Carbon exchange in hot alkaline degradation of glucose

Article abstract of DOI:10.1021/jo025912t

The decomposition of 1-¹³C-D-glucose, 6-¹³C-D-glucose, and 1-¹³C-sodium lactate has been studied in hot (145 ± 3 °C) alkaline (3.5 M) sodium hydroxide solution in order to understand the mechanisms of carbon exchange in th

Full text of DOI:10.1021/jo025912t

Ca r bon Exch a n ge in Hot Alk a lin e Degr a d a tion of Glu cose
Amanda V. Ellis† and Michael A. Wilson*,‡
Department of Chemistry, Materials and Forensic Science, University of Technology Sydney, P.O. Box
123, Broadway, NSW 2007, Australia, and Office of the Dean, College of Science Technology and
Environment, University of Western Sydney, Locked Bag 1797, Penrith South DC, NSW 1797, Australia
ma.wilson@uws.edu.au
Received May 7, 2002
The decomposition of 1-13C-D-glucose, 6-13C-D-glucose, and 1-13C-sodium lactate has been studied
in hot (145 ( 3 °C) alkaline (3.5 M) sodium hydroxide solution in order to understand the
mechanisms of carbon exchange in the alkaline degradation of glucose. The results show that in
the formation of lactate from glucose the carboxylate (COO-) carbon is formed preferentially from
C1 carbons but methyl (CH3) carbon is formed preferentially from C6 carbons. However, on further
decomposition of lactate to ethanol and carbonate, 13C-labeled carboxylate (COO-) is scrambled
equally among carbonate and both carbons in product ethanol molecules. In the production of
glycolate, the labeled C1 carbon mainly ends up as carboxylate (COO-) carbon, while for C6-labeled
glucose the labeled carbon mainly ends up as alcoholic (CH2OH) carbon. In the production of acetate
and formate there is also discrimination between C1 and C6 label.
In tr od u ction
acetate.1-5,11,12 However, the mechanism of formation of
these and other compounds has not been established.
Thus in this work we use 13C-labeled glucose to follow
decomposition pathways using 13C NMR spectroscopy.
Harris11 used the same technique to establish prelimi-
nary mechanistic results at different temperatures and
concentrations not related to Bayer processing. The
results described here are more detailed and quite
different and demonstrate a change in mechanism under
different conditions.
Although the decomposition mechanism of glucose in
alkaline solution has been known for a long time,1-11 little
is known of the pathway by which products are formed
in highly alkaline solution (3.5 M) at high temperatures
(140-150 °C), yet knowledge of this process is important
in industrial processing of plant derived material, for
example, in the Bayer process for separating alumina in
bauxite. In this process organic matter in the bauxite is
known to play a significant role in affecting crystal
growth and quality. Glucose from carbohydrates in plant
detritus plays an important role.12
Resu lts a n d Discu ssion
Isotop e Exch a n ge. It is well-known13 that the con-
centration of the noncyclic open form of glucose and its
furanoic forms in water are negligible, and that equilib-
rium values are 64% â and 36% R for the pyranose forms,
which is confirmed here, (2% for the labeled isomers.
Table 1 shows the 13C distributions of all the products
identified from digestion of unlabeled and labeled 1-13C-
D-glucose and 6-13C-D-glucose after 1 h digestion at 145
°C in 3.5 M NaOH. Table 1 also shows enhanced yields
of 13C in these compounds relative to that in the original
unlabeled mixture of products. It is clear that these
values are greater than one, and that label from 1-13C-
D-glucose and 6-13C-D-glucose ends up in all carbons in
the products. It is shown in Table 1 that labeled lactate
is the predominant compound formed from both 1-13C-
D-glucose and 6-13C-D-glucose under Bayer simulated
laboratory conditions. The labeling occurs at all three
lactate carbons but in different proportions.
The major products from hot alkaline degradation of
glucose are lactic acid, formate, glycolic acid, and
* Author to whom correspondence should be addressed. Phone: 61-
2-4570-1257. Fax: 61-2-4570-1403.
† University of Technology Sydney.
‡ University of Western Sydney.
(1) Yang, B. Y.; Montgomery, R. Carbohydr. Res. 1996, 280, 27-
45.
(2) Luijkx, G. C. A.; van Rantwijk, F.; van Bekkum, H.; Antal, M.
J ., J r. Carbohydr. Res. 1995, 272, 191-201.
(3) de Bruijn, J . M.; Touwslager, F.; Kieboom, A. P. G.; van Bekkum,
H. Starch 1987, 39, 49-52.
(4) de Bruijn, J . M.; Kieboom, A. P. G.; van Bekkum, H. Starch 1987,
39, 23-28.
(5) de Bruijn, J . M.; Kieboom, A. P. G.; van Bekkum, H.; van der
Poel, P. W.; de Visser, N. H. M.; de Schutter, M. A. M. Int. Sugar J .
1987, 89, 208-210.
(6) de Bruijn, J . M.; Kieboom, A. P. G.; van Bekkum, H.; van der
Poel, P. W. Sugar Technol. Rev. 1986, 13, 21-52.
(7) de Wit, G.; Kieboom, A. P. G.; van Bekkum, H. Recl. Trav. Chim.
Pay-Bas 1979, 98, 355-361.
(8) de Wit, G.; Kieboom, A. P. G.; van Bekkum, H. Tetrahedron Lett.
1975, 45, 3943-3946.
The percentage distributions of the label in a given
compound are readily calculated. For 1-13C-D-glucose the
production of C1-labeled lactate (49.5/(49.5 + 1.0 + 37.8)
(9) Sowden, J . C.; Blair, M. G.; Kuenne, D. J . J . Am. Chem. Soc.
1957, 79, 6450-6454.
(10) Sowden, J . C.; Pohlen, E. K. J . Am. Chem. Soc. 1958, 80, 242-
244.
(11) Harris, J . F. Carbohydr. Res. 1972, 23, 207-215.
(12) Ellis, A. V.; Wilson, M. A.; Kannangara, K. Ind. Eng. Chem.
Res. 2002, 41, 2842-2852.
(13) Libnau, F. O.; Christy, A. A.; Kvalheim, O. M. Vib. Spectrosc.
1994, 72, 139-148.
10.1021/jo025912t CCC: $22.00 © 2002 American Chemical Society
Published on Web 11/02/2002
J . Org. Chem. 2002, 67, 8469-8474
8469

Ellis and Wilson
TABLE 1. 13C Distr ibu tion (%) a n d En h a n ced Yield in P r od u cts fr om Un la beled Digested r-D-(+)-Glu cose a n d La beled
1-13C-D-Glu cose a n d 6-13C-D-Glu cose Digestion s in 3.5 M Na OH for 1 h a t 145 °C
unlabeled
R-D-(+)-glucose
digestion (1 h)
enhanced yielda
1-13C-D-glucose
digestion (1 h)
enhanced yielda
6-13C-D-glucose
digestion (1 h)
chemical
1-13C-D-glucose
digestion (1 h)
6-13C-D-glucose
digestion (1 h)
product
lactate
COO-
shift δ (ppm)
185.6
28.7
|
49.5
|
24.2
|
155.2
76.2
|
|
|
|
71.1
|
24.8
|
1.0
|
1.1
|
3.7
3.7
CHOH
|
|
|
23.2
21.6
37.8
33.7
157.9
140.5
CH3
glycolate
COO-
|
184.6
|
2.7
|
3.5
|
2.9
|
116.6
96.4
|
|
66.5
2.8
0.1
31.7
3.7
1019.2
CH2OH
acetate
COO-
|
184.1
|
0.1
|
0.1
|
0.1
|
90.0
|
90.0
|
26.3
0.1
1.3
0.4
1169.8
359.9
CH3
ethanol
CH2OH
|
57.0
|
3.0
|
0.1
|
0.1
|
2.8
|
2.8
|
17.6
3.2
0.3
0.1
8.3
2.8
CH3
formate
HCOO-
carbonate CO3
other products
173.8
170.9
0.1
1.3
0.5
1169.8
449.9
2-
9.7
3.2
0.4
4.6
0.4
4.8
3.7
129.5
3.7
135.0
a
Enhanced yield is defined as the yield of the functional group in 13C-labeled experiment (g/100 g)/yield in unlabeled experiment
(g/100 g) ) [(% distribution in labeled experiment (g/100 g) × fraction of label enhancement in glucose, 0.99)/(% distribution in unlabeled
experiment (g/100 g) × fraction of natural label enhancement (0.011 for 13C in natural glucose))].
SCHEME 1. Mech a n ism for Glu cose Clea va ge To F or m a Ca r ba n ion a n d a n Ald eh yd e
SCHEME 2. Tr iose-En ed iol â-Elim in a tion To F or m Dion es
) 56.1%) is favored over other lactate carbons while from
6-13C-D-glucose the production of C3-labeled lactate (57.1%)
is favored. The differences between labeled C1 and
labeled C3 lactate formation can be explained in terms
of isomerization as shown in Schemes 1 and 2. For C1-
labeled glucose the label is marked with the symbol f
and for C6-labeled glucose the label is marked with the
symbol b.
Initial cleavage of the glucose molecule (1) gives rise
to a 3-carbon carbanion (2) and an aldehyde (3). The
aldehyde 3 can further undergo enolization to 4 (Scheme
2), followed by isomerization. The carbanion 2 may pick
up a proton from water to also give 4 where the labels at
C1 and C6 then become synonymous. Species 4 may
isomerize to form 5 by picking up a proton â to the
original labeled carbon.
8470 J . Org. Chem., Vol. 67, No. 24, 2002

Carbon Exchange in Hot Alkaline Degradation of Glucose
SCHEME 3. F or m a tion of La beled La cta te fr om Dion es
SCHEME 4. Retr oa ld oliza tion of 1-13C-D-Glu cose
in to a La beled 2-Ca r bon Ca r ba n ion a n d a Tetr ose
SCHEME 5. F or m a tion of C2-La beled Glycola te
fr om C1-La beled Glu cose
In Scheme 2 the formation of 1,2-diones 6 and 7 from
4 and 5 involves â-elimination of water. The C3-labeled
intermediate dione 6 rearranges to give only C3- labeled
lactate (Scheme 3, reactions a and b). On the other hand
the C1-labeled dione 7 can rearrange to give either C2-
or C1-labeled lactate through either methyl (Scheme 3,
reaction c) or hydride transfer (Scheme 3, reaction d),
respectively. However, reaction 3c rarely occurs. The
completeness of the isomerization between the trioses 4
and 5 (see Scheme 2, Point A) is reflected in the ratio of
the sum of C1 and C2 to that of C3 label. For example,
for 1-13C-D-glucose this ratio is 1.34 ) [((56.1 + 1.1)/42.8)
or from Table 1, (49.5 + 1.0)/37.8].
SCHEME 6. Retr oa ld oliza tion of 6-13C-D-Glu cose
in to a 2-Ca r bon Ca r ba n ion a n d a La beled Tetr ose
carbanion, Route 4, is protonated then a diol is formed.
If the labeled aldehyde is formed by Route 5, then on
addition of water to the carbonyl group and loss of
hydrogen (Scheme 8, Route 6), the label appears on the
CH2OH group. This is clearly the predominant case for
the C6-labeled glucose and must be an important mech-
anism. Hydrogen has been demonstrated to be formed
readily in strong basic solution.14
For the 6-13C-D-glucose the labeled glycolate produced
is predominately at the C2 (CH2OH, 91.6%) with only
8.4% at the C1 position (COO-). 13C-labeled C2 units can
be formed from C2 carbon retroaldolization of C1-labeled
glucose (Scheme 4). The carbanion is then protonated and
rearrangement forms glycolate labeled at the C2 position
(Route 3, Scheme 5) by hydroxide attack at OH. Table 1
shows that from C1-labeled glucose the predominant form
of the label is at the carboxyl unit in the glycolate. That
is, Route 3 is important.
Interestingly, the acetate produced from 1-13C-D-
glucose and 6-13C-D-glucose has the 13C distribution
almost entirely on the CH3 group, 92.9% and 80.0%,
respectively. Depicted in Scheme 9 (Routes 7 and 8) is
For C6-labeled glucose the label would be on the tetrose
as shown in Scheme 6. This then undergoes further
reaction to give glycolate shown in Scheme 7. If the
(14) Pang, L. S. K.; Vassallo, A. M.; Wilson, M. A. Fuel 1989, 68,
253-254.
J . Org. Chem, Vol. 67, No. 24, 2002 8471

Ellis and Wilson
SCHEME 7. Decom p osition of th e Tetr ose To Give
Tw o 2-Ca r bon F r a gm en ts
Scheme 2. The ethanal would readily oxidize to acetic
acid via addition of water and loss of hydrogen. However,
these are probably not important routes to acetate. The
intermediate 4 yields only labeled CH3 (Route 7). The
intermediate 5 primarily forms labeled formate but not
acetate. It can produce labeled acetate but only if
hydroxide attack is at the unlabeled CdO carbon and
there is methyl migration. The results therefore testify
to the importance of methyl migration by Route 8, if
acetate can form this way. However, the scheme offers
no explanation as to why acetate labeled at COO- differs
in amount from C6- and C1-labeled glucose. It must be
true that reactions from 4 can occur which transfer
hydride and OH so that the labeled carbon becomes
carboxyl and then cleavage occurs. One possibility is
given in Scheme 10.
SCHEME 8. 13C In cor p or a tion a t Hyd r oxyl Ca r bon
in Glycola te
Route 8 yields labeled formate over Route 7 in Scheme
9, and this is reflected in the product distributions since
formate preferentially forms from C1-labeled glucose.
Labeled formate arising from C6-labeled glucose is also
found in significant amounts, although less than from
C1-labeled glucose. A mechanism for this is proposed in
Scheme 11.
Labeled studies were carried out on 1-13C-sodium-L-
lactate to establish the mechanism of formation of
carbonate and ethanol from lactate. Table 2 shows the
13C distribution both before and after digestion in 3.5 M
NaOH at 145 °C.
the formation of CH3-labeled ethanal (17) from two
different labeled propdienones (4 and 5) formed as in
Neither ethanol nor carbonate produced from 1-13C-D-
SCHEME 9. P r op osed Mech a n ism for F or m a te F or m a tion via 1,2-Dica r bon yl Clea va ge of Tr iosesa
a
The dual label symbol is dropped for convenience after depiction of 4 and 5 from Scheme 2.
8472 J . Org. Chem., Vol. 67, No. 24, 2002

Carbon Exchange in Hot Alkaline Degradation of Glucose
SCHEME 10. F or m a tion of Ca r boxyl-La beled Aceta te by Hyd r id e a n d OH Migr a tion
SCHEME 11. Clea va ge of La beled F or m a te fr om a n En d -La beled CH2OH Un it
TABLE 2. 13C Distr ibu tion in Un la beled Digested Sod iu m -L-(+)-la cta te a n d 1-13C-Sod iu m -L-la cta te befor e a n d a fter 1 h
Digestion in 3.5 M Na OH a t 145 ˚C
%
13C distribution
enhanced yielda
1-13C-sodium-L-lactate
digestion (1 h)
chemical
shift
δ (ppm)
original sodium-L-
(+)-lactate digestion (1 h)
1-13C-sodium-L-lactate
before digestion
1-13C-sodium-L-lactate
digestion (1 h)
lactate
COO-
185.6
31.7
|
96.9
|
96.8
|
274.5
|
|
|
71.1
|
30.9
|
1.5
|
1.2
|
3.7
|
CHOH
|
23.2
32.1
1.6
1.2
3.7
CH3
ethanol
CH2OH
57.0
|
1.5
|
0.3
|
18.4
|
not observed
not observed
|
17.6
1.6
0.3
16.5
CH3
2.2b
0.2b
8.3
2-
carbonate CO3
170.9
a
b
See legend to Table 1. Yield of carbonate is not equal to yields of ethanol carbons as some carbon dioxide is in the gas phase.
SCHEME 12. Hyd r id e a n d Hyd r oxyl Sh ifts for C1 a n d C3 Equ iva len ce in La cta te
glucose or 6-13C-D-glucose is selectively 13C enhanced.
Thus when lactate decomposes carbon is scrambled
among all positions.
This must be through hydride and OH shifts such as
that shown in Scheme 12 for C1 and C3 equivalence).
Similarly for C1 and C2 equivalence hydride, hydroxyl,
and methyl migration gives Scheme 13.
be the mechanism for such exchange. Decomposition of
lactate can be influenced by a number of factors. Agents
that inhibit hydride, hydroxy, or methyl transfer may be
effective since the stability of transition states to the
species shown will be sensitive to a number of reaction
conditions. It was thought that labeled carbonate would
arise from the decarboxylation of the C1-labeled carboxyl
group on the lactate. This is the case but the carbon is
equally spread through the ethanol formed as well
Both these include a rather improbable first step, and
a 1,2,3-cyclopropanetriol intermediate would appear to
J . Org. Chem, Vol. 67, No. 24, 2002 8473

Ellis and Wilson
SCHEME 13. Hyd r id e a n d Hyd r oxyl Sh ifts for C1 a n d C2 Equ iva len ce in La cta te
TABLE 3. Ch em ica l Sh ift (p p m ) Da ta a n d 13C-13C Cou p lin g Con sta n ts (J (Hz)) for La beled a n d Un la beled D-Glu cose
(0.416 M) in Distilled Wa ter
chemical shift δa
6-13C-R-D-glucose
chemical
shift δa
6-13C-â-D-glucose
chemical
shift δa
1-13C-R-D-glucose
chemical
shift δa
1-13C-â-D-glucose
chemical
shift δa
R-D-
â-D-
(+)-glucose
carbon
C1
(+)-glucose
J (Hz)
J (Hz)
J (Hz)
J (Hz)
94.70
74.16
75.45
72.25
73.99
63.27
98.52
76.85
78.49
72.13
78.40
63.42
94.70
74.16
75.45
72.29
73.99
63.26
13C6-13C1
J ) 2.75
13C6-13C2
J ) 0
98.51
76.84
78.44
72.25
78.47
63.41
13C6-13C1
J ) 3.75
13C6-13C2
J ) 0
94.80
74.14
75.45
72.29
73.99
63.28
98.51
76.83
78.47
72.25
78.41
63.43
C2
C3
C4
C5
C6
13C1-13C2
J ) 46.1
13C1-13C3
J ) 0
13C1-13C2
J ) 45.8
13C1-13C3
J ) 0
13C6-13C3
J ) 3.13
13C6-13C4
J ) 0
13C6-13C3
J ) 3.96
13C6-13C4
J ) 0
13C1-13C4
J ) 0
13C1-13C4
J ) 0
13C6-13C5
J ) 43.1
13C6-13C5
J ) 43.1
13C1-13C5
J ) 0
13C1-13C5
J ) 0
13C1-13C6
J ) 2.71
13C1-13C6
J ) 3.71
a
δ in ppm.
with nitrogen). All labeled compounds contained 99% 13C. Each
1.0-mL solution was placed in the 2.5-mL Parr reactor and
sealed under 5.2 × 105 Pa of nitrogen then heated in a silicon
oil bath at 145((3) °C. The temperature dropped to 120 °C
and then rose to 145 °C in 30 min. The reaction was then
carried out for 1 h from this time. After each digestion the
bomb was cooled for 5 min under a cold tap water.
An a lysis. Quantitative 13C NMR spectra were obtained
prior to digestion for 0.416 M solutions of unlabeled R-D-(+)-
glucose, unlabeled sodium-L-lactate, 1-13C-D-glucose, 6-13C-D-
glucose, and l-13C-sodium lactate in distilled water to check
purity and 13C distributions (refer to Table 3 for chemical shifts
(ppm) and 13C-13C coupling constants (J (Hz)). It should be
noted that R-D-(+)-glucose undergoes mutarotation and there-
fore there are 12 separate carbon resonances, 6 from the R
anomer and 6 from the â anomer in the pure solutions.
Quantitative inverse gated 13C NMR spectra were obtained
on a Bruker DRX300 instrument, operating at 75.4 MHz for
13C. A sample (400 µL) of each solution was added to a 5 mm
NMR tube and 0.3 M 3-(trimethylsilyl)-1-propanesulfonic acid
sodium salt in D2O (100 µL) was placed in a coaxial tube as
an internal reference (δ 0 ppm). A 45° pulse was used (pulse
width 2.50 µs). The experiment required 1536 transients with
a recycle delay of 30 s to ensure complete relaxation. A short
(0.14 s) acquisition time was used to minimize nuclear Over-
hauser effects (NOE). All data was collected in 5 k data points
and Fourier transformed using a line-broadening factor of 3
Hz. Quantitative 13C NMR spectra were obtained for solutions
both before and after alkaline digestion in 3.5 M NaOH.
Assignments of major product peak chemical shifts (ppm) are
formate 173.8, acetate 26.3, 184.1, lactate 23.2, 71.1, 185.6,
glycolate, 66.5, 184.6, carbonate 170.9, and ethanol 17.6, 57.0.
because of the rapid scrambling process, so that no
preference is found for carbonate or ethanol.
Mech a n ism . Rearrangements in aqueous alkaline
media have been demonstrated and show that alkaline
degradation reactions are a dynamic interconversion of
monosaccharides in which carboxylic acids such as
lactate, acetate, and formate form. The results show that
in the formation of lactate from glucose the carboxylate
(COO-) carbon is formed preferentially from C1 carbons
but methyl (CH3) carbon is formed preferentially from
C6 carbons. The process involves some exchange through
enolysis, hydride, and methyl migration. Once formed
label then becomes indiscriminately scrambled as lactate
decomposes. 1-13C-sodium-L-lactate studies showed that
13C-labeled carboxylate (COO-) is scrambled equally
among carbonate and both carbons in product ethanol
molecules.
The production of glycolate generally results in a high
degree of selectivity for carbon. From C1-labeled glucose,
labeled carbon ends up as carboxylate (COO-) carbon,
but from C6-labeled glucose the labeled carbon ends up
as alcoholic (CH2OH) carbon. A possible explanation is
the position of attack of hydroxy ion at carbonyl rather
than OH functionality.
Formate must arise through a range of end chain
carbon cleavage reactions and especially that from the 6
carbon of glucose. Dienone intermediate routes may be
possible but are not significant. The production of acetate
and formate also discriminates between 1-C- and 6-C-
labeled carbon.
Ack n ow led gm en t. We thank Alcoa Australia and
the Australian Research Council for financial support.
Exp er im en ta l Section
Alk a lin e Digestion of 13C-La beled 1-13C-D-Glu cose,
6-13C-D-Glu cose, a n d 1-13C-Sod iu m La cta te. Alkaline diges-
tions were carried out in a Parr reactor (25 mL) adjusted to
2.5 mL with a Teflon plug. Individual digestions were carried
out on 0.416 M solutions (1.0 mL) of unlabeled R-D-(+)-glucose,
unlabeled sodium-L-lactate, 1-13C-D-glucose, 6-13C-D-glucose,
and l-13C-sodium lactate in 3.5 M NaOH (previously purged
Note Ad d ed a fter ASAP P ostin g. In Table 3, two
coupling constants were reported incorrectly in the
version posted November 2, 2002; the corrected version
was posted November 5, 2002.
J O025912T
8474 J . Org. Chem., Vol. 67, No. 24, 2002