Design, synthesis, and evaluation of curcumin analogues as potential inhibitors of bacterial sialidase

Abstract Sialidases are key virulence factors that remove sialic acid from the host cell surface glycan, unmasking receptors that facilitate bacterial adherence and colonisation. In this study, we developed potential agents for treating bacterial infections caused by Streptococcus pneumoniae Nan A that inhibit bacterial sialidase using Turmeric and curcumin analogues. Design, synthesis, and structure analysis relationship (SAR) studies have been also described. Evaluation of the synthesised derivatives demonstrated that compound 5e was the most potent inhibitor of S. pneumoniae sialidase (IC50 = 0.2 ± 0.1 µM). This compound exhibited a 3.0-fold improvement in inhibitory activity over that of curcumin and displayed competitive inhibition. These results warrant further studies confirming the antipneumococcal activity 5e and indicated that curcumin derivatives could be potentially used to treat sepsis by bacterial infections.


Introduction
Sepsis, a life-threatening organ dysfunction caused by a dysregulated host response to infection, is caused by an overwhelming immune response to an existing bacterial infection 1 . It commonly occurs in the ageing population and results in $20-30 million cases annually worldwide. Overall, sepsis remains one of the top five causes of death worldwide 2 , also the mortality rate is $20-50% for severe sepsis and 40-80% for septic shock. Especially, bacterial sepsis is a major cause of mortality of hospitalised patients, thus the development of drugs for bacterial sepsis is urgently needed and many efforts have been undertaken in the medicinal and pharmaceutical industry.
The Gram-positive bacterium, Streptococcus pneumoniae, is one of the causes of sepsis. It also major human pathogen and causes a variety of diseases, including bacterial meningitis, otitis media, pneumonia, conjunctivitis [3][4][5][6] . Several virulence factors contribute to colonisation and early infection processes, above all sialidases from bacteria are considered key virulence factors 7 . Sialidase removes the terminal sialic acid residues from host cell surface glycans, unmasking receptors that facilitate bacterial adherence and colonisation 8 . This process causes resistance to penicillin and other antibiotics that are used to treat S. pneumoniae infection 9 . According to known literature, all clinical isolates of S. pneumoniae have sialidases activity known to be involved in sepsis 10 . S. pneumoniae sialidase genes in clinical pneumococcal isolates determined that Nan A, Nan B, and Nan C are present in 100%, 96%, and 51% of these strains. Among these sialidases, Nan A has been shown to play an essential role in host-pneumococcal interactions in the respiratory tract and sepsis in mouse models 7,11,12,33 . Therefore, high-affinity inhibitors that can block Nan A are potential agents for prevention and treating sepsis. In the last few years, several studies have reported the discovery of viral or bacterial sialidase inhibitors from an isolated natural product such as flavonoids, coumarins, diplacone, mimulone, pterocarpans, and phlorotannins. However, these compounds are known inhibitors as Clostridium perfringens (Cp-Nan I) or viral sialidase 9,[14][15][16][17] . Recently, some studies have reported to inhibition of S. pneumoniae sialidase such as diazenylaryl sulphonic acids, malabaricone C, Artocarpin, and anthraquinone glycosides 9,10,[18][19][20] . Therefore, to develop novel bacterial sialidase inhibitors, we focused on the natural product, Turmeric, because it had not yet been evaluated.
Turmeric has been used as a traditional medicine for conditions such as liver disease 21 , indigestion 22 , rheumatoid arthritis 23 , and insect bites 24 and is consumed daily by millions of people for the treatment of various diseases. Curcumin is the primary component of Turmeric and has a feruloyl methane group containing methoxy, hydroxyl, and heptadienyl with a 1,3-diketone moiety. Curcumin has been extensively studied in the past few decades as an important therapeutic compound. In addition, it still receives a lot of attention for its biological properties, including anti-inflammatory, anti-viral, anti-bacterial, anti-cancer, anti-oxidant, and anticarcinogenic activities 25 , and its use in debilitating diseases such as Crohn's disease, ulcerative colitis 26 , and Alzheimer's disease 27,28 . Therefore, many studies evaluating the biological activity of curcumin have been performed and potential curcuminoids have been developed for several diseases.
In this study, we report that Turmeric and curcumin derivatives can targeting the S. pneumoniae Nan A. Designed strategies for synthesis of curcumin analogues are shown Scheme 1.

General
All the chemical reagents used in this work and curcumin (4a) were purchased from commercial suppliers (Aldrich, St. Louis, MO; TCI, Japan; Alfa Aesar, Haverhill, MA or Acros Organics, USA companies) and used without further purification. The 1 H and 13 C NMR spectra were recorded using a JEOL ECA-500 spectrometer, Japan at 500 MHz and 125 MHz, respectively, with chemical shift (d) values reported in ppm unit. Multiplicities are describes as singlet (s), doublet (d), doublet of doublet (dd), triplet (t), multiplet (m), and broad exchangeable proton (bs). High resolution mass spectra were obtained on a GC Mate 2, JEOL. A CEM Discover system (No. 908005) fitted with a temperature controller was used for microwave reactions. Irradiation was initiated at 300 W to raise the temperature to the set point (150 C). Reactions were monitored by thin-layer chromatography (TLC) with Merck's DC-Fertigiplatterm Kiegel 60 GE254 plates. Visualisation was accomplished with either UV light or by immersion in a solution of phosphomolybdic acid (PMA) followed by heating on a hot plate for $10 s. The reaction products were purified by open column chromatography using silica gel produced by Merck (Darmstadt, Germany) (Silica gel 60; 63-200 mesh, ASTM) or Cosmosil 140 C-18 OPN produced by Nacalai Tesque, Inc., USA.

Expression, purification, and preparation of S. pneumoniae Nan A
We have synthesised and expressed the full-length genes for the S. pneumoniae sialidase (Nan A) in Escherichia coli. The gene encoding Nan A ( Figure S28, GenBank accession no. COT45929.1, PDB: 2VVZ) of S. pneumoniae TIGR4 was synthesised (Thermo Fisher Scientific GENEART GmbH, Regensburg, Germany). The synthesised gene was inserted into the cloning sites of a pET151/d-TOPO vector (Invitrogen, Carlsbad, CA) containing a 6x His-tag at the C-terminus. S. pneumoniae sialidase was expressed and purified from E. coli BL21 (DE3) (HIT; Real biotech Co., Taipei, Taiwan). The purified Nan A was detected at $56.6 kDa with greater than 90% purity using sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) ( Figure S29). The purified sialidase had specific activities (K m values) of 43.9 mM using 4-methylumbelliferyl-a-D-N-acetylneuraminic acid sodium salt hydrate (MUNANA; Catalogue No. M8639; Sigma) as substrate ( Figure S30).

Enzyme inhibition activity
As described, the inhibitory effects of compounds on S. pneumoniae Nan A were measured using a fluorescence (FL)-based assay. In this assay, the 4-methylumbelliferyl-a-D-N-acetylneuraminic acid (Sigma Chemical Co., St. Louis, MO) was used as a substrate, and the enzyme activity was determined by measuring the increase in fluorescence by continuously monitoring the reactions at 450/ 40 nm with excitation at 365 nm using a SpectraMax M2e Multimode Reader (Molecular Devices Co.). The IC 50 values of the synthesised compounds were measured in a reaction mixture containing enzyme (final concentration of Nan A, 2.2 nM), the test compounds (from 0 to 200 mM), and 50 mM of substrate in 20 mM Sodium phosphate buffer (pH 7.5, containing 300 mM NaCl). To determine the enzyme activity, the experimental data were fit to a logistic curve with Equation (1), a time-drive protocol was used and the initial velocity was recorded over a range of concentrations, and the data were analysed using a nonlinear regression program (Sigma Plot; SPCC Inc., Chicago, IL).
where C is the fluorescence of the control (enzyme, buffer, and substrate) after 60 min of incubation, C 0 is the fluorescence of the control at 0 min, S is the fluorescence of the tested samples (enzyme, sample solution, and substrate) after incubation, and S 0 is the fluorescence of the tested samples at 0 min. The production of 4-methylumbelliferone was measured by monitoring the fluorometric determination at excitation wavelength 365 nm/emission wavelength 450 nm.

Enzyme kinetic study
The inhibition mechanism was determined, and the apparent inhibition constants (K i ) for the respective sialidase (Nan A) were performed on the test compounds, for which the IC 50 values were below 25 mM. The test compounds were studied at three different concentrations that were chosen based on the IC 50 values obtained with each sialidases ($1/2 Â IC 50 , IC 50 , 2 Â IC 50 ). The concentrations of marker substrates were chosen ($1/4 K m , 1/2 K m , K m ) with regard to their Michaelis-Menten kinetics. The K i values were calculated by nonlinear regression analysis by fitting different models of enzyme inhibition to the kinetic data using SigmaPlot Enzyme Kinetics Module 1.3 (SPSS Inc., Chicago, IL). The inhibition mechanism of the compounds was determined by comparing the statistical results, including the Akaike's information criterion values, of different inhibition models and selecting the one with the best fit.
2.6 General procedures for the synthesis of curcumins and the characterisation of synthesised compounds 2.6.1 Synthesis of curcumin derivatives using Pabon's reaction (compounds 3, 4b, 4e, 5a-5p, 5r, 5s) Boron trioxide (43.44 mmol) was added to a solution of 2,4-petadione (65.16 mmol for 3, 4) or monophenyl intermediate (3, 21.7 mmol) in ethyl acetate (100 ml) at ambient temperature. After stirring for 1 h at 90 C, the corresponding benzaldehydes (21.7 mmol) and triethyl borate (21.7 mmol) in ethyl acetate were added to the reaction mixture. The mixture was stirred for 2 h at 90 C, then n-BuNH 2 (21.7 mmol, 1 equiv., 7% solution in ethyl acetate) was slowly added, and the mixture was stirred at 90 C until the aldehydes disappeared on TLC monitoring. The reaction mixture was then cooled to 50 C and 1 M HCl (aq.) was added.
Twenty asymmetrical curcumin derivatives, 5, were prepared using Pabon's reaction with 3 and the corresponding aldehydes (Scheme 3) 39 . Compound 5q was synthesised by deacetylation of 5n using 1 M HCl aqueous solution.
Symmetrical curcumin derivatives 4c, 4d, and 6-8 were prepared by treating the corresponding reagent with curcumin (Scheme 4). Compound 4c was prepared with MeI/K 2 CO 3 in acetone under reflux conditions and compound 4d was produced by a demethylation reaction after treating tribromoborane with dry dichloromethane 41 . Tetrahydrocurcumin (6) was prepared by hydrogenation using 10% Pd-C as a catalyst 42 and acetylated curcumin 7 was prepared using acetic anhydride with DMAP as a catalyst with anhydrous pyridine 34 . Compound 8 was prepared by condensation the 1, 3-diketone in curcumin with hydrazine hydrate under reflux condition 28 . The structures of all compounds were characterised using 1 H NMR, 13 C NMR, and EI-HRMS.

Biological evaluation
Hydrolytic activity of S. pneumoniae Nan A was confirmed using DANA (Neu5Ac2en). The IC 50 value of DANA with respect to S. pneumoniae sialidase inhibition was 4.8 ± 1.1 mM 12 . To identify a sialidase inhibitor of S. pneumoniae, the inhibitory activity of Turmeric ethanol extract and its three major components was compared. For the methanol extract, the Nan A activity was 88% at 30 mg/mL. The sialidase inhibitory activity of the major components of Turmeric was as follows; curcumin (4a, IC 50 ¼ 0.6 ± 0.1 mM), demethoxycurcumin (5p, IC 50 ¼ 0.6 ± 0.2 mM), and bisdemthoxycurcumin (4e, IC 50 ¼ 4.0 ± 1.2 mM). Among these compounds, 4e, in which two of the methoxy groups were removed from curcumin showed diminished inhibitory activity. Based on this result, we predicted that curcumin derivatives would have inhibitory activity against sialidase from S. pneumoniae Nan A and that the methoxy group played an important role in the inhibitory activity. Therefore, we modified curcumin and evaluated the inhibitory effects on S. pneumoniae Nan A (Table 1).   To evaluate the functionalisation of the curcumin skeleton to find a suitable backbone. Hydrogenation of the heptadienyl group (6, IC 50 ¼ 82.1 ± 7.6 mM) resulted in significantly diminished inhibitory activity. Next, to confirm the phenyl group at the 7-position, inhibitory activity after the addition of a pyridinyl (5g) or phenyl ring (5h) was compared. The results indicated that the pyridinyl group (5 g, IC 50 ¼ 4.4 ± 0.1 mM) was more effective than the phenyl group (5 h, IC 50 ¼7.1 ± 0.1 mM), but both compounds showed lower inhibitory activity than curcumin. To investigate the effect of 1,3-diketone, 4-hydroxy or 3-methoxy groups, synthesised and compared with six kinds of curcumins. First, to investigate the effect of the 1, 3-diketone moiety, pyrazole (8, IC 50 ¼ 6.2 ± 1.5 mM) or carbocyclic 1, 3-diketone (4f, IC 50 ¼ 4.7 ± 0.5 mM) was substituted. Second, to investigate the effect of the 4-hydroxyl groups in curcumin, acetyl (7, IC 50 ¼ 2.6 ± 0.6 mM) or methyl groups (4c, IC 50 ¼ 3.4 ± 1.0 mM) were substituted and conversion of the hydroxy at the para position to ester (5i, IC 50 ¼ 1.0 ± 0.5 mM). Third, to investigate the 3-methoxy groups, elimination of the 3-methoxy groups (4d, IC 50 ¼ 1.5 ± 0.7 mM) and conversion of the methoxy at the meta position to ethoxy (5f, IC 50 ¼ 1.5 ± 0.8 mM) were compared. Inhibitory activity of these compounds was diminished than curcumin (4a). This suggested that 1,3-diketone and a heptadienyl group were essential functional groups for sialidase inhibition and either a methoxy or a hydroxyl group was required.
Based on these results, we substituted a methoxy or hydroxyl group at the para-or meta-position of the 1,7-diphenylhepta-1,6diene-3,5-dione backbone. To confirm the positional tendency of the methoxy and hydroxyl groups, we reacted it with feruloyl (3a) or isoferuloyl (3b) acetone or hispolon (3c) with the corresponding aldehyde 2. Because a methoxy or hydroxyl group was  essential for sialidase inhibition, compounds 4c and 4d were excluded from consideration. As shown in Table 1, IC 50 values ranged from 0.2-1.5 mM, indicating there was no positional tendency for inhibition (4a, 4 b, and 5a-5e in Table 1). Among the examined compounds, 5e, containing the catechol with the isoferuloyl moiety, was the most potent inhibitor (IC 50 ¼ 0.2 ± 0.1 mM), with a 3.0-fold improvement in inhibitory activity over that of curcumin. Thereafter, to investigate the electronic effect, we substituted an electron-donating group (EDG) or electron-withdrawing group (EWG) into the para position of the phenyl rings, including nitro (5j), trifluoromethyl (5k), fluoro (5l), bromo (5m), acetamido (5n), methoxy (5o), hydroxyl (5p), amino (5q), N, N-dimethylamino (5r), and piperidinyl (5s) groups. The IC 50 values ranged from 0.6-1.8 mM and the electronic effect did not influence in Nan A inhibition. Based on these observations, we investigated the kinetic mechanisms of inhibitors with IC 50 values of 25 mM or less. We selected the major components of Turmeric (4a, 4e, 5q) and compound 5e as the most potent inhibitors for the kinetic study. We found that the major components of Turmeric (4a, 4e, 5q) showed noncompetitive inhibition characteristics with a K i of 1.3 mM (4a), 1.2 mM (4e), and 0.8 mM (5q), respectively. Conversely, compound 5e exhibited a potent competitive inhibition against Nan A with a K i of 0.14 mM (Figure 1). Synthesised compounds were then evaluated for their inhibitory effect on sialidase from Vibrio cholerae and Clostridium perfringens, which also release sialidase and play a role in the pathogenesis. The inhibitory assay results are summerised in Table 2.
Similar to the above results, methoxy or hydroxyl, heptadienyl and a, b-unsaturated ketone groups played important roles in the inhibitory activity. Although the position of the methoxy and hydroxyl group did not influence the inhibition of the above enzymes, the inhibitory effect increased with the greater substitution of hydroxyl groups in the phenyl rings. Thereafter, we evaluated inhibitor activity to confirm the electronic effect. C. perfringens sialidase inhibitory activity was not affected by the electronic effect. Among the examined compounds, 5e (IC 50 ¼ 0.5 ± 0.07 mM), containing the catechol moiety, was the most potent inhibitor of C. perfringens and displayed a 3.2-fold improvement over curcumin. Conversely, substitution of an electron donating groups at the para position of the phenyl group resulted in better potency than the substitution of an electron withdrawing groups against V. cholera sialidase. Among the examined compounds, compound 5r (IC 50 ¼1.9 ± 0.2 mM), containing the N, N-dimethylamino group, was the most potent and displayed a 2.7-fold improvement in inhibitory activity over that of curcumin. The IC 50 values for sialidase inhibitory activity were 2.7-43.6 mM in V. Choleara and 0.5-269.4 mM in C. perfringens.