Synthesis and In Vitro Anti-Influenza Virus Evaluation of Novel Sialic Acid (C-5 and C-9)-Pentacyclic Triterpene Derivatives

The emergence of drug resistant variants of the influenza virus has led to a great need to identify novel and effective antiviral agents. In our previous study, a series of sialic acid (C-2 and C-4)-pentacyclic triterpene conjugates have been synthesized, and a five-fold more potent antiviral activity was observed when sialic acid was conjugated with pentacyclic triterpene via C-4 than C-2. It was here that we further reported the synthesis and anti-influenza activity of novel sialic acid (C-5 and C-9)-pentacyclic triterpene conjugates. Their structures were confirmed by ESI-HRMS, 1H-NMR, and 13C-NMR spectroscopic analyses. Two conjugates (26 and 42) showed strong cytotoxicity to MDCK cells in the CellTiter-Glo assay at a concentration of 100 μM. However, they showed no significant cytotoxicity to HL-60, Hela, and A549 cell lines in MTT assay under the concentration of 10 μM (except compound 42 showed weak cytotoxicity to HL-60 cell line (10 μM, ~53%)). Compounds 20, 28, 36, and 44 displayed weak potency to influenza A/WSN/33 (H1N1) virus (100 μM, ~20–30%), and no significant anti-influenza activity was found for the other conjugates. The data suggested that both the C-5 acetylamide and C-9 hydroxy of sialic acid were important for its binding with hemagglutinin during viral entry into host cells, while C-4 and C-2 hydroxy were not critical for the binding process and could be replaced with hydrophobic moieties. The research presented herein had significant implications for the design of novel antiviral inhibitors based on a sialic acid scaffold.


Introduction
Influenza is one of the most common and dangerous viral respiratory diseases; it affects three to five million cases of severe illness and leads to approximately 250,000 to 500,000 deaths every year worldwide [1,2]. Based on the antigenic major surface glycoproteins hemagglutinin (HA; 18 subtypes) and neuraminidase (NA; 11 subtypes), it is divided into different subtypes [3]. Currently available options to fight against the respiratory system attack by influenza A viruses include both vaccines and antiviral agents. However, the three types of available vaccines (inactivated influenza vaccine, live and antiviral agents. However, the three types of available vaccines (inactivated influenza vaccine, live attenuated influenza vaccine, and recombinant influenza vaccine) have moderate efficacy that varies seasonally due to constant viral evolution [4]. Currently, two classes of anti-influenza drugs have been developed for interruption of specific processes in influenza infection. Amantadine and rimantadine target the M2 protein which is an ion channel allowing protons to move through the viral envelope to uncoat viral RNA, and thus block the release of viral RNA into the cytoplasm [5]. Oseltamivir (Tamiflu) and zanamivir (Relenza), on the other hand, target neuraminidase (NA) protein, inhibiting its enzymatic activity and causing the tethered progeny virus to be unable to escape from its host cells. However, the emergence of drug-resistant influenza viruses has limited the use of those drugs. [6][7][8][9].
The first step of infection by influenza A virus is mediated by the interaction of sialic acid with the major surface viral glycoprotein haemagglutinin (HA) [10]. The crucial role of HA in the viral lifecycle makes it an attractive target for the development of therapeutics to treat influenza virus infection [11]. Sialic acid (also called as N-acetylneuraminic acid-Neu5Ac), a negatively charged nine-carbon carboxylated monosaccharide, is present at the non-reducing terminal positions of carbohydrate chains of glycoproteins and glycolipids on the cell surface [12]. It is well exploited by many pathogens to attach to and infect cells, and moreover, many pathogens decorate themselves with sialic acid to escape the host immune system [13]. In the last decades, many efforts have been made to develop an anti-influenza drug based on a sialic acid scaffold [14][15][16]. In fact, the analogs and derivatives of sialic acid have been shown to be potent inhibitors of influenza virus sialidase, and are commercially available [17]. Moreover, other bioactive molecules (such as cholesterol, [18] phospholipid, [19] taxol, [20] deazaflavin, [21] cytidine 5′-monophosphate, [22] and anthraquinone, [23]) have been selected to conjugate with sialic acid to prepare new biologically active molecules. Recently, we have explored potential agents against the influenza virus by the modifications at C-2 of the naturally sialic acid moiety (by conjugation of hydrophobic pentacyclic triterpene (Figure 1), which are widely distributed in the plant kingdom and generally believed to enhance the immunity of host plants and to increase plant resistance to pathogens [24]) with one compound exhibited moderate potency for influenza A/WSN/33 (H1N1) virus with IC50 at 41.2 μM [25]. More recently, a five-fold more potent anti-influenza activity was found when sialic acid was conjugated to pentacyclic triterpene via C-4 than to C-2 [26], possibly due to the fact that the big substitute group at C-2 of sialic acid affected the interactions of COOH with the basic amino acid of HA.
As part of our continued interest in the structurally modified pentacyclic triterpene derivatives as an anti-influenza virus entry inhibitor, [25,[27][28][29][30] we thought it of value to prepare a range of C-5 and C-9 modified sialic acid derivatives to better explore the antiviral structure-activity relationship (SAR) of sialic acid-pentacyclic triterpene conjugates. We reported herein the synthesis and antiinfluenza A/WSN/33 virus activity of a series of C-5 and C-9 modified sialic acid conjugates of pentacyclic triterpene.  [25,26]; (b) In this work.
As part of our continued interest in the structurally modified pentacyclic triterpene derivatives as an anti-influenza virus entry inhibitor, [25,[27][28][29][30] we thought it of value to prepare a range of C-5 and C-9 modified sialic acid derivatives to better explore the antiviral structure-activity relationship (SAR) of sialic acid-pentacyclic triterpene conjugates. We reported herein the synthesis and anti-influenza A/WSN/33 virus activity of a series of C-5 and C-9 modified sialic acid conjugates of pentacyclic triterpene.
The C-5 functionalized sialic acid intermediates 9 and 10 were synthesized using the approach shown in Scheme 1. Commercially available sialic acid 1 was used as a starting material, and synthesis of the 4,7,8,9-tetra-O-acetyl-N-acetylneuraminic acid 4 was performed in moderate yield according to published procedures by Tropper et al. (Scheme 1) [32]. In order to synthesize the C-5 azide substituted intermediate 9 required for cycloaddition reaction, the NHAc group at C-5 was first Boc-protected by treatment with di-tert-butyl dicarbonate (Boc 2 O) in THF and then de-O-acetylated under Zemplén conditions [33] (followed by re-O-acetylation to provide derivative 7 in~82% yield over three steps). Transformation to the azide substituted amide 9 was then performed by removal of the Boc group with TFA in DCM, followed by acylation of the amine. Subsequent catalytic reduction of the 2-azido-acetamide at position C-5 of 9 was achieved by exposure to a H 2 atmosphere in the presence of 10% Pd/C in methanol, to afford the N-glycylneraminic acid intermediate 10 in 93% yield (which was used without further purification in the next step).
The C-5 functionalized sialic acid intermediates 9 and 10 were synthesized using the approach shown in Scheme 1. Commercially available sialic acid 1 was used as a starting material, and synthesis of the 4,7,8,9-tetra-O-acetyl-N-acetylneuraminic acid 4 was performed in moderate yield according to published procedures by Tropper et al. (Scheme 1) [32]. In order to synthesize the C-5 azide substituted intermediate 9 required for cycloaddition reaction, the NHAc group at C-5 was first Boc-protected by treatment with di-tert-butyl dicarbonate (Boc2O) in THF and then de-O-acetylated under Zemplén conditions [33] (followed by re-O-acetylation to provide derivative 7 in ~82% yield over three steps). Transformation to the azide substituted amide 9 was then performed by removal of the Boc group with TFA in DCM, followed by acylation of the amine. Subsequent catalytic reduction of the 2-azido-acetamide at position C-5 of 9 was achieved by exposure to a H2 atmosphere in the presence of 10% Pd/C in methanol, to afford the N-glycylneraminic acid intermediate 10 in 93% yield (which was used without further purification in the next step). The synthesis of C-9 functionalized sialic acid derivatives 13 and 14 was accessed from intermediate 4 (Scheme 2). The acetyl groups of 4 were firstly removed under Zemplén conditions, followed by selective mono-tosylation of C-9 hydroxy to afford compound 12. This compound was followed by nucleophilic substitution with sodium azide in DMF to provide the intermediate 13, and further reduction of the azide group by hydrolygenolysis with Pd/C catalysis yielded the corresponding C-9 amine intermediate 14, which was used without further purification in the next step.
Coupling of compound 9 with N-propargyl triterpene amides 15-17 and 27 via click reaction yielded the acetyl-protected conjugates 18, 20, 22, and 28, respectively. All the click reactions were performed in DCM/H2O (1:1, v/v) at room temperature for 12-24 h. In all cases, pentacyclic triterpenesialic acid conjugates were separated by extraction and purified by column chromatography on silica gel, and the yield after the purification was between 88% and 95%. In the next step, the Ac groups were removed under Zemplén conditions to give the corresponding conjugates 19, 21, 23, and 29 as the final products quantitatively (Scheme 3). On the other hand, 13 (which does not contain any protecting groups on the hydroxyl groups) underwent click reaction with N-propargyl triterpene amides 15-17 and 27 under similar conditions to directly provide conjugates 24-26 and 30 in yields ranging from 79% to 87%. The synthesis of C-9 functionalized sialic acid derivatives 13 and 14 was accessed from intermediate 4 (Scheme 2). The acetyl groups of 4 were firstly removed under Zemplén conditions, followed by selective mono-tosylation of C-9 hydroxy to afford compound 12. This compound was followed by nucleophilic substitution with sodium azide in DMF to provide the intermediate 13, and further reduction of the azide group by hydrolygenolysis with Pd/C catalysis yielded the corresponding C-9 amine intermediate 14, which was used without further purification in the next step.
Coupling of compound 9 with N-propargyl triterpene amides 15-17 and 27 via click reaction yielded the acetyl-protected conjugates 18, 20, 22, and 28, respectively. All the click reactions were performed in DCM/H 2 O (1:1, v/v) at room temperature for 12-24 h. In all cases, pentacyclic triterpene-sialic acid conjugates were separated by extraction and purified by column chromatography on silica gel, and the yield after the purification was between 88% and 95%. In the next step, the Ac groups were removed under Zemplén conditions to give the corresponding conjugates 19, 21, 23, and 29 as the final products quantitatively (Scheme 3). On the other hand, 13 (which does not contain any protecting groups on the hydroxyl groups) underwent click reaction with N-propargyl triterpene Alternatively, the 1-benzotriazolyl derivatives of pentacyclic triterpene 31-33 and 43 were coupled with intermediate 10 by route method in DMF. The crude products (compounds 34, 36, 38, and 44) were purified by column chromatography, and the coupling yields were good and ranged from 75% to 85% (Scheme 4). The Ac-removal of 34, 36, 38, and 44 under Zemplén conditions gave the corresponding conjugates 35, 37, 39, and 45, quantitatively. Similarly coupling 31-33 and 43 with 14 gave the final conjugates 40-42 and 46, respectively, with good yields ranging from 75% to 83%.

Biological Evaluation
Since the synthesis of sialic acid derivatives having a pentacyclic triterpene substituent at the C-5 or C-9 positions were accomplished, our attention was directed to the biological activities of the anti-influenza virus. Initially, we examined the cytotoxicity of the newly synthesized compounds 18-26, 28-30, 34-42, and 44-46 in Madin-Darby canine kidney (MDCK) cells using the CellTiter-Glo ® assay. A culture medium containing 0.5% DMSO served as a negative control. 5-Fluorouracil (5-Fu) (a well-known broad spectrum anticancer drug [34]) was selected as a positive control. Except for compounds 26 and 42 (two sialic acid (C-9)-ursolic acid conjugates), the other compounds showed no significant cytotoxicity to MDCK cells at a concentration of 100 µM (Figure 2).  The cytotoxicity of compounds 26 and 42 against human promyelocytic leukemia HL-60, human cervical cancer Hela, and human lung cancer A549, was further examined by the 3-(4,5dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay with tested compound at a concentration of 0.1-10 μM [35]. With the exception of compound 42, which showed weaker cytotoxicity against HL-60 (10 μM, 53%), no significant cytotoxicity was found (see Table S1 in Supplementary Materials). These results revealed that there was almost no cytotoxicity for those two ursolic acid-pentacyclic triterpene conjugates at a concentration of less than 10 μM.
Next, we examined the inhibitory activity of the test compounds against the virus replication in MDCK cells using the influenza A/WSN/33 (H1N1 subtype) virus strain at a concentration of 100 μM. The virus yields as a percent of control were estimated by a plaque titration method, and the results are shown in Figure 3 (including OSV and curcumin-a small-molecule entry inhibitor targeting the influenza virus HA1 domain [36] as positive controls). Four compounds 20, 28, 36, and 44 (two of them are echinocystic acid-sialic acid (C-5)-echinocystic acid conjugates 20 and 36 and the other two are sialic acid (C-5)-betulinic acid conjugates) showed weak anti-influenza A/WSN/33 virus activity with IC50 > 100 μM. All the other compounds displayed no activity against influenza at high concentration. These data indicated that (1) the modification of the C-5 position of sialic acid showed a little more potent antiviral activity than the C-9 position; and (2) the introduction of large hydrophobic pentacyclic triterpene groups onto the C-5 acetamide or C-9 hydroxy of sialic acid might affect the interaction of sialic acid with HA during viral entry. The results of both the study described here and those previously reported by ours [25,26] demonstrated that the C-5 and C-9 positions of sialic acid were important for its binding with the HA protein (the introduction of a large hydrophobic group at those positions would affect their interaction), while the introduction of an appropriate hydrophobic group at the C-2 and C-4 positions of sialic acid could increase the binding with the active site of HA without detriment to binding affinity.
In order to determine their specificity for influenza virus infections, we further tested their inhibition activity on vesicular stomtatis virus G protein pseudo-particle (VSVpp), a negative-strand RNA virus with abroad host range that infects almost all cell lines. The results showed that they displayed no inhibition activity on VSV entry, indicating that they are specific anti-influenza entry inhibitors. The cytotoxicity of compounds 26 and 42 against human promyelocytic leukemia HL-60, human cervical cancer Hela, and human lung cancer A549, was further examined by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay with tested compound at a concentration of 0.1-10 µM [35]. With the exception of compound 42, which showed weaker cytotoxicity against HL-60 (10 µM, 53%), no significant cytotoxicity was found (see Table S1 in Supplementary Materials). These results revealed that there was almost no cytotoxicity for those two ursolic acid-pentacyclic triterpene conjugates at a concentration of less than 10 µM.
Next, we examined the inhibitory activity of the test compounds against the virus replication in MDCK cells using the influenza A/WSN/33 (H1N1 subtype) virus strain at a concentration of 100 µM. The virus yields as a percent of control were estimated by a plaque titration method, and the results are shown in Figure 3 (including OSV and curcumin-a small-molecule entry inhibitor targeting the influenza virus HA1 domain [36] as positive controls). Four compounds 20, 28, 36, and 44 (two of them are echinocystic acid-sialic acid (C-5)-echinocystic acid conjugates 20 and 36 and the other two are sialic acid (C-5)-betulinic acid conjugates) showed weak anti-influenza A/WSN/33 virus activity with IC 50 > 100 µM. All the other compounds displayed no activity against influenza at high concentration. These data indicated that (1) the modification of the C-5 position of sialic acid showed a little more potent antiviral activity than the C-9 position; and (2) the introduction of large hydrophobic pentacyclic triterpene groups onto the C-5 acetamide or C-9 hydroxy of sialic acid might affect the interaction of sialic acid with HA during viral entry. The results of both the study described here and those previously reported by ours [25,26] demonstrated that the C-5 and C-9 positions of sialic acid were important for its binding with the HA protein (the introduction of a large hydrophobic group at those positions would affect their interaction), while the introduction of an appropriate hydrophobic group at the C-2 and C-4 positions of sialic acid could increase the binding with the active site of HA without detriment to binding affinity.
In order to determine their specificity for influenza virus infections, we further tested their inhibition activity on vesicular stomtatis virus G protein pseudo-particle (VSVpp), a negative-strand RNA virus with abroad host range that infects almost all cell lines. The results showed that they displayed no inhibition activity on VSV entry, indicating that they are specific anti-influenza entry inhibitors. In our previous study, we found that certain pentacyclic triterpene displayed broad spectrum anti-influenza activity by blocking virus entry. The synergistic effect of compound 20 (the most potent of the four compounds), combined with OSV (an oral neuraminidase inhibitor), was investigated to explore the potential of the compound for use in cocktail therapy. We found that the addition of compound 20 to OSV increased its efficacy at inhibiting influenza infection. According to the medianeffect equation, [37] moderate synergic anti-influenza effects were observed with the combination index at 0.83. The synergistic effect of pentacyclic triterpene with other anti-influenza virus inhibitors may provide a new option for the treatment of influenza virus infections.
All chemicals were used as supplied without further purification. The synthesis of compounds  2-8, 11-17, 27, 31-33, and 43 has been reported previously [31,32,38]. The intermediates 9 and 10 and the novel sialic acid (C-5 or C-9)-pentacyclic triterpene -Neu5Ac2en conjugates 18-26, 28-30, 34-42, In our previous study, we found that certain pentacyclic triterpene displayed broad spectrum anti-influenza activity by blocking virus entry. The synergistic effect of compound 20 (the most potent of the four compounds), combined with OSV (an oral neuraminidase inhibitor), was investigated to explore the potential of the compound for use in cocktail therapy. We found that the addition of compound 20 to OSV increased its efficacy at inhibiting influenza infection. According to the median-effect equation, [37] moderate synergic anti-influenza effects were observed with the combination index at 0.83. The synergistic effect of pentacyclic triterpene with other anti-influenza virus inhibitors may provide a new option for the treatment of influenza virus infections.

General Procedure A for the Click Reaction
CuSO 4 (48 mg, 0.30 mmol) and sodium ascorbate (119 mg, 0.60 mmol) were added to a solution of alkyne (0.45 mmol) and azide (0.30 mmol) in DCM/H 2 O (1:1 v/v, 12 mL). The resulting solution was vigorously stirred at room temperature for 12 h. The reaction mixture was extracted with DCM (3 × 10 mL). The combined organic layer was dried over Na 2 SO 4 , filtered and concentrated. The residue was purified by column chromatography over silica gel.

General Procedure B for the Deacetylation Reaction
The per-O-acetylated sialic acid (C-5 or C-9)-pentacyclic triterpene conjugate was dissolved in dry MeOH (~5 mL per 100 mg compound), and a solution of MeONa (30% in MeOH, 0.1 equiv mol −1 acetate) was added. The solution was stirred at room temperature for 3 h. After completion (TLC), the reaction mixture was neutralized with Amberlite IR-120 (H + ) ion-exchange resin, filtered, and concentrated. The crude product was purified by column chromatography over silica gel.  Methyl (O-methyl-5-(N-aminoacetyl)-4,7,8,9-tetra-O-acetyl-3,5-dideoxy-D-glycero-α-D-galacto-2-nonulopyranosyl) onate (10). The compound 9 (230 mg, 0.42 mmol) was dissolved in MeOH (15 mL), and palladium-carbon (0.1 equiv.) was added. The suspension was degassed under vacuum and urged with H 2 three times; then it was stirred under an H 2 balloon at room temperature for 24 h. The suspension was filtered through a pad of celite and the pad cake was washed with CH 3 OH. The combined filtrate was concentrated to dryness. The residue was used without further purification in the next step.

CPE Reduction Assay
MDCK cells were seeded into 96-well plates in DMEM, supplemented with 10% FBS and incubated overnight at 37 • C under 5% CO 2 . The culture medium was replaced by the test compound and the influenza virus (MOI = 0.1)-DMEM, supplemented with 1% FBS and 2 µg/mL TPCK-treated trypsin. The final concentration of DMSO was 1%. After 40 h of incubation, CellTiter-Glo reagent (Promega Corp., Madison, WI, USA) was added and the CellTiter-Glo assay was performed.

Conclusions
We synthesized and characterized a series of novel sialic acid (C-5 or C-9)-pentacyclic triterpene conjugates, and their cytotoxicity and anti-influenza A/WSN/33 virus activity were evaluated. Two sialic acid (C-9)-ursolic acid conjugates 26 and 42 showed strong cytotoxicity to MDCK cells at a concentration of 100 µM. Most compounds had no significant activity against the influenza A/WSN/33 strain, except that four compounds 20, 28, 36, and 44 showed weak anti-influenza virus activity. Compound 20 exhibited a synergistic effect (when combined OSV) in inhibiting influenza infection. These results indicated that the positions of C-5 and C-9 of sialic acid were important for its binding with the HA protein during virus entry into host cells, while C-4 and C-2 hydroxy were not critical for the binding process and could be replaced with hydrophobic moieties. Therefore, the synergistic effect of pentacyclic triterpene with other anti-influenza virus inhibitors may provide a new option for the treatment of the influenza virus infection.
Supplementary Materials: Supplementary materials are available online. Table S1: The cytotoxicity of compounds 26 and 42 against HL-60, Hela and A549 cell lines. Selected 1 H-, 13 C-NMR, and HRMS spectra.