Stereoselective synthesis and sialidase inhibition properties of KDO-based glycosyloxathiins

The stereoselective synthesis of KDO-based glyco-1,4-oxathiins is described. Relying on a totally diastereoselective inverse electron demand hetero Diels-Alder, α  α′ -dioxothiones as electron-poor heterodienes, and glycals as electron-rich dienophiles, reacted to give, in high yield, the KDO-based glyco derivatives 11 and 12a-c . Taking into account their structural features, biological tests have been run to evaluate the properties of 11 and 12a as sialidase inhibitors. The synthetic and biological data reported confirmed the versatility of this powerful [4+2] cycloaddition and showed the KDO-based cycloadduct 11 as attractive scaffold for the development of new sialidase inhibitors.


Introduction
Pericyclic reactions represent one of the most powerful tools in synthetic chemistry.These reactions have been widely employed to obtain regio-and stereoselectively complex molecules with high atom economy degree.Among these, the inverse electron demand [4+2] Diels-Alder reactions (iEDDA) gained a great deal of attention [1][2][3][4] proving to be of pivotal importance for the synthesis of complex bioactive molecules and natural products [5][6] .
In 2008 7 iEDDA were suggested as possible metal-free click reactions.In the same year, Fox 8 et al. proposed iEDDA as bioorthogonal ligation reactions for the bioconjugation of proteins.As matter of fact, they can proceed fast and in high yield, in aqueous or cell lysate media and are compatible with many functional groups commonly occurring in biological substrates.More recently, [9][10] iEDDA have also been employed in the synthesis of radioimmunoconjugates and radiolabelled antibodies have been successfully prepared using norbornene-conjugated antibodies as dienophile and radiolabelled tetrazine as diene.
][13] From a mechanistic point of view, extensive kinetic and computational studies on iEDDA have been reported, [14][15][16][17][18] confirming the involvement of the HOMO orbital of the dienophile and the LUMO orbital of the diene.
Relying on the stereochemical outcomes of these iHDA, 12,24 the proper selection of dienes and dienophiles allowed us to obtain an array of synthetically and biologically attracting molecules.6][27][28][29] As matter of fact, αα′-dioxothiones 1a and 4a were successfully employed in iHDA with exo-glycal 6 30 and 1galactal 7 to prepare respectively cycloadducts 8 30 and 9, 28 the first spiro sialyl derivative, the second a mimetic of the mucins Tn antigen.In addition, the micromolar water soluble matrix metalloproteinases inhibitors 10 25

Scheme 4. Synthesis of the cycloadducts 12a-c.
The highly reactive dienes 1a-c were generated in situ in the presence of the dienophile 13 and afforded the corresponding cycloadducts 15a-c in high yield (84-97%) (Scheme 4) and as single diasteroisomers.As expected, all cycloadducts (15a-c) were formed as the α-isomer, that is the isomer obtained from the preferred attack of the thiones 1a-c to the bottom face of the dienophile 13. 24,26 A matter of fact, all the iHDAs were totally chemo-regio-and stereoselective according with our previous data obtained and reported 26 employing 13 as electron-rich dienophile.Hydrolysis of the isopropylidene protecting groups of 15a-c was accomplished in good yield (62-77%, 16a-c) by treating 15a-c with a 1.5/1 (v/v) mixture of acetic acid and water at 50°C.The final deprotection of the carboxylic residues of 16a-c (Scheme 4) with lithium hydroxide (1M in H 2 O) in THF as solvent, provided the desired glycosyl derivatives 12a-c.
Analogously, the synthesis of the glycosyl derivative 11 was accomplished by an iHDA reaction between the αα′-dioxothione 1a and the glycal 17 (Scheme 5).The iHDA of 1a with 17 (Scheme 5) was performed in a 1/2 mixture of CHCl 3 /pyridine at 50°C and the cycloadduct 18, was formed as pure  isomer and in high yield (85%).The analysis of the 1 H NMR spectra of 18 allowed us to ascertain that also in this case, the electron-poor diene (1a) preferentially attacked the lower face of the dienophile (17) affording the thermodynamically more stable α-O-glycosyl derivative (J 3-4 8.4 Hz).Furthermore, the value of the chemical shift of H-3 (2.78 ppm) confirmed the regioselectivite formation of the C 3 -S linkage. 28ydrolysis of the isopropylidene protecting groups of 18 with a solution of HCl (10% in EtOH), followed by the deprotection of the carboxylic residue of 19 with lithium hydroxide (1M in H 2 O) in THF as solvent, provided the desired glycosyl derivative 11 (Scheme 5).
Capitalizing on the biological issues recently reported 27 for 20 (see Scheme 6), a tricyclic thio-mimetic of the melanoma antigen GM 3 lactone ganglioside, characterized by the replacement of the native sialic acid moiety with a KDO-related residue, and taking into account the structural analogy of KDO-like fragment of compound 20 with that of derivatives 11 and 12a-c, we investigated the potential activity of 11 and 12a as inhibitors of influenza virus sialidase.The sialidase enzyme on influenza virus, recognizing sialic acid and its derivative, plays a major role in the virus life cycle by facilitating the release of virus progeny from the infected cells. 32o this end, we measured drug inhibition of the sialidase activity of A/Mississippi/3/2001 wild type H1N1 influenza virus 33 by the MUNANA based fluorescent assay 34 as previously described. 35Both compounds 11 and 12a did inhibit the enzyme activity of the H1N1 virus, but compared to zanamivir, the inhibition curves obtained for compounds 11 and 12a showed unusual steep shifts in percent inhibition over a very narrow range.Because of this, a narrower range of dilutions was used to more accurately determine their IC 50 s, (10, 100, 300, 600, 800, 1000, 3000, 10,000 for compound 11 and 10, 100, 300, 1000, 2000, 3000, 10,000 for compound 12a).The IC 50 s thus assessed were 760 M for compound 11 and 1880 M for 12a, compared to 2.2 nM for zanamivir.

Conclusions
Herein we reported the extension of a powerful diasteroselective iHDA to the synthesis of KDObased glycosyloxathiins 11 and 12a-c (Scheme 3) efficiently prepared in few steps reacting two KDO-based enitols as electron-rich dienophiles.The data obtained confirmed that this peculiar class of iHDA is an efficient and versatile access to structurally heterogeneous diasteropure constructs.
Keeping in mind the recursively pandemic influenza A infections as well as the reported virus resistance to commonly used drugs like oseltamivir which make the discovery of new antiinfluenza drugs compelling, the sialidase inhibition properties of 11 and 12a were evaluated.The enzyme inhibition tests were carried out on the A/Mississippi/03/01 H1N1 virus sialidase, showing that both compounds, 11 and 12a, inhibited the enzyme activity.In particular, compound 11 was about 2-fold more effective than 12a.The striking steep inhibition curves obtained likely reflect a difference in the binding interactions of 11 and 12a vs. the influenza NA active site with respect to zanamivir.In conclusion, though sensibly less effective than the golden standard zanamivir, the KDO-related glycosyloxathiin 11 inhibits sialidase and is characterized by an original and multifunctional scaffold helpful to develop a new generation of drugs.Structural modifications and further binding profile investigations are underway.

Synthesis of compound 17.
To a stirred solution of 13 (0.500 g, 1.85 mmol) in CH 2 Cl 2 (8.0 mL), dry pyridine (5.0 mL) was added.The mixture was warmed at 60°C for 50h then diluted with CH 2 Cl 2 (80 mL) and washed with a saturated solution of NH 4 Cl (2 × 10 mL).The organic phase was dried over Na 2 SO 4 and concentrated to dryness to give a crude which was purified by flash column chromatography on silica gel (AcOEt:Petroleum Ether + NEt 3 0.1%, 1:8) to give 17 (0.300 g, 60%) as a yellow oil.[

Synthesis of compound 18.
To a solution of 17 (0.300 g, 1.11 mmol) in CHCl 3 (2.5 mL) dry pyridine (5.5 mL) and 14a (0.441 mg, 1.435 mmol) were added.The mixture was warmed at 50°C and stirred for 6 h.After this time 14a (0.441 mg, 1.435 mmol) was added and the mixture was stirred at 50°C for 18 h.The reaction mixture was cooled at rt, diluted with CH 2 Cl 2 (70 mL) and washed with a saturated solution of NH 4 Cl (2 × 15 mL).The organic phase was dried over Na 2 SO 4 and concentrated to dryness to give a crude which was purified by flash column chromatography on silica gel (AcOEt:Petroleum Ether + NEt 3 0.1%, 1:7) to give 18 (0.406 g, 85%) as a white solid.[

Synthesis of compound 16b.
To a stirred solution of 15b (0.214 g, 0.418 mmol) in CH 2 Cl 2 (1.5 mL) glacial AcOH (6.0 mL) and H 2 O (4.0 mL), were slowly added.The reaction mixture was warmed to 50°C and stirred for 6h.After this time, the solvent was co-evaporated with toluene under reduced pressure to give a crude which was purified by flash column chromatography on silica gel (AcOEt:MeOH, 20:1) to give 16b (0.111 g, 62%) as a glassy solid.

Enzyme inhibition assays
The A/Mississippi/3/2001 wild type H1N1 influenza virus 34 was used to evaluate susceptibility in an enzyme inhibition assays.25 l of A/Mississippi/3/2001 wild type H1N1 influenza virus was mixed with 25 l of inhibitor (Zanamivir or 11 or 12a), and after preincubation for 30 min at room temperature 50 l of MUNANA was added.After 60 min at 37 o C the reaction was stopped with the addition of 200 mM Na 2 CO 3 .Fluorescence units were quantitated with a BMG FLUOstar with an excitation wavelength of 365 nm and an emission wavelength of 450 nm.Final concentrations in the assay were 50 mM sodium acetate pH 5.5, 5 mM CaCl 2 and 100 M MUNANA.Serial 10-fold dilutions of Zanamivir were prepared in water.Dilutions of the compound 11 and 12a ranged from 10 to 10000 M.