A heterogeneous Pd-Bi/C catalyst in the synthesis of l-lyxose and l-ribose from naturally occurring d-sugars

Article abstract of DOI:10.1039/c1ob06116j

A critical step in the synthesis of the rare sugars, l-lyxose and l-ribose, from the corresponding d-sugars is the oxidation to the lactone. Instead of conventional oxidizing agents like bromine or pyridinium dichromate, it was found that a heterogeneous catalyst, Pd-Bi/C, could be used for the direct oxidation with molecular oxygen. The composition of the catalyst was optimized and the best results were obtained with 5:1 atomic ratio of Pd:Bi. The overall yields of the five-step procedure to l-ribose and l-lyxose were 47% and 50%, respectively. The synthetic procedure is advantageous from the viewpoint of overall yield, reduced number of steps, and mild reaction conditions. Furthermore, the heterogeneous oxidation catalyst can be easily separated from the reaction mixture and reused with no loss of activity.

Full text of DOI:10.1039/c1ob06116j

View Article Online / Journal Homepage / Table of Contents for this issue
Organic &
Biomolecular
Chemistry
Dynamic Article Links
Cite this: Org. Biomol. Chem., 2011, 9, 7720
www.rsc.org/obc
PAPER
A heterogeneous Pd–Bi/C catalyst in the synthesis of L-lyxose and L-ribose
from naturally occurring D-sugars†
Ao Fan,* Stephan Jaenicke* and Gaik-Khuan Chuah*
Received 8th July 2011, Accepted 15th August 2011
DOI: 10.1039/c1ob06116j
A critical step in the synthesis of the rare sugars, L-lyxose and L-ribose, from the corresponding D-sugars
is the oxidation to the lactone. Instead of conventional oxidizing agents like bromine or pyridinium
dichromate, it was found that a heterogeneous catalyst, Pd–Bi/C, could be used for the direct oxidation
with molecular oxygen. The composition of the catalyst was optimized and the best results were
obtained with 5 : 1 atomic ratio of Pd : Bi. The overall yields of the five-step procedure to L-ribose and
L-lyxose were 47% and 50%, respectively. The synthetic procedure is advantageous from the viewpoint
of overall yield, reduced number of steps, and mild reaction conditions. Furthermore, the heterogeneous
oxidation catalyst can be easily separated from the reaction mixture and reused with no loss of activity.
Introduction
Carbohydrates represent the most abundant renewable source for
chemical materials.1 However, their use in fine chemistry is limited
because of the difficulties caused by their high functionalization. In
medicinal chemistry, the epimerization of the naturally occurring
D-form to the corresponding L-form (not naturally occurring) is
Fig. 1 Structure of D-ribose, L-ribose and L-lyxose.
extremely important. Nucleoside analogues containing L-ribose
have been developed as antiviral and antitumour agents.2 These
L-ribose-containing nucleosides frequently show less toxicity, high
metabolic stability and better pharmacological activity than the
D-ribose-containing analogues. However, because L-ribose is not
found in nature, it must be synthesized by elaborate and laborious
chemical or biochemical synthetic methods.3
Similarly, L-lyxose which differs from D-ribose only in the
absolute configuration of C4 (Fig. 1) has interesting biochemical
properties. Reist et al.4 synthesized the L-lyxose derivative, methyl
b-L-lyxopyranose, by inversion of the 3-hydroxyl group of an L-
arabinose derivative. However, the yield for this method was low
(<13%). Kuzuhara et al.5 improved the synthesis method of L-
lyxose by starting from a cheaper chemical, D-glucose, but the
overall yield was only 30%. Brimacombe et al.6 used 1,2:5,6-
di-O-isopropylidene-a-D-gulofuranose as the starting material to
produce L-lyxose. However, this synthetic route was carried out
on a 0.1–0.3 g scale, which does not fulfil the requirements
of industrial production. More recently, access to L-lyxose by
biotransformations using microbial and enzymatic reactions has
been reported.7 Although this synthetic route can give a relatively
highyield(50%)of L-lyxose from ribitol, the following downstream
operation of separating L-lyxose from the mixture of L-ribulose, L-
xylulose and L-lyxose was a difficult and time-consuming process.
Granstro¨m et al.8 used xylitol as the raw material and reduced
the number of steps in the biotransformation. However, the final
equilibrium distribution between L-xylulose, L-xylose and L-lyxose
was 53 : 26 : 21, reducing the overall yield of L-lyxose.
Although current chemical and biochemical processes9 can fulfil
the task of synthesizing the rare L-sugars, they have disadvantages
such as low yields, use of corrosive, toxic, flammable and/or
pyrophoric chemicals such as bromine, pyridinium dichromate,
concentrated acids, lithium aluminum hydride, difficulties with the
separation of the products, control of by-products and disposal of
wastewater. In this regard, heterogeneous catalysts have unique
advantages, e.g., ease of handling, separation from the reaction
mixtures and recovery of the catalysts.10 Moreover, heterogeneous
catalysts often perform well under mild conditions and are
seldom corrosive in nature. However, direct epimerization of
carbohydrates with heterogeneous catalysts is largely unexplored.
Here, we report an efficient and simple synthetic route to
synthesize L-lyxose from the cheap starting material, D-ribose,
through the use of a heterogeneous catalyst and readily available
reagents (Scheme 1). The proposed synthetic route from D-ribose
to L-lyxose consists of five reactions: (i) selective oxidation of
D-ribose, (ii) protection of hydroxyl groups (C2 and C3), (iii)
epimerization reaction, (iv) reduction of the carbonyl group in
Department of Chemistry, National University of Singapore, 3 Science Drive
3, Singapore, 117543. E-mail: chmsj@nus.edu.sg, chmcgk@nus.edu.sg;
Fax: (+65) 6779 1691; Tel: (+65) 6516 2918
† Electronic supplementary information (ESI) available: Experimental and
NMR spectra. See DOI: 10.1039/c1ob06116j
7720 | Org. Biomol. Chem., 2011, 9, 7720–7726
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
Scheme 1 Synthesis route from D-ribose to L-lyxose.
the anomeric carbon to a hydroxyl group, and (v) deprotection
of hydroxyl groups (C2 and C3). The key and also the most
difficult step in the above reactions is the first step where D-ribose is
oxidized to D-ribonolactone. D-Ribonolactone is a very important
intermediate in the synthesis of various organic products, such as
riboflavin (vitamin B2) and naturally occurring lactones.11–13 The
present procedure for producing D-ribonolactone is based on the
oxidation of ribose with bromine.11 Although this method can give
a relatively high yield (80%) within an acceptable time, bromine is
corrosive, volatile and toxic, hence preventing this route from being
a safe and efficient way for the production of D-ribonolactone on
an industrial scale.
surface by acting as a co-catalyst in the oxidative dehydrogenation
mechanism.18 In view of its good oxidation properties in the
oxidation of D-glucose, we decided to investigate its use for the
oxidation of D-ribose to D-ribonolactone in the first step of the
route to the L-sugar. The applicability of the scheme to the
synthesis of L-ribose starting from D-lyxose was also attempted
(Scheme 2).
Results and discussion
Catalyst characterization
Palladium as well as platinum dispersed on active carbon has
been used as a catalyst for the oxidation of D-glucose to D-
gluconate.14 However, the catalyst becomes deactivated in the
course of the reaction resulting in lower activity and selectivity for
D-gluconate.15 The addition of bismuth to the catalyst improves
the activity and selectivity for the oxidation of glucose.16,17 The
bismuth adatoms prevent oxygen poisoning of the palladium
Samples with different bismuth loadings were prepared from a 10
wt% Pd/C catalyst. The atomic ratio of Pd : Bi was varied from 1
to 8 (ESI†). The nitrogen adsorption-desorption isotherms of the
10 wt% Pd/C and the Pd–Bi/C catalysts are given in Fig. 2a. The
steep increase at low P/Po indicates the presence of micropores
while the hysteresis above P/Po of 0.5 shows that mesopores are
also present. The mesopores are between 3.2 to 4.7 nm, irrespective
Scheme 2 Synthesis route from D-lyxose to L-ribose.
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 7720–7726 | 7721
©

View Article Online
Fig. 3 XRD spectra of Pd–Bi/C catalysts.
the peaks in 8Pd-Bi/C and 5Pd-Bi/C are narrower than Pd/C but
broadened again with higher bismuth loadings. Furthermore, the
peak maximum of the (111) reflex in 8Pd–Bi/C and 5Pd–C/Bi
was at 2q of 38.60◦ but shifted to higher angles, 38.63–38.87◦, as
the bismuth loading increased. The shift to higher 2q indicates
that less hydrogen was present in the higher bismuth-containing
samples than 8Pd–Bi/C and 5Pd–Bi/C.
X-ray photoelectron spectroscopy (XPS) measurements con-
firmed that the surface palladium was mostly in the metallic form
with a binding energy of Pd 3d5/2 centred at ~335.2 eV (Fig. 4).
Deconvolution of the peaks showed that Pd(II) was also present.
The fraction of Pd(II) increased with bismuth loading (Table
2). Bismuth was present as Bi(III) and Bi(0), with Bi(III) being
the predominant state for all samples except 5Pd : Bi/C. In this
sample, more bismuth was present in oxidation state 0 than +3.
With higher bismuth loading, the ratio of Bi(III)/Bi(0) increased.
The surface Pd/Bi ratio agrees rather well with the expected ratio
except for 8Pd–Bi/C where the relative surface concentration of
Pd/Bi is higher than the expected 8 : 1. This may be due to the
self-aggregation of bismuth oxide particles rather than spreading
out evenly on the surface at the very low loading.
Fig. 2 (a) Adsorption/desorption isotherms and (b) pore size distribution
of Pd–Bi/C.
of the bismuth loading (Fig. 2b). The Pd/C sample has a high
surface area of 819 m2 g-1 with ~300 m2 g-1 due to micropores. With
the incorporation of bismuth, the surface area decreased (Table 1).
Concomitant with this was a decrease in both the micropore and
mesopore volume, showing that bismuth was deposited into these
pores.
Synthesis of L-lyxose from D-ribose
Oxidation of D-ribose
Powder X-ray diffraction studies showed that metallic palla-
dium (JCPDS 00-051-0681) was present on Pd/C after hydrogen
treatment at 150 ◦C for 2 h. However, for the reduced Pd-Bi
samples, the characteristic peaks of Pd were shifted to smaller
angles (Fig. 3). The peak positions are characteristic of palladium
hydride, PdH0.706 (JCPDS 00-018-0951). To accommodate the
hydrogen, an increase in lattice spacing occurred. The width of
In the first step as outlined in Scheme 1, the oxidation of D-ribose
to D-ribonolactone was investigated using the heterogeneous
5Pd : Bi/C catalyst and oxygen gas as the oxidant. Without any
pH control, the reaction started with a high initial rate but then
levelled off to reach a conversion of 40% after 180 min (Fig. 5). The
pH decreased in the course of the reaction, starting at ~4.4–4.6 and
falling to ~2.8–3.1 at the end of the reaction. The decrease in pH
can be attributed to the formation of ribonic acid. Despite a longer
Table 1 Surface area and pore volume of Pd–Bi/C catalysts
Table 2 XPS results of Pd–Bi/C catalysts
Surf. area
(m2 g-1)
Micropore area
(m2 g-1)
Pore vol.
(cm3 g-1)
Micropore vol.
(cm3 g-1)
Catalyst
Theoretical
Pd : Bi
Pd(0)/
Pd(II)
Bi(III)/
Bi(0)
Experimental
Pd : Bi
Pd
819
708
625
599
575
570
300
260
268
234
231
221
0.81
0.72
0.61
0.60
0.58
0.56
0.16
0.14
0.14
0.13
0.12
0.12
8Pd : Bi
5Pd : Bi
4Pd : Bi
3Pd : Bi
Pd : Bi
8 : 1
5 : 1
4 : 1
3 : 1
1 : 1
40.4
30.0
11.9
7.01
5.93
2.79
0.66
2.32
3.11
4.44
13.64
5.17
4.31
2.21
1.03
7722 | Org. Biomol. Chem., 2011, 9, 7720–7726
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
Table 3 Conversion of D-ribose and selectivity to D-ribonate 2 over Pd-
Bi/C catalysts
Entry
Catalyst
Time (min)
Conv. (%)
Selectivity (%)
1
2
3
4
5
6
Pd
40
80
110
130
130
130
95
94
100
96
85
82
83
8Pd : Bi
5Pd : Bi
4Pd : Bi
3Pd : Bi
Pd : Bi
>95
>95
>95
>95
>95
Reactions condition: D-ribose (0.1 M, 70 cm3), catalyst (50 mg), 50 ◦C,
0.5 L O2 min-1, pH 9
conditions, the gluconate forms instead and could desorb from
the catalyst, thus removing the inhibition.
Maintaining the pH at 5 led to a similar reaction profile as that
without pH control. A limiting conversion of ~48% was reached
after 120 min. However, keeping a constant pH 7 during the
reaction led to a higher rate of reaction and 74% conversion was
obtained after 180 min. A further increase in the reaction pH to
9 resulted in an even higher rate of reaction and 100% conversion
was reached after 110 min. Under the alkaline conditions, the salt,
potassium D-ribonate was formed instead of ribonic acid and the
absence of inhibition clearly shows that ribonic acid poisoned the
catalyst. Hence, there is a very strong influence of pH on the rate
of reaction.
Different Pd : Bi/C catalysts were investigated for the oxidation
of D-ribose keeping the reaction mixture at pH 9 and 50 ◦C.
The rate of reaction was highest when 10% Pd/C was used as
the catalyst. After 40 min, the conversion was 95% (Table 3).
However D-ribose was oxidized to D-ribonate with only 83%
selectivity. By-products of glutarate and oxalate were formed
due to overoxidation and degradation of D-ribose, respectively.
In contrast, when Pd–Bi/C samples were used, the reaction was
slower but the selectivity to D-ribonate improved to >95%. Of
the Pd–Bi catalysts, 5Pd : Bi/C showed the highest activity. Its
initial rate was similar to that of Pd/C but decreased with time
to reach full conversion after 110 min. The main product was
D-ribonate. The promoting effect of bismuth on platinum and
palladium catalysts for carbohydrates has been attributed to a
number of factors. These include a decrease in the size of the
platinum ensembles due to bismuth adatoms, in turn reducing
the irreversible adsorption of the substrate and formation of by-
products as well as a higher affinity of bismuth for oxygen, thus
preventing a high oxygen coverage on the metal.18 Such over-
oxidation on the metal surface can lead to fully oxidized products
and oxygen poisoning of the reaction.
Fig. 4 XPS spectra of Pd–Bi/C catalysts for (a) Pd 3d (b) Bi 4f.
Besson et al.18 proposed that glucose oxidation on Pd–Bi/C
catalyst proceeds via an oxidation dehydrogenation mechanism
where bismuth acts as a cocatalyst to prevent overoxidation of the
palladium surface. From the XPS results, 5Pd : Bi/C has a high
proportion of metallic Pd and Bi at its surface. Chemisorption of
oxygen can occur at the surface of Bi(0). The dehydrogenation of D-
ribose is facilitated by Pd(0), forming PdH. The ease at which PdH
is formed is seen from the XRD spectra of the Pd–Bi/C catalysts as
compared to purely Pd/C. The more Pd(0) sites are available at the
surface of the catalyst, the faster the dehydrogenation reaction can
proceed. Reaction of PdH with chemisorbed oxygen at Bi(0) forms
water and regenerates Pd(0), thus completing the catalytic cycle.
Fig. 5 Oxidation of D-ribose to D-ribonate at different pH. Reaction
conditions: D-ribose (0.1 M, 70 cm3), 5Pd-Bi/C (50 mg), 0.5 L O2 min-1,
50 ◦C.
reaction time, the conversion did not increase, indicating that the
catalyst had been poisoned probably by the ribonic acid. Abbadi
and van Bekkum16 showed that for the oxidation of glucose over
a palladium catalyst, the poisoning was due to the adsorption of
the product, 2-keto-gluconic acid, on the catalyst. Under basic
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 7720–7726 | 7723
©

View Article Online
A high bismuth loading covers the palladium sites, decreasing the
activity of the Pd–Bi catalysts without affecting the selectivity.
Conversely, when the bismuth loading is too low, the palladium
surface can be subjected to oxygen poisoning. Hence, an optimum
balance of available Pd and Bi sites determines the reaction rate.
The temperature also played a significant role in the oxidation
reaction (Fig. 6). At 27 ◦C, the reaction was slow and after
110 min, only 43% conversion was obtained. Increasing the
temperature to 44 ◦C resulted in a higher rate, allowing close
to 97% conversion in the same time. The selectivity to D-ribonate
(>95%) was unchanged despite the different temperatures. The
activation energy was calculated from the Arrhenius equation to
be 71.3 kJ mol-1.
The used catalyst was tested in 5 batch reactions without loss of
activity and selectivity (Fig. 8). Hence, the results showed that Pd–
Bi/C is a reuseable and selective catalyst for D-ribose oxidation.
Fig. 8 Conversion and selectivity of D-ribose oxidation with regenerated
5Pd-Bi/C catalyst.
Transformation of D-ribonate to L-lyxose
The synthesis conditions for the conversion of D-ribonate to 2,3-
O-isopropylidene-D-ribonolactone 3 was investigated (Scheme 3).
First, the potassium D-ribonate 2 can be converted to D-ribonic
acid 2a by exchanging the potassium ions with hydrogen ions.
Due to the dynamic equilibrium between D-ribonic acid and D-
ribonolactone 2b, reaction of the latter with acetone will result in
the desired 2,3-acetonide 3. Thus, a one-pot transformation from
D-ribonate to 3 can be achieved if a suitable acid and reaction
conditions can be found.
Fig. 6 Oxidation of D-ribose to D-ribonate at ( ) 27 (ꢀ) 44 and (᭜)
᭹
50 ◦C. Reaction conditions: D-ribose (0.1 M, 70 cm3), 5Pd-Bi/C (50 mg),
0.5 L O2 min-1, pH 9.
The used catalyst was recovered by filtration of the reaction mix-
ture, regenerated and used in new batch reactions. Washing of the
used catalyst followed by drying led to a reduction of activity (Fig.
7). The catalyst became more active after reducing in H2 at 150 ◦C
for 2 h so that full conversion could be obtained after 400 min.
The best activity was obtained after washing the used catalyst with
◦
KOH and acetone followed by hydrogen reduction at 150 C for
2 h. The activity of the regenerated catalyst was very similar to the
fresh catalyst, showing that no leaching of Pd or Bi had occurred.
Scheme
3
One-pot transformation of D-ribonate to 2,3-O-isopro-
pylidene-D-ribonolactone.
Concentrated HCl was tested as the acid (Table 4). Based on
2.0 g of crude D-ribonate 2, 66% yield of 2,3 acetonide 3 was
obtained by stirring at 25 ◦C for 18 h. The addition of anhydrous
Fig. 7 Activity of (᭜) fresh 5Pd : Bi/C and the used catalyst after (ꢀ)
washing with water and drying, (᭡) washing with water and reducing in
H2, and ( ) washing with KOH, acetone and H2 reduction.
᭺
7724 | Org. Biomol. Chem., 2011, 9, 7720–7726
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
Table 4 One-pot transformation from D-ribonate 2 to 2,3-acetonide 3
Entry
Reagents
Acetone (cm3)
Condition
Yield 3 (%)
1
2
3
4
5
6
7
8
conc. HCl (20 mmol)
15
40
40
40
40
40
40
40
Reflux at 60 ◦C for 4 h, stir at RT for 8 h
Reflux at 60 ◦C for 4 h, stir at RT for 8 h
Stir at RT for 18 h
48
64
66
72
84a
8
conc. HCl (60 mmol)
conc. HCl (60 mmol)
conc. HCl (24 mmol), CuSO4 (12.5 mol)
conc. HCl (24 mmol), CuSO4 (12.5 mol)
conc. H2SO4 (30 mmol)
Reflux at 60 ◦C for 1.5 h
Reflux at 60 ◦C for 4 h
Reflux for 4 h and stir at RT for 8 h
Stir at RT for 18 h
AmberliteIR-120H (5 g)
AmberliteIR-120H (5 g), 2,2
dimethoxypropane (48.8 mmol)
0
Stir at RT for 18 h
30b
Based on 2.0 g of crude potassium D-ribonate. Molar ratio of 2,3-acetonide 3 to3,5-acetonide 3a.a 3 : 1. b 1 : 2.
CuSO4 reduced the reaction time and increased the yield to 72%
when the reaction mixture was refluxed at 60 ◦C for 1.5 h (entry 4).
This was the highest yield of 3 obtained as extending the refluxing
time to 4 h led to the formation of the isomeric 3,5-acetonide 3a
(entry 5). The molar ratio of 2,3-acetonide 3 and 3,5-acetonide
3a was about 3 : 1. The use of conc. H2SO4 resulted in charring
and only 8% yield of 3 was obtained (entry 6). Use of the acidic
anion exchange resin Amberlite IR-120H led to the formation
of more 3,5-acetonide 3a than the desired 2,3-acetonide 3 (molar
ratio 2 : 1).
In the next step, 2,3-acetonide 3 was reacted in pyridine with
methanesulfonyl chloride (MsCl) followed by treatment with
potassium hydroxide.19 2,3-O-Isopropylidene-L-lyxonolactone 4
formed under inversion of the reacting carbon center in 80%
yield. The mechanism of the inversion at C4 has been proposed by
Batra et al.19 to involve an intermediate C4–C5 epoxide which
forms when the D-ribonlactone ring is opened by potassium
hydroxide. Subsequent intramolecular nucleophilic substitution
by the carboxylate ion in a 5-endo-tet process leads to ring opening
of the epoxide and inversion of the configuration at C4. The
carbonyl group on the anomeric carbon of 4 was reduced with
NaBH4 in methanol to give the product 5 as a colorless syrup (95%
yield). Finally, compound 5 was hydrolyzed using acidic Amberlite
IR-120H resin to afford L-lyxose 6 in 95% yield. Hence, L-lyxose
was synthesized from D-ribose in five steps with an overall yield
of 50%. This compares very favorably with yields of 13 to 30% for
the chemical synthesis of L-lyxose starting from L-arabinose4 and
D-glucose and its derivative.5,6 The yield is also comparable to that
obtained from microbial and enzymatic methods.7,8
D-lyxose 7 to D-lyxonate 8 was carried out successfully with the
heterogeneous catalyst, 5Pd : Bi/C, at pH 9. Although the reaction
was slightly slower than for D-ribose, the conversion was >99%
after 180 min. The selectivity to D-ribonate was >92%.
However, different reaction conditions were required to trans-
form D-lyxonate 8 to 2,3-O-isopropylidene-D-lyxonolactone 9
(Table 5). The yield of 9 was <5% even with the use of sufficient
conc. HCl, acetone and CuSO4 in the reaction mixture (entries 1 &
2). The addition of the organic acid, methanesulfonic acid followed
by stirring for 18 h could significantly improve the reaction yield
of 9 to 30% (entry 3). Under reflux conditions, the isomer 3,5-
acetonide 9a was formed (entries 4 & 5). Although the yield of each
◦
batch reaction after stirring at 25 C for 18 h was relatively low
(30%), the significant difference in solubility of D-lyxonolactone
and 2,3-acetonide 9 allows them to be easily separated with liquid–
liquid (ethyl acetate/water) extraction. In turn, it was possible to
recover the unreacted D-lyxonolactone and carry out further batch
reactions to improve the yield. About 70% yield was achieved after
four cycles of the reaction.
Besides this route, a modified synthesis method was used to
produce 2,3-acetonide 9 from D-lyxonate 8 (Scheme 4). Treat-
ment of D-lyxonate 8 with conc. HCl, dimethoxypropane and
MsOH gave D-lyxonolactone 8a and the diacetonide compound
8b which could be easily separated by liquid–liquid extraction
(diethyl ether/H2O). D-Lyxonolactone 8a was reacted with more
dimethoxypropane and MsOH to form 2,3-O-isopropylidene-D-
lyxonolactone 9. For the diacetonide 8b, it was found that acetic
acid/H2O (9 : 1) could selectively deprotect the hydroxyl groups in
C4 and C5.20 Hence, 2,3-O-isopropylidene D-lyxonolactone 9 was
formed with 70% yield with this strategy.
Synthesis of L-ribose from D-lyxose
Compound 9 was next reacted with methanesulfonyl chloride
(MsCl) in pyridine solution followed by potassium hydroxide to
produce the L-ribonolactone 2,3-acetonide 10 (82% yield) with
The generality of the designed route was tested for the synthesis
of L-ribose from D-lyxose (Scheme 2). The aerobic oxidation of
Table 5 One-pot transformation of D-lyxonate 8 to 2,3-acetonide 9
Run
Reagents
Acetone (cm3)
Condition
Yield 9 (%)
1
2
3
4
5
6
conc. HCl (60 mmol)
40
40
40
40
40
40
Reflux for 4 h, stir at RT for 8 h
Reflux for 4 h
Stir at RT for 18 h
Reflux for 4.5 h
Reflux for 12 h
Stir at RT for 18 h
0
conc. HCl (24 mmol), CuSO4 (12.5 mol)
conc. HCl (24 mmol), MsOH (15 mmol)
conc. HCl (24 mmol), MsOH (15 mmol)
conc. HCl (24 mmol), MsOH (15 mmol)
conc. HCl (24 mmol), MsOH (15 mmol),
2,2-dimethoxypropane (48.8 mmol)
< 5
30
30a
36b
40c
Based on 2.0 g of crude potassium D-lyxonate. Molar ratio of 2,3-acetonide 9 to 3,5-acetonide 9a.a 9 : 1. b 3 : 1. c 4 : 1.
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 7720–7726 | 7725
©

View Article Online
Scheme 4 D-Lyxonate 8 to 2,3-O-isopropylidene-D-lyxonolactone 9.
inversion of the reacting carbon center. Reduction of the carbonyl
group on the anomeric carbon of the 2,3-acetonide 10 by NaBH4 in
methanol solution gave product 11 as a colorless syrup (95% yield).
Finally, 11 was subjected to hydrolysis using acidic Amberlite IR-
120H resin to afford L-ribose 12 in 94% yield. The overall yield of
L-ribose was 47% which can be compared with previously reported
yields of 18 to 45%.3,9
7 S. H. Buiyan, Z. Ahmed, M. Utamura and K. Izumori, J. Ferment.
Bioeng., 1998, 86, 513.
8 T. B. Granstro¨m, G. Takata, K. Morimoto, M. Leisola and K. Izumori,
Enzyme Microb. Technol., 2005, 36, 976.
9 T. Ma, S. B. Pai, Y. L. Zhu, J. S. Lin, K. Shanmuganathan, J. Du, C.
Wang, H. Kim, M. G. Newton, Y. C. Cheng and C. K. Chu, J. Med.
Chem., 1996, 39, 2835; L. Petru˘s, M. Petrusˇova and Z. Hricov´ıniova´,
Top. Curr. Chem., 2001, 215, 15; Y. Abe, T. Takizawa and T. Kunieda,
Chem. Pharm. Bull., 1980, 28, 1324; S. Pitsch, Helv. Chim. Acta, 1997,
80, 2286; M. E. Jung and Y. Xu, Tetrahedron Lett., 1997, 38, 4199; H.
Takahashi, Y. Iwai, Y. Hitomi and S. Ikegami, Org. Lett., 2002, 4, 2401;
M. J. Seo, J. An, J. H. Shim and G. Kim, Tetrahedron Lett., 2003, 44,
3051; N. Wymer and E. J. Toone, Curr. Opin. Chem. Biol., 2000, 4, 110;
T. Shimonishi and K. Izumori, J. Ferment. Bioeng., 1996, 81, 493; R.
D. Woodyer, N. J. Wymer, F. M. Racine, S. N. Khan and B. C. Saha,
Appl. Environ. Microbiol., 2008, 74, 2967.
10 J. A. Cusumano, J. Chem. Educ., 1995, 72, 959; R. R. Bader, P.
Baumeister and H.-U. Blaser, Chimia, 1996, 50, 99; R. A. Sheldon,
J. Mol. Catal. A: Chem., 1996, 107, 75.
11 D. W. John, P. K. Vivekanand, E. M. Philip and B. T. Leroy, Org.
Synth., 2005, 82, 75.
Conclusions
The efficient synthesis of the rare sugars L-lyxose and L-ribose was
carried out in a 5-step preparative route starting from naturally
occurring D-ribose and D-lyxose respectively. The aerobic oxida-
tion of D-sugar to lactone was carried out over a heterogeneous
catalyst, 10 wt% Pd-Bi supported on carbon, with high yields of
>95%. The best catalyst contained a Pd : Bi atomic loading of
5 : 1. Under alkaline conditions, the catalyst was not affected by
any poisoning and could be reused for subsequent batch reactions,
with no significant loss of activity and selectivity. The synthetic
route resulted in good yields of L-lyxose and L-ribose of 50% and
47%, respectively.
12 S.-Y. Chen and M. M. Joullie, Tetrahedron Lett., 1983, 46, 5027; A.
Fu¨rstner, K. Radkowski, C. Wirtz, R. Goddard, C. W. Lehmann
and R. Mynott, J. Am. Chem. Soc., 2002, 124, 7061; S. Jiang, B.
Mekki, G. Singh and R. H. Wightman, Tetrahedron Lett., 1994, 35,
5505.
13 L. A. Flexser, Hoffmann-La Roche Inc., US Pat., 2 438 883, 1948, L.
H. Sternbach, Hoffmann-La Roche Inc., US Pat., 2 438 882, 1948, W.
Schmidt and J. Paust, BASF Aktiengesellschaft, US Pat. , 4 294 766,
1981.
14 A. Abbadi, M. Makkee, W. Visscher, J. A. R. van Veen and H. van
Bekkum, J. Carbohydr. Chem., 1993, 12, 573; M. Besson, P. Gallezot, F.
Lahmer, G. Fleche and P. Fuertes, in: Catalysis of Organic Reactions,
ed. J. R. Kosak and T. A. Johnson, M. Dekker Inc., New York, 1994,
169; H. E. van Dam, P. Duijverman, A. P. G. Kieboom and H. van
Bekkum, Appl. Catal., 1987, 33, 373.
Acknowledgements
Financial support from the National University of Singapore
under grant numbers R-143-000-374-112 and R-143-000-418-112
is gratefully acknowledged.
Notes and references
1 F. W. Lichtenthaler and S. Peters, C. R. Chim., 2004, 7, 65.
2 J. Du, Y. Choi, K. Lee, B. K. Chun, J. H. Hong and C. K. Chu,
Nucleosides, Nucleotides Nucleic Acids, 1999, 18, 187; J. S. Cooperwood,
V. Boyd, G. Gumina and C. K. Chu, Nucleosides, Nucleotides Nucleic
Acids, 2000, 19, 219; G. Gumina, Y. Chong, H. Choo, G.-Y. Song and
C. K. Chu, Curr. Top. Med. Chem., 2002, 2, 1065; C. Mathe´ and G.
Gosselin, Antiviral Res., 2006, 71, 276; G. Gumina, Y. Chong and C.
K. Chu, in Cancer Drug Discovery and Development: Deoxynucleoside
Analogs in Cancer Therapy, ed. G. J. Peters, Humana Press, New Jersey,
2007, p. 173; M. Yun, H. R. Moon, H. O. Kim, W. J. Choi, Y.-C. Kim,
C.-S. Park and L. S. Jeong, Tetrahedron Lett., 2005, 46, 5903.
3 K. Okano, Tetrahedron, 2009, 65, 1937.
4 E. J. Reist, L. V. Fisher and D. E. Gueffroy, J. Org. Chem., 1966, 31,
226.
5 H. Kuzuhara, H. Terayama, H. Ohrui and S. Emoto, Carbohydr. Res.,
1971, 20, 165.
6 J. S. Brimacombe, A. M. Mofti and M. Stacey, Carbohydr. Res., 1971,
16, 303.
15 Y. Schuurman, B. F. M. Kuster, K. van der Wiele and G. B. Marin,
Appl. Catal., A, 1992, 89, 47; H. E. van Dam, A. P. G. Kieboom and
H. van Bekkum, Appl. Catal., 1987, 33, 361.
16 A. Abbadi and H. van Bekkum, Appl. Catal., A, 1995, 124, 409.
17 H. E. van Dam, L. J. Wisse and H. van Bekkum, Appl. Catal., 1990,
61, 187; M. Wenkin, C. Renard, P. Ruiz, B. Delmon and M. Devillers,
Stud. Surf. Sci. Catal., 1997, 108, 391.
18 P. Gallezot, Catal. Today, 1997, 37, 405; M. Besson, F. Lahmer, P.
Gallezot, P. Fuertes and G. Fle`che, J. Catal., 1995, 152, 116; T. Mallat
and A. Baiker, Catal. Today, 1994, 19, 247; M. Besson and P. Gallezot,
Catal. Today, 2000, 57, 127.
19 M. Godskesen, I. Lundt, R. Madsen and B. Winchester, Bioorg. Med.
Chem., 1996, 4, 1857; H. Batra, R. M. Moriarty, R. Penmasta, V.
Sharma, G. Stanciuc, J. P. Staszewski, S. M. Tuladhar, D. A. Walsh,
S. Datla and S. Krishnaswamy, Org. Process Res. Dev., 2006, 10,
484.
20 M. Godskesen, I. Lundt and I. Søtofte, Tetrahedron: Asymmetry, 2000,
11, 567.
7726 | Org. Biomol. Chem., 2011, 9, 7720–7726
This journal is
The Royal Society of Chemistry 2011
©