The Hofer-Moest decarboxylation of d-glucuronic acid and d-glucuronosides

Article abstract of DOI:10.1016/j.carres.2006.12.011

Research was undertaken to effect the oxidative decarboxylation of glycuronosides. Experiments with free d-glucuronic acid and aldonic acids were also executed. Both anodic decarboxylation and variants of the Ruff degradation reaction were investigated. Anodic decarboxylation was found to be the only successful method for the decarboxylation of glucuronosides. It was, therefore, proposed that glycuronosides can only undergo a one-electron oxidation to form an acyloxy radical, which decomposes to form carbon dioxide and a C-5 radical, that is, a Hofer-Moest decarboxylation. The radical is subsequently oxidized to a cation by means of a second one-electron oxidation. The cation undergoes nucleophilic attack from the solvent (water), whose product (a hemiacetal) undergoes a spontaneous hydrolysis to yield a dialdose (xylo-pentodialdose from d-glucuronosides).

Full text of DOI:10.1016/j.carres.2006.12.011

Carbohydrate Research 342 (2007) 610–613
Note
The Hofer–Moest decarboxylation of D-glucuronic
acid and D-glucuronosides
Jonathan A. Stapley and James N. BeMiller*
Whistler Center for Carbohydrate Research, Purdue University, 745 Agriculture Mall Drive, West Lafayette, IN 47906-2009, USA
Received 11 September 2006; received in revised form 15 December 2006; accepted 15 December 2006
Available online 21 December 2006
Dedicated to the memory of Professor Nikolay K. Kochetkov
Abstract—Research was undertaken to effect the oxidative decarboxylation of glycuronosides. Experiments with free D-glucuronic
acid and aldonic acids were also executed. Both anodic decarboxylation and variants of the Ruff degradation reaction were inves-
tigated. Anodic decarboxylation was found to be the only successful method for the decarboxylation of glucuronosides. It was,
therefore, proposed that glycuronosides can only undergo a one-electron oxidation to form an acyloxy radical, which decomposes
to form carbon dioxide and a C-5 radical, that is, a Hofer–Moest decarboxylation. The radical is subsequently oxidized to a cation
by means of a second one-electron oxidation. The cation undergoes nucleophilic attack from the solvent (water), whose product (a
hemiacetal) undergoes a spontaneous hydrolysis to yield a dialdose (xylo-pentodialdose from D-glucuronosides).
ꢀ 2007 Elsevier Ltd. All rights reserved.
Keywords: Decarboxylation; D-Glucuronic acids; D-Glucuronosides; xylo-Pentodialdose; Xylitol
von Euler et al.1 reported that uronoside polymers were
decarboxylated by Ruff’s reagent, but our attempts to
decarboxylate poly(galacturonic acid) using Ruff degra-
dation conditions2,3 were unsuccessful. These unsuccess-
ful attempts to apply the Ruff degradation to the
decarboxylation of uronic acid-containing substances
led us to consider the Hofer–Moest reaction4 (Chart 1)
for the oxidative decarboxylation of uronosides.
Both aldonic acids and uronosides were decarboxyl-
ated by the Hofer–Moest reaction.5 In the case of uron-
opyranosides, the ring oxygen atom is electron
withdrawing, and the resulting cation (C-5) is stabilized
as a carboxonium ion. The uronoside structure thus
facilitates the reactions of the Hofer–Moest type. Chart
2 presents the mechanism as applied to a D-glucurono-
pyranoside (I). In this scheme, intermediate II is a hemi-
acetal, which upon hydrolysis (ring opening), by the
concerted mechanism shown, produces III (xylo-pento-
dialdose) spontaneously. Electrolytic decarboxylation,
that is, the Hofer–Moest reaction, has been exploited in
the production of xylitol via reduction of xylo-dialdose
produced by the decarboxylation of glucuronosides.6
Using the crude apparatus described in the Experi-
mental section of this paper, methyl b-D-glucuronopyr-
anoside (aqueous solution) was converted into xylo-
pentodialdose at a rate of 1.5 · 10ꢀ3 mol/h at a current
efficiency of 1.8 kW mol/h. (Decarboxylations of so-
dium D-glucuronate and L-glucuronate in water, metha-
nol, and aqueous methanol were somewhat more facile.)
Kitagawa and Yoshikawa,7 in a review of uronic acid
glycoside cleavage, discussed the anodic decarboxyl-
ation of 2,3,4-tri-O-methyl-D-glucuronopyranosides in
methanol. Their reaction products and proposed mech-
anism are consistent with Chart 1 (R2 = Me). Francisco
et al.8 chemically decarboxylated similar starting materi-
als and reported analogous products, proposing the
same mechanism. Weiper and Scha¨fer9 reported the
decarboxylation of 2,3:4,5-di-O-isopropylidene-b-D-
arabino-hex-2-ulopyranosonic acid and 2,3:4,6-di-O-iso-
*
Corresponding author. Tel.: +1 765 494 5684; fax: +1 765 494
7953; e-mail: bemiller@purdue.edu
Present address: Dynamic Food Ingredients Corp., 127 10th Street,
Kirkland, WA 98033, USA.
0008-6215/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carres.2006.12.011

J. A. Stapley, J. N. BeMiller / Carbohydrate Research 342 (2007) 610–613
611
O
R2
h+
R2
O
O
h+
O
HO
R1
R1
R1
R1
R1
-
O
+
O
H+
CH
CH
X
+
X
X
C
O
X
X
Chart 1. Hofer–Moest decarboxylation; h+ symbolizes an anode.
h+
O-
O
C
O
O
h+
O
O
O
O
HO
HO
O
HO
HO
CH
HO
HO
OH
OH
O
OH
R
R
O
R
O
(I)
HO
H
H+
O
H
O
O
O
CH
HO
HO
HO
HO
HO
HO
O
OH
OH
OH
O
R
O
R
(III)
(II)
R
HO
Chart 2. Hofer–Moest decarboxylation of a D-glucuronopyranoside.
propylidene-b-L-xylo-hex-2-ulofuranosonic acid in metha-
nol to yield ortho-esters. Kitagawa et al.10–12 oxidized D-
glucuronic acid (IV) in methanol anodically and ob-
tained V (Chart 3), which was used as a synthetic inter-
the hydrogen peroxide is not decomposed as a result
of a Ruff degradation type of reaction.14
Both methyl b-D-glucuronopyranoside and methyl a-
D-glucuronopyranoside were subjected to the Cu(II)
variant (the most efficient varient14) of the Ruff degrada-
tion reaction to determine whether uronosides might
form stable complexes that are resistant to chemical
decarboxylation and if the anomeric oxygen atom might
participate in such complexes. In both cases, the reac-
tion times were further extended (250 and 238 min for
a complete consumption of hydrogen peroxide, respec-
tively). Even so, HLPC indicated that no change in the
carbohydrate occurred. It was thus determined that
the configuration of the anomeric carbon atom (orienta-
tion of the aglycon) was not the impediment to the reac-
tion of interest.
Colors of the reaction mixtures when Cu(II) was used
as the catalyst suggested that a decarboxylation-resis-
tant complex is formed. For example, equimolar solu-
tions of sodium D-gluconate, sodium D-glucuronate,
and the sodium salt of methyl D-glucuronopyranoside
containing equimolar concentrations of Cu(II) ion were
made; the pH of the solutions was adjusted to 7, and
hydrogen peroxide was added. Upon the addition of
hydrogen peroxide to the D-gluconate solution, it turned
blue and effervescence occurred immediately. The D-glu-
curonate solution was blue-green; there was no efferves-
cence upon addition of hydrogen peroxide. The methyl
b-D-glucuronopyranoside solution was yellow-green;
mediate. [In water, the ring of
V would open
(analogously to II ! III, the methoxyl of C-1 and the
hydroxyl of C-5 being reversed) yielding xylo-pentodial-
dose.] Renaud et al.13 reported that the electrolytic
decarboxylation of a-hydroxy acids in methanol yields
aldehydes.
First, we examined the efficacy of the standard [Fe(II)]
Ruff reagent for the decarboxylation of D-glucuronic
acid. HPLC chromatograms (not given) indicated that
no decarboxylation occurred, that is, that D-glucuronic
acid is resistant to decarboxylation by both Fe and Cu
variations of the reaction. Extended reaction times
(135 and 146 min for a complete consumption of hydro-
gen peroxide, respectively, more than 10 times the dura-
tion of the Cu(II)–D-gluconate reaction) suggested that
O-
CH3
O
O
O
HO
HO
O
HO
HO
OH
OH
OH
OH
(IV)
(V)
Chart 3. Methanolic Hofer–Moest decarboxylation of D-glucuronic
acid.8–10

612
J. A. Stapley, J. N. BeMiller / Carbohydrate Research 342 (2007) 610–613
O
dilution to ꢁ0.1 mg/mL with double distilled (dd) water
OH
H
and filtering through a 0.2-lm filter. A gradient of (a)
0.100 M sodium hydroxide and (b) 0.100 M sodium
hydroxide with 1.00 M sodium acetate was used for
elution.
OH
O
O
X
H
O
O
Cu
O
O
O
H
O
A 5890 Series II gas chromatograph (Hewlett–Pack-
ard Company, Palo Alto, CA) equipped with a J&W
DB-5 30 m · 0.250-mm, 0.25-lm column (Agilent Tech-
nologies, Palo Alto, CA) was used. Samples were pre-
pared by adding Tri-Sil reagent (200 lL, Pierce
Chemical Co., Rockford, IL) to approximately 0.1 mg
dried sample and heating 20 min at 80 ꢁC. Reaction mix-
tures were then evaporated to dryness with air, dissolved
in hexane (100 lL), dried again, and re-dissolved in
hexane (100 lL).
X
HO
OH
H
O
Chart 4. Potential structure of a mononuclear Cu(II) glucuronoside
complex.20
neither did it effervesce upon the addition of hydrogen
peroxide.
Uronates form complexes quite dissimilar from those
of aldonates.14 While hexuronic acids have a hydroxyl
group adjacent to their carboxylic acid group, that
a-hydroxyl group (C-5) becomes the endocyclic oxygen
atom when the pyranose ring forms. The pyranose ring
form is favored in aqueous solutions but is in constant
equilibrium with its open chain state, as copper ions
catalyze mutarotation.15 Even so, no Ruff degradation
occurs.
While there is conflicting data on D-glucuronate–
Cu(II) systems,16–20 there is consensus that D-glucuro-
nate forms a complex with Cu(II) in such a manner that
only the carboxylate group participates as a ligand.
There is also a substantial amount of similar evidence
that the endocyclic oxygen atom of the pyranose ring,
but no hydroxyl groups, participate in the coordination,
forming a complex similar to the salt bridges formed
when polyuronates gel (Chart 4).
The observations that the Ruff degradation is success-
ful on all aldonic acids and unsuccessful on uronic acids
and their glycosides suggest that uronic acids and their
glycosides cannot form the M1L2Hꢀ2 complex necessary
for decarboxylation.14 We, therefore, turned to the
actual Hofer–Moest reaction (oxidative decarboxylation
at the anode of an electrochemical cell) to decarboxylate
uronosides. As already stated, anodic oxidative decar-
boxylation was quite successful in removing the carbox-
ylate carbon atom as CO2 and producing an aldehydo
group from the a-carbon atom (C-5).
1.2. xylo-Pentodialdose
Aspects of several different approaches21–23 were com-
bined in the preparation of a standard sample of xylo-
pentodialdose. 1,2-O-Isopropylidene-a-D-glucofuranose
(2.20 g) and sodium hydrogen carbonate (0.50 g) were
added to a 25-mL beaker equipped with a stir bar.
The reagents were dissolved in dd water (15 mL), after
which sodium meta-periodate (2.36 g) was added por-
tionwise. After 10 min, the soln remained oxidative
(starch–iodide test). Sodium thiosulfate (0.05 g) was
added until the soln was no longer oxidative. The soln
was then filtered and extracted five times with chloro-
form. The extract was dried over anhydrous sodium sul-
fate and evaporated to dryness at 50 ꢁC in a rotary
evaporator. Trifluoroacetic acid soln (50 mL, 20% v/v)
was added, and the soln was heated at 80 ꢁC with stir-
ring for 1 h and then evaporated to dryness in a rotary
evaporator at 50 ꢁC. Water (dd, 25 mL) was added,
and subsequently evaporated to dryness at 50 ꢁC. The
dry residue was rinsed with 95% ethanol. Another
25 mL of dd water was added, and the soln was again
evaporated to dryness at 50 ꢁC in a rotary evaporator.
One major HPLC peak was present.
Concentrated xylo-pentodialdose contains a mixture
of hemiacetal forms. This preparation and those pro-
duced electrochemically from methyl b-D-glucuronopyr-
anoside were reduced, both using sodium borohydride
and by catalytic (Raney nickel) hydrogenation, to xyli-
tol. Characterization via comparison of HLPC elution
time (single peak), mp, and mixed mp (93–94 ꢁC)24 using
an authentic xylitol standard indicated that the product
was xylitol.
1. Experimental
1.1. Equipment
A Dionex 500 Chromatography System equipped with a
GP40 Gradient Pump, an ED40 Electrochemical Detec-
tor, and PeakNet chromatography systems management
software (Dionex Corporation, Sunnyvale, CA) was
equipped with a PA10 column, a PA10 guard column,
and a 20-lL sample loop. Samples were prepared by
1.3. Electrolytic cell and electrolysis
The electrolytic cell used in this research was fabricated
from a 50-mL polypropylene centrifuge tube. Two cen-
timeters of a 7-cm graphite rod (0.312-12-GR008G,
Graphtek, LLC., Buffalo Grove, IL) were electroplated

J. A. Stapley, J. N. BeMiller / Carbohydrate Research 342 (2007) 610–613
613
with copper. An insulated copper wire was soldered to
the electroplate and secured with epoxy resin. This
graphite rod was placed through two holes at the base
of and through an axis of the centrifuge tube. The
graphite rod was secured with epoxy resin.
was subsequently bubbled through these mixtures at
50 ꢁC for 1.5 h. Then, enough water (dd) was added to
the reaction mixtures to bring them to 1 L. They were
subsequently analyzed by HLPC. Reduction was com-
plete in all cases.
A 5-cm piece of stainless steel rod (2 mm diam) was
soldered to an insulated copper wire, which was secured
with epoxy resin. The steel rod was placed through two
holes close to the top of and though an axis of the cen-
trifuge tube and secured with melted polypropylene. The
inlet and outlet were fashioned from 250-lL pipette tips,
which were affixed to holes in the centrifuge tube with
epoxy resin just above the stainless steel cathode at the
top and the graphite anode at the bottom of the tube.
Glass wool was placed in the cell under the anode and
at the inlet. Graphite pieces were created by crushing
another piece of the graphite rod and sieving; 15 g
of graphite pieces that were collected between No. 9
and No. 2 US sieves were placed into the cell.
Throughout the experiments a silver/silver chloride
reference electrode (RE-5b, Bioanalytical Systems,
Inc., West Lafayette, IN) was placed at a depth of
4 cm from the top of the cell. The anode and cathode
were connected to a SP-2 potentiostat (SP-2, Bioanalyt-
ical Systems, Inc., West Lafayette, IN). The inlet and
outlet were connected with tubing that passed through
a peristaltic pump (Tris; Isco, Inc., Lincoln, NE).
References
1. von Euler, H.; Hasselquist, H.; Eriksson, E. Makromol.
Chem. 1955, 18/19, 375–382.
2. Ruff, O. Ber. Dtsch. Chem. Ges. 1898, 31, 1573–1577.
3. Moody, G. J. Adv. Carbohydr. Chem. 1964, 19, 149–
179.
4. Hofer, H.; Moest, M. Ann. Chim. 1902, 323, 284.
5. Stapley, J. A. Ph.D. Thesis, Purdue University, 2004.
6. Stapley, J. A.; BeMiller, J. N. US Patent Appl. 2005/
0272961 A1, 2005.
7. Kitagawa, I.; Yoshikawa, M. Heterocycles 1977, 8, 783–
811.
8. Francisco, C. G.; Gonzalez, C. C.; Suarez, E. Tetrahedron
Lett. 1997, 38, 4141–4144.
9. Weiper, A.; Scha¨fer, H. J. Angew. Chem. 1990, 101, 228–
230.
10. Kitagawa, I.; Kamigauchi, T.; Ohmori, H.; Yoshikawa,
M. Chem. Pharm. Bull. 1980, 28, 3078–3086.
11. Kitagawa, I.; Kamigauchi, T.; Shirakawa, K.; Ikeda, Y.;
Ohmori, H.; Yoshikawa, M. Heterocycles 1981, 15, 349–
354.
12. Kitagawa, I.; Yoshikawa, M.; Kamigauchi, T.; Shirak-
awa, K.; Ikeda, Y. Chem. Pharm. Bull. 1981, 29, 2571–
2581.
The starting materials of the experiments were
weighed into 150-mL Erlenmeyer flasks to which water
(dd) was added. The reaction solution was then pumped
from the Erlenmeyer flask in which it was prepared
through the input of the electrolytic cell. After transfer-
ring an amount capable of sustaining constant flow, the
pump was connected to the outlet. The amount of solu-
tion transferred was determined gravimetrically (i.e., by
weighing the solution flask before and after the trans-
fer). The reaction solution was circulated through the
cell at a rate of 10 mL/min. The reactions were com-
menced by turning on the potentiostat. The solutions
were electrolyzed for approximately 3 h. Temperature
and pH readings were periodically taken by placing
the potentiostat on a standby and immersing a pH elec-
trode into the cell. In some cases, reaction mixtures were
subjected to catalytic reduction after electrolysis. The
reaction solutions were evacuated from the cell, pooled
with rinses and brought to standard volumes. To
approximately 30 mL of these solutions, freshly pre-
pared Raney Nickel (1 g) was added. Hydrogen gas
13. Renaud, P.; Hurzeler, M.; Seebach, D. Helv. Chim. Acta
¨
1987, 70, 292–298.
14. Stapley, J. A.; BeMiller, J. N. Carbohydr. Res., CARB
4171.
15. Cook, I. B.; Magee, R. J.; Payne, R.; Ternai, B. Aust. J.
Chem. 1986, 39, 1307–1314.
16. Makridou, C.; Cromer-Morin, M.; Scharff, J. P. Bull. Soc.
Chim. Fr. I 1977, 59–63.
17. Aruga, R. Bull. Chem. Soc. Jpn. 1981, 54, 1233–1235.
18. Payne, R.; Magee, R. J. Proc. Indian Acad. Sci.-Chem. Sci.
1982, 91, 31–37.
19. Micera, G.; Dessi, A.; Kozlowski, H.; Radomska, B.;
Urbanska, J.; Decock, P.; Dubois, B.; Olivier, I. Carbo-
hydr. Res. 1989, 188, 25–34.
20. Gyurcsik, B.; Nagy, L. Coord. Chem. Rev. 2000, 203, 81–
149.
21. Sowden, J. C. J. Am. Chem. Soc. 1951, 73, 5496–
5497.
22. Schaffer, R.; Isbell, H. S. J. Res. Nat. Bur. Stand. 1956, 56,
191–195.
23. Kollmann, M.; Kucera, J. Enzyme Microb. Technol. 1986,
8, 293–296.
24. Carson, J. F.; Waisbrot, S. W.; Jones, F. T. J. Am. Chem.
Soc. 1943, 65, 1777–1778.