Scalable, efficient and rapid chemical synthesis of L-Fructose with high purity

Article abstract of DOI:10.1016/j.carres.2019.05.010

An improved process for chemical synthesis of L-fructose with high purity in large scale from readily available L-sorbose is described. In general, this synthetic scheme is characterized by inexpensive and easily available starting materials, simple and s

Full text of DOI:10.1016/j.carres.2019.05.010

Accepted Manuscript
Scalable, efficient and rapid chemical synthesis of l-Fructose with high purity
Jihui Li, Xinjing Song, Tianxiao Wu, Liyu Zhao, Qiaohua Qin, Maosheng Cheng, Yang
Liu, Dongmei Zhao
PII:
S0008-6215(19)30174-0
https://doi.org/10.1016/j.carres.2019.05.010
CAR 7717
DOI:
Reference:
To appear in: Carbohydrate Research
Received Date: 25 March 2019
Revised Date: 16 May 2019
Accepted Date: 22 May 2019
Please cite this article as: J. Li, X. Song, T. Wu, L. Zhao, Q. Qin, M. Cheng, Y. Liu, D. Zhao, Scalable,
efficient and rapid chemical synthesis of l-Fructose with high purity, Carbohydrate Research (2019), doi:
https://doi.org/10.1016/j.carres.2019.05.010.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Graphical Abstract
1

ACCEPTED MANUSCRIPT
Scalable, Efficient and Rapid Chemical Synthesis of
with High Purity
L
-Fructose
Jihui Li, Xinjing Song, Tianxiao Wu, Liyu Zhao, Qiaohua Qin, Maosheng Cheng,
Yang Liu*, Dongmei Zhao*
Key Laboratory of Structure-Based Drug Design & Discovery of Ministry of Education, Shenyang
Pharmaceutical University, Shenyang 110016, China
Abstract：An improved process for chemical synthesis of
L-fructose with high purity in large scale from
readily available -sorbose is described. In general, this synthetic scheme is characterized by inexpensive and
L
easily available starting materials, simple and safe experimental procedures, short time period, low environmental
impact, and great potential for scaling up. The scale-up experiment (100 g ) was carried out to provide 42.7 g of
L
-fructose with high HPLC purity of 99.65 % in total yield of 50.2 %. Consequently, the described improvements
would be helpful for those who may wish to use -fructose and promoting the further evaluation of applications of
-fructose.
L
L
Keywords: L-fructose; scalable synthesis; good yield; high purity; short time period
1. Introduction
Unnatural monosaccharides have many potentially useful applications, such as being chiral building blocks
for synthesis of biologically active compounds1. Given the importance of these monosaccharides, in recent years,
there has been an increasing interest in the development of their synthesis.
degree of sweetness as -fructose, the sweetest among the monosaccharides.
ideally non-nutritive sweetener2 for the patients with diseases related to sugar metabolism, an inhibitor of various
glycosidases,3 and an insecticide for ants and house flies4. Moreover,
-fructose-derived substrates, such as
-fructose-derived ketone, have proved to be effective catalysts with high enantioselectivity in asymmetric
L-Fructose, for example, has the same
D
L-Fructose is well known as an
L
L
epoxidation for a wide range of trans- and tri-substituted olefins, which can bear functional groups such as
tributylsilyl ether, acetal, chloride, and ester5.
2

ACCEPTED MANUSCRIPT
Figure 1. Review of the previous syntheses of L-fructose
L
-Fructose has been prepared by chemical6 and chemoenzymatic7 syntheses (Figure 1).
L-Fructose was first
synthesized by Fischer6a in 1890. Then, DL-fructose was prepared by aldol condensation. DL-Glyceraldehyde
condensation with 1,3-dihydroxy-2-propane, catalyzed by very dilute alkali6b or Dowex-1 (OH-) resin6c, gave a
hexulose mixture. Crystalline DL-fructose was obtained by crystallization of the mixture from methanol, and
following treatment of the DL-fructose mixture with baker’s yeast produced
-fructose was first synthesized by Wolfrom and Thompson from
-arabinonic acid tetraacetate in five steps6d.
Except for the methods mentioned above, -mannose was isomerized to -fructose by an isomerase present in
cell-free extracts of Aerobacter aerogenes grown on -fructose from
-rhamnose7a. The enzymatic synthesis of
dihydroxyacetone phosphate (DHAP) and DL-glyceraldehyde was carried out by a multienzyme system
comprising rhamnulose-1-phosphate aldolase (RhaD) and acid phosphatase (AP)7d-i
-Fructose was also
L-fructose. The single isomer
L
L
L
L
L
L
.
L
synthesized in one step from dihydroxyacetone and DL-glyceraldehyde using RhaD aldolase in the presence of
borate as a phosphate ester mimic7j. However, all these chemical and chemoenzymatic syntheses suffered from the
limitation to offer a feasible route for large-scale synthesis of L-fructose. The chemical synthesis of L-fructose was
characterized by long reaction sequences and low selectivity, and the enzyme-based methods either gave very low
yields or often required relatively expensive starting materials. As a result, the drawbacks of these methods made
the
L-fructose production difficult to scale up and hampered further applications of L-fructose.
3

ACCEPTED MANUSCRIPT
Previous work: total time: 74.3 h, total yield: 31.6 %, scale: 0.83 g (L-Fructose)
selective deprotection
one-pot transformation
two consecutive steps
2. 9 M H2SO4/H2O,
O
O
pH=1, 80
1. 0.115% SnCl2, 2 h
0.25% H2SO4, acetone
O
O
HO
O
O
O
1. 9 M NaOH/H2O
48 h, 70 -80
62.0 % from 2
HO
rt, 20 h
O
2. MsCl, Py, 4 h
3 OMs
OMs
2
51.0 % (two-step)
OH OH
O
OH OH
OH
O
HO
OH
HO
L-SOoHrbose
L-FrOuHctose
1. 10% KOH/H2O
2.5 h, 100
O
O
O
O
1. 0.23% SnCl2, 2.3 h
O
60% HOAc/H2O HO
O
O
45 , 2.5 h
O
HO
2. 9 M H2SO4/H2O, pH=1,
55-60
2. TsCl, Et3N, 6 h
60.3 % (two-step)
OTs
OTs
7
6
83.2 % from 6
Purified by crystallization
Purified by crystallization
This work: total time: 14.0 h, total yield: 50.2 %, 99.65 % HPLC purity,
scale: 42.7 g(L-Fructose)
Scheme 1 Syntheses of
L-fructose from
L-sorbose
OHOH
O
O
O
O
O
OMe
O
O
MeO
MsCl, Py
rt , 4 h
O
O
O
DME, SnCl2
reflux, 2 h
O
HO
OH
OMs
OH
OH
L-Sorbose
1
2
O
O
O
O
O
HO
HO
9 M NaOH/H2O
48 h,70 -80
0.25% H2SO4, acetone
rt, 20 h
NaOH
O
O
℃
℃
HO
OMs
3
4
OHOH
O
O
O
HO
HO
9 M H2SO4/H2O, pH=1
20 min, 70 -80
O
HO
OH
℃
℃
OH
OH
5
L-Fructose
Scheme 2 Original method for the synthesis of L-fructose
In 1988, Chen and Whistler prepared L-fructose from L-sorbose by ketalization and mesylation in order to
protect non-involved hydroxyl groups, formation 3,4-anhydro ring by one-pot acid-base treatment, and further ring
opening to the desired configuration (Scheme 1, upper half and Scheme 2)8. Although this method of preparing
L-fructose from
L-sorbose used available and relatively inexpensive industrial chemicals as its starting materials
and had an acceptable yield, it required much longer reaction time, some steps were sensitive to reaction
conditions, such as the ketalization of
L-sorbose and selective removal of the isopropylidene protecting group, and
4

ACCEPTED MANUSCRIPT
only a small quantity of
L-fructose was prepared (0.83 g). The consequent difficulties in scaling up production
prevented the further evaluation of the applications of
makes the production of -fructose from an economic and environmentally friendly process worthwhile. Recently,
we made efforts to improve the chemical synthesis of -fructose from -sorbose by solving the problems
mentioned above to extensively investigate the possibilities of the synthesis of -fructose on large scale and in high
purity. Herein we wish to report our detailed studies on this subject. (Scheme 1, lower half and Scheme 3). Similar
L-fructose. Also, commercial L-fructose is expensive, which
L
L
L
L
to the results described herein has - to our best knowledge - never been published providing the targeted
in such a scale and purity.
L-fructose
Scheme 3 Our improved method for the synthesis of L-fructose
2. Results and discussion
L
-Fructose was prepared from
one-pot base-acid treatment (Scheme 3). Crystalline 1,2:4,6-di-O-isopropylidene-α-
obtained by ketalization of -sorbose after recrystallization which is likely to be a kinetic product and is easy to
L
-sorbose by sequential ketalization and tosylation, selective deprotection and
L-sorbofuranose (1) was
L
isomerize8. To investigate industrial feasibility, the synthesis of ketal 1 was optimized using the orthogonal
experimental method (Table 1, entries 2-10). The ketalization was very sensitive to reaction conditions such as the
amount of SnCl2, temperature, and reaction time5, 8-9. The presence of an appropriate amount of SnCl2 was
indispensable to the ketalization of L-sorbose (Table 1, entry 1 vs entries 2-10). In entry 1, almost no L-sorbose
was converted to ketal 1 in the absence of SnCl2, resulting in a yield of less than 3 %, while ketal 1 was obtained in
5

ACCEPTED MANUSCRIPT
medium to good yields of 43.3 - 68.8 % (entries 2-10) after crystallization from petroleum ether. To optimize
the most appropriate amount of SnCl2, 0.115 mol % (entries 8-10), 0.230 mol % (entries 5-7), and 0.460 mol %
(entries 2-4) were studied, respectively. The results suggested that, at the same reaction temperature (83 °C or 70
°C), more the SnCl2 was used, more rapidly the L-sorbose was completely consumed (entry 2 vs 5 and 8, entry 4 vs
7 and 10). Moreover, more the SnCl2 was used, the higher conversions were obtained at the same reaction
temperature and for same time (entry 9 vs 3 and 6). As shown in Table 1, although the conversion of entry 3
(98.5 %) was obviously higher than that of entry 6 (90.0 %), the yields of these two entries were almost same
(68.8 % vs 68.5 %). This fact indicated that more SnCl2 could facilitate the isomerization (some unknown isomers
were detected by TLC and LC-MS) of ketal 1 to lead to lower yields. Additionally, a higher reaction temperature
could facilitate the ketalization of
yields, since the high temperature could also induce the isomerization of ketal 1 (entry 2 vs 4, 5 vs 7, 8 vs 10).
Furthermore, -sorbose was completely consumed by prolonging the reaction time (entry 2 vs 3, 5 vs 6), but the
yields decreased with prolonging reaction time, which attributed to the formation of isomers. As the ketalization of
-sorbose was highly sensitive to reaction conditions, we immediately quenched the reaction before the -sorbose
was completely consumed at 70 °C to shorten the reaction time (2.3 h), minimize the isomerization of ketal 1, and
enable the remaining -sorbose to be readily recovered by filtration and reused for the purpose of reducing the
costs (entry 6). Therefore, at the stage of the scale-up (100 g), 1,2:4,6-di-O-isopropylidene-α- -sorbofuranose (1)
L-sorbose to ketal 1. However, raising the temperature not always increased the
L
L
L
L
L
could be facilely prepared in a high yield of 68.5 % (lit8 41 % yield) after crystallization from petroleum ether,
whose optical rotation and melting point were consistent with the values previously reported8.
Table 1 Optimization of the synthesis of ketal 1
Entrya
1
SnCl2 (mol %)
0
Temp. (°C)
83
Time (h)
24
Conv. (%)e
3.00
Yield (%)f
< 3.00
6

ACCEPTED MANUSCRIPT
2
3
0.460
0.460
0.460
0.230
0.230
0.230
0.115
0.115
0.115
70
70
83
70
70
83
70
70
83
3.0b
2.3c
1.5b
4.0b
2.3c
3.0b
9.0d
2.3c
9.0b
100
98.5
100
100
90.0
100
66.0
42.0
100
56.3h
68.8g
58.3h
59.7h
68.5g
50.0h
55.0g
43.3g
56.6h
4
5
6
7
8
9
10
a
Reaction conditions: L-sorbose (1.0 g), 2,2-dimethoxypropane (DMP, 6.0 mL), 1,2-dimethoxyethane
(DME, 0.2 mL), SnCl2. b The reaction was stopped once the solution became clear. c Other isomers of ketal
1 were detected by TLC. d Entry 8 was quenched at the very moment when the substrate (L-sorbose) of
f
entry 10 was completely consumed. e The conversion was calculated from the recovery of
L-sorbose.
Obtained by crystallization from petroleum ether. g The yield was calculated based on the recovery of
-sorbose. h The yield was calculated based on
-sorbose.
L
L
Table 2 Optimization of the synthesis of tosylate 6
Entrya
TsCl (eq.)
1.5
Base (eq.)
Solvent
DCM
DCM
DCM
DCM
DCM
DCM
DCM
Time (h)
48b
Yield (%)
57.9c
1
2
3
4
5
6
7
Et3N (3.0)
Et3N (4.0)
Et3N (6.0)
Et3N (8.0)
Et3N (10.0)
Et3N (5.0)
DIPEA(10.0)
2.0
36
79.6c
3.0
21
79.3d
78.3d
83.3d
49.5c
4.0
12
5.0
6.0
5.0
6.0b
6.0b
5.0
NDe
7

ACCEPTED MANUSCRIPT
8
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
Imidazole (10.0)
Py (10.0)
DCM
DCM
DCM
THF
6.0b
6.0b
6.0b
6.0b
6.0b
6.0b
6.0b
6.0b
6.0b
6.0g
NDe
53.1c
NDe
9
10
11
12
13
14
15
16f
17
2,6-Lutidine (10.0)
Et3N (10.0)
Et3N (10.0)
Et3N (10.0)
Et3N (10.0)
Et3N (10.0)
K2CO3 (10.0)
Py (10.0)
55.6c
35.0c
36.3c
42.3c
30.2c
NDe
acetone
MTBE
MeCN
DMF
DMF
DMF
NDe
a Reaction conditions: ketal 1 (0.52 g, 2.0 mmol, 1.0 eq.), solvent (5.0 mL), DMAP (12 mg, 0.1 mmol,
c
0.05 eq.), base. b Ketal 1 was not completely converted. Isolated yield after column chromatography. d
e
f
Crystallized yield in EtOH. Not detected by the TLC analysis. Considering the solubility of K2CO3,
DMF was used. g Ketal 1 was completely converted, but the product was not the desired tosylate 6.
In the original procedure (Scheme 2), the formed ketal 1 was used directly for mesylation to give mesylate 2.
As we all know, methanesulfonyl chloride can be easily decomposed by water and requires relatively strict
reaction conditions, which makes it unsuitable beyond laboratory scale. Considering the limitions mentioned above,
-
p-methylbenzenesulfonyl chloride (TsCl) was chosen to replace MsCl in its role. In addition, the OTs is a better
leaving group than the -OMs. Therefore, the 3,4-anhydro ring of compound 4 can be formed much more easily by
the action of alkali. We thus then optimized the synthesis of tosylate 6, particularly the equivalents of TsCl which
had a significant influence on the reaction time (Table 2, entry 1 vs 2-5). As shown in Table 2, ketal 1 was not
completely consumed with 1.5 equivalents of TsCl, and the yield was only 57.9 %, whereas the reaction completed
in the presence of 2.0 equivalents of TsCl with a high yield of 79.6 % (entry 1 vs 2). Next, higher equivalents of
TsCl were used to further shorten the reaction time. The results demonstrated that the reaction time was
dramatically shortened with increasing the equivalents of TsCl (entry 2 vs 3, 4 and 5), for example, 6.0 times
shorter reaction time attributed to 2.5 times more TsCl (entry 2 vs 5). The experimental result was particularly
encouraging when 5.0 equivalents of TsCl were applied (entry 5). In this condition, crystalline tosylate 6 was
8

ACCEPTED MANUSCRIPT
obtained with a good yield and in high purity by removing the redundant TsCl with absolute ethyl alcohol. Other
equivalents of TsCl were not investigated for the following reasons: 1) an acceptable yield was obtained by the
method described in entry 5, and 2) a greater excess of TsCl led to higher costs and, more importantly, more
difficult removal of the remaining TsCl from tosylate 6. Further as shown in Table 2, lowering the equivalents of
Et3N led to the yield significant decreasing (entry 5 vs 6). In addition, we checked several other organic and
inorganic bases (entries 7-10 and entry 16), and found that these bases had less effect on the reaction. Therefore,
10.0 equivalents of Et3N were adopted. Meanwhile, in view of the safety of DCM in industrial application, we
investigated the solvent effects on the reaction (entries 11-15) in order to make it more suitable for the scale-up.
Unfortunately, the yields dramatically decreased (30.2 - 55.6 %) when THF, acetone, MTBE, MeCN and DMF
were used, respectively. It should be emphasized that no target compound 6 was obtained when Et3N was replaced
by Py in DMF (entry 17), but quantitative formylation of 1 was observed. In summary, tosylate 6 was synthesized
from ketal 1 in the presence of TsCl, DMAP, and Et3N in DCM at 30 °C (Scheme 3, entry 5). In the stage of the
scale-up, commercially available L-sorbose underwent sequential ketalization with 2,2-dimethoxypropane, and
following tosylation with TsCl, tosylate 6 was generated in 60.3 % over two steps.
In the original procedure (Scheme 2), it took a long reaction time (about 3 days) to achieve the one-pot
transformation of mesylate 2 to L
-fructose8. We made efforts to improve this process for large scales and found that
tosylate 6 was converted to compound 7 by the action of 0.25 % aqueous sulfuric acid at room temperature for 24 -
48 h. Consequently, this rather long reaction time led to the decomposition of compound 7 into a by-product, most
likely 3-O-tosyl-L
-sorbose8. Increasing the concentration of sulphuric acid did shorten the retention time but
increased the generation of by-products5. We speculated that this selective deprotection reaction was sensitive to
the concentration and strength of acids, as well as the reaction time. Thus, to shorten the reaction time and
minimize the generation of by-products, we attempted to operate the reaction at a higher temperature in the
presence of a weak acid. Finally, aq. 60 % HOAc at 45 °C for 2.5 h was obtained as an effective condition
(Scheme 3). The colorless crystalline 7 was collected by pouring the reaction mixture into ice water followed by
vacuum suction filtration and used directly in the next step without purification.
According to the literature, compound 2 could be converted to L-fructose by a one-pot transformation in water
with no organic solvent5, 8-9. The formation of the 3,4-anhydro ring was easily achieved in alkaline conditions such
as aq. 2 M NaOH, but the opening of the anhydro ring by base catalysis was somewhat difficult. Ring opening was
9

ACCEPTED MANUSCRIPT
completed in a strong, aqueous alkaline solution, such as aq. 9 M NaOH, after heating for 2 to 3 days at
70-80 °C8-9. A higher concentration of base could accelerate the reaction, but the reaction mixture became more
complex5. Given the reaction efficiency and product stability, a much stronger but more dilute alkali was utilized
to achieve the one-pot transformation in our experiments. As we expected, after the solution of 7 was made
alkaline using aq. 5 % KOH (0.9 M), the conversion of compound 7 to triol 5 via epoxide 4 was efficiently
achieved at 100 °C in 4.5 h, which the reaction time was more than 10 times (4.5 h vs 2-3 days) shorter than that of
the original method8-9. At the same time, the use of aq. 10 % KOH (1.8 M) at 100 °C was also investigated, and the
conversion of compound 7 to triol 5 occurred in only 2.5 h, which further shortened the reaction time about 2.0
times (2.5 h vs 4.5 h). According to the literature8, the 1,2-O-isopropylidene group of triol 5 was then easily
removed after being acidified to pH=1 with 9 M H2SO4 and heated at 70 - 80 °C for 20 min. But at work up stage,
the reaction should be carefully monitored by TLC and quenched as soon as the reaction was finished. Otherwise,
heating the reaction mixture for a long time caused problems, possibly polymerization of the L-fructose5. Worse,
the polymers of
possibility of polymerization of
The lower temperature (55 - 60 °C) and slightly longer reaction time (40 min) were adapted to give a syrup from
which a crude -fructose was obtained by several extractions with ethanol. Under these conditions, notably, no
polymers of the -fructose were monitored by TLC. The crude product was purified by chromatography to give
-fructose as a syrup with a high yield of 83.2 %, and then the syrup crystallized in absolute ethyl alcohol to give
L-fructose were difficult to remove from crude
L-fructose. Therefore, in order to minimize the
L-fructose and investigate the scalable feasibility, a few parameters were revised.
L
L
L
1
solid
L
-fructose as a white powder whose the H-NMR, 13C-NMR and optical rotation were very consistent to
7a, 8,
those values previously reported6d,
10. Large-scale applicability of this newly improved method was
successfully demonstrated on multigram scale to obtain
With abundant -fructose in hand, a -fructose-derived alcohol 8, a highly enantioselective catalyst, could be
L-fructose with 99.65 % HPLC purity in 50.2 % total yield.
L
L
conveniently synthesized in large scale in laboratory for the asymmetric epoxidation of a wide range of trans- and
tri-substituted olefins9, 11-12. The values of 1H-NMR and 13C-NMR, optical rotation, and melting point agreed with
those in the literature5, which confirmed the structure of our synthesized
L-fructose as well.
10

ACCEPTED MANUSCRIPT
Figure 2. The L-fructose-derived alcohol 8
3. Conclusions
In summary, we have described a scalable, efficient and rapid synthesis of high quality
-sorbose within a short time period. After extensive optimization of the reaction conditions, the prominent
features of our improved strategy included the following: first, 1,2:4,6-di-O-isopropylidene-3-O-tosyl-α- -sorbo-
furanose (6) was efficiently obtained from -sorbose by sequential ketalization and tosylation with a high yield of
L-fructose from
L
L
L
60.3 %; second, the selective deprotection of the 4,6-O-isopropylidene group of 6 was easily achieved by action of
aq. 60 % HOAc at 45 °C for 2.5 h, which significantly shortened the reaction time (8.0 times, 2.5 h vs 20.0 h) and
minimized by-product generation; finally, but most importantly, the one-pot transformation proceeded efficiently
and rapidly in water with no organic solvent via the action of aq. 10 % KOH followed by aq. 9 M H2SO4, which
dramatically shortened the reaction time (15 times, 3.2 h vs 48.3 h). Notably, the intermediates were either easily
purified by recrystallization or used without any purification. This synthetic scheme is characterized by
inexpensive and easily available starting materials, simple and safe experimental procedures, short reaction time,
low environmental impact, and great potential for scaling up. With the process adaptions, multigram quantities of
L
-fructose have been prepared with 99.65 % HPLC purity in total yield of 50.2 % (lit8 31.6 % yield) within short
time period (14.0 h vs 74.3 h). Consequently, the described improvements would be helpful for those who may
wish to use -fructose and promoting the further evaluation of applications of -fructose.
L
L
4. Experimental section
4.1. General information
All reagents and solvents were obtained from commercial suppliers and used without further purification unless
11

ACCEPTED MANUSCRIPT
otherwise indicated. Column chromatography was carried out on silica gel (200-300 mesh). All reactions were
monitored by thin-layer chromatography (TLC) on silica gel plates (TLC; HG/T2354-92, GF254). Detection
involved spraying the chromatogram with a solution of p-anisaldehyde (10 mL) and conc. H2SO4 (10 mL) in 95 %
1
EtOH (200 mL), followed by heating the plate for several minutes with an alcohol lamp. H-NMR and 13C-NMR
spectra were recorded in CDCl3, DMSO-d6 or D2O on a Bruker ARX-600 NMR spectrometer with TMS as an
internal standard and reported in ppm (δ). High-resolution accurate mass spectrometry (HRMS) was performed on
a Bruker Micromass Time of Flight mass spectrometer equipped with an electrospray ionization (ESI) detector.
Melting points were determined with a Buchi B-540 melting point apparatus and are uncorrected. Optical rotation
was determined at room temperature with an MCP-200 polarimeter (Anton-Paar). The purity of L-fructose was
determined by HPLC using a Agilent LC-1260 series instrument equipped with differential refractive
index detector (YRD-880 midget). The HPLC analysis data was reported in relative area % and was not adjusted to
weight %.
4.2. 1,2:4,6-Di-O-isopropylidene-α-L-sorbofuranose (1)
To a suspension of L-sorbose (pulverized) (100.0 g, 0.56 mol) in 2,2-dimethoxypropane (300 mL, 3.0 vol)
was added a solution of 1,2-dimethoxyethane (20 mL, 0.2 vol) containing SnCl2 (250 mg, 1.28 mmol). The
mixture was refluxed gently with vigorous stirring under N2 at 70 °C (bath temperature) for 2.3 h until other
isomers of compound 1 were carefully detected by TLC with ethyl acetate-petroleum ether (1:4, v/v) as a solvent.
At this point, the reaction was quenched immediately with Et3N (1.26 mL) and filtered to remove the unreacted
L-sorbose (10.0 g), and the solid was washed with CH2Cl2 (20 mL). The filtrate was concentrated to give a dark
brown syrup. The syrup was then dissolved in CH2Cl2 (500 mL) and washed with water (100 mL) and brine (100
mL×3). The CH2Cl2 fraction was dried over anhydrous Na2SO4, filtered, and evaporated to a brown syrup. The
syrup was then dissolved with hot petroleum ether (360 mL) followed by filtration to remove the undissolved
substance. Crystallization occurred when the hot petroleum ether extract was cooled slowly to room temperature.
The precipitated product was collected by filtration, washed with cool n-hexane (200 mL), and dried in vacuum to
25
afford 1 (89.0 g, 68.5 % based on the recovered
L
-sorbose) as a pale yellow solid. M.p. 71.3-73.7 °C, [α]D = –
25.7° (c 1.010, acetone) [lit6 M.p. 71-73 °C, [α]D25 = – 24.7° (c 1.029, acetone)]; 1H NMR (600 MHz, DMSO-d6) δ
5.22 (d, J = 5.2 Hz, 1H), 4.11 (dd, J = 3.4, 1.9 Hz, 1H), 4.01 (m, 1H), 3.94 (s, 2H), 3.92 (dd, J = 12.8, 3.3 Hz, 1H),
3.81 (dd, J = 5.2, 1.9 Hz, 1H), 3.67 (dd, J = 12.8, 3.1 Hz, 1H), 1.40 (s, 3H), 1.34 (s, 3H), 1.32 (s, 3H), 1.24 (s, 3H);
12

ACCEPTED MANUSCRIPT
13C NMR (151 MHz, CDCl3) δ 111.29, 110.64, 97.50, 79.50, 74.36, 72.96, 72.58, 60.51, 28.64, 26.45, 25.41, 19.19;
HRMS (ESI): m/z calcd for C12H20O6Na [M+Na]+ 283.1158, found 283.1167.
4.3. 1,2:4,6-Di-O-isopropylidene-3-O-tosyl-α-L-sorbofuranose (6)
(a) The solid 1 (88.8 g, 0.34 mol) was dissolved in CH2Cl2 (700 mL, 7.9 vol). After cooling with an ice bath,
Et3N (472 mL, 3.4 mol) and DMAP (2.07 g, 17 mmol) were added with stirring, and then
p-methylbenzenesulfonyl chloride (324.1 g, 1.7 mol) was added in portions. After stirring at 0 °C for 30 min, the
reaction mixture was allowed to warm to room temperature for additional 6 h. The precipitate was filtered off and
washed with CH2Cl2 (300 mL), and then the filtrate was washed with water (200 mL) and brine (200 mL×3), dried
over anhydrous Na2SO4, filtered and concentrated to afford the crude product. The crude product was stirred in
EtOH (280 mL, 3.0 vol) for several hours (6 h), and then the precipitate was filtered off and washed with cold
EtOH (100 mL) to give 6 (117.3 g) as a white solid. Yield 83.3 %; M.p. 114.1-117.2 °C, [α]D25 = – 24.9° (c 1.006,
acetone); 1H NMR (600 MHz, CDCl3) δ 7.84 (d, J = 8.3 Hz, 2H), 7.34 (d, J = 8.0 Hz, 2H), 4.65 (d, J = 1.8 Hz, 1H),
4.31 (dd, J = 3.4, 2.0 Hz, 1H), 4.18 (m, 1H), 4.12 (m, 2H), 3.94 (dd, J = 12.8, 3.6 Hz, 1H), 3.83 (dd, J = 12.9, 3.5
Hz, 1H), 2.45 (s, 3H), 1.47 (s, 3H), 1.30 (s, 3H), 1.27 (s, 3H), 1.24 (s, 3H). 13C NMR (151 MHz, CDCl3) δ 145.08,
133.33, 129.74 (2C), 128.20 (2C), 111.59, 109.29, 98.21, 83.93, 73.12, 72.89, 71.91, 60.23, 27.77, 25.89, 25.62,
21.66, 20.01; HRMS (ESI): m/z calcd for C19H26O8SNa [M+Na]+ 437.1246, found 437.1266.
(b) To a suspension of pulverized L-sorbose (100.0 g, 0.56 mol) in 2,2-dimethoxypropane (300 mL, 3.0 vol)
was added a solution of 1,2-dimethoxyethane (20 mL, 0.2 vol) containing SnCl2 (250 mg, 1.28 mmol). The
mixture was refluxed gently with vigorous stirring under N2 at 70 °C (bath temperature) for 2.3 h until other
isomers of compound 1 were carefully detected by TLC with ethyl acetate-petroleum ether (1:4, v/v) as a solvent.
At this point, the reaction was quenched immediately with Et3N (1.26 mL) and filtered to remove the unreacted
L-sorbose (6.3 g), and the solid was washed with CH2Cl2 (20 mL). The filtrate was concentrated to give a dark
brown syrup.The syrup was then dissolved in CH2Cl2 (500 mL) and washed with water (100 mL) and brine (100
mL×3). The CH2Cl2 fraction was dried over anhydrous Na2SO4, filtered, and evaporated to a brown syrup.
The above syrup was dissolved in CH2Cl2 (1000 mL, 10 vol). After cooling with an ice bath, Et3N (776.7 mL,
5.6 mol) and DMAP (3.4 g, 28.0 mmol) were added with stirring, and then p-methylbenzenesulfonyl chloride
(533.8 g, 2.8 mol) was added in portions. After stirring at 0 °C for 30 min, the reaction mixture was allowed to
13

ACCEPTED MANUSCRIPT
warm to room temperature for additional 6 h . The precipitate was filtered off and washed with CH2Cl2 (300 mL),
and the filtrate was washed with water (200 mL) and brine (200 mL×3), dried over anhydrous Na2SO4, filtered and
concentrated to afford the crude product. The crude product was stirred in EtOH (430 mL, 4.3 vol) for several
hours (6 h), and then the precipitate was filtered off and washed with cold EtOH (100 mL) to give 6 (130.0 g,
60.3 % over two steps based on the recovered L-sorbose) as a white solid.
4.4. L-Fructose
A suspension of tosylate 6 (118.3 g, 0.28 mol) in 60 % (w) HOAc/H2O (1.18 L, 10 vol) was stirred at 45 °C
until all the starting material was consumed, as monitored by TLC (2.5 h). The colorless crystalline product 7 was
collected by pouring the reaction mixture into ice water (aq. 60 % HOAc: ice water = 1:10, v/v) followed by
vacuum suction filtration and used directly in the next step with no purification.
A suspension of the crude product 7 in 10 % (w) KOH/H2O (1.58 L, 2.8 mol, 10 eq.) was heated to reflux
until the reaction was completed, as monitored by TLC (2.5 h). After acidification to pH 1.0 with aq. 9 M H2SO4
(≈ 80 mL), the reaction mixture was heated at 55-60 °C for exact 40 min, then neutralized to pH 6.0-7.0 with aq. 9
M NaOH, and concentrated to give a residue. The resulting residue was then extracted with ethanol (4 × 500 mL),
and the ethanol solution was concentrated to give the crude product
product was chromatographed on a silica gel column. Elution with ethyl acetate-methanol-water (40:10:7, v/v)
gave -fructose (42.7 g) as a syrup which crystallized from absolute ethyl alcohol to give solid -fructose as a
L-fructose as a yellow syrup. The crude
L
L
white powder with a HPLC purity of 99.65 % (relative area). Yield: 83.2 %; M.p. 89.6-92.6 °C, [α]D25 = + 89.1° (c
1.548, H2O) ; 1H NMR (600 MHz, D2O) δ 4.02 – 3.86 (m, 2H), 3.78 (dd, J = 10.1, 3.5 Hz, 1H), 3.68 (d, J = 10.0
25
1
Hz, 1H), 3.62 – 3.56 (m, 2H), 3.44 (d, J = 11.8 Hz, 1H); [lit6 [α]D = + 88° (c 1.5, H2O); H-NMR (D2O: δ
4.13-3.37 (m)]. 13C NMR (151 MHz, D2O) δ 98.00, 69.60, 69.12, 67.45, 63.79, 63.28. HRMS (ESI): m/z calcd for
C6H12O6Na [M+Na]+ 203.0532, found 203.0535.
4.5. Synthesis of alcohol 8
To the above L-fructose syrup (2.2 g, 12.22 mmol) in acetone (100 mL) was added concentrated H2SO4 (70
µL, 1.22 mmol) under N2 after cooling to 0 °C. The resulting reaction mixture was stirred overnight at 25 °C. After
neutralizing with 10 % NaOH/H2O, the reaction mixture was concentrated to a solid, which was then dissolved in
14

ACCEPTED MANUSCRIPT
CH2Cl2 (150 mL). The resulting solution was washed with water (2 × 50 mL) and brine (50 mL), dried over
anhydrous Na2SO4, filtered, and concentrated to give a white solid. After recrystallization in hexane-CH2Cl2 (4:1,
25
v/v), alcohol 8 (1.7 g) was obtained as white needles. Yield: 54.8 %; M.p. 116.4-117.8 °C, [α]D = + 141.8° (c
0.314, acetone) [lit3 M.p. 115-116 °C, [α]D25 = + 142.4° (c 1.05, CDCl3)]; 1H NMR (600 MHz, CDCl3) δ 4.22 (dd,
J = 5.8, 2.1 Hz, 1H), 4.19 (d, J = 8.8 Hz, 1H), 4.15 – 4.10 (m, 2H), 4.01 (d, J = 13.4 Hz, 1H), 3.99 (d, J = 8.8 Hz,
1H), 3.67 (d, J = 6.9 Hz, 1H), 1.94 (brs, 1H), 1.54 (s, 3H), 1.52 (s, 3H), 1.45 (s, 3H), 1.37 (s, 3H); 13C NMR (151
MHz, CDCl3) δ 111.84, 109.44, 104.50, 77.30, 73.31, 72.34, 70.43, 60.75, 27.94, 26.45, 26.24, 25.95. HRMS
(ESI): m/z calcd for C12H20O6Na [M+Na]+ 283.1158, found 283.1158.
ASSOCIATED CONTENT
1
Supporting Information. The following file is available free of charge; The copies of HRMS, H and 13C NMR
spectra of all compounds and the HPLC information with chromatogram (PDF).
AUTHOR INFORMATION
Corresponding Authors
Dongmei Zhao*, e-mail: medchemzhao@163.com; Fax: +86-24-23995043; Tel: +86-24-43520219.
Yang Liu*, e-mail:y.liu@syphu.edu.cn.
ACKNOWLEDGEMENTS
We gratefully acknowledge the Program for Innovative Research Team of the Ministry of Education and the
Program for Liaoning Innovative Research Team in University.
REFERENCES
1.
Degradation, and Epimerization; Academic Press: San Diego, 1998.
2. Levin, G. V.; Zehner, L. R.; Saunders, J. P.; Beadle, J. R., Am. J. Clin. Nutr. 1995, 62 (5 Suppl), 1161S.
3. Muniruzzaman, S.; Pan, Y. T.; Zeng, Z.; Atkins, B.; Izumori, K.; Elbein, A. D. Glycobiology 1996, 6, 795.
Györgydeák, Z, Pelyvás, I. F. Monosaccharide Sugars-Chemical Synthesis by Chain Elongation,
4.
Levin, G. V.; Zehner, L. R. Eur. Pat. EP 397,027 (Cl. A01N43/08), 14. Nov 1990; US Patent 6,166,193, 24.
15

ACCEPTED MANUSCRIPT
Nov 1992; Chem. Abstr. 1991, 114, 242827q.
5.
6.
Zhao, M. X.; Shi, Y., J. Org. Chem. 2006, 71 (14), 5377.
(a) Fischer, E., Ber.dtsch.chem.ges 1890, 23, 370-394; (b) Schmitz, E., Eur. J. Inorg. Chem. 2010, 46 (2),
2327-2335; (c) Morgenlie, S., Carbohydr. Res. 1982, 107 (1), 137-141; (d) Wolfrom, M. L.; Thompson, A., J. Am.
Chem. Soc. 1946, 68 (5), 791-793.
7.
(a) Mayo, J. W.; Anderson, R. L., Carbohydr. Res. 1968, 8 (3), 344-347; (b) Dhawale, M. R.; Szarek, W. A.;
Hay, G. W.; Kropinski, A. M. B., Carbohydr. Res. 1986, 155, 262-5; (c) Itoh, H.; Izumori, K., J. Ferment. Bioeng.
1996, 81 (4), 351-353; (d) Alajarin, R.; Garciajunceda, E.; Wong, C. H., J. Org. Chem. 1995, 60 (13), 4294-4295;
(e) Henderson, I.; Sharpless, K. B.; Chi, H. W., Cheminform 1994, 25 (22), 558-561; (f) Fessner, W. D.; Badía, J.;
Eyrisch, O.; Schneider, A.; Sinerius, G., Tetrahedron Lett. 1992, 33 (36), 5231-5234; (g) Hubschwerlen, C.,
Synthesis 1986, (11), 962-4; (h) Baer, E.; Fischer, H. O. L., J. Am. Chem. Soc. 1939, 128 (2), 761-765; (i) Franke,
D.; Machajewski, T.; Hsu, C. C.; Wong, C. H., J. Org. Chem. 2003, 68 (17), 6828-31; (j) Sugiyama, M.; Hong, Z.;
Whalen, L. J.; Greenberg, W. A.; Wong, C. H., Adv. Synth. Catal. 2006, 348 (18), 2555-2559.
8.
9.
Chen, C. C.; Whistler, R. L., Carbohydr. Res. 1988, 175 (2), 265-271.
Wang, Z. X.; Tu, Y.; Frohn, M.; Jianrong Zhang, A.; Shi, Y., J. Am. Chem. Soc. 1997, 119 (46), 11224-11235.
10. Gizaw, Y.; BeMiller, J. N., Application of a phase transfer reaction to the synthesis of L-fructose. Carbohydr.
Res. 1995, 266 (1), 81-5.
11. (a) Tu, Y.; Wang, Z. X.; Shi, Y., J. Am. Chem. Soc. 1996, 118 (40), 9806-9807; (b) Shu, L.; Shi, Y.,
Tetrahedron 2001, 57 (24), 5213-5218.
12. Tu, Y.; Frohn, M.; Wang, Z.-X.; Shi, Y., Org. Synth. 2003, 80, 1-8.
16

ACCEPTED MANUSCRIPT
Highlights:
※
L-Fructose can be efficiently and rapidly prepared in large scales via a concise and consecutive procedure, a
selective deprotection and a one-pot base-acid transformation in high HPLC purity (99.65 %) and with good total
yield (50.2%) .
※
This synthetic scheme is characterized by inexpensive and easily available starting materials, simple and safe
experimental procedures, short reaction time, low environmental impact, and great potential for scaling up.
※
The improved process is potentially of considerable importance as a facile, efficient and scalable route to
L-fructose.