Synthesis of New Cytotoxic Aminoanthraquinone Derivatives via Nucleophilic Substitution Reactions

Aminoanthraquinones were successfully synthesized via two reaction steps. 1,4-Dihydroxyanthraquinone (1) was first subjected to methylation, reduction and acylation to give an excellent yield of anthracene-1,4-dione (3), 1,4-dimethoxyanthracene-9,10-dione (5) and 9,10-dioxo-9,10-dihydroanthracene-1,4-diyl diacetate (7). Treatment of 1, 3, 5 and 7 with BuNH2 in the presence of PhI(OAc)2 as catalyst produced seven aminoanthraquinone derivatives 1a, b, 3a, and 5a–d. Amination of 3 and 5 afforded three new aminoanthraquinones, namely 2-(butylamino)anthracene-1,4-dione (3a), 2-(butylamino)anthracene-9,10-dione (5a) and 2,3-(dibutylamino)anthracene-9,10-dione (5b). All newly synthesised aminoanthraquinones were examined for their cytotoxic activity against MCF-7 (estrogen receptor positive human breast) and Hep-G2 (human hepatocellular liver carcinoma) cancer cells using MTT assay. Aminoanthraquinones 3a, 5a and 5b exhibited strong cytotoxicity towards both cancer cell lines (IC50 1.1–13.0 µg/mL).

It has also been reported that amino-substituted anthraquinones show significantly increased antiproliferative activities against human/mammalian cancer cell lines [21,22] and are known to have potential antitumor activity, but are less toxic to normal cells and display low cardiotoxicity [23][24][25][26]. A study on 4-(N-cyclohexylamino)-emodin implied that it can discriminate well between hepatoma cells and primary hepatocytes and it retained the capacity to reverse the multi-drug-resistance phenotype [27]. Other aminoanthraquinone derivatives such as Reactive Blue 2 (RB-2), Acid Blue 25 (AB-25) and Acid Blue 129 (AB-129) also known as a good nucleotide-binding proteins [28] where RB-2 is one of the most widely used P2-receptor antagonists [29]. The potential of RB-2 was driven by its hydrophobic interactions of aromatic -electron systems, and the hydrogen bonds with nitrogen as donor and acceptor atoms as these properties could block the P2-recptor.
Based on these promising bioactivities, the aim of this research was therefore to synthesise some new aminoantharaquinone derivatives with high potential as anticancer and antimicrobial agents. The effects of substrate amount, catalyst amount, reaction time and reaction temperature were studied. According to a study, shorter amines would result in lower cytotoxic effects [24] whereas the use of diamines or longer amine chains allow the possibility of forming side products due to their reactive properties [30]. Therefore, a simple straight chain amine containing four carbons (BuNH 2 ) was chosen for this study, while diacetoiodobenzene [PhI(OAc) 2 ] was used as catalyst because of it selectivity towards the substitution reaction [24]. Synthesised aminoanthraquinones were then further tested for cytotoxicity and antimicrobial activity.

Reduction, Methylation and Acylation
The reduction of compound 1 using NaBH 4 (1:1 equiv.) was performed successfully in 30 min to give a mixture of 4-hydroxyanthracene-1,10-dione (2) and anthracene-1,4-dione (3) in 69% and 21% yields, respectively. Increasing the amount of NaBH 4 to 3 equiv. led to the formation of only compound 3 in excellent yield (Scheme 1 and Table 1). Further NaBH 4 increases resulted in a lower yield of 3 as the single product obtained (Table 1, Entry I, II and III).

Amination
Amination reactions were attempted using procedure of Teich et al. [24]. It was stated that the formation of aminoanthraquinones can be achieved in higher yield using 1.1 equiv. of the catalyst PhI(OAc) 2 . Compounds 1, 3, 5, and 7 (1 equiv.) were then used in a substitution reaction to produce aminoanthraquinones derivatives. The reaction was performed with excess BuNH 2 in the presence of PhI(OAc) 2 for 15 min to 6 d of reaction, depending on the hindrance of the starting material. BuNH 2 also acted as a solvent for the reaction (Scheme 2). The catalytic effect was studied by performing the reaction either with or without the catalyst to observe what is the possible reaction or product that might occur or form. Therefore, direct amination of 1 with higher equivalents of butylamine in the presence of 1.1 equiv. of catalyst at RT gave a mixture of anthraquinone butylamines 1a and 1b in 1:1 ratio (78% yield). Reducing the equivalents of amine to half either with or without the catalyst produced a single product 1a in higher yield (70%-90%, Table 2, Entry I, II and III). However, increasing the reaction temperature gave a mixture of products again, where 1a was isolated as the major one (70%). The butylamino group was observed to substitute at the ortho position or replace the OH group. The results proved that the amination occurred via a nucleophilic substitution reaction and in agreement with reported works [21,25]. The 1 H-and 13 C-NMR spectra of 1a and 1b are also consistent with their structures and the reported data.   In the presence of PhI(OAc) 2 , the reaction proceeded with the interaction of the OH group of compound 1 with the catalyst to give the intermediate, O-IPh(OAc). Further attack by BuNH 2 at the next carbon was produced C=O and removed the PhIOAc. Proton elimination then restored the double bond and OH of 1a (Scheme 2). In the absence of catalyst, the electron-donating nature of the OH group together with the electron delocalization made compound 1 easily attacked by the BuNH 2 at the ortho-position to form 1a (Scheme 3).

Scheme 3.
Proposed mechanism of formation of 1a without using a catalyst. The compounds 1a, 1b and the optimised amount of BuNH 2 (225 equiv.) and PhI(OAc) 2 (1.1 equiv.) were used as a comparison of nucleophilic substitution for amination of compounds 3, 5 and 7. Therefore, the effect of different 1,4-disubstituted of anthraquinones and the mechanisms were further investigated.

Amination of 3
Amination of compound 3 produced a higher yield of 2-(butylamino)anthracene-1,4-dione (3a) when treated with 255 equiv. of BuNH 2 (Scheme 4). It was observed that the yield decreased when the reaction was performed without a catalyst or if equivalents of amine were reduced ( The 1 H-NMR spectrum of 3a showed aromatic proton singlets at 8.50 and 8.48 ppm that are attributed to the aromatic carbons at positions C9 and C10, respectively. A proton singlet observed at 5.80 ppm indicated the presence of an unsaturated proton, whereas a broad singlet observed at 5.99 ppm referred to the chelated proton of the amine. The 13 C-NMR spectrum showed the presence of 18 signals correspond to the 18C in the structure. Higher chemical shifts were observed at 182.7 and 181.4 ppm that represent the two C=O groups, whereas lower signals for C2 and C3 were observed at 130.2 and 134.0 ppm. In comparison to the starting material, only one signal was observed for both C=O ( 184.7), and for C2 and C3 ( 140.1) due to the symmetrical structure of 3. The position of the butylamino group of 3a was confirmed by an HMBC experiment which showed a 3 J correlation as indicated in Figure 1. The proposed mechanism showed that C3 is positively charged and readily attacked by BuNH 2 due to the resonance that drives by the interaction between the oxygen at C1 with the iodine of the catalyst. Further elimination of a proton at position C3, PhI and AcOH then restored the double bonds between C-C and C-O to give the product 3a (Scheme 5). A similar resonance effect is assumed to drive the BuNH 2 to attack the C3 to produce a negatively charged oxygen for the reaction without catalyst. The reaction proceeded through a keto-enol tautomerism followed by elimination of H to give 3a (Scheme 6).
The 1 H-NMR spectra of 5a and 5c displayed signals attributed to methoxy protons at δ 3.98 and δ 3.98 ppm, respectively. A broad singlet signal at δ 7.27 ppm in the spectrum of 5a was assigned to the non-chelated amine proton, whereas for 5c, the spectrum show a higher chemical shift at δ 9.84 ppm which is due to the hydrogen bond between the amine hydrogen and carbonyl oxygen. Both compounds shared the same molecular formulae, C 19 H 19 NO 3 , as the mass spectra exhibited a molecular ion peak at m/z 309. Selected HMBC correlations were shown in Figure 1.
The 1 H-NMR of 5d showed the disappearance of the signal for the methoxy protons, which clearly indicated that nucleophilic aromatic substitution had occurred. The broad singlet at δ 10.73 ppm was assigned to the chelated amine proton. The 13  The reaction occurred as the methoxy group interacts with the catalyst. The positive charge on oxygen is then stabilized by the C1-O bond cleavage driven by the formation of PhI(OAc)(OCH 3 ) and leaves the aromatic carbocation. With increasing heat, the carbocation is readily attacked by BuNH 2 to produce 5c and 5d. In contrast a slow reaction at RT favored the carbocation rearrangement to produce 5a and 5b (Scheme 7). This aromatic substitution of methoxy groups by amines is also supported by literature precedents [21,25,28].

Amination of 7
2-(Butylamino)-1,4-dihydroxyanthracene-9,10-dione (1a) was successfully obtained in 83% yield under a similar amination approach applied to the acylated anthraquinone 7 (Scheme 4). The mechanism suggested that the acetate groups were reduced by the conditions and reagent used to give 1,4-dihydroxyanthraquinone (1) as an intermediate before it further reacted with BuNH 2 /PhI(OAc) 2 to give 1a (Scheme 8). It was observed that 1a can still be obtained when the reaction was performed without catalyst or at 80 °C but gave lower yield. Heat applied to the reaction induced the reaction to be faster as the reaction time reduced from 6 h to 4 h ( Formation of 5b and 5d were followed the same step of mechanisms. The 1 H-NMR for 1a showed the disappearance of the COOCH 3 signal which proved that the acyl groups were reduced. This is also proven by the existence of strong chemical shifts at δ 14.23 and δ 13.90 ppm which represented the two chelated OH. A broad singlet proton at δ 5.60 ppm was attributed to the amine proton. All three new aminoanthraquinones 3a, 5a and 5b were subjected to cytotoxic activity assays against MCF-7 (estrogen receptor positive human breast) and Hep-G2 (human hepatocellular liver carcinoma) cancer cell lines with compounds 3a and 5a being the most cytotoxic (IC 50 1.1-3.0 µg/mL). Increasing the number of amino substituents (compound 5b) on the structure of anthraquinone basically reduced the activity against both MCF-7 and Hep-G2 cancer cell lines with IC 50 of 3.0 and 13.0 µg/mL, respectively (Table 3). Saha et al. [34] reported that the starting material, quinizarin 1 (free hydroxy groups) was not active towards MCF-7 but the methylated hydroxyl derivative of quinizarin showed some potential with IC 50 values 70-80 µM. The results of this study also supported their finding where the presence of methoxy or amino groups enhanced the cytotoxic activity of the compounds. Compounds 3a, 5c and 5d have been screened for antimicrobial activity using four different types of microbe, methicillin-resistant Staphylococcus aureus (MRSA), Pseudomonas aeruginosa, Candida albicans and Escherichia coli. The other two new compounds, 5a and 5b were replaced by compounds 5c and 5d due to insufficient materials and based on the similarity of their structures. All four compounds have the same methoxy and amino substituents and only differ at their amino positions. Therefore, it was assumed that all four compounds might give similar activity results. However, none of compounds had any effect on the growth of all four bacteria or the fungus.

General
All chemicals are commercially available and were of analytical grade and used without purification unless otherwise stated. All reactions were monitored by TLC, using Silica gel 60 F245 (Merck KGaA) precoated aluminium backed plates and visualised by UV and H 2 SO 4 solution. All organic extracts were dried over Na 2 SO 4 and evaporated using rotary evaporator. Column chromatography was performed on silica gel 60. Melting points were recorded by digital melting point equipment (Electrothermal IA9000 Series). The IR spectra were obtained by using Perkin-Elmer FT-IR Model Spectrum 100 series spectrophotometer using UATR techniques and the adsorption bands were measured in the range of 280-4000 cm −1 . MS spectra were recorded using equipped Shimadzu model QP5050A series. The 1D NMR and 2D NMR spectra were run on a JEOL machine at 400 MHz or 500 MHz. All chemical shifts, δ were recorded in ppm relative to TMS signal. The coupling constants J are given in Hz.

Acylation
A mixture of compound 1 (1 equiv.), Ac 2 O (33 equiv.), and NaOAc (1 equiv) was stirred at 100 °C for 2 h. The reaction mixture were added to ice cold water and extracted with CH 2 Cl 2 (3 × 25 mL). The combined organic layers was dried and evaporated under vacuum. The crude product was purified by column chromatography (CH 2 Cl 2 -petroleum ether, 3:1) to produce compounds 6 and 7.

Cytotoxic Assays
The cytotoxic assay was carried out according to the method described by Sukari et al. [35]. The MCF-7 (estrogen receptor positive human breast) and Hep-G2 (human hepatocellular liver carcinoma) cancer cells were purchased from ATCC. The cells were grown and maintained in RPMI 1640 media, supplemented with 10% fetal calf serum (FCS) and 1% antibiotic penicillin-streptomycin in an atmosphere of 5% CO 2 at 37 °C. The medium was used to dilute the cells to a concentration of 5 × 10 5 cells/mL. From this cells suspension, 100 μL of various concentrations of the synthesised compounds were pipetted into a 96-well micro titer plate and incubated in 37 °C, 5% CO 2 incubator for 72 h. The various concentration used were 100, 50, 25, 12.5, 6.25, 3.125, 1.56 μg/mL. The assay of each concentration of synthesised compounds was performed in triplicate and the control wells of untreated population were also included. After three d, the fraction of surviving cells was determined relative to the untreated cells population by the colorimetric MTT (3-(4,5-dimethylthiazol-2-yl)-2,5diphenyltetrazolium bromide) method where the viability of cells was measured by 20 μL of blue formazan crystals of MTT solution (5 mg/mL in phosphate-buffered saline, PBS) added to each well followed by incubation in 37 °C, 5% CO 2 incubator for 3-4 h. 100 μL of cells suspension or cells monolayer in each micro titer was removed from each well. The plate was left at room temperature for 30 min before reading the absorbance. The absorbance was read with the multiwell scanning spectrophotometer (ELISA reader) test wavelength of 570 nm and reference wavelength of 630 nm. The cytotoxic index used was IC 50 which is the concentration that yield 50% inhibition of the cells compared with untreated control.

Antimicrobial Assay
The modified disc diffusion method was followed a procedure by Garba and Okeniyi [36]. Nutrient agar (20 g) was suspended in 1 L of distilled water and stirred. It was boiled to dissolve homogenously and autoclaved at 121 °C for 20 min. The agar was allowed to cool to 50 °C and poured into sterile disposable petri dishes. Four microbes, methicillin-resistant Staphylococcus aureaus (MRSA), Pseudomonas aeruginosa, Candida albicans and Escherichia coli were inoculated into prepared nutrient broth and incubated at 37 °C for overnight. The suspension of the microbes in the broth was inoculated on the nutrient agar using sterile cotton bud. The sterile 6 mm paper discs were impregnated with the synthesised compounds in concentrations of 20, 10, 5, 2, 1, 0.5 and 0.1 mg/mL and allowed to soak for 1 min. The paper discs were removed, dried and placed on the surface of agar plates inoculated with the microbial cultures. Each synthesised compound was tested in triplicate. Paper discs impregnated with acetone were used as a negative control. The petri dishes were incubated in inverted position at 37 °C for 24 h. The zones of inhibitions (clear area without bacterial growth) were measured in cm.