C-2 and C-3 oxidation of D-Glc, and C-2 oxidation of D-Gal by pyranose dehydrogenase from Agaricus bisporus

Article abstract of DOI:10.1016/S0008-6215(98)00151-7

The quinone-dependent sugar oxidoreductase, pyranose dehydrogenase purified from the basidiomycete Agaricus bisporus catalyzed C-2 and C-3 oxidation of D-Glc to D-arabino-hexos-2-ulose (2-keto-D-Glc) and, preferentially, D-ribo-hexos-3-ulose (3-keto-D-Glc). The two aldoketoses accumulated only transiently in the reaction mixture being continually converted to the same end-product, D-erythro-hexos-2,3-diulose (2,3-diketo-D-Glc). D-Gal was oxidized by pyranose dehydrogenase and 1,4-benzoquinone exclusively at C-2 to produce D-lyxo-hexos-2-ulose (2-keto-d-Gal). Structures of the enzyme reaction products were deduced from their N,N-diphenylhydrazone derivatives. Copyright (C) 1998 Elsevier Science Ltd.

Full text of DOI:10.1016/S0008-6215(98)00151-7

Carbohydrate Research 310 (1998) 151±156
Note
C-2 and C-3 oxidation of d-Glc, and C-2
oxidation of d-Gal by pyranose dehydrogenase
from Agaricus bisporus
Jindrich Volca,*, Petr Sedmeraa, Petr Haladaa, Vera Prikrylovaa,
Geo€rey Danielb
a
Â
Institute of Microbiology, Academy of Sciences of the Czech Republic, VõdenÏska 1083,
Â
142 20 Prague 4, Czech Republic
b Department of Forest Products, Swedish University of Agricultural Sciences, Box 7008,
S-750 07 Uppsala, Sweden
Received 16 March 1998; accepted 13 June 1998
Abstract
The quinone-dependent sugar oxidoreductase, pyranose dehydrogenase puri®ed from the basi-
diomycete Agaricus bisporus catalyzed C-2 and C-3 oxidation of d-Glc to d-arabino-hexos-2-ulose
(2-keto-d-Glc) and, preferentially, d-ribo-hexos-3-ulose (3-keto-d-Glc). The two aldoketoses
accumulated only transiently in the reaction mixture being continually converted to the same
end-product, d-erythro-hexos-2,3-diulose (2,3-diketo-d-Glc). d-Gal was oxidized by pyranose
dehydrogenase and 1,4-benzoquinone exclusively at C-2 to produce d-lyxo-hexos-2-ulose (2-keto-
d-Gal). Structures of the enzyme reaction products were deduced from their N,N-diphenylhy-
drazone derivatives. # 1998 Elsevier Science Ltd. All rights reserved
Keywords: 2-Keto-d-Glc, 3-Keto-d-Glc, 2,3-Diketo-d-Glc, 2-Keto-d-Gal, d-arabino-Hexos-2-ulose, d-ribo-
Hexos-3-ulose, d-erythro-Hexos-2,3-diulose, d-lyxo-Hexos-2-ulose, Pyranose dehydrogenase, Agaricus bisporus
Characteristic for wood degrading fungi seems
to be an increasingly growing family of pyranose
oxidases (EC 1.1.3.10), enzymes catalyzing the C-2
oxidation of several aldoses [1,2]. Recently, we
reported a novel dicarbonyl sugar-producing oxi-
doreductase, the quinone dependent pyranose
dehydrogenase (PDH) isolated from the fungus
Agaricus bisporus [3]. With 1,4-benzoquinone as an
electron acceptor, this ¯avoglycoprotein catalyzes
oxidation of numerous monosaccharides and oli-
gosaccharides. The time course TLC analysis of d-
Glc transformation by PDH revealed that next to a
product of Rf 0.28 (blue spot), identi®ed as d-ara-
bino-hexos-2-ulose (1) [3], another sugar derivative,
2 (Rf 0.40, brownish yellow spot) was formed
simultaneously with 1. Both products accumulated
only transiently in the reaction mixture and dis-
appeared after prolonged incubation. We have now
identi®ed the second of the two products and clar-
i®ed the metabolic fate of both the products. With
respect to the complexity of the PDH reaction
Abbreviations: PDH, pyranose dehydrogenase, DPH,
diphenylhydrazone.
* Corresponding author. Fax: +420 2 4752384; e-mail:
volc@biomed.cas.cz
0008-6215/98/$Ðsee front matter # 1998 Elsevier Science Ltd. All rights reserved
PII S0008-6215(98)00151-7

152
J. Volc et al./Carbohydrate Research 310 (1998) 151±156
mixture with d-Glc (several products in tautomeric
equilibrium), spectroscopic examination of their
well de®ned hydrazone derivatives was used to
determine the position of carbonyl group(s).
The multiplicity of PDH/glucose oxidation pro-
ducts was con®rmed by HPLC analyses demon-
strating that peak of 1 (IV, T=12.25 min) was
followed by a further two peaks (V, VI) of
unknown product(s), probably partially separated
tautomeric forms of 2 [Fig. 1(B)±(D)]. Peak VII
seemed to correspond to a decomposition product
because it was strongly repressed in conversions at
4 ꢀC or at lower pHs. The total area of the product
peaks IV±VII at maximum was signi®cantly less
than that of starting Glc, indicating that the pri-
mary products were continually transformed into
the end-product which co-eluted substantially with
bu€er peak I, thereby increasing its area. In control
experiments where 1 was substituted for Glc in the
same PDH reaction mixture, it was completely
utilized, while in the PDH blank mixture 1 was not
consumed after overnight incubation (not shown).
1
FAB MS together with H and 13C NMR spec-
tra showed that the reaction products isolated as
Fig. 1. (A)±(D) HPLC monitoring of d-Glc oxidation to corresponding hexosuloses by pyranose dehydrogenase at incubation
times of 0, 0.5, 2 and 4 h, respectively. Peaks: I, citrate bu€er; II, d-erythro-hexos-2,3-diulose; III, d-Glc; IV, d-arabino-hexos-2-
ulose; V,VI, d-ribo-hexos-3-ulose (tautomeric forms); VII, unknown, probably decomposition product; (E)±(H) monitoring of d-
Gal oxidation at times of 0, 20, 40 and 90 min, respectively. Peaks: I, citrate bu€er; II, d-Gal; III, d-lyxo-hexos-2-ulose; IV, 1,4-
benzoquinone; V, hydroquinone. Reaction conditions are described in Experimental.

J. Volc et al./Carbohydrate Research 310 (1998) 151±156
153
N,N-diphenylhydrazone (DPH) derivatives were
Table 3
13C NMR data (chemical shift, multiplicity) of hydrazones 4±8
three bis-DPH (5, 6, 7) and one mono-DPH deri-
vative (4). AB and ABCD spin systems found in
1H NMR spectra of 4 (Tables 1 and 2) indicate
that C-3 carries no hydrogen. The only carbon
signal assignable to this atom (Table 3, a singlet at
104.92 ppm) implies an acetal or hemiacetal group.
As C-6 resonates down®eld with respect to free
sugars, it forms a 6!3 bond completing a furanose
ring (4, Fig. 2). Further supporting evidence is
provided by the elimination of the side chain with
hydrogen transfer in FAB MS (Fig. 2) producing
the m/z 226 ion. Therefore, the parent free sugar
was d-ribo-hexos-3-ulose (2, 3-keto-d-Glc, Fig. 3).
Singlet H-1 and an (OCH)3CH2OH moiety present
in 5 identi®es it as bis-DPH of d-arabino-hexos-2-
ulose (1, 2-keto-d-Glc). This is in accord with
previous identi®cation of one of the PDH/d-Glc
reaction products as 1 through its peracetylated 1-
mono-DPH derivative [3]. Singlet H-1 and four-
spin system of H-4 to H-6 in 1H NMR spectra of 6
and 7 suggest that these compounds are derived
from d-erythro-hexos-2,3-diulose (3, Fig. 3). The
position of the free carbonyl was inferred from
mass spectral fragmentation (Fig. 2) and from the
examination of ®ne structure of multiplets of the
respective quaternary carbon atoms in proton-
coupled 13C NMR spectra. The tricarbonyl sugar 3
was recently identi®ed as a new sugar metabolite
resulting from oxidation of 1 by another fungal
enzyme, pyranose 2-oxidase [4]. On HPLC, 3 in the
Carbon
4
5
8
6
7
1
2
3
4
5
6
138.04 d 128.41 d 133.51 d 128.83 d 128.33 d
75.75 d 161.21 s 201.16 s 188.31 s 143.25 s
104.92 s
72.15 d
72.07 d
72.88 t
70.66 d
75.06 d
73.39 d
65.61 t
73.32 d 145.03 s 201.14 s
73.55 d
72.40 d
64.68 t
67.73 d
72.97 d
63.71 t
76.14 d
75.33 d
63.97 t
PDH reaction mixture had the same retention time
[Fig. 1(B)±(D), II, co-eluting substantially with
bu€er peak] as authentic 3 giving a broad peak of
its tautomeric forms.
Based on the above deduced structures of free
di- and tricarbonyl sugars and results of TLC and
HPLC analyses, we conclude that d-Glc is with
PDH and 1,4-benzoquinone simultaneously con-
verted to aldoketoses 1 and 2 which are subse-
quently oxidized by the same enzyme to the
identical end-product, aldodiketose 3 (Fig. 3).
Thus, PDH acts alternatively at d-Glc C-2 and C-
3, whether or not one of these positions adjacent to
the site of attack bears a ketonic carbonyl group.
Time course analyses of the PDH reaction showed
that the molar ratio of the two aldoketoses varied
depending on the individual enzyme preparation
and reaction time. Under conditions corresponding
to Fig. 1(C) it was ca. 1:3 for 1 and 2, respectively,
during the time of their maximum accumulation.
Generally, compound 2 was initially accumulated
faster, suggesting that position C-3 is the principle
site of Glc oxidation. The molar ratio of 1 and 2
was, however, a€ected by the relative rate of their
consumption during oxidation to 3.
Table 1
1H NMR data (chemical shift, multiplicity) of hydrazone
derivatives 4±8
Proton
4
5
8
6
7
Dehydrogenation of another PDH preferred
substrate, d-Gal, resulted in almost quantitative
conversion into a single product [Fig. 1(F)±(H)]. Its
chromatographic behaviour corresponded to that
of authentic d-lyxo-hexos-2-ulose. Consistently, H-
1 of its mono-DPH derivative (8, Fig. 2) resonates
as a singlet, C-3 to C-6 form a contiguous spin
system and C-2 is free (Table 3). Consequently,
PDH catalyses dehydrogenation of d-Gal exclu-
sively at C-2.
1
2
3
4
5
6u
6d
6.640 d
4.287 d
Ð
7.017 s
Ð
5.546 d
6.679 s
Ð
5.393 d
8.009 s
Ð
Ð
6.422 s
Ð
Ð
4.063 d
4.091 dd 3.985 dd 3.490 d
5.448 d
4.160 ddd 3.951 ddd 3.915 ddd 3.757 ddd 4.177 ddd
3.813 dd 3.752 dd 3.637 dd 3.130 dd 3.671 dd
3.900 dd 3.942 dd 3.656 dd 3.470 dd 3.741 dd
Table 2
Proton±proton coupling constants [Hz] of hydrazones 4±8
Oxidation of d-Glc to 2 has been earlier
reported for another microbial enzyme, d-gluco-
side 3-dehydrogenase (EC 1.1.99.13) isolated from
Agrobacterium tumefaciens and Flavobacterium
saccharophilum, and characterized as an 85 and
66 kDa iron-sulfur ¯avoprotein, respectively [5,6].
Similar to PDH, this enzyme has high anity for a
J(i,j)
4
5
8
6
7
4,5
5.5
5.0
3.7
9.5
8.1
7.0
3.2
1.8
6.7
6.0
8.8
6.0
3.3
5.1
6.4
4.5
5,6u
5,6d
6u,6d
11.9
11.0
11.5
11.4
Additional couplings: 4, J1,2=5.6 Hz; 5, J3,4=1.4 Hz; 8,
J3,4=7.4 Hz.

154
J. Volc et al./Carbohydrate Research 310 (1998) 151±156
Fig. 2. Structures and FAB MS fragmentation of hydrazone derivatives used for identi®cation of the oxidation products of d-Glc
and d-Gal by pyranose dehydrogenase.
variety of oligomeric sugars and glucosides but, in
contrast, is much more selective for mono-
saccharides and oxidizes d-Gal at C-3. To our
knowledge, there has been no earlier report of a
fungal oxidoreductase acting at C-3 of d-Glc. On
the other hand, C-2 oxidation of d-Gal is a com-
mon catalytic feature of PDH (monomeric extra-
cellular 79 kDa protein [3]) with pyranose 2-
oxidase, typically a tetrameric intracellular ca.
300 kDa protein of fungal origin [7,8].
Considering the above demonstrated lower spe-
ci®city of the Agaricus's sugar oxidoreductase for
the site of action on d-Glc (dehydrogenation at
both C-2 and C-3, the latter prevailing), we pro-
pose pyranose dehydrogenase (PDH) as a more
suitable trivial name for this enzyme compared
with the name pyranose 2-dehydrogenase (P2D) we
used in our preliminary report on the enzyme [3].
1. Experimental
Enzyme source and puri®cation.ÐAgaricus bis-
porus (Lange) Imbach, strain U3, obtained from
the culture collection at Horticulture Research
International (Wellesbourne, UK) was grown sta-
tionary on a liquid medium [3] for 14 days. Pyr-
anose dehydrogenase (PDH) was puri®ed from
mycelial extracts to apparent homogeneity (16 U/
mg protein) and assayed by monitoring the reduc-
tion of 1,4-benzoquinone to hydroquinone
1
1
("290 ˆ 2:3 mM cm ) by Glc [3]. One U of the
1
enzyme activity oxidizes d-Glc at 1 ꢀmol min
under the assay conditions.
PDH catalyzed oxidation of d-Glc and d-Gal.Ð
The reaction mixture (20 mL) for semipreparative
transformations contained d-Glc [Fig. 1(A)±(B)] or
d-Gal (180 mg, 50 mM), 1,4-benzoquinone [(Sigma,

J. Volc et al./Carbohydrate Research 310 (1998) 151±156
155
Fig. 3. The proposed scheme for two step oxidation of d-Glc to d-erythro-hexos-2,3-diulose (3, 2,3-diketoGlc) via d-arabino-hexos-
2-ulose (1, 2-ketoGlc) and d-ribo-hexos-3-ulose (2, 3-ketoGlc), by pyranose dehydrogenase from Agaricus bisporus.
St. Louis, MO) 108 mg, 50 mM] and PDH (10 U) in
10 mM Na citrate, pH 5.5. Aerobic incubations
proceeded under gentle stirring in darkness at 30 ꢀC
for 2 h. Samples (0.1 mL) withdrawn at time inter-
vals for HPLC analyses were puri®ed by passing
through Ultrafree-MC 30,000 NMWL Filter Units
(Millipore, Bedford, MA) and diluted 10 times
with water. Analytical scale conversions [Fig. 1(E)±
(H), Gal conversion] were followed in reaction
mixtures (5 mL) consisting of the respective sugar
(10 mM), 1,4-benzoquinone (10 mM) and PDH
(1 U) in 5 mM Na citrate, pH 5.5.
were each supplemented successively with the same
volume of ethanol, 1 mL 99% acetic acid and
270 ꢀL freshly distilled N,N-diphenylhydrazine
(Koch-Light Lab., Colnbrook, UK), stirred for 1 h
and left overnight at 25 ꢀC. The dark residue
obtained after vacuum evaporation of the solvents
was dissolved in 1.5 mL ethanol and subjected to
TLC as above. Pale-yellow to orange streaks of
four major hydrazones (Rf 0.13, 0.25, 0.39 and
0.41) derived from d-Glc oxidation products were
cut out, combined separately, and extracted into
methanol. The individual components were further
puri®ed by rechromatography to substantial
homogeneity. The hydrazone of the d-Gal oxida-
tion product (Rf 0.10, one major component) crys-
tallized directly from the reaction mixture at
15 ꢀC.
1H and 13C NMR data (400 and 100 MHz,
respectively) of isolated hydrazones (Tables 1±3)
were recorded on a Varian INOVA-400 spectro-
meter in CD3OD at 30 ꢀC. Residual solvent signals
(ꢁH 3.33, ꢁC 49.3) were used as internal standards.
2D NMR experiments (COSY, ROESY, HET-
COR) were performed using the manufacturer's
software. Positive-ion FAB mass spectra were
recorded on a Finnigan MAT 95 double-focusing
instrument (Finnigan MAT, Bremen, Germany).
The matrix was m-nitrobenzyl alcohol (Aldrich-
Chemie, Steinheim, Germany) and the calibration
was performed with Ultramark 1600F (PCR Inc.,
USA) as a standard. The products of CIDs in the
®rst ®eld-free region of the instrument were ana-
lysed by daughter-ion linked scan (B/E constant)
using the manufacturer's software.
Chromatographic analysis of sugar oxidation
products.ÐTLC of free sugars was conducted on
SigmaCell Type 100 (Sigma) cellulose sheets
using 140:100:33:80 1-BuOAc±AcOH±Me2CO±
H2O and detection by diphenylamine reagent [9].
Standard d-lyxo-hexos-2-ulose and d-arabino-
hexos-2-ulose were prepared enzymatically using
pyranose oxidase from Oudemansiella mucida [10].
Aldosulose hydrazones were separated on Silica
Gel 60 F254 TLC aluminium sheets (Merck,
Darmstadt, Germany) using 12:1 CHCl3±MeOH.
HPLC was performed on a SP 8800 liquid chro-
matograph (Spectra Physics, San Jose, CA) ®tted
with a refractive index detector, using Ostion LG
KS 0800 Na+ (d-Glc oxidation products) or
Ostion LG KS 0800 Ca2+ (d-Gal) column
250Â8 mm (Watrex, Prague, Czech Republic)
ꢀ
eluted at 80 C with deionized water at a ¯ow rate
1
of 0.5 mL min .
Derivatization and identi®cation of d-Glc and d-
Gal oxidation products.ÐThe mixtures (20 mL) of
sugar semipreparative transformation products

156
J. Volc et al./Carbohydrate Research 310 (1998) 151±156
d-ribo-Hexos-3-ulose-(3,6) 1-(N,N-diphenyl-
d-lyxo-Hexos-2-ulose 1-(N,N-diphenylhydra-
zone) (8). Compound of Rf 0.10, recrystallized
twice from ethanol, pale-yellow ®ne needels
(109 mg), m.p. 148±149 ꢀC, UV/VIS: lmax (MeOH)
229, 348, 412 nm. FAB MS, m/z (% rel.int.): 345
[M+H]+, 327 (100), 254 (10), 223 (16), 183 (7),
169 (24), 168 (20), 121 (2).
hydrazone) (4). Compound of Rf 0.13, obtained as
a pale-yellow sirup after rechromatography on
Silica Gel 60 F254 coated HPTLC aluminium sheets
(Merck) in 40:3 CHCl3±MeOH (24 mg). UV/VIS:
lmax (MeOH) 271, 416 nm. FAB MS, m/z (% rel.
int.): 345 [M+H]+, (daughter ions of the proto-
nated molecule) 327 (100), 267 (3), 226 (12), 195
(1), 184 (1), 168 (3).
References
d-arabino-Hexos-2-ulose 1,2-bis(N,N-diphenyl-
hydrazone) (5). Compound of Rf 0.25, rechroma-
tographed using 40:1 CHCl3±MeOH/20 Silufol
(Kavalier, Votice, Czech Republic) silica gel foils,
dark yellow solid (17 mg). UV/VIS: lmax (MeOH)
234, 295, 346, 411 nm. FAB MS, m/z (% rel. int.):
511 [M+H]+, 493 (20), 449 (5), 419 (52), 342 (56),
327 (100), 252 (20), 223 (1), 195 (2), 169 (8).
[1] P. Oliveira, N. Rodeia, A. Clemente, and
A. Carmali, Glucose-2-oxidase production by
white rot fungi, in J.F. Kennedy, G.O. Phillips and
P.A. Williams (Eds.), Ligno-Cellulosics: Science,
Technology, Development and Use, Ellis Horwood,
New York, 1992, pp 33±40.
[2] A. Huwig, H.J. Danneel, and F. Gi€horn, J. Bio-
technol., 32 (1994) 309±315.
d-erythro-Hexos-2,3-diulose 1,2-bis(N,N-di-
phenylhydrazone) (6). Compound of Rf 0.39,
rechromatographed on Silica Gel 60 TLC alumi-
nium sheets (Merck) in 20:1 CHCl3±MeOH and
crystallized from methanol (yellow needles, 8 mg),
mp 118±119 ꢀC. UV/VIS: lmax (MeOH) 250,
413 nm. FAB MS, m/z (% rel. int.): 509 [M+H]+,
491 (100), 461 (2), 447 (100), 418 (1), 392 (1), 360
(1), 325 (2), 285 (8), 250 (1), 223 (2), 184 (1), 169 (2).
d-erythro-Hexos-2,3-diulose 1,3-bis(N,N-di-
phenylhydrazone) (7). Compound of Rf 0.41, pur-
i®ed on Silica Gel 60 sheets (Merck) in 40:1
CHCl3±MeOH (orange sirup, 66 mg). UV/VIS:
lmax (MeOH) 237, 288, 412 nm. FAB MS, m/z (%
rel. int.): 509 [M+H]+, 491 (100), 417 (1), 390 (1),
340 (62), 325 (17), 323 (9), 280 (3), 222 (2), 195 (1),
183 (2), 169 (7).
[3] J. Volc, E. Kubatova, D.A. Wood, and G. Daniel,
Arch. Microbiol., 167 (1997) 119±125.
[4] P. Sedmera, J. Volc, V. Havlõcek, S. Pakhomova, and
A. Jegorov, Carbohydr. Res., 297 (1997) 375±378.
[5] J. Van Beeumen and J. De Ley, Eur. J. Biochem., 6
(1968) 331±343.
[6] M. Takeuchi, N. Asano, Y. Kameda, and
K. Matsui, Agric. Biol. Chem., 52 (1988) 1905±
1912.
[7] A. Schafer, S. Bieg, A.Huwig, G.-W. Kohring, and
F. Gi€horn, Appl. Environ. Microbiol., 62 (1996)
2586±2592.
[8] J. Volc, E. Kubatova, P. Sedmera, G. Daniel, and
J. Gabriel., Arch. Microbiol. 156 (1991) 297±301
[9] J. Kocourek, M. Ticha, and J. Kostõr, J. Chroma-
togr., 24 (1966) 117±124.
[10] J. Gabriel, J. Volc, P. Sedmera, G. Daniel, and
E. Kubatova, Arch. Microbiol. 160 (1993) 27±34.