A robust method for the synthesis and isolation of β-gluco- isosaccharinic acid ((2R,4S)-2,4,5-trihydroxy-2-(hydroxymethyl)pentanoic acid) from cellulose and measurement of its aqueous pKa

Article abstract of DOI:10.1016/j.carres.2011.11.022

In alkaline pulping wood pulp is reacted with concentrated aqueous alkali at elevated temperatures. In addition to producing cellulose for the manufacture of paper, alkaline pulping also generates large amounts of isosaccharinic acids as waste products. I

Full text of DOI:10.1016/j.carres.2011.11.022

Carbohydrate Research 349 (2012) 6–11
Contents lists available at SciVerse ScienceDirect
Carbohydrate Research
journal homepage: www.elsevier.com/locate/carres
A robust method for the synthesis and isolation of b-gluco-isosaccharinic
acid ((2R,4S)-2,4,5-trihydroxy-2-(hydroxymethyl)pentanoic acid) from
cellulose and measurement of its aqueous pKa
⇑
Paul B. Shaw, Glenn F. Robinson, Craig R. Rice, Paul N. Humphreys, Andrew P. Laws
Department of Chemical and Biological Sciences, School of Applied Sciences, University of Huddersfield, Queensgate Huddersfield HD1 3DH, United Kingdom
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 18 October 2011
Received in revised form 18 November 2011
Accepted 22 November 2011
Available online 3 December 2011
In alkaline pulping wood pulp is reacted with concentrated aqueous alkali at elevated temperatures. In
addition to producing cellulose for the manufacture of paper, alkaline pulping also generates large
amounts of isosaccharinic acids as waste products. Isosaccharinic acids are potentially useful raw mate-
rials: they are good metal chelating agents and, in their enantiomerically pure form, they are valuable
carbon skeletons with predefined stereochemistry that can be easily functionalised for use in synthesis.
Despite this, there is no simple procedure for isolating pure beta-(gluco)isosaccharinic acid and very lim-
ited work has been undertaken to determine the chemical and physical properties of this compound. We
report here a very simple but effective method for the synthesis of a mixture containing equal portions of
the two isosaccharinic acids ((2S,4S)-2,4,5-trihydroxy-2-(hydroxymethyl)pentanoic acid and (2R,4S)-
2,4,5-trihydroxy-2-(hydroxymethyl)pentanoic acid) and the separation of the two as their tribenzoate
esters. We also report for the first time the aqueous pKa of beta-(gluco)isosaccharinic acid (3.61).
Ó 2011 Elsevier Ltd. All rights reserved.
Keywords:
Isosaccharinic acid
Glucoisosaccharinic acid
Cellulose
Beta-isosaccharinic acid
3-Deoxy-2-C-hydroxymethyl-
D-pentonic
acid
1. Introduction
in developing new uses for isosaccharinic acids. Isosaccharinic
acids are potentially useful raw materials: they are capable of act-
It has been known since the 1930s that treatment of cellulose
with alkaline solution results in a change in morphology of cellu-
lose and that prolonged exposure at elevated temperatures leads
to a reduction in the degree of polymerisation of the cellulose
chains.1,2 Machell and Richards3 proposed a mechanism for the
degradation pathway applying the ‘peeling’ reaction, originally
proposed by Isbell4 to account for the alkaline degradation of
substituted monosaccharides, to cellulose. In the peeling reaction
hexose units, present at available reducing chain ends, are trans-
ing as metal chelating agents11–16 and, in their enantiomerically
pure form, they are valuable carbon skeletons with predefined ste-
reochemistry that can be easily functionalised for use in synthe-
sis.17–21 Their metal chelating properties are also of interest to
organisations involved in the storage of cellulosic waste including
municipal waste facilities11 and the repositories currently being
considered for the storage of nuclear waste;22–27 it has been shown
that isosaccharinic acids generated in cellulosic wastes can
promote leaching of metals into water courses. Isosaccharinates
are particularly effective at chelating radioactive actinides and
increase their aqueous solubility, potentially threatening the integ-
rity of nuclear waste repositories by providing a mechanism for the
leaching of radioisotopes into the biosphere.12–15,28,29
formed initially into 4-deoxy-3-oxo-D
-glycero-2-hexulose (1)5,6
which subsequently undergoes a benzylic acid type rearrangement
to generate (gluco)isosaccharinic acids: (2R,4S)-2,4,5-trihydroxy-
2-(hydroxymethyl)pentanoic acid (2, b-GISA) and (2S,4S)-2,4,5-tri-
hydroxy-2-(hydroxymethyl)pentanoic acid (3,
a
-GISA) and a new
It is now more than 50 years since the first methods were intro-
reducing chain end. The peeling reaction continues at the new
chain end, slowly reducing the degree of polymerisation of the cel-
lulosic substrate. The alkaline degradation of cellulosic materials
generates a large number of hydroxyaliphatic acids with isosac-
charinic acids being the most abundant.7 Unfortunately ‘peeling’
is an unwanted side reaction in alkaline pulping8–10 where it gen-
erates large amounts of by-products and this waste currently has
little economic value. There is considerable commercial interest
duced for the direct synthesis of saccharinic acids,30 Whistler and
BeMiller demonstrated that a mixture of
a-GISA & b-GISA could
be prepared by reacting 1,4-linked oligosaccharides and polysac-
charides with lime water.30–32 Over the preceding decades these
methods have been widely employed to prepare
a-GISA for use
in studies of its complexation of actinides.12–15,28,29,33 In contrast,
very little work has been carried out with b-GISA. The reason for
this is that it is very difficult to isolate b-GISA free from
a-GISA.
A number of methods for the preparation of the calcium salt of
b-GISA have been reported and these include the treatment of
either lactose34–36 cellulose36,37 or guaran32 with calcium
⇑
Corresponding author. Tel.: +44 (0) 1484 472668; fax: +44 (0) 1484 472182.
E-mail address: a.p.laws@hud.ac.uk (A.P. Laws).
0008-6215/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carres.2011.11.022

P. B. Shaw et al. / Carbohydrate Research 349 (2012) 6–11
7
hydroxide for extended periods and the separation of small quan-
tities of b-GISA from -GISA using either normal phase or anion ex-
change chromatography. Both Whistler and BeMiller34 and Feast
et al.35 have reported the production of
-GISA & b-GISA from
lactose, the conversion of the acids to their corresponding lactones
and finally, the derivatisation of an analytical sample of the
lactones as their tribenzoate esters (4 and 5). Feast et al.35 also
reported that they were able to isolate an analytical sample of
the b-GISA tribenzoate ester using preparative thin layer
chromatography.
pound via retro-aldol type reactions38 to give a range of C1–C4
hydroxyaliphatic acids in preference to GISAs. In the present stud-
ies, when a saturated solution of calcium hydroxide was used as a
base, the concentration of the GISA acids reached a maximum and
then fell; this was more noticeable when reactions were performed
under aerobic conditions (results not shown). Similar findings have
been reported by Glaus and Van Loon39 who observed that when
calcium hydroxide is used as a base GISAs are lost from the system
by mechanisms involving absorption onto solid calcium hydroxide
and via oxidation. Another reason for using sodium instead of cal-
cium is that when calcium hydroxide is reacted with cellulose
a
a
We were therefore interested to see if we could optimise this
sequence of steps with a view to providing a robust method for
isolating gram quantities of b-GISA which could be used for
determining its chemical and physical properties (Scheme 1).
approximately equal amounts of a-GISA and b-GISA are produced,
however, when sodium hydroxide was used as the base three
times more of the desired b-GISA is present compared to that of
the a-GISA.
Using the optimal reaction conditions (NaOH 0.5 M, 90 °C) the
reaction was complete after 16 h and heating for extended periods
of time did not increase the yield of GISAs. Interestingly, when a
similar reaction was performed but using microwave heating, it
was found that the reaction was complete after just 15 min. It is
known that the peeling reaction takes place at accessible reducing
chain ends and that these are normally within the amorphous
regions of cellulosic substrates. The rate of the peeling reaction is
reduced when the availability of chains end falls as amorphous
cellulose reacts leaving crystalline cellulose. Moharram and
Mahmooud40 have reported that microwave heating reduces the
time taken for the mercerization of cellulose; microwave heating
is suspected to promote the transformation of cellulose (I) to cellu-
lose (II) and the swelling of cellulose fibres. In the present system,
the swelling of the cellulose fibres would be expected to increase
the availability of the reducing ends and increase the rate of
production of the isosaccharinic acids.
2. Results and discussion
In our initial experiments, attempts were made to optimise the
yield of a- and b-GISA available from the treatment of micro-crys-
talline cellulose with a solution of sodium hydroxide. During these
optimisation reactions, a range of reaction conditions were varied
including: the reaction temperature (30, 50 and 90 °C), the length
of the reaction (1–52 days) and a limited number of provisional
experiments were undertaken using saturated solutions of calcium
hydroxide. The progress of reactions was monitored using high
performance anion exchange chromatography coupled with pulsed
amperometric detection (HPAEC-PAD) and the maximum solution
concentration of a- and b-GISA acids was obtained when a suspen-
sion of cellulose was treated with a solution of sodium hydroxide
(0.5 M) at 90 °C for 16 h under an atmosphere of nitrogen (Fig. 1).
Historically, calcium hydroxide has been used as a catalyst for
the synthesis of
a-GISA. It has been reported that calcium pro-
motes the benzylic acid rearrangement3 of the 4-deoxy-3-oxo-
D
-
Unfortunately, the use of microwave heating does not increase
the yield; using conventional and microwave heating the final con-
centrations of the two acids were very similar with the b-GISA
being present in larger amount (8.1 gLꢀ1 (2) & 2.7 gLꢀ1 (3)). The
glycero-2-hexulose intermediate (1) at the same time it has also
been reported that the use of sodium hydroxide at elevated tem-
peratures (>90 °C) results in fragmentation of the dicarbonyl-com-
(1)
(2)
CO2Na
(3)
CO2Na
OH
CH2OH
O
O
H
NaOH (0.5M)
HO
H
H
CH2OH
H
OH
HOCH2
Cellulose
H
H
H
OH
+
90 oC, 16 h or
H
H
microwave 15 min.
OH
CH2OH
β-(gluco)Isosaccharinic Acid
CH2OH
α-(gluco)Isosaccharinic Acid
CH2OH
Amberlite-H+
O
C
O
C
HO
H
H
CH2OH
H
HOCH2
OH
H
H
H
O
O
CH2OH
CH2OH
(β-2L)
PhCOCl/Pryridine
DMAP
CO2Na
O
O
C
BzOCH2
C
HO
H
H
CH2OH
H
OH
i) NaOH (2M)
BzO
H
H
CH2OBz
H
OBz
H
H
H
+
ii) MeOH/Cat. NaOMe
O
O
CH2OH
CH2OBz
(4)
Scheme 1. Synthesis and isolation of the sodium salt of b-GISA (2).
CH2OBz
β-(gluco)Isosaccharinic Acid
(5)

8
P. B. Shaw et al. / Carbohydrate Research 349 (2012) 6–11
Figure 1. Time course for the production of b-GISA (2) expressed as a percentage of
the final yield, for the sodium hydroxide catalysed decomposition of cellulose using
conventional heating at 90 °C (diamonds) and microwave heating at reflux
(squares).
yield of the isosaccharinic acid reaches
a maximum after
approximately a quarter of the cellulosic material has reacted.
Under the conditions of the experiments being reported here, the
reaction slows and stops. It is well known that cellulose that has
been treated with alkaline solution becomes increasingly resistant
to further degradation and this is explained by the so called
stopping reactions: the number of available reducing chain ends
falls as chains enter inaccessible crystalline regions or as chains
are terminated by a stopping reaction in which meta-saccharinic
is generated at the chain end.
After 16 h, the crude reaction mixture was hot filtered to remove
unreacted cellulose and insoluble salts. The filtrate was passed
through a cation exchange resin, in its H+ form, in order to convert
the sodium salts of a-GISA and b-GISA into their free acids which,
under the prevailing conditions (pH <2) undergo a rapid cyclization
to form their corresponding isosacharino-1, 4-lactones.41
The isosaccharino-1,4-lactones were converted into their tri-
benzoate esters using the procedures described by Whistler and
BeMiller.30 The key step to isolating pure b-GISA was the separa-
tion of the two tribenzoate esters using normal phase chromatog-
raphy. The crude mixture of triesters was applied to a large silica
column and the products eluted using a linear gradient starting
from 95% hexane and 5% ether and slowly rising to 30% ether
(8 L) collecting 400 fractions (20 mL). NMR analysis identified that
Figure 3. X-ray captions. Top: Solid-state structure of b-GISA-triester (thermal
ellipsoids are shown at 50% probability level). Bottom: Solid-state structure of
GISA-triester (thermal ellipsoids are shown at 50% probability level).
a-
1
80.5
80.0
79.5
79.0
the early fractions contained pure
tions contained pure b-tribenzoate 4; a number of the intermediate
fractions contained mixtures of both - and b-tribenzoates. Frac-
a-tribenzoate 5 and the last frac-
1
a
tions containing single isomers were pooled and the solvent vol-
ume reduced to approximately one third to give crystals that
were suitable for X-ray analysis. The crystal structures confirmed
the cis and trans arrangement of the 2,5-hydroxymethyl groups
1
1
in the
a and b-GISA tribenzoate esters, respectively (Fig. 3).
178.5
Hydroxide catalyzed ring opening of the lactone ring followed
by debenzoylation using Zemplén42 methods afforded the sodium
salt of the GISAs in greater than 95% purity.
178.0
The ability of b-GISA to complex metals will be influenced to a
large extent by the basicity of the carboxylate group. Determina-
tion of the solution state properties of isosaccharinic acids is
complicated by the fact that in acidic solution (pH <4) they
undergo acid catalysed ring closure to form cyclic lactones and this
prevents the direct measurement of the pKa by acid–base
titration.43 In order to determine the pKa of b-GISA we have mea-
sured the variation in the chemical shift of the carbons in b-GISA
1
77.5
2
3
4
5
6
7
8
9
pD
Figure 2. Variation of the 13C chemical shifts of b-GISA’s carbonyl carbon as a
function of pD. Solid line represents the fit to the equation:
dðpDÞ ¼ dacid þ dbase½10ꢀpD=ðKa þ 10ꢀpDÞꢁ giving a pKa in D2O of 4.01 0.03.

P. B. Shaw et al. / Carbohydrate Research 349 (2012) 6–11
9
as a function of pH using 13C NMR spectroscopy following the pro-
3.4. Determination of the aqueous pKa of b-GISA
cedures described by Cho et al.43 Between pD 2 and pD 8 the res-
onance position of the carbonyl carbon follows a sigmoidal
dependence on the hydrogen ion concentration (Fig. 2) with an
apparent pKa in D2O of 4.01. The measured pKa (H2O) of
3.61 0.03 (pKa D2O—0.4) is higher than the value reported in
The calcium salt of b-GISA (30 mg, 7.5 mmol) was dissolved in
D2O (1 mL) and the pH of the solution was varied between pD 1
and 8 using either DNO3 or NaOD and reporting pD as the mea-
sured pH +0.4. Samples were quickly transferred to NMR tubes
the literature for the
same technique and this is consistent with the elution of the b-
GISA after the -GISA in the HPAEC. These results suggest that b-
GISA anion is more basic than the -GISA anion. It should be noted
-GISA have been reported using
a
-GISA (3.27–3.36)43 measured using the
and spectra were recorded on a Bruker Avance 500 MHz
(125 MHz 13C) spectrometer using quantitative carbon pulse pro-
grammes. Each spectrum was recorded using 4000 scans employ-
a
a
ing a wait time of 12 s between scans. A capillary insert,
that higher values for the pKa for
a
containing CDCl3, was used as an internal reference. The solution
pH was measured in the NMR tube using a Beckman ø40 pH Meter
(High Wycombe, UK) in combination with a Hanna instruments
glass long reach NMR pH electrode (Mannheim, Germany). The
pKa of the acid was determined using the Henderson–Hasselbach
equation: pKa = pD +log{[AD]/[Aꢀ]}
potentiometric methods.41
In summary, we have presented an optimised method for the
synthesis and isolation of pure b-GISA and reported the pKa of
the acid group. It is hoped that these procedures will make b-GISA
readily accessible and lead to studies of its complexation reactions
with metals and with tetravalent actinides.
3.5. X-ray structure determination
3. Experimental
Single crystal X-ray diffraction data were collected on a Bruker
Apex Duo diffractometer equipped with a graphite monochromat-
ed Mo(Ka) radiation source. Preliminary scans were employed to
3.1. Materials and chemicals
assess crystal quality, lattice symmetry, ideal exposure time etc.
prior to collecting a full sphere of diffraction intensity data using
SMART operating software (SMART Diffractometer Control Software,
Bruker Analytical X-ray Instruments Inc., Madison, WI, 1998).
Intensities were then integrated from several series of exposures
All organic solvents were dried before use, employing standard
methods. The solvents used for column chromatography were GPR
grade. Microcrystalline cellulose-Avicel PH-101 (Lot BCBB5909)
was purchased from Fluka Analytical, all other reagents were pur-
chased from Aldrich and were used without further purification.
(each exposure covering 0.3° in
x), merged and corrected for Lor-
entz and polarisation effects using SAINT software (SAINT Integra-
tion Software, Siemens Analytical X-ray Instruments Inc.,
Madison, WI, 1994). Solutions were generated by direct methods
and refined by full-matrix non-linear least squares on all F2 data,
using SHELXS-97 and SHELXL software, respectively (as imple-
mented in the SHELXTL suite of programmes (SHELXTL Program
System, Vers. 5.1, Bruker Analytical X-ray Instruments Inc., Madi-
son, WI, 1998)). Empirical absorption corrections were applied
based on multiple and symmetry-equivalent measurements using
SADABS (Sheldrick, G. M. SADABS: A Program for Absorption Correc-
tion with the Siemens SMART System, University of Göttingen, Ger-
many, 1996). All structures were refined until convergence (max
shift/esd <0.01) and in each case, the final Fourier difference map
showed no chemically sensible features.
3.2. General methods
Analytical TLC was performed on Silica Gel 60-F254 (Merck)
and detection with either charring (lactones) following immersion
in 5% H2SO4/H2O and/or fluorescence. NMR spectra were recorded
on either a Bruker Avance 400 MHz or 500 MHz spectrometer
using Bruker pulse sequences, samples were dissolved in either
D2O or CDCl3 and referenced to acetone (1H) or internal CDCl3-
(capillary inset in D2O samples). Chemical shifts are given in parts
per million. High resolution mass spectra (HRMS) were recorded
on a Bruker micrOTOF-Q with an ESI interface and operating in po-
sitive ion mode.
3.3. Monitoring production of GISA
3.6. Preparation of 2,5,6-tri-O-benzoyl-2-hydroxymethyl-
2R,4S,5-trihydroxypentanoic acid-1,4-lactone (4) & 2,5,6-tri-O-
benzoyl-2-hydroxymethyl-2S,4S,5-trihydroxypentanoic acid-
1,4-lactone (5)
The production of the a- and b-GISA in the alkali (sodium or cal-
cium hydroxide) catalysed decomposition of cellulose was moni-
tored using high performance anion exchange chromatography
using the procedures described by Glaus el al.36 At appropriate
intervals (each hour for reactions performed using conventional
heating and every minute for the reaction performed with micro-
wave heating), small aliquots (5 mL) were removed from the reac-
tion mixture. These samples were then centrifuged (25,000 g,
30 min) and a small portion of the supernatant liquid was diluted
Avicel (200 g dry weight) was suspended in an aqueous solution
of sodium hydroxide (2 L, 0.5 M) in a large round bottom flask (3 L)
and an atmosphere of nitrogen was maintained above the reaction
solution; the reaction mixture was stirred whilst heating at 90 °C
for 16 h during which time the appearance of the suspension
turned dark brown. After 16 h the reaction mixture was filtered
hot to remove residual insoluble cellulose (142 g dry weight—
29% conversion to soluble products) and the supernatant liquid,
(ꢂ40) with an aqueous solution containing
D-ribonic acid as an
internal standard (5 mg Lꢀ1). The resulting solutions were then
analysed by high performance anion exchange chromatography
using pulsed amperometric detection on a Dionex 3000 Ion chro-
matography system (Dionex, Camberly, UK) employing a Dionex
containing the crude mixture of the sodium salts of a- and b-GISA
and impurities, was passed in several portions through a cation ex-
change column (Amberlite, 6 ꢂ 40 cm) which had previously been
conditioned with dilute HCl (1 M, 6 bed volumes). NMR analysis of
the material eluting from the amberlite column indicated that in
addition to the exchange of Na+ for H+ the free acids had also been
converted to their corresponding lactones (data not shown). After
removal of water under vacuum at 40 °C, a crude mixture of prod-
ucts (57 g-estimated to contain approximately 18 g of GISAs based
on analysis of the crude product by HPAEC 0.13 mol, 10.4% yield)
Carbopac PA20 column (3 ꢂ 150 mm, 6.5
lm particle size) and
eluting with aqueous sodium hydroxide (50 mM) operating at a
flow rate of 0.5 mL minꢀ1. The column temperature was main-
tained at 25 °C. The isosacharinic acids were detected using a
pulsed amperometric detector using a gold working electrode
and a Ag/AgCl reference electrode. The retention times of two iso-
saccharinic acids were determined by comparison with analytical
standards; see below for the preparation of standards.

10
P. B. Shaw et al. / Carbohydrate Research 349 (2012) 6–11
was obtained as a dark brown solid which was used directly in the
next synthetic step.
F(000) = 992; crystal dimensions 0.60, 0.20, 0.10 mm;
(MoK ) = 0.71073 mmꢀ1, T = 296.(2) K. A total of 10405 reflec-
l
a
To convert the lactones into their corresponding tribenzoate
esters, the crude mixture was dissolved in pyridine (300 mL) a
catalytic amount of 4-N,N-(dimethylamino)pyridine (1 g,
8.2 mmol) was added and the solution was cooled to 0 °C before
slowly adding a large excess of benzoyl chloride (150 g, 1.2 mol).
The mixture was stirred for 2 h at 0 °C and then at room tempera-
ture for a further 16 h. After 16 h the reaction mixture was poured
onto an equal volume of water and the mixture stirred at room
temperature for a further 1 h during which time the excess benzoyl
chloride hydrolysed. The tribenzoate esters were extracted into
chloroform (3 ꢂ 50 mL) and excess benzoic acid was removed from
the chloroform layer by repeated extraction with a saturated solu-
tion of sodium hydrogen carbonate (3 ꢂ 50 mL). The chloroform
layer was dried over anhydrous sodium sulfate and a large number
of the highly coloured polar impurities were removed from the es-
ters by passing the chloroform extract through a thin-bed of silica
(6 cm deep ꢂ 20 cm od) and washing the bed with chloroform
(100 mL). No attempt was made to recover the coloured impurities.
Subsequent evaporation of the combined chloroform layers gave a
pale-yellow syrup (27 g) which was shown by 1H NMR to be rich in
the tribenzoate esters.
tions measure in the range 2 ꢃ h ꢃ 27.87 (hkl range indexes:
ꢀ16 ꢃ h ꢃ 17, ꢀ25 ꢃ k ꢃ 25 ꢀ13 ꢃ l ꢃ 11), 4784 unique reflections
(Rint = 0.0347). The structure was refined to F2 to Rw = 0.0864,
R = 0.0381 (3619 reflections with I >2r
(I)) and GOF = 1.014 on F2
for 316 refined parameters, 2 restraints. Largest peak and hole
0.155 and ꢀ0.131 eAꢀ3. The X-ray coordinates have been lodged
with the Cambridge Crystallographic Data Centre (CCDC ID:
822401).
Crystal data for
M = 474.45, triclinic P-1, a = 9.7994(9), b = 10.4549 (10),
c = 11.7106 (11) Å, = 74.014 (2), b = 81.033 (2), = 86.059 (2)°,
V = 1138.86 (19), Z = 2,
calc = 1.384 Mg mꢀ3, F(000) = 496; crystal
dimensions 0.08, 0.06, 0.05 mm;
(MoK ) = 0.71073 mmꢀ1
a-GISA-triester (2nd structure) C27H22O8:
a
c
q
l
,
a
T = 162.(2) K. A total of 16452 reflections measure in the range
1.83 6 h 6 25.68 (hkl range indexes: ꢀ11 6 h 6 11, ꢀ12 6 k 6 12
ꢀ14 6 l 6 14), 4310 unique reflections (Rint = 0.0347). The struc-
ture was refined to F2 to Rw = 0.0849, R = 0.0378 (4310 reflections
with I >2r
(I)) and GOF = 1.010 on F2 for 316 refined parameters,
0 restraints. Largest peak and hole 0.195 and ꢀ0.204 eAꢀ3. The X-
ray coordinates have been lodged with the Cambridge Crystallo-
graphic Data Centre (CCDC ID: 822402).
The key step to isolating pure b-gluco-isosaccharinic acid was
the separation of the two tribenzoate esters using normal phase
chromatography. Part of the crude mixture (16 g) was applied to
a large silica column (50 ꢂ 7 cm) and the products were eluted
using a linear gradient starting with 95% hexane and 5% ether
slowly rising to 30% ether (8 L) collecting 400 fractions each of
approximately 20 mL. NMR analysis identified that a number of
3.8. Preparation of sodium (2R,4S)-2,4,5-trihydroxy-2-
(hydroxymethyl)pentanoate (b-gluco-isosaccharinic acid b-
GISA, 2)
Deprotection of the esters was accomplished in two stages: ring
opening of lactone was achieved by dissolving the tribenzoylated
lactone (4, 12 g) in an acetonitrile water mix (100:30 v/v) to which
was added a solution of sodium hydroxide (2 M) maintaining the
pH between 12 and 13. After the consumption of one equivalent
of base the pH of the solution was brought back to pH 7 and the
solvent removed under vacuum to give a solid product, this was di-
rectly dissolved in methanol (500 mL) and a catalytic amount of
sodium methoxide (1 g) was added to remove the benzoate groups.
After stirring for 2 h, the methanol was evaporated and the solid
product was dissolved in water and extracted with ether to remove
methyl benzoate. Removal of water under vacuum generated the
sodium salt of b-GISA in greater than 95% purity (4.79 g, 94%).
1H NMR (400 MHz, D2O) 3.69 ppm (m 1H) 3.56 ppm (d, 1H,
J = 11.5) 3.43 ppm (d 1H J = 11.5) 3.37 ppm (dd 1H, J = 11.7 Hz,
3.8 Hz) 3.26 ppm (dd 1H, J = 11.7 Hz, 6.8 Hz) 1.68 ppm (dd 1H,
J = 14.7 Hz, 3.5 Hz) 1.68 ppm (dd 1H, J = 14.7 Hz, 8.5 Hz). 13C NMR
(100 MHz D2O) 180.2, 80.1, 69.5, 67.8, 65.9, 37.5.
the early fractions contained pure
a-tribenzoate 5 (<1 g, tubes
225–260), these were closely followed by mixed fractions contain-
ing increasing amounts of the b-tribenzoate 3 (<2 g, tubes 261–
295), the third set of fractions contained only the b-tribenzoate 4
(<7 g, tubes ꢀ296–320, 14.7, 24%) a final fraction contained
b-tribenzoate 4 contaminated with increasing amounts of the ben-
zoate ester of 3,4-dihydroxybutanoic acid-1,4-lactone (<5 g).
3.6.1. 2,5,6-Tri-O-benzoyl-2-hydroxymethyl-2R,4S,5-
trihydroxypentanoic acid-1,4-lactone (4)
1H NMR (400 MHz, CDCl3) 8.17–8.04 ppm (m, 6H) 7.72–
7.60 ppm (m, 3H,) 7.52–7.44 ppm (m, 6H) 5.00–4.94 ppm (1H, m)
4.92 ppm (1H, d, J = 11.1 Hz) 4.73 ppm (1H, d J = 11.1 Hz) 4.74–
4.65 (m 2H) 2.56 ppm (d, J = 8.1 Hz). 13C NMR (100 MHz CDCl3)
171.7, 166.3, 165.7, 164.8, 134.1, 133.9, 133.4, 130.1, 130.0,
129.8, 129.3, 128.8, 128.7, 128.5, 128.3, 78.7, 74.4, 65.6, 65.4, 32.2.
High Resolution Mass Spectrometry (HRMS) (ESI) Exact mass
calculated for C27H22O8 [MNa+] 497.1213 found 497.1222.
HRMS (ESI) Exact mass calculated for C6H12O6[MꢀH]ꢀ 179.0561
found 179.05.
3.6.2. 2,5,6-Tri-O-benzoyl-2-hydroxymethyl-2S,4S,5-
3.9. Preparation of 2-hydroxymethyl-2R,4S,5-
trihydroxypentanoic acid-1,4-lactone (b-2L)
trihydroxypentanoic acid-1,4-lactone (5)
1H NMR (400 MHz, CDCl3) 8.11–8.03 ppm (m, 6H) 7.67–
7.56 ppm (m, 3H,) 7.52–7.42 ppm (m, 6H) 5.42–5.34 ppm (1H, m)
4.91 ppm (1H, d, J = 11.8 Hz) 4.68 ppm (1H, d J = 11.8 Hz) 4.64
(dd 1H J = 12.3 Hz, 3.4 Hz) 4.51 (dd 1H J = 12.3 Hz, 6.4 Hz)2.82 ppm
A pure sample of the b-lactone was prepared by passing the so-
dium salt of (2R,4S)-2,4,5-trihydroxy-2-(hydroxymethyl)pentano-
ate (1 g, 5 mmol) through a cation exchange column (Amberlite,
6 ꢂ 40 cm) which had previously been conditioned with dilute
HCl (1 M, 6 bed volumes).The crude product was extracted into
ether, dried with anhydrous sodium sulfate and the solvent re-
moved under vacuum to give the lactone as a sticky clear liquid
(0.8 g).
(dd 1H J = 15.0 Hz, 8.6 Hz) 2.61 ppm (dd 1H J = 15.0 Hz, 7.0 Hz) 13
C
NMR (100 MHz CDCl3) 172.1, 166.3, 165.9, 165.6, 134.4, 134.0,
133.7, 130.3, 130.0, 129.9, 128.9 128.7, 78.6, 75.6, 66.2, 65.5, 32.9.
HRMS (ESI) Exact mass calculated for
497.1213 found 497.1170.
C
27H22O8 [MNa+]
1H NMR (400 MHz, CDCl3) 4.52 ppm (m 1H) 3.72 ppm (dd, 1H,
J = 13.0 Hz, 2.6 Hz) 3.61 ppm (d, 1H J = 11.7 Hz) 3.49 ppm (dd, 1H
J = 13.0 Hz, 5.6 Hz) 2.43 ppm (dd 1H, J = 13.5 Hz, 6.7 Hz) 1.98 ppm
(dd 1H, J = 13.5 Hz, 9.2 Hz). 13C NMR (125 MHz D2O) 179.8, 79.0,
77.0, 64.9, 64.9, 62.7, 34.0.
3.7. X-Ray data for b-GISA-triester and a-GISA-triester
Crystal data for b-GISA-triester (1st structure) C27H22O8: M =
474.45, monoclinic Cc, a = 13.581(5), b = 19.487(7), c = 10.066(7)
Å, b = 118.715(6)°, V = 2336.(2), Z = 4,
q ,
calc = 1.349 Mg mꢀ3

P. B. Shaw et al. / Carbohydrate Research 349 (2012) 6–11
11
HRMS (ESI) Exact mass calculated for C6H10O5 [MNa+] 185.0426
found 185.0474.
17. Trinh, M. C.; Florent, J. C.; Monneret, C. J. Chem. Soc., Chem. Commun. 1987,
615.
18. Trinh, M. C.; Florent, J. C.; Monneret, C. Tetrahedron 1988, 44, 6633.
19. Bertounesque, E.; Millal, F.; Meresse, P.; Monneret, C. Tetrahedron: Asymmetry
1998, 9, 2999.
3.10. Preparation of a crude mixture of (gluco)isosaccharinic
acids using microwave heating
20. Kim, J.; Hong, J. H. Carb. Res. 2003, 338, 705.
21. Thomassigny, C.; Bennis, K.; Gelas, J. Synthesis 1997, 1997, 191.
22. Glaus, M. A.; Van Loon, L. R. Environ. Sci. Technol. 2008, 42, 2906.
23. Glaus, M. A.; Van Loon, L. R.; Schwyn, B.; Vines, S.; Williams, S. J.; Larsson, P.;
Puigdomenech, I. Sci. Basis Nucl. Waste Manage. XXXI 2008, 1107, 605.
24. Van Loon, L. R.; Glaus, M. A.; Laube, A.; Stallone, S. J. Environ. Polym. Degrad.
1999, 7, 41.
25. Van Loon, L. R.; Glaus, M. A.; Laube, A.; Stallone, S. Radiochim. Acta 1999, 86,
183.
26. Van Loon, L. R.; Glaus, M. A. J. Environ. Polym. Degrad. 1997, 5, 97.
27. IAEA Low and Intermediate Level Waste Repositories: Socioeconomic Aspects
and Public Involvement, International Atomic Energy Agency, Vienna, 2007,
IAEA-TECDOC-1553, pp 1–152.
In a further attempt to optimise production of the cellulose
decomposition products, a small scale reaction, 20 g of Avicel and
200 mL of 0.5 M NaOH, was heated using microwave heating in
an adapted domestic microwave fitted with a condenser and an
insulated magnetic stirrer bar (750 W). The production of
a- and
b-GISA was monitored using HPAEC-PAD and the reaction was
complete after 15 min.
28. Rai, D.; Rao, L. F.; Moore, D. A. Radiochim. Acta 1998, 83, 9.
29. Rai, D.; Yui, M.; Moore, D. A.; Rao, L. J. Solution Chem. 2009, 38, 1573.
References
30. Whistler, R. L.; BeMiller, J. N.
a-D-Isossacharino-1,4-Lactone In Methods in
1. Davidson, G. F. J. Text. Inst., Trans. 1932, 23, T95.
2. Richtzenhain, H.; Lindgren, B. O.; Abrahamsson, B. A.; Holmberg, K. Sven.
Papperstidn. 1954, 57, 363.
3. Machell, G.; Richards, G. N. J. Chem. Soc. 1960, 1938.
4. Isbell, H. S. J. Res. Natl. Bur. Stand. 1944, 32, 45.
5. Whistler, R. L.; BeMiller, J. N. J. Am. Chem. Soc. 1959, 82, 3705.
6. Lindstrom, L. A.; Samuelson, O. Acta Chem. Scand. B 1977, 31, 479.
7. Knill, C. J.; Kennedy, J. F. Carbohydr. Polym. 2003, 51, 281.
8. Green, J. W.; Pearl, I. A.; Hardacker, K. W.; Andrews, B. D.; Haigh, F. C. Tappi
1977, 60(10), 120.
9. Niemelä, K.; Alén, R. In Sjöström, E., Alén, R., Eds.; Analytical Methods in Wood
Chemistry, Pulping, and Papermaking; Springer: Berlin, 1999; p 193.
10. Sjöström, E. Tappi 1977, 60(9), 151.
Carbohydrate Chemistry, Vol. 2 Reactions of Carbohydrates; Wolfrom, M. L.,
BeMiller, J. N., Eds.; Academic Press: New York, 1963; pp 477–479.
31. Whistler, R. I.; BeMiller, J. N. Adv. Carbohydr. Chemi. Biochem. 1958, 13, 289.
32. Whistler, R. L.; BeMiller, J. N. J. Org. Chem. 1961, 26, 2886.
33. Wieland, E.; Tits, J.; Dobler, J. P.; Spieler, P. Radiochim. Acta 2002, 90, 683.
34. Whistler, R. L.; Medcalf, D. G. J. Org. Chem. 1962, 27, 3560.
35. Feast, A. A. J.; Lindberg, B.; Theander, O. Acta Chem. Scand. 1965, 19, 1127.
36. Glaus, M. A.; van Loon, L. R.; Achatz, S.; Chodura, A.; Fischer, K. Anal. Chim. Acta
1999, 398, 111.
37. Greenfield, B. F.; Harrison, W. N.; Robertson, G. P.; Somers, P. J.; Spinder, W. W.
Mechanistic Studies on The Alkaline Degradation of Cellulose in Cement. In
AEA-D&R-0219; AEA Technology: Harwell, 1993; pp 1–31.
38. Richards, G. N.; Sephton, H. H. J. Chem. Soc. 1957, 4, 4492.
39. Glaus, M. A.; VanLoon, L. R. Chemical Reactivity of alpha-Isosaccharinic Acid in
Heterogeneous Alkaline Systems. In PSI Bericht Nr 08-01; Paul Scherrer Institut:
Villigen, Switzerland, 2009; pp 1–77.
11. Svensson, M.; Berg, M.; Ifwer, K.; Sjöblom, R.; Ecke, H. J. Hazard. Mater. 2007,
144, 477.
12. Vercammen, K.; Glaus, M. A.; Van Loon, L. R. Radiochim. Acta 1999, 84, 221.
13. Vercammen, K.; Glaus, M. A.; Van Loon, L. R. Radiochim. Acta 2001, 89, 393.
14. Warwick, P.; Evans, N.; Hall, T.; Vines, S. Radiochim. Acta 2003, 91, 233.
15. Warwick, P.; Evans, N.; Hall, T.; Vines, S. Radiochim. Acta 2004, 92, 897.
16. Gaona, X.; Montoya, V.; Colas, E.; Grive, M.; Duro, L. J. Contam. Hydrol. 2008,
102, 217.
40. Moharram, M. A.; Mahmoud, O. M. J. Appl. Polym. Sci. 2008, 107, 30.
41. Ekberg, S.; Ekberg, C.; Albinsson, Y. J. Solution Chem. 2004, 33, 465.
42. Zemplén, G.; Kunz, A. Ber. Dtsch. Chem. Ges. 1923, 56, 1705.
43. Cho, H.; Rai, D.; Hess, N. J.; Xia, Y. X.; Rao, L. F. J. Solution Chem 2003, 32,
691.