. The sample concentration of the crude extract was 10 mg/mL in case of the CPC fractions 1 mg/mL, and for the isolated compounds it was 0.1 mg/mL. UV spectra of compounds 22 and 23 were recorded with a U-2900 spectrophotometer (Hitachi, Tokyo, Japan) in methanol.