Synthesis of 1-hydroxy-3-O-substituted xanthone derivatives and their structure-activity relationship on acetylcholinesterase inhibitory effect

Abstract This study focused on the synthesis of 1,3-dihydroxyxanthone (1) and its new derivatives with alkyl (2a–2f), alkenyl (2 g–2k), alkynyl (2 l–2n), and alkylated phenyl (2o–2r) groups at C3 position. The structures of these compounds were confirmed by MS, NMR, and FTIR spectroscopic data. All the substituted xanthones (2a–2r) showed significantly stronger acetylcholinesterase (AChE) inhibitory activities than 1. Compounds 2g and 2j exhibited the strongest activities with the IC50 values of 20.8 and 21.5 μM and their enzyme kinetic analyses indicated a mixed-mode inhibition. Molecular docking study revealed that 2g binds favourably to the active site of AChE via π–π stacking and hydrogen bonding from the xanthone ring, in addition to π-alkyl interaction from the substituent group. These xanthone derivatives are potential lead compounds to be further developed into Alzheimer’s disease drugs. Graphical Abstract


Introduction
Processes in the human body, including sensory and motoric, are coordinated by sophisticated nervous system. The nervous system is vulnerable as slight interferences to the neuron's structural pathway can induce dysfunction in the nervous system (Feigin et al. 2020). Among neurological disorders, Alzheimer's disease (AD) has the highest prevalence, with an estimated 47 million cases worldwide in 2020 (Alzheimer's Association 2020). The widely recognised pathological mechanisms of AD include progressive loss of cholinergic neurotransmission, aggregation of beta-amyloid plaques (Ab-plaques), and neurofibrillary tangles (NFT) from abnormal hyperphosphorylations of tau proteins (Hoenig et al. 2018). The most successful approach to manage AD is based on the enhancement of cholinergic activity by administering acetylcholinesterase (AChE) inhibitors (Colovi c et al. 2013;Haake et al. 2020). These agents inhibit the hydrolysis of a neurotransmitter, acetylcholine, to improve the cognitive function of AD patients (Grossberg 2017). Unfortunately, the current cholinesterase inhibitors, donepezil, galantamine and rivastigmine, are associated with adverse effects, including nausea, dizziness, blurred vision, loss of appetite, vomiting and diarrhoea (Craig et al. 2011;Birks and Harvey 2018). Tacrine, another cholinesterase inhibitor and the first anti-Alzheimer's drug approved by the FDA (Crismon 1994), has been discontinued clinically due to concerns in acute liver toxicity (Blackard et al. 1998;Lou et al. 2015). Hence, due to the increased prevalence and irreversible consequences of AD, the development of drugs for AD is highly imperative.
Extensive studies on new leads of AChE inhibitors were recently done by screening bioactive natural compounds (Ee et al. 2015;Teh et al. 2016), synthetic sources (Kwong et al. 2017;Svobodova et al. 2019), and through computational-aided approaches (Menendez et al. 2017). Among the bioactive compounds, xanthones are notable for their pharmacological benefits including anti-cancer , antibacterial (Narasimhan et al. 2017), anti-inflammatory (Gunter et al. 2020), anti-malarial (Syahri et al. 2017) and anti-cholinesterase (Loh et al. 2021;Vanessa and Mah 2021). In particular, xanthones as secondary plant metabolites (Nag et al. 2015;Saenkham et al. 2020) and their synthetic derivatives (Luo et al. 2017;Chi et al. 2020) have received attention due to their AChE inhibitory activity. Xanthones are heterocyclic compounds with a symmetrical dibenzo-c-pyrone as a basic scaffold, which is deemed 'privileged' due to its planar tricyclic nucleus with a carbonyl moiety in the central skeleton that can carry a wide variety of substituents ( Figure 1). Its structural features allow xanthones to bind to various biomolecular targets, giving rise to a plethora of pharmacological benefits.
The current study aimed to synthesise a series of new xanthones derivatives from 1,3-dihydroxyxanthone with different substituents, including alkyl (2a-2f), alkenyl (2g-2k), alkynyl (2l-2n) and alkylated phenyl (2o-2r) groups (Scheme 1). The AChE inhibition effects of these xanthone derivatives were evaluated, and their structure-activity relationship (SAR) was elucidated. Furthermore, enzyme kinetic analysis was carried out on the xanthone derivatives with the strongest activities, followed by molecular docking simulations to illustrate their modes of inhibition and binding interactions with AChE.

Results and discussion
2.1. Synthesis & characterisation of 1,3-dihydroxyxanthone (1) 1,3-Dihydroxyxanthone (1) was obtained as a yellow crystal with 82% yield through the acylation-dehydration reaction between salicylic acid and phloroglucinol (Eaton et al. 1973). The structure of 1 was confirmed by the spectral data obtained from FT-IR, MS, NMR and FT-IR spectroscopies. The mass spectrum revealed a molecular ion peak at m/z 228, which coincides with the molecular weight of 1. The FTIR spectrum demonstrated multiple essential peaks, including absorptions of free O-H at 3398 to 3478 cm À1 , -OHÁÁÁO at 2600 to 3100 cm À1 , sp 2 C-H stretch at 3073 cm À1 , sp 3 C-H stretch at 2940 cm À1 , C ¼ O stretch at 1651 cm À1 , aromatic C ¼ C stretch at 1610 and 1454 cm À1 , C-O-H bend at 1341 cm À1, C-O stretch at 1312 and 1076 cm À1 . The 1 H-NMR spectrum revealed the presence of two -OH groups at C-1 and C-3 as two singlet peaks at d 12.90 and 9.92 ppm, respectively. The electron-withdrawing effects of the carbonyl group at C-9 lead to a de-shielding effect, resulting in a higher chemical shift for 1-OH than 3-OH. Six signals were observed in the range of d 6.21-8.39 ppm belonging to the aromatic protons present in the tricyclic skeleton of xanthone. The 13 C NMR revealed the presence of thirteen carbons which are in agreement with the molecular formula of 1, C 13 H 8 O 4 . A highly de-shielded quaternary signal at d 180.5 was assigned to the carbonyl carbon at C-9, followed by the hydroxylated carbons at C-3 (d 165.7) and C-1 (163.9). Two signals at d 158.1 and 156.0 were assigned to the oxygenated aromatic carbons at C-4a and C-5a. Six methine carbons and two quaternary aromatic carbons were found in the range of d 103.0 to 135.5 ppm. The spectral data assignment for 1 are consistent and in good agreement with the reported literature values (Frahm and Chaudhuri 1979;Qin et al. 2013)

Synthesis & characterisation of derivatives (2a-2r)
The parent compound, 1 was reacted with alkyl, alkenyl, alkynyl and alkylated-phenyl bromides through nucleophilic substitution reaction to afford eighteen new xanthone derivatives, 2a-2r. Sixteen derivatives were obtained in good yield of greater than 70%, except for 2l and 2p, obtained at 67.0% and 44.6% of yield.
The mass spectra of 2a-2r revealed molecular ion peaks that correspond with the molecular weight of the respective compound. From the FT-IR, the structural characteristics of 2a-2r are similar to 1, except for the broad O-H peak at the region of 3398-3478 cm À1 were absent in 2a-2r. The results indicate that the hydroxyl group at C-3 was successfully replaced by the substituents groups.
The etherification of 2a-2r was further confirmed by 1 H NMR, where a singlet at d 9.9 ppm corresponding to 3-OH in the structure of 1 has disappeared. Additional signals of alkoxy protons of side chains at C-3 were found around d 0.9-4.1 ppm across four series of derivatives (2a-2r). Specifically, the presence of C ¼ C groups in the derivatives of alkenyl-series (2g-2k) was confirmed by two signals located at a higher chemical shift region of d 5.1-5.9 ppm. The shielding effect of CC in the alkynyl-series (2l-2n) resulted in a slightly downfield chemical shift at d 4.7-4.9 ppm. For the alkylated-phenyl series (2o-2r), the presence of an aromatic ring from the side chain was confirmed by three additional methine proton peaks at d 7.1-7.3 ppm. Notably, the substitution of 1-OH was omitted due to a robust intramolecular hydrogen bonding with the adjacent carbonyl group (Freitas and Ribeiro da Silva 2018). This is evident through the presence of a downfield singlet at around d 12.8 ppm, which corresponds with the chelated hydroxyl group at C-1 in the 1 H NMR spectra of 1 and 2a-2r.
The alkoxy peaks that appeared at the region of d 57-75 ppm in 13 C NMR spectra of xanthone derivatives 2a-2r validate the successful substitution reaction through the formation of an ether bonding. In addition, the C ¼ C carbon in 2g-2k was confirmed by the observation of two signals at d 115-136 ppm. Another two signals at around d 85-73 ppm are associated with the CC group in 2l-2n. In 2o-2r, three additional signals related to the methine carbon were detected at d 126, 128 and 139 ppm, validating the presence of an additional aromatic ring in their structures.

Acetylcholinesterase inhibitory activities
The parent compound (1) and the derivatives (2a-2r) were investigated for their AChE inhibitory activities using slightly modified Ellman's method (Ellman et al., 1961). The standard drug, tacrine, inhibited AChE with an IC 50 value of 0.87 lM. The result is in good Scheme 1. Synthetic of 1,3-dihydroxyxanthone (1) and its derivatives (2a-2r). agreement with the reported IC 50 value of tacrine against AChE in past literature (Jang et al. 2018;Svobodova et al. 2019). The xanthones, 1 and 2a-2r were firstly evaluated for their AChE inhibition activities at a concentration of 45 lM. The parent compound 1 did not show any inhibition activity towards AChE at this concentration. Interestingly, all derivatives, 2a-2r demonstrated moderate-to-good inhibitory effects. Most of the derivatives could inhibit 50% of AChE at 45 lM, except for 2c, 2l, 2m and 2n. Subsequently, the IC 50 values of the 1 and 2a-2r were determined, as presented in Table 1. The IC 50 value obtained for 1 was 157.47 lM, close to that of published data of 150.61 lM (Thongchai et al. 2014). The activity of 1 against AChE is significantly weaker than 3-hydroxyxanthone (IC 50 ¼ 2.4 lM) (Loh et al. 2021). This finding shows that the hydroxyl group at C-1 of xanthone failed to contribute favourable effects to anti-AChE activities. This might be due to its close distance to the adjacent carbonyl group, resulting in a significant chelating effect (Freitas and Ribeiro da Silva 2018;Vanessa and Mah 2021).
As expected, the derivatives (2a-2r) showed significantly stronger anti-AChE effects than 1 with IC 50 values ranging from 20.81 to 71.22 lM. The results indicate that substituting hydrophobic moiety at C-3 of 1 improved AChE inhibition significantly. Among the derivatives, 2g and 2j appeared as the strongest AChE inhibitors with IC 50 values of 20.81 and 21.51 lM. Notably, the inhibition effects of these two derivatives are approximately seven-fold stronger than 1. In agreement with our findings, Qin et al. (2013) revealed that additional hydrophobic nature is exceptionally favourable for the anti-AChE effects of hydroxy xanthones, evidenced by having an allyloxyl group at C-3 resulted in an approximately eighteen times stronger activity. The activity was further increased by forty and forty-five times with substitution of prenyloxyl and methoxyl groups at C-3, respectively (Qin et al. 2013).
Subsequently, the effects of different types of xanthone substituents at C-3 against AChE were elucidated. The derivatives, substituted with alkenyl (2g-2k) and alkylatedphenyl (2o-2r) groups, were significantly stronger than alkyl (2a-2f) and alkynyl (2l-2n) groups (Table 1). The pronounced AChE inhibitory effects of 2g-2k clearly showed the importance of C ¼ C in contributing to the bioactivity. The results are consistent with a recent report that revealed formation of favourable p-r and p-alkyl interaction between the C ¼ C moieties in prenylated and geranylated xanthones with Trp82 in the choline-binding pocket and Trp86 in the anionic active site of the AChE, respectively (Saenkham et al. 2020).
Besides that, the positive effect of the other aromatic ring in 2o-2r is apparent ( Table 1). The xanthones bearing phenylpropyl (2o) and phenylbutyoxyl (2p) groups showed a significant increase in AChE inhibition levels, approximately 80%-85% stronger than 1. Similar findings were observed in previous study, whereby substituting hydroxyl group at C-3 with a phenylpropxyl or phenylbutoxyl group resulted in 48%-63% stronger anti-AChE effects compared to its parent, 3-hydroxyxanthone (Loh et al. 2021). This observation is in agreement with several reports indicating that arene groups in xanthones allow stronger binding affinity to AChE through hydrophobic p-p interactions with Trp84 in the choline-binding pocket (Rampa et al. 2001;Belluti et al. 2005;Seca et al. 2014).

Enzyme kinetics analysis
The results of enzyme kinetics analysis of 2g and 2j were presented by the Lineweaver-Burk plot in Figures 2 and 3. Both the reciprocal plots of 2g and 2j showed the lines that intersect at the secondary quadrant, indicating a mixed-mode inhibition. This assumption was supported by the Michaelis-Menten parameter tabulated in Tables S1 and S2, which showed both the maximal velocity of the AChE-ATCI enzyme-substrate reaction (V max ) and affinity (K m ) were affected by the addition of 2g and 2j. A mixed- Figure 4. Orientation of 1-hydroxy-3-(but-3-en-1-yloxy)-9H-xanthen-9-one (2g) (carbon atoms are colored green) in the active sites of (A) Electrophorus electricus AChE (PDB ID: 1C2O) and (B) human AChE (PDB ID: 4PQE). The atoms in the residues are coloured as follows: carbons in cyan, oxygen in red, and nitrogen in blue. Distance is indicated in angstroms, Å. mode inhibition allows the binding of a substrate (ATCI) to an enzyme (AChE) but with reduced affinity. This indicate that the xanthone derivatives are likely to bind to either the active site of AChE, including the esteratic site (ES) and anionic catalytic site (AS), and allosteric sites such as peripheral active site (PAS).

Binding interactions of 2g with AChE
The flexible docking results, as shown in Figure 4, suggest that 2g could bind favourably (-9.0 kcal/mol) to the active site of Electrophorus electricus AChE. The planar geometry of the xanthone ring allowed it to enter the gorge and bind deeply into AS. The ring forms ᴨ-ᴨ stacking with the phenol side-chain of Tyr341. On the other hand, the carbonyl group directed towards the hydration site comprised of D74, T83, W86, N87, and S125 residues and formed hydrogen bonding interactions with one of the buried water molecules and hydroxyl side-chain of Tyr124. As expected, the hydroxyl group in the first position of the xanthone ring failed to improve the binding interactions with either the crystal water molecules or the adjacent residues. On the other hand, we observed that 2g exhibited a similar binding pose in the human AChE (-13.9 kcal/ mol). The but-1-ene group appeared to be flexible and could interact with Leu289 and Ser293 (Electrophorus electricus AChE) and Tyr72, Ser298, and Tyr124 (human AChE), further validating the importance of C ¼ C group in the AChE inhibition effects of xanthones.

Conclusions
Overall, 1,3-dihydroxyxanthone (1) and a known xanthone derivative (2a), along with seventeen new derivatives (2b-2r) were successfully synthesised. Remarkably, all derivatives (2a-2r), having four types of different hydrocarbon substituents, including alkyl, alkenyl, alkynyl, and alkylated-phenyl groups possessed stronger AChE inhibitory effect than 1. Particularly, 2g and 2j exhibited excellent inhibitory effects amongst the derivatives with IC 50 values of 20.8 and 21.5 lM, respectively. Kinetic analysis suggested that 1-hydroxy-3-O-substituted xanthones inhibited AChE in a mixed-mode manner, indicating the possibility of binding to both AS and PAS of AChE. SAR analysis revealed that the hydrophobic interactions of the substituent groups at C-3 of xanthone gave positive impacts on the inhibition effects. Specifically, xanthone derivatives substituted with C ¼ C or alkylated phenyl in linear chains are favourable for the activity. The molecular docking study of 2g further confirmed the importance of the hydrophobic substituents by forming p-alkyl interactions with AChE. Two new xanthone derivatives (2g and 2j) with promising AChE inhibitory properties are strongly recommended to be further studied for hit-to-lead optimisation of new Alzheimer's drug development.

Disclosure statement
No potential conflict of interest was reported by the authors.

Funding
This research is supported by the Ministry of Higher Education Malaysia (MOHE) under the Fundamental Research Grant Scheme (FRGS/1/2019/STG01/TAYLOR/02/1). Taylor's University was also acknowledged in providing the funding for Vice-Chancellor Award Programme.