A unique synthetic method to convert a D-sugar into 2-deoxy-L-ribitol through carbonyl translocation

Article abstract of DOI:10.1002/ejoc.201201489

A unique and highly efficient radical process was developed to prepare 2-deoxy-L-ribitol. The radical cyclization precursor was prepared over several steps from D-sugars. Radical cyclization followed by fragmentation gave exclusively the desired product, with carbonyl translocation in an acyclic carbohydrate system. The radical process was optimized to avoid the formation of byproducts. The reaction mechanism was studied by the concentration effect, which indicated that both the radical cyclization and the fragmentation steps are relatively rapid. The stereochemistry of the 2-deoxy-L-ribitol was confirmed by X-ray crystal structure. By using this synthetic approach, a D-sugar derivative could be transformed into 2-deoxy-L-ribose in a single radical reaction. A unique and highly efficient radical process (complete within minutes) was developed to prepare of 2-deoxy-L-ribitol. Radical cyclization followed by fragmentation gave exclusively the desired product, with carbonyl translocation in an acyclic carbohydrate system. By using this synthetic approach, a D-sugar derivative could be transformed into a 2-deoxy-L-sugar in a single radical reaction. Copyright

Full text of DOI:10.1002/ejoc.201201489

SHORT COMMUNICATION
DOI: 10.1002/ejoc.201201489
A Unique Synthetic Method to Convert a
D
-Sugar into 2-Deoxy- -ribitol
L
Through Carbonyl Translocation
Nai-Yun Hsu[a] and Che-Chien Chang*[a]
Keywords: Carbohydrates / Nucleosides / Radical reactions / Cyclization / Fragmentation
A unique and highly efficient radical process was developed
to prepare 2-deoxy- -ribitol. The radical cyclization precur-
sor was prepared over several steps from -sugars. Radical
cyclization followed by fragmentation gave exclusively the
desired product, with carbonyl translocation in an acyclic
carbohydrate system. The radical process was optimized to
avoid the formation of byproducts. The reaction mechanism
was studied by the concentration effect, which indicated that
both the radical cyclization and the fragmentation steps are
L
D
relatively rapid. The stereochemistry of the 2-deoxy-
was confirmed by X-ray crystal structure. By using this syn-
thetic approach, a -sugar derivative could be transformed
into 2-deoxy- -ribose in a single radical reaction.
L-ribitol
D
L
fragments to re-form the carbonyl upon delivering a radical
species[7] or gives a more stable radical through a rearrange-
ment,[8] thus providing methods for the construction of
carbocyclic systems. The straight reduction product may be
obtained if the cyclization reaction is not an efficient pro-
cess. With an oxy substituent positioned α to the carbonyl
group, radical cyclization provides a cycloalkoxy radical in-
termediate. This radical intermediate can undergo hydrogen
abstraction to form a cyclic alcohol or fragmentation to
give products with carbonyl translocation (Scheme 1).[9]
Introduction
2-Deoxy-l-sugars are not abundant in nature relative to
sugars in the d configuration. Among the 2-deoxy-l-sugars,
2-deoxy-l-ribose is an important carbohydrate because it
can be used as the sugar backbone in preparing l-DNA,
which can be used in studies involving nucleic acid recogni-
tion.[1] The synthesis of modified l-deoxynucleosides, such
as l-thymidine (l-T) and l-2Ј-fluoro-5-methyl arabinofur-
anosyl uracil (l-FMAU), from 2-deoxy-l-ribose has been
reported, as such compounds exhibit antiviral activities and
are less toxic that their corresponding d isomers.[2] Because
of the interest in such derivatives, a number of methods
have been reported for the synthesis of 2-deoxy-l-ribose.[3–5]
Although its synthesis from less abundant l-sugars has
been explored,[3] the use of inexpensive d-sugars[4] as start-
ing materials would be more practical. However, a synthesis
starting from inexpensive d-sugars would be complex and
would require multiple steps. It would be necessary to first
convert the sugar from the d configuration to the l configu-
ration, followed by selective deoxygenation of the C-2 hy-
droxy group to afford 2-deoxy-l-ribose. Hence, a practical
synthetic method for the synthesis of 2-deoxy-l-ribose from
inexpensive d-sugars would be highly desirable to organic
chemists.
Processes involving radical cyclization followed by frag-
mentation have been extensively studied for decades.[6] The
radical cyclizes onto the carbonyl group and subsequently
Scheme 1. Radical cyclization of α-oxy carbonyl groups.
[a] Department of Chemistry, Fu Jen Catholic University,
New Taipei City, Taiwan 24205, R.O.C.
Fax: +886-2-2902-3209
We report herein on a highly efficient radical process that
exclusively provides the product with carbonyl translocation
in an acyclic carbohydrate system. After carbonyl translo-
cation, the original stereogenic carbon of the d-sugar is no
longer used for the designation of the newly formed 2-de-
oxysugar; the product is now a 2-deoxy-l-sugar. This radi-
E-mail: 080686@mail.fju.edu.tw
Homepage: http://140.136.176.3/joom/data/eng/Faculty/
C-C%20Chang.htm
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.201201489.
658
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2013, 658–661

Conversion of a d-Sugar into 2-Deoxy-l-ribitol
cal process is so unique that a d-sugar derivative can be in five steps from known intermediate 1 and was ready for
converted into a 2-deoxy-l-sugar in a single radical cycliza- use in the radical cyclization/fragmentation process.
tion/fragmentation reaction (Scheme 2).
A series of standard radical conditions (AIBN/Bu3SnH)
were tested for converting d-sugar derivative 6 into 2-deoxy-
l-ribose 7 in a single radical reaction. The product ratios
as a function of concentration and reaction time are shown
in Table 1. The first experiment (0.1 m) involved the use of
1
a slow addition technique (Table 1, entry 1). The H NMR
spectrum of the crude reaction mixture indicated that de-
sired product 7 was formed alongside two other side prod-
ucts (i.e., 8 and 9) in a ratio of 1:0.19:1 (7/8/9). Unfortu-
nately, it was not possible to isolate product 7, as decompo-
sition occurred during column chromatography. Byproduct
8 was produced by an elimination reaction involving prod-
uct 7. Byproduct 9 was produced as a result of intramolecu-
lar hydrogen abstraction from the aldehyde C–H bond
(bond dissociation energy: 86 kcalmol–1),[12] followed by
the loss of a molecule of carbon monoxide (CO). Although
slow addition is an experimental technique used to avoid
the formation of the straight reduction product, we found
that a solution of azobisisobutyronitrile (AIBN) and
Bu3SnH, when added in one portion at the beginning of the
reaction (0.03 m), provided the best results (Table 1, en-
tries 2–5). The shorter the reaction time, the better was the
product ratio. The results showed that no straight reduction
products and no cyclic alcohols were produced. Only the
products with carbonyl translocation were observed. Pre-
vious studies showed that to mainly produce a product with
carbonyl translocation, the radical process needs to be in
conjunction with a rigid system[13] or assisted by the gemi-
nal dialkyl effect.[9c] At this point, our unique radical pro-
cess for the conversion of d-sugar derivative 6 into 2-deoxy-
l-ribose 7 was successful. Surprisingly, the reaction reached
completion within 7 min[14] (Table 1, entry 5), which indi-
cates that this is a highly efficient radical process.
Scheme 2. Synthetic plan.
Results and Discussion
The synthesis started with known intermediate 1, which
was readily prepared in three steps (overall yield of 68%)
from d-mannitol[10] (Scheme 3). Intermediate 1 was depro-
tected under acidic conditions (THF/2 m HCl) to provide
tetraol 2 in 76% yield. To differentiate the hydroxy groups
for further elaboration, a trityl (Tr) group was used as the
protecting group for the primary hydroxy group of 2. To
avoid possible 1,5-hydrogen transfer, all other secondary
alcohol groups were protected with benzoyl (Bz) groups.
Fully protected 3 was obtained in 99% yield after treating
tetraol 2 with trityl chloride (TrCl) followed by benzoyl
chloride (BzCl). Removal of the trityl group without mi-
gration of the benzoyl groups was successfully performed
in a cosolvent system containing p-toluenesulfonic acid
(pTsOH) to give compound 4 in 69% yield.[11] Direct bro-
mination with tetrabromomethane (CBr4) and tri-
phenylphosphane (PPh3) afforded bromide 5 in 89% yield.
Subsequent ozonolysis provided desired cyclization precur-
sor 6 in 99% yield. This cyclization precursor was prepared
Table 1. Radical cyclization/fragmentation reactions of precursor 6.
Entry Concentration Slow addition Reaction time
7/8/9
[m]
[min]
[min]
ratio[a]
1
2
3
4
5
0.1
120
0
0
0
0
60
120
60
30
7
1:0.19:1[b]
1:0.36:0
1:0.20:0
1:0.13:0
1:0.02:0
0.03
0.03
0.03
0.03
Scheme 3. Reagents and conditions: (a) Ref.[10]; (b) THF/2 m HCl
(5:1), r.t., 10 h, 76%; (c) TrCl, pyridine, 90 °C, 2 h, then BzCl, r.t.,
10 h, 99%; (d) pTsOH, MeOH/THF (1:1), r.t., 14 h, 69%; (e) CBr4,
PPh3, CH2Cl2, r.t., 1 h, 89%; (f) O3, CH2Cl2, –78 °C, then Me2S,
–78 °C to r.t., 99%.
[a] Product ratio was determined by integration of the signals in the
1H NMR spectrum of the crude reaction mixture. [b] Compound 7
could not be isolated; compounds 8 and 9 were isolated in 15 and
47% yield, respectively.
Eur. J. Org. Chem. 2013, 658–661
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
659

N.-Y. Hsu, C.-C. Chang
SHORT COMMUNICATION
A mechanism for this unique and highly efficient radical fore, hydrogen abstraction from the aldehyde C–H bond (12
process is proposed in Scheme 4. A tributyltin radical, gen- Ǟ 14) partially occurs to give byproduct 9 upon delivering
erated from AIBN and Bu3SnH, reacts with precursor 6 a CO molecule. However, when Bu3SnH is added in one
to give alkyl radical 10. Cyclization of radical 10 affords portion at the beginning of the reaction (concentrated
cycloalkoxy radical 11. In our system, β-fragmentation oc- Bu3SnH conditions; Table 1, entries 2–5), hydrogen abstrac-
curs exclusively to give α-oxy radical 12 rather than hydro- tion from Bu3SnH (12 Ǟ 7) is more efficient, and desired
gen abstraction to give cyclic alcohols. Subsequently, α-oxy product 7 is produced exclusively. This concentration effect
radical 12 abstracts a hydrogen atom from Bu3SnH to give indicates that byproduct 9 is obtained from radical 12
desired benzoyl-protected 2-deoxy-l-ribose 7, with regener- through the second fragmentation of radical 14 and not
ation of the tributyltin radical. When a longer reaction time from radical 10. The results also indicate that the rates of
is used, an elimination reaction occurs and conjugated com- both cyclization (10 Ǟ 11) and fragmentation (11 Ǟ 12)
pound 8 is produced. The reason for the formation of 8 are relatively rapid, as no straight product and no cyclic
could be attributed to a retro-Michael reaction of product alcohols are observed.
7, caused by the Lewis acidity of tributyltin bromide
Although the direct deprotection of product 7 to give 2-
(Bu3SnBr).[15] An alternative method to reduce the forma- deoxy-l-ribose by using NaOMe or K2CO3 in methanol
tion of compound 8 involving the use of tristrimethylsilylsil- was not successful, the presence of an aldehyde group and
ane (TTMSS)[16] was not successful. In radical cyclization the benzoyl protecting groups of 7 would permit the prepa-
reactions of carbonyl groups, hydrogen abstraction from the ration of other derivatives, such as dithiolacetal sugars[19]
aldehyde C–H bond is a well-known process.[17] In our sys- and l-deoxynucleosides. To verify that desired product 7
tem, both radical intermediates 10 and 12 could also intra- was indeed 2-deoxy-l-ribose, direct reduction of the crude
molecularly abstract a hydrogen atom from the aldehyde residue with sodium triacetoxyborohydride [NaB(OAc)3H]
C–H bond to from acyl radicals 13 and 14, respectively. was performed, which gave corresponding alcohol 15 in
Byproduct 9 is then obtained by the loss of a molecule of
CO from the acyl radicals.[18]
Scheme 5. Reagents and conditions: (a) Bu3SnH, AIBN, benzene,
reflux, 7 min; (b) NaB(OAc)3H, HOAc, THF, 0 °C, 1 h, 70%, over
two steps from 6.
Scheme 4. Proposed mechanism.
In the case of slow addition (dilute Bu3SnH conditions;
Table 1, entry 1), the formation of byproduct 9 could be
explained by the abstraction of a hydrogen atom from
Bu3SnH (12 Ǟ 7), which is not an efficient process. There- Figure 1. X-ray crystal structure of compound 15.
660
www.eurjoc.org
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2013, 658–661

Conversion of a d-Sugar into 2-Deoxy-l-ribitol
camone, M. Capobinaco, A. Carcuro, F. Colonna, S. Lotti, A.
Garbesi, J. Med. Chem. 1992, 35, 4214–4220.
70% (over two steps from 6, Scheme 5). The stereochemis-
try of 15 was confirmed by X-ray crystal structure analysis.
The obtained crystal structure indicated that product 7 was
indeed an acyclic form of 2-deoxy-l-ribose with a (3R,4S)
stereochemistry (Figure 1).
[3] From l-sugars: a) M. L. Sznaidman, M. R. Almond, P. Amir,
Nucleosides Nucleotides 2002, 21, 155–163; b) M. E. Jung, Y.
Xu, Org. Lett. 1999, 1, 1517–1519; c) W. Zhang, K. S. Ramas-
amy, D. R. Averett, Nucleosides Nucleotides 1999, 18, 2357–
2365; d) E. Moyroud, P. Strazewski, Tetrahedron 1999, 55,
1277–1284; e) J. Du, Y. Choi, K. Lee, B. K. Chun, J. H. Hong,
C. K. Chu, Nucleosides Nucleotides 1999, 18, 187–195.
[4] For syntheses from d-sugars, see: a) Q. Ji, M. Pang, J. Han, S.
Feng, X. Zhang, Y. Ma, J. Meng, Synlett 2006, 15, 2498–2500;
b) M. E. Jung, Y. Xu, Tetrahedron Lett. 1997, 38, 4199–4202;
c) M. E. Jung, C. J. Nichols, O. Kretscik, Y. Xu, Nucleosides
Nucleotides 1999, 18, 541–546.
[5] For syntheses from other non-sugar materials, see: a) B. H.
Cho, J. H. Kim, H. B. Jeon, K. S. Kim, Tetrahedron 2005, 61,
4341–4346; b) M. E. Jung, C. J. Nichols, Tetrahedron Lett.
1998, 39, 4615–4618; c) F. Fazio, M. P. Schneider, Tetrahedron:
Asymmetry 2000, 11, 1869–1876; d) F. Fazio, M. P. Schneider,
Tetrahedron: Asymmetry 2001, 12, 2143–2145.
[6] a) D. P. Curran, Synthesis 1988, 417–439; b) C. P. Jasperse,
D. P. Curran, T. L. Fevig, Chem. Rev. 1991, 91, 1237–1286; c)
M. Ramaiah, Tetrahedron 1987, 43, 3541–3676.
[7] a) D. P. Curran, H. Liu, J. Org. Chem. 1991, 56, 3463–3465; b)
D. P. Curran, U. Diederichsen, M. Palovich, J. Am. Chem. Soc.
1997, 119, 4797–4804; c) S. Kim, S. Y. Jon, Chem. Commun.
1996, 1355–1336; d) S. Kim, S. Y. Jon, Chem. Commun. 1998,
815–816.
[8] a) Y.-M. Tsai, C. D. Cherng, Tetrahedron Lett. 1991, 32, 3515–
3518; b) Y.-M. Tsai, H.-C. Hieh, J.-S. Pan, D.-D. Hsiao, Chem.
Commun. 1996, 2469–2470; c) C.-H. Huang, S.-Y. Chang, N.-
S. Wang, Y.-M. Tsai, J. Org. Chem. 2001, 66, 8983–8991.
[9] a) R. Tsang, B. Fraser-Reid, J. Am. Chem. Soc. 1986, 108,
8102–8104; b) B.-W. A. Yeung, R. Alonso, G. D. Vite, B. Fra-
ser-Reid, J. Carbohydr. Chem. 1989, 8, 413–427; c) M. A. Ciu-
folini, S. Zhu, J. Org. Chem. 1998, 63, 1668–1675.
[10] V. Yadav, D. Argawal, Chem. Commun. 2007, 5232–5234.
[11] C.-C. Chang, Y.-H. Kuo, Y.-M. Tsai, Tetrahedron Lett. 2009,
50, 3805–3808.
Conclusions
In summary, a unique and highly efficient radical process
(complete within minutes) was developed to prepare 2-de-
oxy-l-ribitol. Radical cyclization followed by fragmentation
of an acyclic carbohydrate system gave exclusively the prod-
uct with carbonyl translocation. This is one of the few re-
ports in which radical cyclization of an α-oxy carbonyl
compound involves a 1,4-CHO migration without any as-
sistance, such as the need for a rigid system or acceleration
assisted by the geminal dialkyl effect. Moreover, this radical
process could be applied to the field of carbohydrate chem-
istry to transform a d-sugar derivative into a 2-deoxy-l-
sugar in a single radical reaction. Application of this unique
radical process to give other rare l-sugars from inexpensive
d-sugars is currently underway.
Supporting Information (see footnote on the first page of this arti-
cle): Experimental procedures; characterization data; 1H–13C/
DEPT NMR spectra of 2–6, 8–9, and 15; 1H NMR spectrum of 7;
X-ray data of 15.
CCDC-895176 (for 15) contains the supplementary crystallo-
graphic 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.
[12] J. A. Kerr, in: CRC Handbook of Chemistry and Physics, CRC,
Boca Raton, FL, 1986, p. F-178.
[13] M. E. Jung, S. W. T. Choe, Tetrahedron Lett. 1993, 34, 6247–
6250.
Acknowledgments
[14] When a reaction time of 5 min was used, the ratio of product/
starting material (7/6) was observed to be 1:0.15 (an estimated
13% of the starting material was not consumed).
[15] C. H. Yoder, J. C. Otter, A. Grushow, T. F. Ganunis, B. G. En-
der, A. I. Zafar, J. N. Spencer, J. Organomet. Chem. 1990, 385,
33–37.
The authors thank the National Science Council (NSC), Taiwan,
and the Department of Chemistry, Fu Jen Catholic University for
financial support. N.-Y. H. held a Department of Chemistry Mas-
ter Student Fellowship.
[16] C. Chatgilialoglu, Chem. Eur. J. 2008, 14, 2310–2320.
[17] a) A. L. J. Beckwith, B. P. Hay, J. Am. Chem. Soc. 1989, 111,
2674–2681; b) A. L. J. Beckwith, K. D. Raner, J. Org. Chem.
1992, 57, 4954–4962.
[18] C. Chatgilialoglu, D. Crich, M. Komatsu, I. Ryu, Chem. Rev.
1999, 99, 1991–2070.
[19] A. Tauss, A. J. Steiner, A. E. Stütz, C. A. Tarling, S. G. Withers,
T. M. Wrodnigg, Tetrahedron: Asymmetry 2006, 17, 234–239.
Received: November 7, 2012
[1] D. D’Alonzo, A. Guaragna, G. Palumbo, Chem. Biodiversity
2011, 8, 373–413.
[2] a) Z.-D. Shi, B.-H. Yang, Y.-L. Wu, Tetrahedron 2002, 58,
3287–3296; b) S. B. Pai, S.-H. Liu, Y.-L. Zhu, C. K. Chu, Y.-
C. Cheng, Antimicrob. Agents Chemother. 1996, 40, 380–386; c)
C. K. Chu, T. W. Ma, K. Shanmuganathan, C. G. Wang, Y. J.
Xiang, S. B. Pai, G.-Q. Yao, J.-P. Sommaclossi, Y.-C. Cheng,
Antimicrob. Agents Chemother. 1995, 39, 979–981; d) S. Spad-
ari, G. Maga, F. Focher, G. Ciarroocchi, R. Manservigi, F. Ar-
Published Online: December 20, 2012
Eur. J. Org. Chem. 2013, 658–661
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
661