Diterpenes Synthesized from the Natural Serrulatane Leubethanol and Their in Vitro Activities against Mycobacterium tuberculosis

Seventeen new derivatives of the natural diterpene leubethanol, including some potential pro-drugs, with changes in the functionality of the aliphatic chain or modifications of aromatic ring and the phenolic group, were synthesized and tested in vitro by the MABA technique for their activity against the H37Rv strain of Mycobacterium tuberculosis. Some compounds showed antimycobacterial selectivity indices higher than leubethanol.


Introduction
Tuberculosis (TB) caused by Mycobacterium tuberculosis (Mtb) remains a worldwide health problem. According to the latest report from the World Health Organization (WHO), the number of people suffering from TB in 2013 was 9.0 million and approximately 1.5 million people died from the disease [1]. Moreover, Mtb is highly infectious; it is reported that about one third of the world's population is latently infected, and that 10% of this population will develop active disease. A further significant problem is the association with HIV infection, from 1.5 million deaths caused by TB in 2013, 0.36 million deaths were in HIV positive people. Therefore, TB-HIV co-infection is a major public health problem [1].
The current first line drugs for TB treatment (isoniazid (INH), rifampicin (RIF), pyrazinamide and ethambutol) were discovered decades ago and are increasingly becoming less useful due to emerging multidrug-resistant (MDR), extended (XDR) or extremely (XXDR) resistant strains [2]. There were an estimate of 480,000 new cases of multidrug-resistant tuberculosis (MDR-TB) and 170,000 MDR-TB-related deaths. Treatment of MDR-TB disease requires high resources and at least the administration of RIF and INH in combination with second-line drugs. Second line drugs are more expensive, more toxic, and less effective than drugs used in standard therapy [3]. These facts have motivated the search for new drugs and treatment strategies. New drug candidates should shorten conventional chemotherapy and be effective against MDR-TB.
Natural products have played an important role in the discovery of new drugs, as is the chemotherapy of tuberculosis [4,5], with the discovery of streptomycin, capreomycin, cycloserine, or development of rifamycin derivatives. Now, there is a re-emerging interest in natural products as novel template for the development of new drugs and particularly suitable as antibacterial leads [6][7][8]. Compounds bearing a serrulatane skeleton have been reported as good antimycobacterial agents [9][10][11].
Waksman et al. isolated leubethanol (Leub), a natural serrulatane diterpenoid, from Leucophyllum frutescens methanolic extracts [10] and introduced different modifications on the skeleton to improve antimycobacterial activity [12]. Taking into account previous results of the research group on leubethanol structure modification related to antimycobacterial activity [12] and the information of similar structures reported in the literature [9,11], we intended to explore in the present study new possibilities to obtain better antimycobacterial compounds by introducing a glycosylated fragment on Leub phenolic hydroxyl, substitutions on the aromatic ring A, and new side-chain functionalizations ( Figure 1).
Inversion of the anomeric carbon configuration of the 8-O-β-D-glucopyranosyl Leub derivative was established by a study of the H-1' coupling constant, with a value of 3.7 Hz in the glucopyranosyl trichloroacetimidate and of 7.9 Hz in the Leub aduct. Afterwards, the obtained Leub derivative was treated with sodium ethoxide in ethanol to provide compound 1 in 52% yield. Unequivocal structural assignment was achieved by 2D-NMR experiments ( 1 H, 1 H-COSY, HMQC, HMBC).
We introduced electron withdrawing groups (NO2, Br) on ring A. Leub treatment with 65% HNO3 gave a mixture of 7-nitro and 5-nitro derivatives, (compounds 2 and 3, respectively) in 1.5:1.0 ratio. The unambiguous structural assignment of the nitrated derivatives was established by NOE experiments. Irradiation of the proton at 6.50 ppm on compound 2 showed a NOE effect on two of the methyls, the aromatic methyl at 2.21 ppm and the methyl doublet on the aliphatic chain at 0.83 ppm. On the other hand, irradiation of the proton at 6.63 ppm on compound 3 showed a NOE effect on the hydroxylic proton at 11.27 ppm and on the aromatic methyl at 2.57 ppm. The aliphatic double bond of Leub was hydrogenated with H2 in the presence of Pd-C catalyst, and the resulting 14,15-dihydroleubethanol treated with N-bromosuccinimide (NBS) to provide 7-bromoleubethanol (4) in 67% yield (Scheme 1). As just mentioned in the nitro derivatives, the bromine atom introduction at Leub position 7 was unequivocally established by NOE experiments.

Scheme 1.
Modifications of leubethanol on the phenolic hydroxyl and aromatic ring A.
Then, some modifications on the side chain were performed (Scheme 2). First, we obtained the allylic alcohol 5 and the α,β-unsaturated aldehyde 6 (in 2.2:1 proportion) by Leub treatment with selenium dioxide. Treatment of compound 5 under Sharpless asymmetric epoxidation conditions (with L-(+)diethytartrate) gave the corresponding α-oxirane derivative 7 in 65% yield and ~91% ee. The allylic alcohol 5 was condensed with pyrazinecarboxylic acid in the presence of N,N-dicyclohexyl-carbodiimide (DCC) as coupling reagent to give the pyrazine ester derivative 8 in 53% yield. Compound 8 showed 1 H-NMR signals at 9.29 ppm (1H, d, J = 1.6 Hz), 8.75 ppm (1H, d, J = 2.4 Hz) and 8.73 (1H, dd, J = 2.4; 1.6 Hz) of the pyrazine ring. Attempts to obtain other esters by treatment with 2-amino-isonicotinic acid or with 6-aminopyridine-3-carboxylic acid reaction did not progress, even by changing the coupling reagent (1-ethyl-3-(3-dimethylaminopropyl)carbodiimide, EDCI), the reaction solvent, or by increasing reaction time or the temperature. Furthermore, the α,β-unsaturated aldehyde 6 was used to obtain compounds 9-13. Treatment of compound 6 with hydroxylamine hydrochloride in dichloromethane/pyridine provided the hydroxyimine 9, which showed two singlet signals at 7.66 and 8.09 ppm in its 1 H-NMR for the hydroxyimine group. The dihydropyranone 10 was obtained by reaction of 6 with trans-methoxy-trimethylsilyloxy-1,3butadiene in the presence of boron trifluoride diethyl ether with 48% yield [19]. Compound   [20]. Compound 11 showed in its 1 H-NMR a one olefinic proton singlet at 7.00 ppm (H-16) and a methyl at 3.75 ppm of the CH3-N. Two moles of 6 were treated with one mole of ethylenediamine in anhydrous toluene and a large excess of anhydrous sodium sulfate at room temperature to give a diimine 12a, which was reacted with sodium borohydride in ethanol to provide the ethylenediamine 12b, in a total yield of 86%. Finally, compound 6 was treated with sodium chlorite [21] to oxidize the aldehyde to an acid. The additional oxidation on the aromatic ring was observed. Compound 13 showed a quinone fragment with chemical shifts at 188.4 and 187.1 ppm of carbons C-5 and C-8, respectively, in addition to a new carbonyl at 173.00 ppm of the acid group. Only one aromatic proton was appreciated in 1 H-NMR at 6.52 ppm. To compare the influence of the quinone fragment on the antimycobacterial activity, we decided to obtain the quinone of Leub, compound 14.
Promising antimycobacterial activity against Mtb H37Rv strain has been reported in the literature [9] for pseudopteroxazole and its analogue seco-pseudopteroxazole (MIC values of 97% and 66% mycocacterial growth inhibition at 12.5 μg/mL). Then, we decided to obtain the tricyclic diterpene analogue of Leub (Scheme 3). The phenol group was protected as an acetate by reaction of Leub with acetate anhydride in the presence of pyridine. Compound 15 showed a methyl group at 2.12 ppm in 1 H-NMR and two acetyl group signals at 169.7 and 21.2 ppm in the 13 C-NMR. Then, the allylic alcohol 16 was obtained in 56% yield by compound 15 treatment with phenylseleninic acid (obtained in situ from diphenyl deselenide and hydrogen peroxide) followed by tert-butyl-hydroperoxide. Compound 16 displayed two singlet methyl groups at 1.22 and 1.23 ppm, and two olefinic protons at 5.30 ppm as a multiplet, and at 5.51 ppm as a doublet (J = 15.8 Hz) in the 1 H-NMR. Cyclization of compound 16 was achieved using methanesulfonic acid in CH2Cl2 at −40 °C. Although the cyclization proceeded quantitatively, the product was a mixture of diastereomers 17a and 17b in a ratio of about 2:1. Compound 17a could be separated by silica gel chromatography, while compound 17b was obtained as a mixture of 17a and 17b.
In the case of 17a, the vinyl proton attached to C(14) appeared at δ 4.99 as a broad doublet, J = 9.2 Hz while in 17b it was observed at δ 5.14 (doublet, J = 9.2 Hz). Compound 17a saponification with KOH/MeOH finally gave the phenol derivative 18. The overall yield from Leub to 18 was 41%.

Antimycobacterial Activity
The anti-MTB activity was assessed in vitro against the H37Rv strain (ATTC 27294) [22] susceptible to all SIREP anti-TB drugs, according to a modified Microplate Alamar Blue Assay (MABA) [23]. Ethambutol (EMB) was used as the reference drug and the assays were performed by triplicate independent experiments. Cytotoxicity on Vero cells [24] was also measured for those compounds with an appreciable in vitro anti-MTB effect.
The in vitro antimycobacterial results of the Leub and seventten derivatives are shown in Table 1. The introduction of a glycoside fragment on the phenolic group induces the loss of the activity. We knew that the introduction of acyl or alkyl fragments at this position were not convenient for the activity, but we expected that the introduction of other hydroxyl groups (as is a glycoside) would not modify the activity results, so this indicates that phenol acidic properties are crucial for good activity. Substitutions at position 5 or 7 on Leub were also relevant for the antimycobacterial activity. We had previously found that the introduction of aromatic substituents on position 5 maintained Leub activity [12]. Taken into account the antimycobacterial activity results found for the pseudopterosins [9], compounds with an oxazole between positions 7 and 8 of the serrulatane skeleton, we expected to see good activity results for compounds substituted at position 7, like compounds 2 and 4, but they did not show activity. The antimycobacterial activity of compound 3 was a slightly lower than that of Leub, with small increase in the selectivity index. We planned to obtain the bromine derivative at position 5 to compare the antimycobacterial activity with compound 3, for this we had to previously protect the phenol group with a bulky group to force the introduction of the bromine at position 5, but in the process of removing the protecting group we got a mixture of compounds of difficult separation. As we indicated in the previous work functionalization of the C-16 methyl, compounds 5 and 6 showed similar potency and selectivity index to the natural compound leubethanol, Modifications of the aliphatic chain by introduction of an oxirane on the 14,15-double bond, (compound 7), or modifications on the alcohol or aldehyde functions, pyrazine ester 8, the hydroxyimine 9, or the heterocyclic compounds 10 and 11 led to the loss of activity. Only Leub dimer 12a, tricyclic derivative 18 and especially the ethylenediamine 12b maintained the antimycobacterial activity. Regarding the Vero cells/Mtb selectivity, compounds 3, 5, 6 and 18 showed similar selectivity to leubethanol, and compound 12b twice as much as leubethanol.

General Information
All commercial chemicals and solvents used were reagent grade. Flash column chromatography was done using Merck Silica Gel 60 (0.04-0.063 mm). Reactions were monitored by TLC using Merck 60F254 silica gel plates. Compounds were detected visually under UV irradiation (254 nm) and by spraying with sulfuric acid and phosphomolybdic acid reagents followed by heating at 100 °C. 1 H-NMR    (1) 2,3,4,6-Tetra-O-acetyl-β-D-glucopyranosyl leubethanol (160 mg, 0.25 mmol) was dissolved in dry EtOH (5 mL) and a solution of sodium ethoxide (28 mg Na, 1.25 mmol) in EtOH was added. After stirring the mixture at room temperature for 2 h the reaction was completed. The solution was taken to dryness to obtain a residue that was purified by column chromatography with CH2Cl2/MeOH   (3) To a solution of leubethanol (56 mg, 0.19 mmol) in n-hexane (2 mL) 65% HNO3 (16 μL, 0.23 mmol) was added and the mixture stirred for 8 h. Then, ethyl acetate (30 mL) was added and the organic layer was washed with 5% NaHCO3, water and dried over Na2SO4. The solvent was removed under vacuum to provide a crude (60 mg), that was purified by silica flash chromatography using n-hexane/ethyl acetate (95:5) as eluent, to give 23 mg (37%) of 7-nitroleubethanol (2) and 16 mg (26%) of 5-nitroleubethanol (3).

In Vitro Antimycobacterial Evaluation
The antimycobacterial activity was assessed against M. tuberculosis H37Rv ATTC 27294 susceptible to all five first-line anti-TB drugs (streptomycin, isoniazid, rifampin, ethambutol, and pyrazinamide) in a modified Microplate Assay Blue Alamar [22,23]. The compounds for M. tuberculosis bioassays were prepared at a concentration of 1 mg/mL in 2.5% DMSO in Middlebrook 7H9 (Becton Dickinson and Co., Sparks, MD, USA) broth. All solutions were sterilized by filtration using 13 mm diameter PTFE acro-discs (0.22 μm pore size, Millipore Co., Bedford, MA, USA). The concentrations for organic compounds used ranged from 100 μg/mL to 0.78 μg/mL, results are reported as minimal inhibition concentration (MIC). Ethambutol (EMB) was used as positive control. All biological assays were developed at least by triplicate.

Cytotoxicity Assay
Cytotoxicity was determined according to the MTT method [24] using African green monkey kidney epithelial cells (Vero cells) grown in RPMI-1640 medium supplemented with penicillin (100 units/mL), streptomycin (100 μg/mL), and fetal bovine serum (10%) at 37 °C under 5% CO2. In the confluence, 4000 cells per well was applied to a 96-well microtitre plate. After incubation for 24 h (37 °C, 5% CO2), 10 μL of a solution containing concentrations ranging from 500 to 0.5 μg/mL of the compound under evaluation was added and incubated for 48 h. 25 μL of MTT (4 mg/mL in PBS) was added to each well and incubated for another 3 h. The solution in each well was removed and DMSO (200 μL) was then added to each well. The absorbance was recorded on a microplate reader at a wavelength of 575 nm. The CC50 value was calculated by linear regression as the concentration of the compound inhibiting 50% cellular viability.

Conclusions
The results found in this study regarding the antimicrobial activity indicate that there are few possibilities of modification on the leubethanol molecule. The dimer of leubethanol, compound 12b showed a higher selectivity index than the natural compound.