Synthesis, structural characterization and biological evaluation of 4′-C-methyl- and phenyl-dioxolane pyrimidine and purine nucleosides

Nucleoside analogues play an important role in antiviral, antibacterial and antineoplastic chemotherapy. Herein we report the synthesis, structural characterization and biological activity of some 4′-C-methyl- and -phenyl dioxolane-based nucleosides. In particular, α and β anomers of all natural nucleosides were obtained and characterized by NMR, HR-MS and X-ray crystallography. The compounds were tested for antimicrobial activity against some representative human pathogenic fungi, bacteria and viruses. Antitumor activity was evaluated in a large variety of human cancer cell-lines. Although most of the compounds showed non-significant activity, 23α weakly inhibited HIV-1 multiplication. Moreover, 22α and 32α demonstrated a residual antineoplastic activity, interestingly linked to the unnatural α configuration. These results may provide structural insights for the design of active antiviral and antitumor agents.


Introduction
The development of nucleoside and nucleotide analogues has been a very active research area over the past few decades (Jordheim 2013;Romeo 2010). Modified nucleosides have been employed to inhibit essential enzymes involved in the replication pathways of many viruses (Jordheim 2013;De Clercq 2013), parasites (Lawton 2005) and bacteria (Van Calenbergh 2012). Moreover, nucleoside-based drugs have proved to be useful in chemotherapy due to their antitumor activities (Jordheim 2013;Mitsuya 1985). Among the useful modifications on nucleosides, variation of the carbohydrate moiety has been extensively studied and several deoxynucleosides (I) and dideoxynucleosides (II) have been discovered (Chart 1) (Mitsuya 1985;Mitsuya 1986;Yarchoan 1988;Hamamoto 1987). However, the emergence of drug resistance and toxicity issues remain a major cause for treatment failure, prompting the scientific community to keep on searching for new effective and low toxic drugs.
On the basis of these findings, we extended the series of 4 0 -C-methyl dioxolane-based nucleosides and we prepared the 4 0 -C-phenyl analogues to verify the effect on activity of a bulkier substituent. Herein, we report the synthesis, structural characterization and biological evaluation of novel 4 0 -C-methyl and -phenyl dioxolane-based pyrimidine and purine nucleosides (V, Chart 1).

General
All reagents, solvents and other chemicals were used as purchased from Sigma-Aldrich without further purification unless otherwise specified. When air-or moisture-sensitive reactants were employed the reactions were carried out under nitrogen atmosphere and dry solvents, unless otherwise noted. Flash column chromatography purifications (medium pressure liquid chromatography) were carried out using Merck silica gel 60 (230-400 mesh, ASTM). The purity of the compounds was determined by elemental analysis (C, H, N) on a Carlo Erba 1106 Analyzer. Melting points were determined with a Stuart SMP3 and they are uncorrected. The structures of all the isolated compounds were ensured by nuclear magnetic resonance (NMR) and Mass spectrometry. 1 H and 13 C NMR (1D and 2D experiments) spectra were recorded on a DPX-200 Avance (Bruker) spectrometer at 200 MHz and on a DPX-400 Avance (Bruker) spectrometer at 400 MHz. Chemical shifts are expressed in d (ppm). 1 H NMR chemical shifts are relative to tetramethylsilane (TMS) as internal standard. 13 C NMR chemical shifts are relative to TMS at d 0.0 or to the 13 C signal of the solvent: CDCl 3 d 77.04, CD 3 OD d 49.8, DMSO-d 6 d 39.5. 1 H NMR data are reported as follows: chemical shift, multiplicity (s, singlet; d, doublet; t, triplet; q, quartet; m, multiplet; br, broadened), coupling constants (Hz), number of protons, and assignment. 1 H-1 H correlation spectroscopy and nuclear overhauser effect spectroscopy (NOESY), 1 H-13 C heteronuclear single quantum coherence and heteronuclear multiple bond connectivity (HMBC) NMR 2D experiments were recorded for determination of 1 H-1 H and 1 H-13 C correlations respectively. The purities of the final compounds were confirmed by elemental analysis, on a Carlo Erba 1106 Analyzer, and the values obtained are within ±0.4 % of the calculated ones. The exact mass of the compounds was determined by using a Q-TOF mass spectrometer equipped with an electrospray ESI or atmospheric-pressure chemical ionization source. MS(?) spectra were acquired by direct infusion (0.1 mL/min) of a solution containing the sample (10 pmol/mL), dissolved in a 0.1 % acetic acid, with a mobile phase methanol/water 50:50, at the optimum ion voltage of 4800 V. The yields reported are based on a single experiment and are not optimized. All the analytical data are reported in the Supplementary Materials.

X-ray crystallography
Single-crystal X-ray structure determinations on racemates 32a and 32bÁH 2 O were carried out at 140(2) K and at room temperature, respectively, on a Bruker-Nonius X8APEX diffractometer equipped with Mo-K a generator, area detector and Kryoflex liquid nitrogen cryostat. The structures were solved in space group P-1 and successfully refined on F o 2 by standard methods, using SIR92 (Altomare et al. 1993) and SHELXL-97 (Sheldrick 2008). softwares included in the WINGX v2013.3 suite (Farrugia 2012). Geometrical calculations were carried out using PLATON (V-290610) (Spek 2003). All nonhydrogen atoms were refined anisotropically. The asymmetric unit in 32a was found heavily disordered even after expansion to space group P1. The possibility of overlooked reciprocal lattice layers and of an incorrectly-determined unit cell metrics was excluded by careful examination of precession images reconstructed from data collection frames. The two disordered components correspond to the two enantiomers of the compound and are present in a 1:1 ratio; within experimental resolution, they share the same positions for the phenyl atoms. Refinement was carried out by restraining the two parts to have flat thymine moieties Chart 1 Working hypothesis (within 0.05 Å 3 ) and to share the same geometry (1,2-and 1,3-distances equal within 0.01-0.02 Å , respectively). Approximate isotropic behaviour was imposed (within 0.01/0.02 Å 2 for nonterminal/terminal atoms) and neighbouring atoms were forced to have similar U ij components (within 0.01 Å 2 ). The phenyl ring in 32bÁH 2 O was found disordered over two equally-populated positions rotated by ca. 26.5°around the C ipso -C-2 0 bond. The two parts were constrained to share the same C ipso atom, to be flat (as above) and to have the same geometry (as above). The U ij components along each rings were also restrained to be similar (as above).
The disordered hydroxyl group in 32a was subject to rotating group refinement (AFIX 147 instruction, with thermal parameter U(H) = 1.5U eq (O)). Water and hydroxyl hydrogen atoms in 32bÁH 2 O were located in DF maps and fully refined with isotropic displacement parameters. The remaining hydrogens of both compounds were set in idealized positions and refined isotropically, with U(H) = 1.5U eq (X) for methyl groups and U(H) = 1.2U eq (X) for the remaining hydrogen atoms (X is the attached C or N atom). Torsion angle was refined for methyl groups (AFIX 137 instruction). Crystal data and refinement parameters for compounds 32a and 32bÁH 2 O are given in Table S1 in Supplementary Material. Partiallylabelled ORTEP-3 (Farrugia 1997). plots are displayed. The geometry of H-bond interactions is detailed in Tables S2 and S3, while Table S4 gathers conformational data for 32a, 32b H 2 O and related compounds. CCDC-1055778 and 1055779 contain the supplementary crystallographic data for this paper. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.

Biological evaluation methods
Compounds were dissolved in DMSO at 200 mM and then diluted in culture medium.
Cell lines were purchased from american type culture collection (ATCC). The absence of mycoplasma contamination was checked periodically by the Hoechst staining method. Cell lines supporting the multiplication of HIV-1 and BVDV were CD4 ? human T-cells containing an integrated HTLV-1 genome (MT-4) and Madin Darby Bovine Kidney (MDBK), respectively. Viruses representative of positive-sense, single-stranded RNAs (ssRNA ? ) were the III B laboratory strain of human immunodeficiency virus type-1 (HIV-1), obtained from the supernatant of the persistently infected H9/III B cells, and bovine viral diarrhoea virus (BVDV).

Cytotoxicity assays
Exponentially growing MT-4 cells (analogous to those which support the replication of HIV-1) were seeded in 96-well plates, at an initial density of 1 9 10 5 cells/mL, in RPMI-1640 medium supplemented with 10 % fetal bovine serum (FBS), 100 units/mL penicillin G and 100 lg/mL streptomycin. Cell cultures were then incubated at 37°C in a humidified, 5 % CO 2 atmosphere, in the absence or presence of serial dilutions of test compounds. Cell viability was determined after 96 h at 37°C by the 3-(4,5dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide MTT method (Pauwels et al. 1998). MDBK cells were seeded in 96-well plates, at an initial density of 6 9 10 5 cells/mL, in minimum essential medium with Earle's salts (MEM-E) containing L-glutamine (1 mM), sodium pyruvate (5 mM) and kanamycin (25 mg/L), supplemented with 10 % FBS. Cell cultures were then incubated at 37°C in a humidified, 5 % CO 2 atmosphere in the absence or presence of serial dilutions of test compounds. Cell viability was determined after 48 h at 37°C by the MTT method.

Antiviral assays
The anti-HIV-1 activity was based on inhibition of virusinduced cytopathogenicity in exponentially growing MT-4 cell acutely infected at a multiplicity of infection (m.o.i.) of 0.01. Briefly, 50 lL of RPMI containing 1 9 10 4 MT-4 cells were added to each well of flat-bottom microtitre trays containing 50 lL of RPMI without or with serial dilutions of test compounds. Then, 20 lL of a HIV-1 suspension containing 100 CCID 50 were added. After 4 days of incubation at 37°C cell viability was determined by the MTT method (Pauwels et al. 1998). The activity against BVDV was based on inhibition of the virus-induced cytopathogenicity in MDBK cells acutely infected at a m.o.i. of 0.01. Briefly, MDBK cells were seeded in 96-well plates at a density 3 9 10 4 cells/well and were allowed to form confluent monolayers by incubating overnight in growth medium at 37°C in a humidified CO 2 (5 %) atmosphere. Cell monolayers were then infected with 50 lL of a proper virus dilution in maintenance medium (MEM-Earl supplemented with 0.5 % inactivated FBS) to give an m.o.i of 0.01. After 1 h, 50 lL of maintenance medium, without or with serial dilutions of test compounds, were added. After a 3 day incubation at 37°C, cell viability was determined by the MTT method. Efavirenz and 2 0 -Cmethyl-guanosine were used as reference inhibitors of HIV-1 and BVDV, respectively. Synthesis, structural characterization and biological evaluation… 539

Antibacterial assays
Staphylococcus aureus and Salmonella spp. were recent clinical isolates (obtained from Clinica Dermosifilopatica, University of Cagliari). Assays were carried out in nutrient broth (DIFCO), pH 7.2, with an inoculum of 10 3 bacterial cells/tube. Minimum inhibitory concentrations (MICs) were determined after incubation at 37°C for 18 h in the presence of serial dilutions of test compounds (Artico et al. 1999).

Antifungal assays
Yeast inocula were obtained by properly diluting cultures of ATCC strains incubated at 37°C for 30 h in Sabouraud dextrose broth to obtain 5 9 10 3 cells/mL. Dermatophyte inocula were obtained from cultures grown at 37°C for 5 days in Sabouraud dextrose broth by finely dispersing clumps with a glass homogenizer and then diluting to 0.05 OD 590 /mL. Then, 20 lL of the above suspensions were added to each well of flat-bottomed microtitre trays containing 80 lL of medium with serial dilutions of test compounds and were incubated at 37°C. Growth controls were visually determined after 2 days (yeast) or 3 days (dermatophytes) (Menozzi et al. 2004). MIC was defined as the compound concentration at which no macroscopic sign of fungal growth was detected. The minimal bactericidal/fungicidal concentrations (MBC/MFC) were determined by subcultivating in Sabouraud dextrose agar samples from cultures with no apparent growth. Miconazole and Streptomycin were used as reference compounds in antimycotic and antibacterial assays, respectively.

Linear regression analysis
The extent of cell growth/viability or viral multiplication at each drug concentration tested was expressed as percentage of untreated or uninfected controls. Concentrations resulting in 50 % inhibition (CC 50 or EC 50 ) were determined by linear regression analysis.

Sulforhodamine B (SRB) assay
Cells (5 9 10 4 cells/mM) were treated with various concentrations of compounds in 96-well culture plates for 72 h. After incubation, cells were fixed with 10 % trichloroacetic acid, dried, and stained with 0.4 % SRB in 1 % acetic acid. The unbound dye was washed out, and the stained cells were dried and resuspended in 10 mM Tris (pH 10.0). The absorbance at 515 nm was measured, and cell proliferation was determined as follows: cell proliferation (%) = (average absorbance compound -average absorbance day zero ) / (average absorbance control -average absorbance day zero ) 9 100 (Lee 2002). IC 50 values were calculated by nonlinear regression analysis using Table Curve 2D v 5.01 (Systat Software Inc., Richmond, CA, USA).

Results and discussion
Chemistry Compounds 19223 and 31235 were prepared through the coupling of a nucleophilic pyrimidine or purine with a suitable 4 0 -C-substituted sugar-like dioxolane, electrophilic at the anomeric carbon (position 1 0 -C) for the presence of the acetoxy group, as described by Vorbrüggen (Vorbrueggen et al. 1981).
Thus, depending on the reactivity of the silylated nucleobases the Vorbrüggen reaction was carried out at room temperature in dichloromethane (uracil, thymine), or by heating in 1,2-dichloroethane (cytosine and purine). Significantly, in the case of cytosine, heating favours the formation of the b anomer.
For the 4 0 -C-methyl series, the separation of the two couples of isomers was accomplished by silica gel flash column chromatography directly on the Vorbrüggen reaction mixture to yield pure a and b anomers of uracil 8, thymine 10, 6-chloropurine 15 and 2-amino-6-chloropurine 17 (Scheme 2).
In the case of the cytosine derivative 12, benzoylation of the aminic function, to give 13, was required in order to separate a from b. In this series, to the first eluting isomer was assigned the a configuration based on 1 H-1 H NMR experiments (see section ''Structural characterization'').
For the 4 0 -C-phenyl series, the separation was achieved on the acetoxy derivatives 26-30, obtained from the bromomethyl derivatives 9, 11,14 and 24, 25 (Scheme 3). In the case of 24 and 25, the conversion of the 6-chloropurine and 2-amino-6-chloropurine derivatives 16 and 18 into the corresponding adenine and guanine derivatives was required before the substitution reaction. Unlike the 4 0 -Cmethyl series, the configuration assignment was opposite and the first eluting isomer was the b anomer. As previously reported, the glycosylation reaction with the purine nucleobases led to the formation of two regioisomers (N-7/ N-9) (Franchini et al. 2012). Reaction time, temperature and solvent were optimized in order to guide the reaction towards N-9 natural nucleosides. In particular, we observed that the conversion of the probably kinetically favored N-7 into the thermodynamically more stable N-9 was achieved by refluxing in 1,2-dichloroethane for 4-8 h. Differently, in the case of pyrimidine nucleobases, only N-1 natural regioisomers were obtained. Lastly, deprotection of the hydroxyl group with NH 3 /MeOH at room temperature afforded the final a and b nucleosides 19-21 (Scheme 2) and 31-35 (Scheme 3). For the 4 0 -C-methyl series, the final purine nucleosides 22 and 23 were obtained from 15 to 17 after heating in a reactor in the presence of NH 3 /MeOH or NaOH/MeOH, respectively (Scheme 2).
In particular, within the methyl series a cross-peak between H-4 0 and the protons of the methyl group at 2 0 position is indicative of a b configuration. By contrast, a correlation between H-4 0 and the protons of the hydroxymethyl group is observed in a anomers. The latter correlation was also used to assign the configuration within the phenyl series. These assignments were confirmed by inspection of the chemical shifts observed in the 1 H-NMR spectra (Table 1).
In particular, for the a anomers of the methyl series, methyl and hydroxymethyl protons undergo downfield and upfield shifts, respectively, as compared with the corresponding b anomers. These differences are attributed to the deshielding effect exerted by the nucleobase, which is located on the same side of the dioxolane ring as the methyl/hydroxymethyl group in a/b anomers. These findings are in accordance with those reported for other nucleoside analogues (Kim et al. 1992a(Kim et al. , b, 1993. For the phenyl series the relative chemical shifts of the proton H-4 0 (dioxolanyl) and H-6 (pyrimidinyl) or H-8 (purinic) were particularly informative (Table 2). In fact, in the b-anomers the signal of the dioxolanyl proton (H-4) occurs at higher field than in the corresponding a-anomers, due to the shielding effect exerted by the phenyl ring. For the same reason, in the a anomers the pyrimidinyl proton H-6 or the purinic proton H-8 undergo upfield shifts with respect to the corresponding b-anomers.
In particular, for the pyrimidine derivatives, the correlation between the proton H-4 0 and the carbon C-6 of the nucleobase was diagnostic of N-1 substitution. Similarly, for the purine derivatives, the correlation between the proton H-4 0 and the quaternary carbon C-4 of the nucleobase was employed to identify N-9 regioisomers. The assignment of N-9 or N-7 substitution was also confirmed by UV analysis: k max of N-9 and N-7 isomers was 265 and 273 nm, respectively, in accordance with previously reported data (Franchini et al. 2012;Jeong et al. 1993).

X-ray crystallography
In order to confirm the structural and configurational data, the structures of the racemates 32a and 32b were determined by X-ray crystallography, as shown in Fig. 2. From water solution, 32a grows as colourless plates belonging to triclinic space group P-1. The structure is heavily disordered and the asymmetric unit comprises the two enantiomers of the compound in 1:1 ratio and sharing the same crystallographic site (Fig. 2a). The two molecules (A and B) have coinciding phenyl substituents (within experimental Scheme 2 Reagents and conditions: (i) NH 3 /MeOH, r.t., 12 h, 60 % for 19a and 64 % for 19b, 74 % for 20a and 72 % for 20b, 66 % for 21a and 95 % for 21b; (ii) NH 3 /MeOH, reactor, 90°C, 12 h, 52 % for 22a and 60 % for 22b, or NaOH sol., MeOH, reflux, 12 h, 44 % for 23a and 73 % for 23b resolution), slightly shifted thymine groups and 2-(hydroxymethyl)-1,3-dioxolan-4-yl moieties with significantly different conformations (Table S4). The five-membered ring of molecule A is twisted, with the largest endocyclic torsion angle around the C-5 0 -O-1 0 bond (m 2 = 32.4(8)°). By contrast, molecule B features an envelope conformation on C-5 0 , with a torsion angle m 4 = -3.6(9)°around C-2 0 -O-3 0 bond. In both molecules, the thymine and hydroxymethyl groups are in axial positions, whereas the phenyl substituent is bisectional. The glycosidic bond lengths average to 1.480(5) Å and glycosidic torsion angles are v A = -138.1(3)°and v B = -174.8(4)°, indicating -ac and -ap orientations of the thymine group (the sign of torsion angles in this section is appropriate for enantiomers with an R configuration at the glycosidic carbon atom, as in Fig. 2b) (Norbeck et al. 1989;Kim et al. 1992a, b).
Compound 32b was isolated from H 2 O solution as colorless needles of the monohydrate. Crystals of 32bÁH 2 O are triclinic, space group P-1, and the unit cell comprises two molecules related by an inversion center (Fig. 2b). The 1,3-dioxolane ring adopts a twisted conformation (Table S4), the largest endocyclic torsion angle being found around the C-5 0 -O-1 0 bond (m 2 = 39.34(14)°), as in molecule A of 32a. The thymine substituent is in axial position, with a glycosidic bond length of 1.4812(19) Å , hence identical to that found in 32a. The hydroxymethyl and phenyl substituents are in bisectional and axial positions, respectively. The torsion angle at the glycosidic bond (v = -82.97(16)°) is smaller (in absolute value) than in 32a, indicating a -sc orientation for the thymine substituent. This contrasts with the anti-conformation found in the b isomer of Dioxolane-T in both enantiomerically-pure (Kim et al. 1992a, b) and racemic form (Norbeck et al. 1989). The hydroxymethyl substituent in the two compounds is found in ± sc orientation with |c|*60°, indicating the typical gauche-gauche conformation (Norbeck et al. 1989;Kim et al. 1992a, b).
In spite of the aforementioned disorder effects, the crystal lattice of 32a entails a well-defined pattern of hydrogen bonds. Base-pairing interactions link molecules into dimers across inversion centers. Complementary hydrogen bonds between carbonyl and hydroxyl groups afford chains of coplanar dimers, which interact laterally via hydrophobic contacts to give sheets parallel to (121) planes. Neighboring sheets are linked primarily by basestacking interactions.
In the crystal lattice of 32bÁH 2 O, molecules are extensively hydrogen bonded to give sheets oriented parallel to the (001) Fig. 2b). These hydrogenbonded sheets alternate along the c axis with hydrophobic layers comprising phenyl substituents. Additional details on intermolecular contacts are provided in Supplementary Materials.

Biological activity
The compounds were tested in cell-based assays against human immunodeficiency virus type-1 (HIV-1) and BVDV. Efavirenz and 2 0 -C-methyl-guanosine were used as reference inhibitors, respectively. The cytotoxicity against the cells lines supporting the viral replication was evaluated in parallel with the antiviral activity. All compounds showed no cytotoxicity against MT4 or MDBK cells. However, only 23a weakly inhibited the HIV-1 multiplication (EC 50 = 52 lM) whereas none of them showed anti-BVDV activity (Table S5). The title compounds were also evaluated in vitro against representatives of Gram negative (Salmonella spp) and Gram positive (S. aureus) bacteria and human pathogenic fungi (Candida albicans). Miconazole and Streptomycin were used as reference compounds in antimycotic and antibacterial assays, respectively. None of the compounds showed antimicrobial activity (Table S6 in Supplementary  Materials).
Lastly, the compounds were assayed for cytotoxic effects in several human cancer cell lines such as colon cancer (HCT116), lung cancer (A549), stomach cancer (SNU638), prostate cancer (PC-3) and human liver cancer (SK-Hep-1) cells, using the sulforhodamine B (SRB) protein staining method as reported in Table 3 (Lee et al. 2002).

Theoretical calculations
In order to gain insight into the scarce antiviral activity observed in the cell-based assays, molecular orbital calculations were performed. It is known that kinases display a substrate preference for South (S) type conformation while DNA polymerases almost exclusively incorporate the North (N) triphosphate (Marquez et al. 2004(Marquez et al. , 2006. Moreover, it has been reported that syn-and anti-conformation around the N-glycosyl bond plays a critical role for the binding of ligands to the proper enzyme (El Kouni et al. 1996). In particular, several crystal structures of nucleoside kinases show the nucleobase in the anti conformation (Kim et al. 1992a, b;Norbeck et al. 1989). Thus, the relative energies of N and S conformers were determined as a function of the torsion angle (v) around the glycosidic bond for the reference compound Dioxolane-T and its analogues, 20b and 32b (Fig. 3).
In particular a single enantiomer (2 0 R, 4 0 R) of Dioxolane-T, 20b and 32b was employed (in this section we adopt the same atom numbering scheme as in Structural Cell proliferation was determined as follows: cell proliferation (%) = (average absorbance compound -average absorbance day zero )/(average absorbance control -average absorbance day zero ) 9 100 IC 50 values were calculated by nonlinear regression analysis using TableCurve 2D v5.01 (Systat Sofrware Inc., Richmond, CA, USA) a Franchini and Brasili (1997) characterization). In the N-and S-type conformers the dihedral angle m 2 = C4 0 -C5 0 -O1 0 -C2 0 was fixed at 42°and -42°, respectively, as previously reported (Norbeck et al. 1989;Kubota et al. 2013). Rotational energies around the N-glycosyl bond O3 0 -C4 0 -N1-C2 were then calculated for both conformers at the HF/3-21G* level employing Gaussian 09 program. As expected the North or South conformation of the ribose influences the conformation of the nucleobase (syn or anti) relative to the sugar moiety (Saenger 1984). According to our calculations Dioxolane-T adopts N-type puckering and anti-orientation (v*180°) as the preferential conformation. This result is qualitatively consistent with the solid state structure of Dioxolane-T in both enantiomerically-pure and racemic forms (Norbeck et al. 1989) (see Table S4). Significantly, the introduction of the methyl or phenyl substituent has no influence on the preferential conformation of the dioxolane ring obtained by theoretical calculations, which remains the same as for Dioxolane-T. Notice, however, that the crystal structure of 32bÁH 2 O reveals an N-type puckering but a syn orientation, even though with a large |v| value (Table S4). The conformational data obtained for 20b and 32b thus reveal a similar conformational landscape as in Dioxolane-T. In particular, for the new compounds, the energy difference between the S-anti and N-anti conformations is comparable or lower than found in Dioxolane-T.
However, although the S-anti conformation appears to be stable enough to be processed by the cellular kinases, the monophosphate formation, which has often been found to be the rate-limiting step, could be strongly inhibited by the steric hindrance of the substituent at position 4 0 . In this regard it would be interesting to prepare some monophosphate-prodrugs to see if, by overcoming the first phosphorylation step, these compounds would gain some biological activity (Perrone et al. 2007;McGuigan et al. 2009).
In summary, we efficiently synthesized 4 0 -C-methyl and -phenyl 1,3-dioxolane-based pyrimidine and purine analogues. The synthetic approach has provided good yields of the desired products both as a and b anomers. Additionally, we have reported conformational and configurational studies, supported by NMR and X-ray crystallography, for this new class of compounds. The evaluations of the biological activity revealed only a weak anti-HIV-1 activity of compound 23a. Furthermore 22a and 32a demonstrated residual anti-proliferative activity, interestingly linked to the unnatural a configuration. Although active compounds were not obtained, there is no doubt that the accumulation of SAR studies, including this, may provide structural insights for the design of active antiviral and antitumor agents. Moreover, the synthetic strategy presented here could represent a versatile approach to obtain both a and b anomers of different 4 0 -C-substituted dioxolane-based nucleosides.