Stereoselective Synthesis of Selenium-Containing Glycoconjugates via the Mitsunobu Reaction

A simple and efficient route for the synthesis of new glycoconjugates has been developed. The approach acts as a model for a mini-library of compounds with a deoxy-selenosugar core joined to a polyphenolic moiety with well-known antioxidant properties. An unexpected stereocontrol detected in the Mitsunobu key reaction led to the most attractive product showing a natural d-configuration. Thus, we were able to obtain the target molecules from the commercially available d-ribose via a shorter and convenient sequence of reactions.


Introduction
In the last decade, a large amount of evidence has suggested a link to oxidative stress (OS) and ageing-related neurodegenerative diseases (NDs) [1,2], which are characterized by progressive damage in neural cells as well as by a selective loss of neuronal populations. Since OS refers to an imbalance between the production and removal of reactive oxygen species (ROS), antioxidant therapies have been proposed to prevent or attenuate their induced damage [3,4].
Recently, the therapeutic effects of several small molecules, normally taken with the diet, have been studied and their potential applications proposed. In more detail, several selenium-containing compounds have been synthesized, and they turned out to be potent neuroprotective agents with a modest effect on normal tissues and are clinically well tolerated [5,6]. Among them, the best known is the Ebselen [7] (Figure 1), a small organoselenium compound that acts as a glutathione peroxidase (GPx) mimetic. Likewise, several polyphenolic molecules have shown a strong antioxidant activity [8,9], including caffeic acid, which has the potential to modulate oxidation and inflammation, as demonstrated in several studies [10]; curcumin, with proven anticancer and anti-inflammatory characteristics [11,12]; and dopamine, whose loss is involved in the early phase of Alzheimer's Disease (AD) development [13].

Introduction
In the last decade, a large amount of evidence has suggested a lin (OS) and ageing-related neurodegenerative diseases (NDs) [1,2], whi by progressive damage in neural cells as well as by a selective loss o tions. Since OS refers to an imbalance between the production and oxygen species (ROS), antioxidant therapies have been proposed to p their induced damage [3,4].
Recently, the therapeutic effects of several small molecules, norm diet, have been studied and their potential applications proposed. In selenium-containing compounds have been synthesized, and they tur neuroprotective agents with a modest effect on normal tissues and ar erated [5,6]. Among them, the best known is the Ebselen [7] (Figu noselenium compound that acts as a glutathione peroxidase (GPx) several polyphenolic molecules have shown a strong antioxidant acti caffeic acid, which has the potential to modulate oxidation and inflam strated in several studies [10]; curcumin, with proven anticancer and characteristics [11,12]; and dopamine, whose loss is involved in the heimer's Disease (AD) development [13].   The introduction of selenium into organic structures has bee , regio-, and stereoselective methods, such as nucleophile, electr niques only during the past three decades, challenging the we unusual behavior of early derivatives [14]. The results of Matsu Jeong [17] represent the starting approach towards the chemistry pounds.
Inspired by these studies and taking into account that, to th no examples of 4-selenosugar-5-conjugates have been described we wish to report their first synthesis here.

Results and Discussion
As outlined in the retrosynthetic path (Scheme 1), we started available D-ribose. In fact, the latter was converted into the corre sugar derivative 1, modifying a procedure reported by Jeong et a Firstly, we protected D-ribose with orthogonal groups, with moving them during the synthetic strategy. In detail, isopropylid hydroxyl groups in C-2 and C-3 through a protocol [19] already by smooth reaction conditions with a low environmental impact with a 95% yield. The O-isopropilydene derivative, without any work up or purification procedures and after treatment with t ylsilane (TBDPSCl), triethylamine (TEA), and dimethylaminopy drous CH2Cl2, afforded the silyl ether 3, which in turn provided lent yield (>99%) under reductive conditions (sodium borohydri the diol 4 was converted to the corresponding dimesylate 5 allow ing groups on C-1 and C-4 so that the subsequent treatment of presence of sodium borohydride in EtOH-THF at 60 °C gave th rivative 6 at a 95% yield. It is noteworthy that the reaction takes the configuration [16][17][18], thus leading to the corresponding L-d The introduction of selenium into organic structures has been performed via chemo-, regio-, and stereoselective methods, such as nucleophile, electrophile, and radical techniques only during the past three decades, challenging the well-known instability and unusual behavior of early derivatives [14]. The results of Matsuda [15], Pinto [16], and Jeong [17] represent the starting approach towards the chemistry of selenium-ribose compounds.
Inspired by these studies and taking into account that, to the best of our knowledge, no examples of 4-selenosugar-5-conjugates have been described in the literature to date; we wish to report their first synthesis here.

Results and Discussion
As outlined in the retrosynthetic path (Scheme 1), we started from the commercially available D-ribose. In fact, the latter was converted into the corresponding deoxyselenosugar derivative 1, modifying a procedure reported by Jeong et al. [18].
Firstly, we protected D-ribose with orthogonal groups, with a view of selectively removing them during the synthetic strategy. In detail, isopropylidene was used to protect hydroxyl groups in C-2 and C-3 through a protocol [19] already described, characterized by smooth reaction conditions with a low environmental impact leading to compound 2 with a 95% yield. The O-isopropilydene derivative, without any further time consuming work up or purification procedures and after treatment with tert-butyl(chloro)diphenylsilane (TB-DPSCl), triethylamine (TEA), and dimethylaminopyridine (DMAP) in anhydrous CH 2 Cl 2 , afforded the silyl ether 3, which in turn provided the diol 4 with an excellent yield (>99%) under reductive conditions (sodium borohydride). The next step where the diol 4 was converted to the corresponding dimesylate 5 allowed us to enter good leaving groups on C-1 and C-4 so that the subsequent treatment of 5 with selenium in the presence of sodium borohydride in EtOH-THF at 60 • C gave the deoxy-selenosugar derivative 6 at a 95% yield. It is noteworthy that the reaction takes place with inversion of the configuration [16][17][18], thus leading to the corresponding L-deoxy-selenosugar. Using dry conditions allowed the optimization of the procedure [18], providing 6 with a 99% yield (Scheme 2). Finally, the removal of the TBDPS protecting group on C-5 exploiting the great affinity of fluoride to silicon afforded the seleno building block 7 with a high yield (75% overall yield).
To perform the reaction giving the glycoconjugate, we attempted to convert 7 to an iododerivative using a triphenylphosphine polymer-bound/iodine complex already reported on different substrates [20], but the only product found was the dimer, probably resulting from the quick formation of the corresponding iodide that was attacked by the OH group of the remaining 7. It was not possible to evaluate the efficiency of mesyl or tosyl derivatives, since both were prepared in very low yields. The failed results prompted us to take a different approach based on the Mitsunobu reaction, which is well-known and To perform the reaction giving the glycoconjugate, we attempted to convert 7 to an iododerivative using a triphenylphosphine polymer-bound/iodine complex already reported on different substrates [20], but the only product found was the dimer, probably resulting from the quick formation of the corresponding iodide that was attacked by the OH group of the remaining 7. It was not possible to evaluate the efficiency of mesyl or tosyl derivatives, since both were prepared in very low yields. The failed results prompted us to take a different approach based on the Mitsunobu reaction, which is well-known and carried out with an activated alcohol prepared in situ using triphenylphosphine (TPP) and an azodicarboxylate as promoter.
Therefore, our investigation began with the coupling of the seleno derivative 7 TPPcomplex with monoacetylated hydroquinone [21] chosen as a model compound. The reaction in dry THF after three days gave a 43%-yield compound 8 as a major and unexpected product (Scheme 3). Therefore, our investigation began with the coupling of the seleno derivative 7 TPPcomplex with monoacetylated hydroquinone [21] chosen as a model compound. The reaction in dry THF after three days gave a 43%-yield compound 8 as a major and unexpected product (Scheme 3).

Scheme 2. Synthesis of building block 7.
To perform the reaction giving the glycoconjugate, we attempted to convert iododerivative using a triphenylphosphine polymer-bound/iodine complex alre ported on different substrates [20], but the only product found was the dimer, p resulting from the quick formation of the corresponding iodide that was attacked OH group of the remaining 7. It was not possible to evaluate the efficiency of m tosyl derivatives, since both were prepared in very low yields. The failed results pr us to take a different approach based on the Mitsunobu reaction, which is well and carried out with an activated alcohol prepared in situ using triphenylphosphin and an azodicarboxylate as promoter.
Therefore, our investigation began with the coupling of the seleno derivative complex with monoacetylated hydroquinone [21] chosen as a model compound. action in dry THF after three days gave a 43%-yield compound 8 as a major an pected product (Scheme 3). NMR was used to assign the configuration at C4 in the latter. It is well-kno the stereostructure of the furanose moiety is often difficult to analyze due to the ergy barriers between various puckering states [22]. It has already been reporte literature that the bulky selenium atom in the 4′-selenoribonuclesosides led to (C2′-endo) pucker state [17]. However, since significant NOE effects were not dete the hydrogens at C2 and C4 and the coupling constants of 8 were not supportive NMR was used to assign the configuration at C4 in the latter. It is well-known that the stereostructure of the furanose moiety is often difficult to analyze due to the low energy barriers between various puckering states [22]. It has already been reported in the literature that the bulky selenium atom in the 4 -selenoribonuclesosides led to a south (C2 -endo) pucker state [17]. However, since significant NOE effects were not detected for the hydrogens at C2 and C4 and the coupling constants of 8 were not supportive for the stereochemistry assignment with certainty, we decided to evaluate the outcome of the Ddeoxy-selenosugar derivative 11 coming from the commercially available D-ribonolactone (Scheme 4). Thus, we employed the procedure reported by Batra et al. [23] to accomplish a first inversion of the configuration leading to the intermediate 9, which, in turn, was converted with a high yield into the corresponding D-target 11, using the same route previously described for his isomer 7.

PEER REVIEW 4 of 11
stereochemistry assignment with certainty, we decided to evaluate the outcome of the Ddeoxy-selenosugar derivative 11 coming from the commercially available D-ribonolactone (Scheme 4). Thus, we employed the procedure reported by Batra et al. [23] to accomplish a first inversion of the configuration leading to the intermediate 9, which, in turn, was converted with a high yield into the corresponding D-target 11, using the same route previously described for his isomer 7.

Scheme 4. Synthesis of the isomer of 7.
A comparative analysis of 1 H NMR coupling constants in compounds 6 and 10 as well as 7 and 11 (Table 1) was therefore carried out to investigate the range of the vicinal coupling constants. The 3 J3,2 medium value was 5.6 Hz, the 3 J2,1b was 4.8 Hz, and the 3 J2,1a was 2.0 Hz. More interestingly, however, was that for 6 and 7, the 3 J3,4 medium value was 4.3 Hz, while for 10 and 11, the same coupling value was 1.6 Hz. Taking this into account, in glycoconjugates, the 3 J3,4 medium value of 1. 6 Hz is consistent with the 3,4-anti-configuration [15]. The selectivity seen in the model reaction, whose mechanism is under investigation, was also detected when the scope of our approach was explored. Indeed, all polyphenolic substrates used gave the sole corresponding D-deoxy-seleno glycoconjugate (8, and 12-15, Table 2). Moreover, when we performed the Mitsunobu reaction on the same polyphenols using the isomer 11 as the starting seleno substrate, the spectral and analytical data clearly indicated we had obtained the same product 8 and 12-15 with corresponding comparable yields ( Table 2). These results, presumably resulting from the already reported unusual behavior of selenium [24], support the use in future of the "L-route" (seven steps up to isomer 7, overall yield 75%) to obtain further products, thereby avoiding the alternative time and products consuming "D-route" (eight steps up to isomer 11, overall yield 47%). A comparative analysis of 1 H NMR coupling constants in compounds 6 and 10 as well as 7 and 11 (Table 1) was therefore carried out to investigate the range of the vicinal coupling constants. The 3 J 3,2 medium value was 5.6 Hz, the 3 J 2,1b was 4.8 Hz, and the 3 J 2,1a was 2.0 Hz. More interestingly, however, was that for 6 and 7, the 3 J 3,4 medium value was 4.3 Hz, while for 10 and 11, the same coupling value was 1.6 Hz. Taking this into account, in glycoconjugates, the 3 J 3,4 medium value of 1. 6 Hz is consistent with the 3,4-anti-configuration [15]. stereochemistry assignment with certainty, we decided to evaluate the outcome of the Ddeoxy-selenosugar derivative 11 coming from the commercially available D-ribonolactone (Scheme 4). Thus, we employed the procedure reported by Batra et al. [23] to accomplish a first inversion of the configuration leading to the intermediate 9, which, in turn, was converted with a high yield into the corresponding D-target 11, using the same route previously described for his isomer 7.

Scheme 4. Synthesis of the isomer of 7.
A comparative analysis of 1 H NMR coupling constants in compounds 6 and 10 as well as 7 and 11 (Table 1) was therefore carried out to investigate the range of the vicinal coupling constants. The 3 J3,2 medium value was 5.6 Hz, the 3 J2,1b was 4.8 Hz, and the 3 J2,1a was 2.0 Hz. More interestingly, however, was that for 6 and 7, the 3 J3,4 medium value was 4.3 Hz, while for 10 and 11, the same coupling value was 1.6 Hz. Taking this into account, in glycoconjugates, the 3 J3,4 medium value of 1. 6 Hz is consistent with the 3,4-anti-configuration [15]. The selectivity seen in the model reaction, whose mechanism is under investigation, was also detected when the scope of our approach was explored. Indeed, all polyphenolic substrates used gave the sole corresponding D-deoxy-seleno glycoconjugate (8, and 12-15, Table 2). Moreover, when we performed the Mitsunobu reaction on the same polyphenols using the isomer 11 as the starting seleno substrate, the spectral and analytical data clearly indicated we had obtained the same product 8 and 12-15 with corresponding comparable yields ( Table 2). These results, presumably resulting from the already reported unusual behavior of selenium [24], support the use in future of the "L-route" (seven steps up to isomer 7, overall yield 75%) to obtain further products, thereby avoiding the alternative time and products consuming "D-route" (eight steps up to isomer 11, overall yield 47%). 6 1. The selectivity seen in the model reaction, whose mechanism is under investigation, was also detected when the scope of our approach was explored. Indeed, all polyphenolic substrates used gave the sole corresponding D-deoxy-seleno glycoconjugate (8, and 12-15, Table 2). Moreover, when we performed the Mitsunobu reaction on the same polyphenols using the isomer 11 as the starting seleno substrate, the spectral and analytical data clearly indicated we had obtained the same product 8 and 12-15 with corresponding comparable yields ( Table 2). These results, presumably resulting from the already reported unusual behavior of selenium [24], support the use in future of the "L-route" (seven steps up to isomer 7, overall yield 75%) to obtain further products, thereby avoiding the alternative time and products consuming "D-route" (eight steps up to isomer 11, overall yield 47%).  In the preliminary attempts performed on hydroquinone used as a model compound, the presence of a hydroxyl group different from the one involved in the reaction appeared to be a limit of reactivity for the methodology, thus affording a very low yield of the desired conjugated. On the other hand, the reaction on the monoacetylated hydroquinone proceeded with fair yields.
A different approach was reserved for curcumin, whose hydroxyl groups in C-4 have not been intentionally protected due to known handling problems of the molecule. Nevertheless, the yield was reasonable with the challenging purification procedure currently under improvement.
Concerning the polyphenols having the COOH group in addition to the phenolic one (caffeic and ferulic acids), as expected, due to the lower pka of the carboxyl group, they proved to be excellent substrates for the reaction (yields 67-72%), taking into account that we used a protecting-free approach, which is very useful in multistep programs.  In the preliminary attempts performed on hydroquinone used as a model compound, the presence of a hydroxyl group different from the one involved in the reaction appeared to be a limit of reactivity for the methodology, thus affording a very low yield of the desired conjugated. On the other hand, the reaction on the monoacetylated hydroquinone proceeded with fair yields.
A different approach was reserved for curcumin, whose hydroxyl groups in C-4 have not been intentionally protected due to known handling problems of the molecule. Nevertheless, the yield was reasonable with the challenging purification procedure currently under improvement.
Concerning the polyphenols having the COOH group in addition to the phenolic one (caffeic and ferulic acids), as expected, due to the lower pka of the carboxyl group, they proved to be excellent substrates for the reaction (yields 67-72%), taking into account that we used a protecting-free approach, which is very useful in multistep programs.  In the preliminary attempts performed on hydroquinone used as a model compound, the presence of a hydroxyl group different from the one involved in the reaction appeared to be a limit of reactivity for the methodology, thus affording a very low yield of the desired conjugated. On the other hand, the reaction on the monoacetylated hydroquinone proceeded with fair yields.
A different approach was reserved for curcumin, whose hydroxyl groups in C-4 have not been intentionally protected due to known handling problems of the molecule. Nevertheless, the yield was reasonable with the challenging purification procedure currently under improvement.
Concerning the polyphenols having the COOH group in addition to the phenolic one (caffeic and ferulic acids), as expected, due to the lower pka of the carboxyl group, they proved to be excellent substrates for the reaction (yields 67-72%), taking into account that we used a protecting-free approach, which is very useful in multistep programs.  In the preliminary attempts performed on hydroquinone used as a model compound, the presence of a hydroxyl group different from the one involved in the reaction appeared to be a limit of reactivity for the methodology, thus affording a very low yield of the desired conjugated. On the other hand, the reaction on the monoacetylated hydroquinone proceeded with fair yields.
A different approach was reserved for curcumin, whose hydroxyl groups in C-4 have not been intentionally protected due to known handling problems of the molecule. Nevertheless, the yield was reasonable with the challenging purification procedure currently under improvement.
Concerning the polyphenols having the COOH group in addition to the phenolic one (caffeic and ferulic acids), as expected, due to the lower pka of the carboxyl group, they proved to be excellent substrates for the reaction (yields 67-72%), taking into account that we used a protecting-free approach, which is very useful in multistep programs.  In the preliminary attempts performed on hydroquinone used as a model compound, the presence of a hydroxyl group different from the one involved in the reaction appeared to be a limit of reactivity for the methodology, thus affording a very low yield of the desired conjugated. On the other hand, the reaction on the monoacetylated hydroquinone proceeded with fair yields.
A different approach was reserved for curcumin, whose hydroxyl groups in C-4 have not been intentionally protected due to known handling problems of the molecule. Nevertheless, the yield was reasonable with the challenging purification procedure currently under improvement.
Concerning the polyphenols having the COOH group in addition to the phenolic one (caffeic and ferulic acids), as expected, due to the lower pka of the carboxyl group, they proved to be excellent substrates for the reaction (yields 67-72%), taking into account that we used a protecting-free approach, which is very useful in multistep programs.  In the preliminary attempts performed on hydroquinone used as a model compound, the presence of a hydroxyl group different from the one involved in the reaction appeared to be a limit of reactivity for the methodology, thus affording a very low yield of the desired conjugated. On the other hand, the reaction on the monoacetylated hydroquinone proceeded with fair yields.
A different approach was reserved for curcumin, whose hydroxyl groups in C-4 have not been intentionally protected due to known handling problems of the molecule. Nevertheless, the yield was reasonable with the challenging purification procedure currently under improvement.
Concerning the polyphenols having the COOH group in addition to the phenolic one (caffeic and ferulic acids), as expected, due to the lower pka of the carboxyl group, they proved to be excellent substrates for the reaction (yields 67-72%), taking into account that we used a protecting-free approach, which is very useful in multistep programs.  In the preliminary attempts performed on hydroquinone used as a model compound, the presence of a hydroxyl group different from the one involved in the reaction appeared to be a limit of reactivity for the methodology, thus affording a very low yield of the desired conjugated. On the other hand, the reaction on the monoacetylated hydroquinone proceeded with fair yields.
A different approach was reserved for curcumin, whose hydroxyl groups in C-4 have not been intentionally protected due to known handling problems of the molecule. Nevertheless, the yield was reasonable with the challenging purification procedure currently under improvement.
Concerning the polyphenols having the COOH group in addition to the phenolic one (caffeic and ferulic acids), as expected, due to the lower pka of the carboxyl group, they proved to be excellent substrates for the reaction (yields 67-72%), taking into account that we used a protecting-free approach, which is very useful in multistep programs.  In the preliminary attempts performed on hydroquinone used as a model compound, the presence of a hydroxyl group different from the one involved in the reaction appeared to be a limit of reactivity for the methodology, thus affording a very low yield of the desired conjugated. On the other hand, the reaction on the monoacetylated hydroquinone proceeded with fair yields.
A different approach was reserved for curcumin, whose hydroxyl groups in C-4 have not been intentionally protected due to known handling problems of the molecule. Nevertheless, the yield was reasonable with the challenging purification procedure currently under improvement.
Concerning the polyphenols having the COOH group in addition to the phenolic one (caffeic and ferulic acids), as expected, due to the lower pka of the carboxyl group, they proved to be excellent substrates for the reaction (yields 67-72%), taking into account that we used a protecting-free approach, which is very useful in multistep programs.   1 Reaction carried out on derivative 7 (1 eq), TPP (1.5 eq), diisopropyl azodicarboxyate (DIAD, 1.5 eq), and polyphenol (1.5 eq) in dry THF under N2 atmosphere for three days. 2 Percentage yield; isolated yield by column chromatography; double reaction products of the polyphenol were never detected. 3 Percentage yield; reaction starting from derivative 11 under same condition.
In the preliminary attempts performed on hydroquinone used as a model compound, the presence of a hydroxyl group different from the one involved in the reaction appeared to be a limit of reactivity for the methodology, thus affording a very low yield of the desired conjugated. On the other hand, the reaction on the monoacetylated hydroquinone proceeded with fair yields.
A different approach was reserved for curcumin, whose hydroxyl groups in C-4 have not been intentionally protected due to known handling problems of the molecule. Nevertheless, the yield was reasonable with the challenging purification procedure currently under improvement.
Concerning the polyphenols having the COOH group in addition to the phenolic one (caffeic and ferulic acids), as expected, due to the lower pka of the carboxyl group, they proved to be excellent substrates for the reaction (yields 67-72%), taking into account that we used a protecting-free approach, which is very useful in multistep programs.  In the preliminary attempts performed on hydroquinone used as a model compound, the presence of a hydroxyl group different from the one involved in the reaction appeared to be a limit of reactivity for the methodology, thus affording a very low yield of the desired conjugated. On the other hand, the reaction on the monoacetylated hydroquinone proceeded with fair yields.
A different approach was reserved for curcumin, whose hydroxyl groups in C-4 have not been intentionally protected due to known handling problems of the molecule. Nevertheless, the yield was reasonable with the challenging purification procedure currently under improvement.
Concerning the polyphenols having the COOH group in addition to the phenolic one (caffeic and ferulic acids), as expected, due to the lower pka of the carboxyl group, they proved to be excellent substrates for the reaction (yields 67-72%), taking into account that we used a protecting-free approach, which is very useful in multistep programs. 30 (28) 3 1 Reaction carried out on derivative 7 (1 eq), TPP (1.5 eq), diisopropyl azodicarboxyate (DIAD, 1.5 eq), and polyphenol (1.5 eq) in dry THF under N 2 atmosphere for three days. 2 Percentage yield; isolated yield by column chromatography; double reaction products of the polyphenol were never detected. 3 Percentage yield; reaction starting from derivative 11 under same condition.
In the preliminary attempts performed on hydroquinone used as a model compound, the presence of a hydroxyl group different from the one involved in the reaction appeared to be a limit of reactivity for the methodology, thus affording a very low yield of the desired conjugated. On the other hand, the reaction on the monoacetylated hydroquinone proceeded with fair yields.
A different approach was reserved for curcumin, whose hydroxyl groups in C-4 have not been intentionally protected due to known handling problems of the molecule. Nevertheless, the yield was reasonable with the challenging purification procedure currently under improvement.
Concerning the polyphenols having the COOH group in addition to the phenolic one (caffeic and ferulic acids), as expected, due to the lower pka of the carboxyl group, they proved to be excellent substrates for the reaction (yields 67-72%), taking into account that we used a protecting-free approach, which is very useful in multistep programs.
The removal of the isopropylidene group on C-2 and C-3 (Scheme 5) under acid conditions provided, except for the curcumin, the corresponding derivatives 12a-14a (yield 68-80%) that will be investigated for their reducing and scavenging abilities consistent with the mechanism generally accepted for antioxidant compounds.
Molecules 2021, 26, x FOR PEER REVIEW The removal of the isopropylidene group on C-2 and C-3 (Scheme 5) under ditions provided, except for the curcumin, the corresponding derivatives 12a-1 68-80%) that will be investigated for their reducing and scavenging abilities c with the mechanism generally accepted for antioxidant compounds.

Materials and Methods
All the reagents were acquired (Aldrich, Fluka, Sigma) at the highest purity and used without further purification. Thin-layer chromatography was perform silicagel plates Merck 60 F254, and the display of the products on TLC was accom with UV lamp lighting, molecular iodine, and in H2SO4-EtOH (95:5) with furthe until the development of color. The column chromatographies were carried out u 230 mesh silica gel (Merck). The NMR spectra (Supplementary Materials) were on Bruker DRX (400MHz) and Varian Inova Marker (500MHz) spectrometers solution unless otherwise specified. The chemical shifts are reported in ppm (δ and 13 C NMR full characterization of the products was obtained based on 2D N 1D selective NOESY. The TopSpin 4.0.6 software package was used for the analy Preparation of 2,3-O-isopropylidene-b-D-ribofuranose (2): To a magnetically stirred su of anhydrous triphenylphosphine (0.525 g, ca. 2.0 mmol) in anhydrous acetone at r.t., a solution of I2 (0.508 g, 2.0 mmol) in the same solvent (18.0 mL) was add wise in the dark and under dry N2 atmosphere. After 15 min, solid D-ribose (0.1 mmol) was added in one portion to the suspension. TLC monitoring (CHCl3-Me showed that the starting sugar was completely consumed within 30 min. The mixture was then filtered, washed with acetone, and dried (Na2SO4). The evapo the solvent under reduced pressure gave a crude residue that was used directly in step without further purification (0.181 g, 0.95 mmol, 95%). 1  The NMR spectral data were identical to values in the literature [19].

Materials and Methods
All the reagents were acquired (Aldrich, Fluka, Sigma) at the highest purity available and used without further purification. Thin-layer chromatography was performed with silicagel plates Merck 60 F 254 , and the display of the products on TLC was accomplished with UV lamp lighting, molecular iodine, and in H 2 SO 4 -EtOH (95:5) with further heating until the development of color. The column chromatographies were carried out using 70-230 mesh silica gel (Merck). The NMR spectra (Supplementary Materials) were recorded on Bruker DRX (400MHz) and Varian Inova Marker (500MHz) spectrometers in CDCl 3 solution unless otherwise specified. The chemical shifts are reported in ppm (δ). The 1 H and 13 C NMR full characterization of the products was obtained based on 2D NMR and 1D selective NOESY. The TopSpin 4.0.6 software package was used for the analysis.