Laccase-Catalyzed 1,4-Dioxane-Mediated Synthesis of Belladine N-Oxides with Anti-Influenza A Virus Activity

Belladine N-oxides active against influenza A virus have been synthetized by a novel laccase-catalyzed 1,4-dioxane-mediated oxidation of aromatic and side-chain modified belladine derivatives. Electron paramagnetic resonance (EPR) analysis confirmed the role of 1,4-dioxane as a co-oxidant. The reaction was chemo-selective, showing a high functional-group compatibility. The novel belladine N-oxides were active against influenza A virus, involving the early stage of the virus replication life cycle.


Introduction
Belladine 1 and norbelladine 2 (firstly extracted from the Amaryllidaceae family [1]) are bioactive precursors in the synthesis of drugs acting on the central nervous system, such as galantamine 3, lycorine 4, and haemanthamine 5 ( Figure 1, Panel a) [2,3]. They are natural substances emerging in therapy, showing a cholinesterase inhibitory activity comparable to that of 3 in the treatment of Alzheimer's disease [4,5]. In addition, a computational study suggested that quaternary belladine derivatives can interact with the neuroaminidase (NA) protein of influenza A virus, inhibiting viral release from host cell [6].
Recently, the use of N-formyl-2-bromo-O-methynorbelladine 7a in the total synthesis of 3 by a laccase (benzenediol: oxygen oxidoreductases, EC 1. 10.3.2) [7]-mediator system has been reported, focusing on the formation of the spyrocyclohexadienone 6 as a tri-cyclic intermediate ( Figure 1, Panel b) [8]. In this latter case, undesired side-products were produced depending on the nature of the N-substituent. Electron withdrawing group EW (R = CHO, 7a) favored the formation of phenoxy radicals and successive oxidative coupling, and the hydrolysis of the iminium ion (I) to side-chain degradation products was the only observed side-process [8,9] (Figure 1; Panel b, pathway a). Conversely, the oxidative coupling was not operative with electron-donating group ED (R = CH 3 , 7b), in which case the isoindoline 8 was produced by an iminium-ion Polonovski transformation of the N-oxide intermediate (II) (not isolated) [10,11] (Figure 1; Panel b, pathway b).
Amine N-oxides are widely diffused in nature [12,13], and they play an important role as chiral ligands, organo-catalysts, and synthons [14,15]. These compounds are synthetized using hazardous stoichiometric oxidants, or in the alternative, heavy metal catalysts and hydrogen peroxide (H 2 O 2 ), which leads to the formation of toxic wastes and undesired by-products [16]. As an alternative, dioxygen (O 2 ) is an effective green oxidant [17][18][19]

Optimization of the Reaction Conditions
The Polonovski transformation [11] of amine N-oxides occurs by two successive steps: (a) the protonation of the quaternary N-oxide moiety [31,32]; and (b) the cleavage of the iminium ion to corresponding aminium radical cation, followed by α-hydrogen elimination and skeletal rearrangement [33,34]. In order to avoid the occurrence of the Polonovski reaction in the laccase catalyzed synthesis of belladine N-oxides, the critical reaction step was expected to be the protonation of the N-oxide moiety. We started our investigation using 7b (Scheme 1) as a model substrate (general procedures are in Supplementary Materials (SM) #1, and the synthesis of 7b in SM #2).

Optimization of the Reaction Conditions
The Polonovski transformation [11] of amine N-oxides occurs by two successive steps: (a) the protonation of the quaternary N-oxide moiety [31,32]; and (b) the cleavage of the iminium ion to corresponding aminium radical cation, followed by α-hydrogen elimination and skeletal rearrangement [33,34]. In order to avoid the occurrence of the Polonovski reaction in the laccase catalyzed synthesis of belladine N-oxides, the critical reaction step was expected to be the protonation of the N-oxide moiety. We started our investigation using 7b (Scheme 1) as a model substrate (general procedures are in Supplementary Materials (SM) #1, and the synthesis of 7b in SM #2).

EPR Studies
Compound 10 is reported to be the ring-opening product of 2-hydroperoxy-1,4dioxane III (not isolated, Scheme 1) [39][40][41]. EPR studies with 5,5-dimethyl-1-pyrroline N-oxide (DMPO) confirmed the presence of III in the reaction mixture. As reported in Figure 2  was detected during the oxidation of 7b with laccase (line c). The III/DMPO-adduct was common to all cases studied with increasing intensity after the addition of laccase. From a spectroscopic point of view, the spin-trapping approach allows to trap the first radical species formed in the reaction, and the increase of intensity of the signal at 348-353 mT confirms the role of 1,4-dioxane as co-oxidant in the oxidation. after the addition of laccase (line b). In addition, the same signal was detected during the oxidation of 7b with laccase (line c). The III/DMPO-adduct was common to all cases studied with increasing intensity after the addition of laccase. From a spectroscopic point of view, the spin-trapping approach allows to trap the first radical species formed in the reaction, and the increase of intensity of the signal at 348-353 mT confirms the role of 1,4dioxane as co-oxidant in the oxidation.

Synthesis and Characterization of Belladine N-Oxide Derivatives 12a-h
The procedure was generalized to derivatives 11a-h, covering a large panel of substituents in the aromatic ring and side chain (the synthesis of compounds 11a-h is in SM #2; NMR data of 11a-h are in SM #3). Compounds 11a-h (0.1 mmol) were treated with laccase (100 U·mmol −1 ) under O2 atmosphere in 10:1 ratio 1,4-dioxane/sodium acetate buffer (2.50 mL; pH 5.6) at 25 °C for 3.0 h to afford N-oxides 12a-h from good to high yield (53-78%), besides to unreacted substrate ( Table 2, entries 1-9) (NMR data of 12a-h are in SM #3). Isoindolines were not detected in the reaction mixture. All type of substituent patterns and side-chain length were allowed, highlighting the high chemo-selectivity and functional-group compatibility of the procedure. A further evidence of the III/DMPO-adduct is reported in Figure 2 (line d) where the EPR signal recorded during the oxidation of 11h is paired to its simulation. The intensity of the signal is higher than in the previous case. The magnetic parameters obtained from the fitting are: g = 2.0061 ± 0.0001, AN = 1.36 mT, AH = 1.01 mT and AH = 0.117 mT. These parameters are typical of peroxyl radical adduct with the DMPO in organic solvents [42]. The intensity of this signal was higher than that previously observed in the oxidation of 7b, in accordance with the higher yield of 12h with respect to 9.

Synthesis and Characterization of Belladine N-Oxide Derivatives 12a-h
The procedure was generalized to derivatives 11a-h, covering a large panel of substituents in the aromatic ring and side chain (the synthesis of compounds 11a-h is in SM #2; NMR data of 11a-h are in SM #3). Compounds 11a-h (0.1 mmol) were treated with laccase (100 U·mmol −1 ) under O 2 atmosphere in 10:1 ratio 1,4-dioxane/sodium acetate buffer (2.50 mL; pH 5.6) at 25 • C for 3.0 h to afford N-oxides 12a-h from good to high yield (53-78%), besides to unreacted substrate ( Table 2, entries 1-9) (NMR data of 12a-h are in SM #3). Isoindolines were not detected in the reaction mixture. All type of substituent patterns and side-chain length were allowed, highlighting the high chemo-selectivity and functional-group compatibility of the procedure. A further evidence of the III/DMPOadduct is reported in Figure 2 (line d) where the EPR signal recorded during the oxidation of 11h is paired to its simulation. The intensity of the signal is higher than in the previous case. The magnetic parameters obtained from the fitting are: g = 2.0061 ± 0.0001, A N = 1.36 mT, A H = 1.01 mT and A H = 0.117 mT. These parameters are typical of peroxyl radical adduct with the DMPO in organic solvents [42]. The intensity of this signal was higher than that previously observed in the oxidation of 7b, in accordance with the higher yield of 12h with respect to 9.
X-ray data confirmed the structure of 12b, which was the only product isolated as a crystal ( Figure 3). The compound crystallizes in a centric space group (C2/c) containing both the enantiomers (for details of the X-ray analysis and crystallization procedure see the Section Materials and Methods).
The reactions were performed using 100 U·mmol −1 of laccase for 0.1 mmol of the substrate in a solvent mixture of 1,4-dioxane (2.25 mL) and sodium-acetate buffer 0.1 M pH 5.6 (0.25 mL) under O2 atmosphere at 25 °C for 3. All the reactions were conducted in triplicate. Reactions were performed in the presence of 1,4-dioxane deprived by distillation of the commercial additive butyl hydroxytoluene (BHT). 2 The yield was calculated on the basis of starting mmol of the substrate. 3 Optical rotations were recorded on a JASCO P−1000 series at 589 nm. X-ray data confirmed the structure of 12b, which was the only product isolated as a crystal ( Figure 3). The compound crystallizes in a centric space group (C2/c) containing both the enantiomers (for details of the X-ray analysis and crystallization procedure see the Section Materials and Methods). As a selected case of study, the 1 H-NMR of 12b with chiral lanthanide shift reagent Eu(hfc)3 (700 µL MeOD, 13.9 mM Eu(hfc)3) [43] showed the expected asymmetric shift pattern of the AB quartet system (4.60-4.20 ppm) for the resolution of the two enantiomers, with an enantiomeric excess (ee) of 10% ( Figure 4). Polarimetric analyses of 9, 12a-e and 12g-h are reported in Table 2.
The reactions were performed using 100 U·mmol −1 of laccase for 0.1 mmol of the substrate in a solvent mixture of 1,4-dioxane (2.25 mL) and sodium-acetate buffer 0.1 M pH 5.6 (0.25 mL) under O2 atmosphere at 25 °C for 3. All the reactions were conducted in triplicate. Reactions were performed in the presence of 1,4-dioxane deprived by distillation of the commercial additive butyl hydroxytoluene (BHT). 2 The yield was calculated on the basis of starting mmol of the substrate. 3 Optical rotations were recorded on a JASCO P−1000 series at 589 nm. X-ray data confirmed the structure of 12b, which was the only product isolated as a crystal ( Figure 3). The compound crystallizes in a centric space group (C2/c) containing both the enantiomers (for details of the X-ray analysis and crystallization procedure see the Section Materials and Methods). As a selected case of study, the 1 H-NMR of 12b with chiral lanthanide shift reagent Eu(hfc)3 (700 µL MeOD, 13.9 mM Eu(hfc)3) [43] showed the expected asymmetric shift pattern of the AB quartet system (4.60-4.20 ppm) for the resolution of the two enantiomers, with an enantiomeric excess (ee) of 10% ( Figure 4). Polarimetric analyses of 9, 12a-e and 12g-h are reported in Table 2. As a selected case of study, the 1 H-NMR of 12b with chiral lanthanide shift reagent Eu(hfc) 3 (700 µL MeOD, 13.9 mM Eu(hfc) 3 ) [43] showed the expected asymmetric shift pattern of the AB quartet system (4.60-4.20 ppm) for the resolution of the two enantiomers, with an enantiomeric excess (ee) of 10% ( Figure 4). Polarimetric analyses of 9, 12a-e and 12g-h are reported in Table 2.

Antiviral Activity of Compound 7b and 12a-h against Influenza A Virus
Compounds 9 and 12a-h were tested against influenza A/Puerto Rico/8/34 H (PR8) virus in order to evaluate previously reported computational hypothesis abou inhibition of viral NA [6]. Influenza is responsible for large epidemics and pandem

Antiviral Activity of Compound 7b and 12a-h against Influenza A Virus
Compounds 9 and 12a-h were tested against influenza A/Puerto Rico/8/34 H1N1 (PR8) virus in order to evaluate previously reported computational hypothesis about the inhibition of viral NA [6]. Influenza is responsible for large epidemics and pandemics causing severe health problems [44]. The influenza A virus (Orthomyxoviridae family) is characterized by the release of eight viral RNA segments associated with the nucleoprotein (NP) and the viral RNA-dependent RNA polymerase (RdRp) complex responsible for replication and transcription cycles [45]. Among the inhibitors of the influenza A virus, the compounds active against NA received great attention being involved in the release of viral particles from infected cells [46,47]. In the first set of experiments, A549 cells infected with 0.001 MOI of PR8 were treated with different concentrations (range 10-40 mg/mL) of compounds 9 and 12a-h for 24 h. The expression of Hemagglutinin (HA) was analyzed by means of In Cell Western (ICW) assay (as described in the Materials and Methods section) on cell monolayers. As control of cytotoxicity, cell monolayers were also treated with the same concentrations of compounds 9 and 12a-h and stained with a Cell tag (as described in the Materials and Methods section). The supernatants of the infected A549 cells were recovered and used to newly infect a fresh monolayer of MDCK (Madin-Darby canine epithelial kidney) cells, in order to evaluate whether viral particles released from the infected cells were still infective. Table 3 shows the values of IC50, CC50, and relative SI obtained on A549 and MDCK cells. Compound 12h was the most effective against viral replication in both cell lines (IC50 range 70-73 µg/mL) with the highest SI.  50 are expressed in micromolar units as mean ± SD. All experiments were conducted in triplicate. The Selectivity Index (SI) of each compound was calculated as the ratio CC 50 /IC 50 . 2 ND = Not Determined.
As an example, 12h significantly reduced the HA protein expression on A549 cells ( Figure 5, panel a). The released viral particles in the supernatants of A549 cells were then used to infect new monolayers of MDCK cells. After 24 h infection, the ICW assay confirmed a dose-dependent reduction of HA protein expression on MDCK cell monolayers ( Figure 5, panel b), suggesting the occurrence of a block in the release of viral particles from the infected cells.
To evaluate whether the compound 12h was able to impair the cell-to-cell virus spread, higher concentrations (40 and 80 mg/mL) of 12h were added to A549 cell monolayers after viral challenge, and HA protein expression was analyzed directly on these monolayers after 24 h. The ICW assay showed a significant reduction of relative fluorescence intensity of HA protein (~50% inhibition with 80 mg/mL). Furthermore, the reduction of foci of infection in cell monolayers treated with the compound 12h compared to DMSO-treated cells [48], suggested a block in the release of viral particles probably due to an interference with the viral NA ( Figure 6).
As an example, 12h significantly reduced the HA protein expression on A549 cells ( Figure 5, panel a). The released viral particles in the supernatants of A549 cells were then used to infect new monolayers of MDCK cells. After 24 h infection, the ICW assay confirmed a dose-dependent reduction of HA protein expression on MDCK cell monolayers ( Figure 5, panel b), suggesting the occurrence of a block in the release of viral particles from the infected cells. To evaluate whether the compound 12h was able to impair the cell-to-cell virus spread, higher concentrations (40 and 80 mg/mL) of 12h were added to A549 cell monolayers after viral challenge, and HA protein expression was analyzed directly on these monolayers after 24 h. The ICW assay showed a significant reduction of relative fluorescence intensity of HA protein (~50% inhibition with 80 mg/mL). Furthermore, the reduction of foci of infection in cell monolayers treated with the compound 12h compared to DMSO-treated cells [48], suggested a block in the release of viral particles probably due to an interference with the viral NA ( Figure 6).

Conclusions
Laccase was able to activate 1,4-dioxane as co-oxidant in the selective synthesis of belladine N-oxides 9 and 12a-h, as confirmed by the EPR detection of the corresponding DMPO/III adduct. Other organic solvents were not effective in the transformation, high- Figure 6. The expression of Hemagglutinin (HA) by means of ICW assay at late steps of viral replication. A549 cells were infected with PR8 and treated or not with 40 or 80 µg/mL of compound 12h. After 24 infection, cells were fixed and stained for HA protein, as described in the Materials and Methods section. The expression of viral HA was analyzed by ICW assay, using LI-COR Image Studio Software. The percentage of relative fluorescence intensity (RFI) was calculated in comparison to untreated infected cells (considered 100%). Values are the mean ± S.D. of two experiments, each performed in duplicate (n = 4). Statistical significance of the data vs untreated infected cells was defined as ** p < 0.001 and *** p < 0.0001.

Conclusions
Laccase was able to activate 1,4-dioxane as co-oxidant in the selective synthesis of belladine N-oxides 9 and 12a-h, as confirmed by the EPR detection of the corresponding DMPO/III adduct. Other organic solvents were not effective in the transformation, highlighting the specific role of the formation of III in the oxygen atom transfer process. This reaction is an alternative to the widespread reported laccase/mediator procedure for the oxidation of amines [9,49,50]. Irrespective of the experimental conditions, the oxidation proceeded from good to high yield, showing high functional-group compatibility and chemo-selectivity avoiding the undesired formation of reactive quinone species and oligomeric products [51][52][53]. In addition, an appreciable stereoselectivity was observed, probably due to partial inhibition of the pyramidal inversion at the nitrogen center as a consequence of steric hindrance of the substituents. Compounds 12a-c and 12h were the most active derivatives against influenza A virus. The highest values of IC 50 and SI were observed in the case of 12h, which is characterized by three carbon atoms in the side-chain and only one hydroxy moiety on the aromatic rings. As a general trend, the presence of at least one hydroxy moiety on the aromatic rings and two or three carbon atoms in the side-chain were required to obtain significant antiviral activity. Finally, the presence of an electron-withdrawing substituent on the aromatic ring (12d) deprived the molecule of antiviral activity.

Materials
Reagents and laccase from Trametes versicolor were obtained from commercial suppliers (Sigma-Aldrich Srl, Milan, Italy).

Enzyme Activity Assay
The enzyme activity was assayed by using 2,2 -azino-bis(3-ethylbenzothiazoline-6sulfonic acid)diammonium salt (ABTS) procedure. ABTS (5.0 mM), sodium acetate buffer (2.0 mL, pH 5.0), and the enzyme solution (200 µL) were used as a standard solution. The formation of the cation radical was detected by measuring the increase of absorbance at 420 nm (ε 420 = 36,000 M −1 cm −1 ). One unit of laccase activity has been defined as the amount of enzyme that catalyzed the oxidation of 1.0 µmol of ABTS in a 200 µL reaction mixture at 25 • C during 1.0 min.

EPR Analysis
The reaction solution was prepared adding 7b and 11h (40 mM), DMPO (60 mM), and laccase (0.12 mM) in 1,4-dioxane/sodium acetate buffer (9:1 ratio). To perform the EPR experiments, capillaries of 1.2 mm diameter were filled in and inserted in a quartz tube of 3 × 3.5 I.D. × O.D. CW (continuous wave) X-band (9 GHz). Experimental condition: 9.86 GHz 123 microwave frequency, 0.1 mT modulation amplitude, and 0.2 mW microwave power. EPR spectra were recorded at room temperature with a Bruker E580 Elexsys Series, using the Bruker ER4122 SHQE cavity. A simulation was carried out with the Easyspin simulation program 5.2.28 version, using the "garlic function".

X-Ray Crystallography Data for Compound 12b
Compound 12b was crystallized in an NMR tube, adding 5 mg of the compound in 300µL of deuterated methanol (CD 3 OD). Compound 12b was completely dissolved heating the system, and the solution was slowly cooled overnight. A single crystal of 12b was submitted to X-ray data collection on an Oxford-Diffraction Xcalibur Sapphire 3 diffractometer with a graphite monochromated Mo-Kα radiation (λ = 0.71073 Å) at 293 K. The structure was solved by direct methods implemented in the SHELXS program (Version 2013/1) [54]. The refinement was carried out by full-matrix anisotropic least-squares on F 2 for all reflections for non-H atoms by means of the SHELXL program [55]. The structure crystallizes in the monoclinic crystal system, space group C2/c. Crystallographic data have been deposited with the Cambridge Crystallographic Data Centre as supplementary publication no. CCDC 2,045,206. Copies of the data can be obtained, free of charge, on application to CCDC, 12 Union Road, Cambridge CB2 1EZ, UK; (fax: +44-(0)-1223- ; or e-mail: deposit@ccdc.cam.ac.uk).

Procedure for the Synthesis of Amine N-Oxides 9 and 12a-h
Compounds 7b and 11a-h (1.0 eq., 0.1 mmol) were dissolved in a solvent mixture of 1,4-dioxane (2.25 mL) and sodium acetate buffer 0.1 M, pH = 5.6 (0.25 mL). Laccase (100 U·mmol −1 ) was added and the mixture was gently stirred at 25 • C under O 2 atmosphere (balloon) for 3 h. After this period, the mixture was filtered and the solvent was evaporated under reduced pressure. The crude mixture was purified by silica gel column chromatography (ethyl acetate/methanol 7:1) to afford the desired products 9 and 12a-h. Chemical reactions were monitored using thin-layer chromatography on precoated aluminum silica gel Merck 60 F254 plates and a UV lamp was used for visualization. Merck silica gel 60 (230-400 mesh) was used for chromatography. All products were dried in a high vacuum (10-3 mbar). 1 H NMR and 13 C-NMR