Properties of family 79 β-glucuronidases that hydrolyze β-glucuronosyl and 4-O-methyl-β-glucuronosyl residues of arabinogalactan-protein

Article abstract of DOI:10.1016/j.carres.2008.03.004

The carbohydrate moieties of arabinogalactan-proteins (AGPs), which are mainly composed of Gal, l-Ara, GlcA, and 4-Me-GlcA residues, are essential for the physiological functions of these proteoglycans in higher plants. For this study, we have identified two genes encoding family 79 β-glucuronidases, designated AnGlcAase and NcGlcAase, in Aspergillus niger and Neurospora crassa, respectively, based on the amino acid sequence of a native β-glucuronidase purified from a commercial pectolytic enzyme preparation from A. niger. Although the deduced protein sequences of AnGlcAase and NcGlcAase were highly similar, the recombinant enzymes expressed in Pichia pastoris exhibited distinct substrate specificity toward 4-Me-GlcA residues of AGPs: recombinant AnGlcAase (rAnGlcAase) substantially liberated both GlcA and 4-Me-GlcA residues from radish AGPs, whereas recombinant NcGlcAase (rNcGlcAase) activity on the 4-Me-GlcA residues of AGPs was very low. Maximum activity of rAnGlcAase hydrolyzing PNP β-GlcA occurred at pH 3.0-4.0, whereas the maximum rNcGlcAase activity was at pH 6.0. The apparent K_m values of rAnGlcAase were 30.4 μM for PNP β-GlcA and 422 μM for β-GlcA-(1→6)-Gal, and those of rNcGlcAase were 38.3 μM and 378 μM, respectively. Similar to the native enzyme, rAnGlcAase was able to catalyze the transglycosylation of GlcA residues from PNP β-GlcA to various monosaccharide acceptors such as Glc, Gal, and Xyl. We propose that both AnGlcAase and NcGlcAase are instances of a novel type of β-glucuronidase with the capacity to hydrolyze β-GlcA and 4-Me-β-GlcA residues of AGPs, although they differ significantly in their preferences.

Full text of DOI:10.1016/j.carres.2008.03.004

Available online at www.sciencedirect.com
Carbohydrate Research 343 (2008) 1191–1201
Properties of family 79 b-glucuronidases that hydrolyze
b-glucuronosyl and 4-O-methyl-b-glucuronosyl residues
of arabinogalactan-proteinI;II;H
Tomoyuki Konishi,a Toshihisa Kotake,a Dina Soraya,a Koji Matsuoka,b Tetsuo Koyama,b
Satoshi Kaneko,c Kiyohiko Igarashi,d Masahiro Samejimad and Yoichi Tsumurayaa,*
aDivision of Life Science, Graduate School of Science and Engineering, Saitama University,
255 Shimo-okubo, Sakura-ku, Saitama 338-8570, Japan
bDivision of Material Science, Graduate School of Science and Engineering, Saitama University,
255 Shimo-okubo, Sakura-ku, Saitama 338-8570, Japan
cFood Biotechnology Division, National Food Research Institute, 2-1-12, Kannondai, Tsukuba, Ibaraki 305-8642, Japan
dGraduate School of Agricultural and Life Sciences, The University of Tokyo, Yayoi 1-1-1, Bunkyo-ku, Tokyo 113-8657, Japan
Received 3 December 2007; received in revised form 3 March 2008; accepted 4 March 2008
Available online 18 March 2008
Abstract—The carbohydrate moieties of arabinogalactan-proteins (AGPs), which are mainly composed of Gal, L-Ara, GlcA, and 4-
Me-GlcA residues, are essential for the physiological functions of these proteoglycans in higher plants. For this study, we have iden-
tified two genes encoding family 79 b-glucuronidases, designated AnGlcAase and NcGlcAase, in Aspergillus niger and Neurospora
crassa, respectively, based on the amino acid sequence of a native b-glucuronidase purified from a commercial pectolytic enzyme
preparation from A. niger. Although the deduced protein sequences of AnGlcAase and NcGlcAase were highly similar, the recom-
binant enzymes expressed in Pichia pastoris exhibited distinct substrate specificity toward 4-Me-GlcA residues of AGPs: recombi-
nant AnGlcAase (rAnGlcAase) substantially liberated both GlcA and 4-Me-GlcA residues from radish AGPs, whereas recombinant
NcGlcAase (rNcGlcAase) activity on the 4-Me-GlcA residues of AGPs was very low. Maximum activity of rAnGlcAase hydrolyz-
ing PNP b-GlcA occurred at pH 3.0–4.0, whereas the maximum rNcGlcAase activity was at pH 6.0. The apparent Km values of
rAnGlcAase were 30.4 lM for PNP b-GlcA and 422 lM for b-GlcA-(1?6)-Gal, and those of rNcGlcAase were 38.3 lM and
378 lM, respectively. Similar to the native enzyme, rAnGlcAase was able to catalyze the transglycosylation of GlcA residues from
PNP b-GlcA to various monosaccharide acceptors such as Glc, Gal, and Xyl. We propose that both AnGlcAase and NcGlcAase are
instances of a novel type of b-glucuronidase with the capacity to hydrolyze b-GlcA and 4-Me-b-GlcA residues of AGPs, although
they differ significantly in their preferences.
Ó 2008 Elsevier Ltd. All rights reserved.
Keywords: Arabinogalactan-protein; Aspergillus niger; Family 79 glycoside hydrolase; b-Glucuronidase; Neurospora crassa; Transglycosylation
1. Introduction
q Sugars mentioned in this paper belong to the D-series unless
Arabinogalactan-proteins (AGPs) are a family of prote-
otherwise noted.
qq The nucleotide sequences reported in this paper have been
submitted to the DNA Data Bank of Japan (DDBJ) with accession
number An02g11890 for AnGlcAase and AB293989 for NcGlcAase.
w Experimental details concerning cDNA cloning, expression of b-
glucuronidase in Pichia pastoris, and purification of recombinant
enzymes are given in Supplementary data.
oglycans involved in many physiological processes such
as cell-to-cell signaling, cell adhesion, cell elongation,
cell death, and stress responses in plants.1–4 AGPs are
characterized by large carbohydrate components rich
in galactose and L-arabinose, and protein components
(core proteins: generally less than 10% of total weight)
rich in hydroxyproline, serine, threonine, alanine, and
*
Corresponding author. Tel.: +81 48 858 3401; fax: +81 48 858
3384; e-mail: tsumura@molbiol.saitama-u.ac.jp
0008-6215/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carres.2008.03.004

1192
T. Konishi et al. / Carbohydrate Research 343 (2008) 1191–1201
glycine. The carbohydrate moieties of AGPs have a
common overall structure of b-(1?3)-linked galactosyl
backbones to which side chains of b-(1?6)-linked galac-
tosyl residues are attached through O-6. The b-(1?6)-
linked galactosyl chains are further substituted with L-
arabinose, glucuronic acid, and 4-O-methyl-glucuronic
acid and lesser amounts of other sugars such as L-rham-
nose and L-fucose.1 In the case of radish root AGPs,
uronic acids including b-glucuronic acid and 4-O-
methyl-b-glucuronic acid constitute more than 4% of
total weight of the carbohydrate moieties of AGPs (see
Fig. 1 for a schematic illustration of radish AGP
structure).5
It seems that in planta, the carbohydrate moieties of
AGPs are structurally modified and degraded through
the action of several endogenous glycosidases, such as
b-galactosidase (EC 3.2.1.23)6,7 and a-L-arabinofura-
nosidase (a-L-Arafase; EC 3.2.1.55).8,9 Fungi and bacte-
ria appear to have various glycosidases that hydrolyze
the carbohydrate moieties of AGPs in order to utilize
the sugars as a carbon source or as nutrients. Examples
include exo-b-(1?3)-galactanase (EC 3.2.1.145)10,11 and
endo-b-(1?6)-galactanase (EC 3.2.1.164).12,13 We have
previously purified a b-glucuronidase (EC 3.2.1.31) from
a commercial pectolytic enzyme preparation from
Aspergillus niger. This was the first enzyme shown to
hydrolyze the b-glucuronosyl and 4-O-methyl-b-glucur-
onosyl residues substituted at O-6 of non-reducing ter-
minals of b-(1?6)-linked galactosyl side chains, which
are generally found in AGPs.14,15
Glycoside hydrolases are classified into more than 100
families based on amino acid sequence and structural
similarity,16,17 and most b-glucuronidases characterized
so far have been classified as glycosyl hydrolase family
2, but their properties and biological functions have
not yet been extensively investigated. A small number
of b-glucuronidases are categorized as family 1 or 79
glycoside hydrolases. A family 79 b-glucuronidase iso-
lated from skullcap (Scutellaria baicalensis) has been
shown to hydrolyze baicalein 7-O-b-glucuronide (a gly-
cosylated flavone).18 However, it is unknown if these
b-glucuronidases act on AGPs.
For the present study, novel genes encoding b-glucu-
ronidases that may play a role in the metabolism of
AGPs were identified in A. niger and Neurospora crassa.
We have characterized the enzymatic properties and
investigated the actions on the carbohydrate moieties
of AGPs of the corresponding proteins via recombinant
enzymes expressed in Pichia pastoris.
2.1. Materials
A. niger van Tieghem (ATCC 22343) and N. crassa
OR74A were used for this study. Mycelia of A. niger
were cultured in liquid medium containing 2% (w/v)
malt extract, 2% (w/v) glucose, and 0.1% (w/v) peptone
(pH 6.0, MGP medium) at 25 °C for 2 days, then the
submerged culture was inoculated into the same medium
containing 1.5% (w/v) agar at 25 °C for 5 days. N. crassa
was cultured in a wheat bran medium containing 2 g of
wheat bran and 3 mL of water at 25 °C for 7 days.
The b-(1?3)- and b-(1?6)-linked galactobioses used
were prepared from larch wood arabinogalactan19 and
b-(1?4)-galactobiose was prepared from soybean arab-
inan–galactan.6 Acacia gum and the sap of the lac tree,
Rhus vernicifera, were the sources for preparation of the
following glucuronic acid (GlcA)- or 4-O-methyl-glucu-
ronic acid (4-Me-GlcA)-containing acidic oligosaccha-
rides: b-GlcA-(1?6)-Gal, b-GlcA-(1?6)-b-Gal-(1?6)-
Gal, 4-Me-b-GlcA-(1?6)-Gal, 4-Me-b-GlcA-(1?6)-b-
Gal-(1?6)-Gal,
and
4-Me-b-GlcA-(1?6)-b-Gal-
(1?6)-b-Gal-(1?3)-Gal.14 The preparation of b-GlcA-
(1?3)-Gal was made from a polysaccharide produced
by the cultivation of Pseudomonas viscogena TS-1004
as described.14 Laminaribiose, cellobiose, and heparan
sulfate from bovine kidney were purchased from Sei-
kagaku Co. (Tokyo, Japan). Gentiobiose, p-nitrophenyl
b-glucopyranoside (PNP b-Glc), PNP b-galactopyrano-
side (PNP b-Gal), PNP a-L-arabinofuranoside (PNP a-
L-Ara), PNP b-xylopyranoside (PNP b-Xyl), PNP b-
glucopyranosiduronic acid (PNP b-GlcA), PNP b-galac-
topyranosiduronic acid (PNP b-GalA), PNP a-L-fuco-
pyranoside (PNP a-L-Fuc), laminarin from Laminaria
digitata, chitosan from crab shells, guar gum, locust
bean gum, and xylan from birchwood were from Sigma
(St. Louis, MO, USA). (1?3)(1?4)-b-Glucan from bar-
ley, CM-cellulose 4 M, and b-(1?4)-galactan from lupin
were purchased from Megazyme (Wicklow, Ireland).
CM-curdlan was from Wako (Osaka, Japan). Pustulan
from Umbilicaria papullosa was from Calbiochem (San
Diego, CA, USA). AGPs from radish leaves (AGP R-
II) and roots (AGP IV) and their a-L-Arafase-treated
derivatives were prepared in our laboratories.5,20
2.2. Sugar and protein analyses
Total sugar content was determined by the phenol–sul-
furic acid method.21 Total uronic acids in AGPs were
determined by a modified carbazole–sulfuric acid meth-
od22 using GlcA as the standard. The concentration of
protein was determined by the method of Bradford23
with BSA as the standard.
2. Experimental
Native b-glucuronidase was purified by conventional
chromatographic techniques from Pectinex Ultra SP-L,
a commercial pectolytic enzyme preparation from A. ni-
ger (Novo Nordisk Ferment Ltd, Chiba, Japan) as de-
Experimental details of cDNA cloning, expression of b-
glucuronidase in P. pastoris, and purification of recom-
binant enzymes are given in Supplementary data.

T. Konishi et al. / Carbohydrate Research 343 (2008) 1191–1201
1193
scribed previously.14 The purified b-glucuronidase was
2.4. Transglycosylation reaction
separated on SDS-PAGE,24 blotted onto a PVDF-Plus
membrane (Osmonics, Moers, Germany), and subjected
to N-terminal amino acid analysis with a protein se-
quencer (HP G1000A, Hewlett Packard, Palo Alto,
CA, USA). The N-terminal amino acid sequence of re-
combinant A. niger b-glucuronidase (rAnGlcAase) was
determined as well.
Transglycosylation assays of rAnGlcAase used reaction
mixtures (20 lL) consisting of the enzyme (67 ng rAnG-
lcAase), 20 mM PNP b-GlcA, 0.5 M acceptor monosac-
charide [Glc, Gal, Xyl, GlcNAc, N-acetylgalactosamine
(GalNAc), Man, L-Fuc, L-rhamnose (L-Rha), D-Ara, L-
Ara, Rib, or sorbitol], and 7 mM sodium acetate buffer,
pH 4.0. This mixture was incubated at 37 °C and the
reaction was terminated by heating to 100 °C for
5 min. The amounts of p-nitrophenol and GlcA released
by transglycosylation and hydrolysis reactions were
quantitated colorimetrically as described above. Trans-
glycosylation products in the reaction mixture were ana-
lyzed by paper chromatography on Whatman 3MM
filter paper using 6:4:3 1-butanol–pyridine–water and
5:2:3 1-butanol–AcOH–water as the eluents. Sugar spots
on the paper chromatograms were visualized with an
alkaline AgNO3 reagent.
2.3. Enzyme assays
Hydrolytic activity was determined in mixtures
(200 lL) consisting of the enzyme, 2 mM PNP b-GlcA,
and 50 mM sodium acetate buffer, pH 4.0, for rAnG-
lcAase (32 ng enzyme) or pH 6.0 for recombinant N.
crassa b-glucuronidase (rNcGlcAase; 39 ng). After
incubation at 37 °C, the reaction was terminated by
the addition of 200 mM sodium carbonate (800 lL)
and monitored at 420 nm. One unit of enzyme activity
liberates 1 lmol of p-nitrophenol per min. Activity to-
ward oligosaccharides was assayed in reaction mixtures
(8 lL) containing enzyme (33 ng rAnGlcAase or 63 ng
rNcGlcAase), 5 mM oligosaccharide, and 50 mM so-
dium acetate buffer, pH 4.0 (for rAnGlcAase) or 6.0
(for rNcGlcAase). After incubation at 37 °C for an
appropriate reaction time, the liberated GlcA was as-
sayed by the reductometrical method of Milner and
Avigad,25 while the amount of released 4-Me-GlcA
was determined reductometrically by the neocuproine
assay26 using GlcA as the standard. One unit of en-
zyme activity liberates 1 lmol of GlcA or 4-Me-GlcA
per min.
2.5. Linkage analysis of transglycosylation product
In order to identify oligosaccharides formed through
transglycosylation by rAnGlcAase, a large-scale reac-
tion (500 lL) with 20 mM PNP b-GlcA and 0.5 M Gal
as acceptor substrate as above was incubated at 37 °C
for 3 h. The reaction was terminated by heating and
the reaction products (a main product and a minor
one) were separated on paper chromatography using
Whatman 3MM filter paper with 5:2:3 1-butanol–
AcOH–water as the solvent, extracted from the paper,
and lyophilized (the main product, yield 5.3 mg).
Substrate specificity of the recombinant enzymes to-
ward polysaccharides was investigated using reaction
mixtures (100 lL) consisting of the enzyme (64 ng
rAnGlcAase or 79 ng rNcGlcAase), 0.5% (w/v) polysac-
charide, and 50 mM sodium acetate buffer, pH 4.0 or
6.0. After incubation at 37 °C for a suitable reaction per-
iod, the amount of reducing sugars was determined by
the method of Nelson27 and Somogyi.28 The hydrolytic
activity toward AGPs was determined for reaction mix-
tures (20 lL) consisting of the enzyme (12–155 ng
rAnGlcAase or 21–436 ng rNcGlcAase), 0.25% (w/v)
AGP, and 50 mM sodium acetate buffer, pH 4.0 or
6.0. After reaction at 37 °C, the mixture was heated to
100 °C for 5 min to terminate the reaction, and then cen-
trifuged at 18,000g for 10 min to remove insoluble mate-
rials. The resulting supernatant was filtered through a
membrane filter (0.45 lm; Chromatodisk, Krabou,
Osaka, Japan) and was analyzed by high perfor-
mance anion-exchange chromatography with pulsed
amperometric detection (HPAEC-PAD) using a Dionex
DX-500 liquid chromatograph fitted with a CarboPac
PA-1 column (4 Â 250 mm; Dionex Japan, Osaka,
Japan) and a pulsed amperometric detector as described
previously.29
A part of the transglycosylation product was incu-
bated with rAnGlcAase in 40 lL of a reaction mixture
containing 3 mM product (based on the molecular mass
obtained by the analysis described below), 7 mM so-
dium acetate buffer, pH 4.0, and 4 milliunits rAnGlcA-
ase at 37 °C for 10 h. The reaction products were
analyzed on HPAEC-PAD.
Matrix-assisted laser desorption/ionization time-of-
flight mass spectrometry (MALDI-TOFMS) of the
transglycosylation product was performed with a KOM-
PACT MALDI IV tDE (Shimadzu, Kyoto, Japan) as
described previously.14 For the mass calibration,
cyclomaltoheptaose was used. Structural elucidation of
the transglycosylation product was performed by a
combination of 1H and 13C NMR spectroscopy. 1H
and 13C NMR spectra were recorded at 400 and
100 MHz, respectively, with a Bruker DPX-400 spec-
trometer. The sample was dissolved in D2O, and MeOD
1
was used as the internal standard (3.3 ppm for H and
49.5 ppm for 13C). Ring-carbon assignments in NMR
were made by a combination of 1D and 2D NMR
experiments. Bruker standard pulse sequences were
used for DEPT135, 1D-total correlation spectroscopy,

1194
T. Konishi et al. / Carbohydrate Research 343 (2008) 1191–1201
and 2D homonuclear shift correlation spectroscopy
experiments.
hydrolases. They did exhibit low but significant similar-
ity to several family 79 enzymes such as b-glucuronidase
from skullcap (S. baicalensis, identity, 13%),18 heparan-
ase from human (Homo sapiens, 11–16%),31 heparanase
from mole rat (Spalax golani, 10–15%),32 and heparan-
ase from Norway rat (Rattus norvegicus, 14–16%).33
The b-glucuronidase from skullcap has been shown to
possess an acid–base catalyst (Glu residue) and a cata-
lytic nucleophile (Glu residue) in the active site.18
Although the overall similarity of AnGlcAase and
NcGlcAase to family 79 enzymes was low, the catalytic
regions of family 79 enzymes appeared to be conserved
in AnGlcAase (Glu189 and Glu324 residues) and
NcGlcAase (Glu175 and Glu312 residues) (Fig. 1A), sug-
gesting that AnGlcAase and NcGlcAase are family 79
glycoside hydrolases. Although the skullcap b-glucuron-
idase possesses a putative ATP/GTP binding motif and
a leucine-zipper motif,18 neither was found in AnGlcA-
ase nor in NcGlcAase. Phylogenetic analyses of AnG-
lcAase and NcGlcAase and other family 79 members
indicated that AnGlcAase and NcGlcAase belong to a
subfamily of fungal glycosidases. This group is appar-
ently set apart from plant b-glucuronidases, vertebrate
heparanases, and bacterial glycosidases (Fig. 1B). How-
ever, the functions of the fungal glycosidases, other than
AnGlcAase and NcGlcAase, have not yet been exam-
ined. We suspect that other fungal glycosidases will
show a pattern of substrate specificity similar to that
of AnGlcAase and NcGlcAase described below.
3. Results and discussion
3.1. Amino acid sequences of b-glucuronidases
The native b-glucuronidase purified from Pectinex Ultra
SP-L was partially digested with V8 proteinase and
yielded the internal peptide sequence SAGLPVLHPF
VSFSIEFVFF PDYAG upon analysis with a protein se-
quencer. This peptide sequence showed significant simi-
larity to an open reading frame (ORF, An02g11890) of
A. niger, which had not been characterized so far. Based
on the internal peptide sequence, we cloned cDNA for
An02g11890 by reverse transcriptase-PCR with specific
primers, obtaining AnGlcAase (A. niger b-glucuroni-
dase). The amino acid sequence deduced from AnGlcA-
ase did not completely match the peptide sequence
determined for the native b-glucuronidase in that four
residues differed (the 1st, 3rd, 18th, and 23rd residues
among 25 determined for the native enzyme; see
Fig. 1A). Because we could find no other highly similar
sequence in the genomic database of A. niger, we believe
that the mismatch in the peptide sequences arises be-
cause the strain used for the preparation of the purified
enzyme and the strain used for the cDNA cloning were
different. However, we cannot rule out the possibility
that the ORF encodes a closely related enzyme distinct
from the native enzyme. Blast search revealed that an
ORF (EAA35527) from N. crassa was 47% identical to
the amino acid sequence deduced from An02g11890,
and we thus refer to this ortholog of An02g11890 as
NcGlcAase (N. crassa b-glucuronidase).
AnGlcAase and NcGlcAase appeared to encode poly-
peptides of 541 and 559 amino acids (molecular mass
58,859 and 60,235 Da), respectively (Fig. 1A). Using
the SignalP program,30 putative signal sequences were
found in both AnGlcAase and NcGlcAase, indicating
that these enzymes are secreted proteins. The calculated
molecular masses of the mature proteins of AnGlcAase
and NcGlcAase were 56,682 and 58,123 Da, respec-
tively. The calculated isoelectric point for mature AnG-
lcAase was 5.10, while that of NcGlcAase was 6.83. The
cDNA deduced sequences of both proteins have 12
putative N-glycosylation sites; eight of these are con-
served between AnGlcAase and NcGlcAase. Indeed,
treatment of the native enzyme of A. niger with endo-gly-
canase Endo-F (Seikagaku Kogyo Co., Tokyo, Japan)
reduced the molecular mass of the native enzyme on
SDS-PAGE from 68 to 55 kDa (data not shown), con-
firming that the native enzyme is highly glycosylated.
The amino acid sequences deduced from AnGlcAase
and NcGlcAase were not similar to any known b-glucu-
ronidases that belong to the family 1 or 2 glycoside
3.2. Properties of rAnGlcAase and rNcGlcAase
The ORFs of AnGlcAase and NcGlcAase, except for
the signal peptides, were fused to a yeast secretion signal
sequence (a-factor), and introduced into the methylo-
trophic yeast P. pastoris. Recombinant b-glucuronidases
(rAnGlcAase and rNcGlcAase) were induced under the
control of the alcohol oxidase promoter and purified
from the culture medium by conventional chromatogra-
phies (Table S1). The molecular masses of rAnGlcAase
(85.7 kDa) and rNcGlcAase (90.1 kDa) determined on
SDS-PAGE were larger than those expected from the
protein sequences (56.7 kDa for AnGlcAase, 58.1 kDa
for NcGlcAase). The purified rAnGlcAase specimen
was accompanied by a trace amount of impurity with
a slightly lower molecular mass, which could not be re-
moved (Fig. S1). It is well known that molecular masses
of recombinant proteins expressed in P. pastoris with a-
factor occasionally increase due to incomplete process-
ing of the a-factor presequence at N-terminals of the
expressed proteins.34 We therefore sequenced the
N-terminal amino acid residues of rAnGlcAase with a
protein sequencer and found SMNSHPTA. This
sequence corresponds to the predicted N-terminal
sequence (Fig. 1A) preceded by a 4 residue long exten-
sion (SMNS), which derives from the ClaI and EcoRI

T. Konishi et al. / Carbohydrate Research 343 (2008) 1191–1201
1195
A
B
Figure 1. Amino acid sequence alignment and phylogenetic analysis of AnGlcAase and NcGlcAase. (A) The amino acid sequences of AnGlcAase
and NcGlcAase were aligned using ClustalW. The amino acid residues are numbered from the first methionine. Gaps (À) were introduced to achieve
optimal alignment. Identical amino acid residues among these sequences are highlighted in black. The solid line indicates the amino acid sequence
corresponding to that determined for the native enzyme purified from Pectinex Ultra SP-L. The arrowheads indicate the putative cleavage sites of the
signal peptide, and asterisks indicate potential N-glycosylation sites. Based on a sequence alignment of AnGlcAase and NcGlcAase with skullcap b-
glucuronidase,18 the potential catalytic acid/base and nucleophile residues conserved among family 79 glycoside hydrolases were identified as
indicated by closed circles. (B) Phylogenetic relationships of AnGlcAase, NcGlcAase, and other family 79 enzymes were analyzed by using ClustalW.
The phylogram is constructed based on the CAZy database (http://www.cazy.org). The bar indicates substitutions per site.

1196
T. Konishi et al. / Carbohydrate Research 343 (2008) 1191–1201
sites between the a-factor and the predicted mature
cDNA insert (see Methods in Supplementary data). This
observation indicates that post-translational modifica-
tion of rAnGlcAase cleaving the a-factor has occurred.
This is likely also the case for rNcGlcAase. We also trea-
ted rAnGlcAase and rNcGlcAase with endo-glycosidase
Endo-H (Fig. S1) to find that their respective molecular
masses were reduced to 60 kDa and 62 kDa by the enzy-
matic digestion. This indicates that the higher than ex-
pected molecular masses of the recombinant enzymes
expressed in P. pastoris were most likely caused by large
high-mannose type glycans attached to the secreted pro-
teins in the yeast.35
Activities of rAnGlcAase and rNcGlcAase were
examined with PNP b-GlcA as the substrate. While
the specific activity (30.6 unit/mg protein) of rAnG-
lcAase was comparable to that (29.3 unit/mg protein)
of the native b-glucuronidase of A. niger,14 the activity
(23.4 unit/mg protein) of rNcGlcAase was slightly
lower. The maximum activity of rAnGlcAase occurred
at pH 3.0 to 4.0, but that of rNcGlcAase occurred at
6.0 when examined in reaction mixtures with the fol-
lowing buffers at final concentrations of 50 mM: gly-
cine–HCl (pH 1.0–3.5); sodium acetate (pH 3.5–6.0);
potassium phosphate (pH 6.0–8.0); Tris–HCl (pH
7.0–9.0) (data not shown). We found that rAnGlcAase
was stable within the pH range of 1–7 when exposed
to various pHs at 4 °C for 24 h, whereas rNcGlcAase
was stable between pH 6 and 9, but its activity grad-
ually decreased below pH 6. The optimum reaction
temperature was 45 °C for both recombinant enzymes.
The thermal stability of both enzymes was also almost
identical; both lost 50% of activity when heated to
50 °C for 30 min. Repeated freezing and thawing did
not impair rAnGlcAase, whereas rNcGlcAase proved
rather unstable and was thus kept at 4 °C after
purification.
3.3. Substrate specificity of recombinant enzymes
Substrate specificity of rAnGlcAase and rNcGlcAase to-
ward sugar groups was determined using various PNP
glycosides, and oligo- and polysaccharides (Table 1).
Both recombinant enzymes specifically acted on PNP
b-GlcA and b-GalA; the other PNP glycosides tested
did not serve as substrates at all. We examined the
hydrolytic activities of rAnGlcAase and rNcGlcAase to-
ward b-GlcA-(1?6)-Gal, b-GlcA-(1?6)-b-Gal-(1?6)-
Gal, 4-Me-b-GlcA-(1?6)-Gal, 4-Me-b-GlcA-(1?6)-b-
Gal-(1?6)-Gal,
and
4-Me-b-GlcA-(1?6)-b-Gal-
(1?6)-b-Gal-(1?3)-Gal, as these occur in acidic side
chains of AGPs. Both recombinant enzymes substan-
tially acted on b-GlcA residues substituted at non-reduc-
ing terminals of b-galactooligosaccharides, whereas b-
GlcA-(1?3)-Gal, a regioisomer of b-GlcA-(1?6)-Gal,
was not susceptible to the enzymes. This agrees with
the observations for the native enzyme of A. niger.14
The recombinant enzymes did, however, exhibit a signif-
icant difference in hydrolytic activity toward 4-Me-b-
GlcA residues of the oligosaccharides: 4-Me-b-GlcA res-
idues served as a good substrate for rAnGlcAase, but
not for rNcGlcAase. Several neutral oligosaccharides
tested were not hydrolyzed by the enzymes.
Although AnGlcAase and NcGlcAase are somewhat
similar to family 79 endo-b-glucuronidases/heparanases,
the enzymes hydrolyzed heparan sulfate only very
weakly (Table 1). A likely explanation is that the recom-
binant enzymes act on b-(1?4)-linked GlcA residues
Table 1. Substrate specificity of recombinant b-glucuronidases toward PNP glycosides, oligosaccharides, and polysaccharides
Substratea
Relative activityb
rAnGlcAase
rNcGlcAase
PNP glycosides
PNP b-GlcA
PNP b-GalA
100.0 0.0
4.7 0.1
100.0 0.0
3.0 0.3
Acidic oligosaccharides
b-GlcA-(1?6)-Gal
b-GlcA-(1?6)-b-Gal-(1?6)-Gal
4-Me-b-GlcA-(1?6)-Gal
4-Me-b-GlcA-(1?6)-b-Gal-(1?6)-Gal
4-Me-b-GlcA-(1?6)-b-Gal-(1?6)-b-Gal-(1?3)-Gal
b-GlcA-(1?3)-Gal
85.7 11.5
94.6 11.8
21.4 1.5
42.6 0.7
35.5 3.0
7.7 0.1
49.9 1.5
67.4 2.2
3.3 0.5
4.8 0.8
3.2 0.3
5.4 0.4
Polysaccharide
Heparan sulfate
0.25 0.1
0.43 0.1
a The enzyme was incubated with substrates at a concentration of 2 mM (PNP glycosides) or 5 mM (oligosaccharide substrates). For polymer
substrates, a concentration of 0.5% (w/v) was used. Enzyme activity was determined under the optimal conditions for each enzyme.
b Activity is expressed in % of that toward PNP b-GlcA (rAnGlcAase, 30.6 unit/mg protein; rNcGlcAase, 23.4 unit/mg protein). The values are the
averages of three assays performed separately. The following sugars did not serve as substrates at all: PNP glycosides including b-Glc, b-Gal, a-L-
Ara, b-Xyl, and a-L-Fuc groups; oligosaccharides including b-(1?3)-, b-(1?4)-, and b-(1?6)-galactobioses, laminaribiose, cellobiose, and gen-
tiobiose; polysaccharides including chitosan from crab shells, CM-cellulose, CM-curdlan, galactomannans from guar and locust bean, b-(1?4)-
galactan from lupin, b-(1?3)(1?4)-glucan from barley, laminarin, pustulan, and xylan from birchwood.

T. Konishi et al. / Carbohydrate Research 343 (2008) 1191–1201
1197
only if they are located at the non-reducing terminals of
heparan sulfate. Xylan from birchwood, which contains
small amounts of GlcA and 4-Me-GlcA residues
attached through a-linkages, and the other poly-
saccharides tested, did not serve as substrates for rAnG-
lcAase and rNcGlcAase at all.
charides, rAnGlcAase showed the highest affinity to-
ward 4-Me-b-GlcA-(1?6)-b-Gal-(1?6)-Gal.
3.5. Action of recombinant enzymes toward arabino-
galactan-proteins
The mode of action of rAnGlcAase and rNcGlcAase on
AGPs from radish leaves and roots, and their modified
AGPs treated with a-L-Arafase5,20 was analyzed using
HPAEC-PAD (Fig. 2). Initial velocities of recombinant
3.4. Kinetics of the recombinant enzymes
The effect of substrate concentration on the activity of
the recombinant enzymes was examined using PNP b-
GlcA, PNP b-GalA, and GlcA- or 4-Me-GlcA-contain-
ing acidic oligosaccharides. The resulting Km, kcat, and
catalytic efficiency (kcat/Km) values are summarized in
Table 2. The Km values of rAnGlcAase (30.4 lM) and
rNcGlcAase (38.3 lM) for PNP b-GlcA were almost
the same. These values are apparently smaller than that
of the native enzyme (230 lM),14 whereas the values of
rAnGlcAase (422 lM) and of rNcGlcAase (378 lM)
for b-GlcA-(1?6)-Gal were comparable to that
(710 lM) of the native enzyme.14 Consistent with the
weak activity of rAnGlcAase and rNcGlcAase toward
PNP b-GalA, the Km values of the enzymes for PNP
b-GalA (9.80 mM and 3.21 mM, respectively) were
much higher than those for PNP b-GlcA, giving quite
low catalytic efficiency values. Both enzymes exhibited
higher affinity for b-GlcA-(1?6)-b-Gal-(1?6)-Gal than
that for b-GlcA-(1?6)-Gal, consistent with the sub-
strate specificity summarized in Table 1. The Km values
of rAnGlcAase were higher for 4-Me-GlcA-containing
oligosaccharides than that for GlcA-containing sub-
strates. Consequently, the catalytic efficiency values of
rAnGlcAase for 4-Me-GlcA-containing oligosaccha-
rides were much lower than those for GlcA-containing
substrates. Among the 4-Me-GlcA-containing oligosac-
A
Root AGP
100
80
60
40 rAnGlcAase
rNcGlcAase
20
No enzyme
0
B
Leaf AGP
100
80
60
40
20
0
rAnGlcAase
rNcGlcAase
No enzyme
10
15
20
25
30
Elution time (min)
Figure 2. Action of recombinant b-glucuronidases on radish AGPs.
The enzymes were incubated at 37 °C for 24 h with 0.25% (w/v) a-L-
Arafase-treated AGPs prepared from radish roots (A) or leaves (B).
Control experiments were done without enzymes. The amount of
uronic acids released from the AGPs was determined with HPAEC-
PAD. The arrowheads indicate the peaks corresponding to 4-Me-GlcA,
and the arrows indicate the GlcA peaks. The amounts of 4-Me-GlcA
and GlcA released from the AGPs are summarized in Table 3.
Table 2. Kinetics of recombinant b-glucuronidases
Substratea
Km (lM)
kcat (sÀ1
)
kcat/Km (MÀ1 sÀ1
)
rAnGlcAase
PNP b-GlcA
PNP b-GalA
b-GlcA-(1?6)-Gal
b-GlcA-(1?6)-b-Gal-(1?6)-Gal
4-Me-b-GlcA-(1?6)-Gal
4-Me-b-GlcA-(1?6)-b-Gal-(1?6)-Gal
4-Me-b-GlcA-(1?6)-b-Gal-(1?6)-b-Gal-(1?3)-Gal
30.4
9800
422
26.9
6.70
20.1
32.4
5.35
12.9
17.9
8.86 Â 105
6.83 Â 102
4.78 Â 104
1.73 Â 105
3.09 Â 103
9.19 Â 103
4.74 Â 103
187
1730
1410
3790
rNcGlcAase
PNP b-GlcA
PNP b-GalA
b-GlcA-(1?6)-Gal
b-GlcA-(1?6)-b-Gal-(1?6)-Gal
38.3
3210
378
13.5
1.14
5.27
6.34
3.54 Â 105
3.56 Â 102
1.40 Â 104
4.90 Â 104
129
a Reactions were carried out at various concentrations of substrates (0.025–2.0 mM of PNP b-GlcA, 0.1–25 mM of PNP b-GalA, and 0.06–10 mM of
acidic oligosaccharides) under the optimal conditions for each recombinant enzyme. The assays were done three times independently. The Km and
kcat values were calculated by the least squares method from the Hanes–Woolf plot using the obtained activities. Because of the low activity, no
kinetic analysis of rNcGlcAase toward 4-Me-GlcA-containing oligosaccharides was performed.

1198
T. Konishi et al. / Carbohydrate Research 343 (2008) 1191–1201
Table 3. Action of recombinant b-glucuronidases on radish AGPs
AGPa
Initial velocityb
Total uronic acid released from
1 mg AGPc
% of hydrolysisd
4-Me-GlcA (lmol/min/mg protein)
GlcA (lmol/min/mg protein)
4-Me-GlcA (lg)
GlcA (lg)
rAnGlcAase
Leaf (+)
1.526 0.215
0.010 0.001
7.455 1.299
0.669 0.028
0.287 0.040
0.004 0.000
0.282 0.053
0.019 0.003
18.30 0.26
0.82 0.01
51.68 0.87
12.82 0.06
2.07 0.04
0.31 0.11
1.80 0.04
0.77 0.00
20.2
1.8
35.0
10.5
Leaf (À)
Root (+)
Root (À)
rNcGlcAase
Leaf (+)
0
0
0
0
0.015 0.002
0.004 0.000
0.014 0.002
0.012 0.002
0.72 0.02
0
3.25 0.03
0.23 0.01
1.63 0.04
0.28 0.06
1.36 0.03
0.46 0.01
2.3
0.5
3.0
0.5
Leaf (À)
Root (+)
Root (À)
a Radish leaf and root AGPs treated with a-L-Arafase (+) or untreated (À) were used as substrates.
b The enzyme was incubated with AGPs at a concentration of 0.25% (w/v, total 1 mg) and the initial velocity of 4-Me-GlcA and GlcA release was
determined. The data are means SE of three independent experiments.
c The amounts of 4-Me-GlcA and GlcA residues released after exhaustive (24 h) digestion with the recombinant enzymes (0.5 milliunits of each
recombinant enzyme) are shown. The data are means SE of three independent experiments.
d Total uronic acid contents of AGPs are as follows: Leaf (+), 10.1% (w/w; based on total sugar content); leaf (À), 6.2%; root (+), 15.3%; root (À),
13.0%. Values indicate the ratio of released uronic acids to total uronic acids.
b-glucuronidases releasing 4-Me-GlcA and GlcA resi-
dues from radish AGPs were determined (Table 3). As
implied by the kinetic values of rAnGlcAase and
rNcGlcAase for 4-Me-GlcA- or GlcA-containing oligo-
saccharides (Tables 1 and 2), rAnGlcAase released both
uronosyl groups from native and a-L-Arafase-treated
AGPs, while rNcGlcAase acted largely only on GlcA
groups. The initial velocity of rAnGlcAase toward both
uronic acids of the leaf and root AGPs markedly in-
creased after a-L-Arafase treatment, while that of
NcGlcAase did not increase as much following treat-
ment. After exhaustive digestion, rAnGlcAase split
20% and 35% of total uronic acids from a-L-Arafase-
treated radish leaf and root AGPs, respectively, whereas
rNcGlcAase split only 2.3% and 3.0% from a-L-Arafase-
treated leaf and root AGPs, respectively (Fig. 2, Table
3). Recombinant NcGlcAase liberated only small
amounts of 4-Me-GlcA residues, while the amounts of
GlcA released from the AGPs by rNcGlcAase were
comparable to those released by rAnGlcAase. These re-
sults indicate the higher capability of rAnGlcAase to re-
lease uronosyl residues attached to the carbohydrate
moieties of AGPs.
Considering the structure of the sugar moieties of rad-
ish AGPs (Chart 1), it seems highly probable that the re-
moval of L-Ara residues attached to longer b-(1?6)
galactosyl side chains of AGPs by digestion with a-L-
Arafase renders the uronosyl residues more accessible
to rAnGlcAase, resulting in effective release of the uron-
osyl residues. Apparently, the b-glucuronidases could
not liberate all uronosyl residues attached to AGPs,
even after a-L-Arafase treatment. According to our pre-
vious results,15 the uronosyl residues located at short
side chains such as (4-Me-)b-GlcA-(1?6)-b-Gal-
(1?6)-b-Gal-(1? and (4-Me-)b-GlcA-(1?6)-b-Gal-
(1? sequences, and at the inner regions of AGPs close
to the core peptides may be resistant to the b-glucuro-
nidases due to steric hindrance. Since the peptide se-
quences of AnGlcAase and NcGlcAase are highly
similar (Fig. 1A), we suggest that the distinct substrate
specificity toward 4-Me-b-GlcA residues is the result
of a difference in a few amino acid residues involved in
substrate recognition.
3.6. Transglycosylation activity of rAnGlcAase
The transglycosylation activity of rAnGlcAase was
examined in a reaction mixture containing 20 mM
PNP b-GlcA as donor and 0.5 M Gal as acceptor. Accu-
mulation of p-nitrophenol, GlcA, and transglycosyla-
tion products in the reaction mixture was monitored
up to 25 h (Fig. 3A). Most of the PNP b-GlcA in the
reaction mixture was consumed, and the accumulation
of transglycosylation products reached a maximum at
3 h. The transglycosylation products gradually de-
creased during prolonged incubation, possibly due to
hydrolysis of the products by the enzyme. The transgly-
cosylation products were analyzed by paper chromato-
graphy (Fig. 3B), which showed the formation of at least
two products, a main product (indicated by a white ar-
row) and a minor product (arrowhead). The migration
of the main product corresponded to b-GlcA-(1?6)-
Gal (S1) and its structure was confirmed (see below),
while that of the minor product corresponded to b-
GlcA-(1?3)-Gal (S2). There is also the possibility that
the minor product was formed by the successive trans-
glycosylation of two GlcA residues onto the acceptor
Gal. The kinetic parameters of rAnGlcAase for Gal as
acceptor substrate were determined with a reaction mix-
ture consisting of 20 mM PNP b-GlcA and various con-

T. Konishi et al. / Carbohydrate Research 343 (2008) 1191–1201
1199
Chart 1. A schematic model of a partial structure of radish AGPs. Radish leaf and root AGPs contain L-Fuc, L-Ara, Gal, 4-Me-GlcA, and GlcA in a
molar proportion of 5:33:57:4:1, and L-Ara, Gal, 4-Me-GlcA (24:62:14), respectively. This shows that root AGP has a rather simpler architecture,
lacking L-Fuc and containing only 4-Me-GlcA as the sole uronic acid constituent.5,20 Like other AGPs, the carbohydrate moieties of radish AGPs
have b-(1?3)-linked galactosyl backbone chains to which b-(1?6)-linked galactosyl side chains are attached through O-6. The length of the main
chains of radish AGPs is not precisely known, while galactosyl side chains are known to vary in length from single to (at least) 20 consecutive
groups.10 The side chains of radish AGPs are occasionally substituted with uronic acids at their non-reducing terminals, and with single residues and/
or short chains of L-Ara residues along the side chains. A distinctive feature of leaf AGP is the presence of L-Fuc residues attached to L-Ara residues
through a-(1?2) linkages. Treatment of radish AGPs with a-L-Arafase exposes the uronosyl residues by removing almost all L-Ara residues attached
to Gal residues. This renders them more vulnerable to rAnGlcAase and rNcGlcAase.14,15 The a-L-Fuc-(1?2)-a-L-Ara-(1? sequences of leaf AGP,
however, resist a-L-Arafase digestion.
centrations of Gal (Fig. 3C). Apparent Km and kcat val-
ues for acceptor Gal were 391 mM and 51.8 sÀ1, respec-
tively. The high Km value of rAnGlcAase for the
acceptor Gal indicates that the transglycosylation rarely
occurs under natural conditions.
dium adduct. Incubation of the transglycosylation prod-
uct with rAnGlcAase resulted in complete digestion of
the product into GlcA and Gal, indicating that GlcA
residues of the product are linked to the acceptor Gal
via b-linkages (Fig. S2A). Further structural analysis
of the transglycosylation product by 13C NMR spectro-
scopy (Fig. S2B) confirmed that it was indeed b-GlcA-
(1?6)-Gal.
Glycoside hydrolases are classified as either retain-
ing or inverting enzymes. The transglycosylation ob-
served for rAnGlcAase implies that the enzyme is a
retaining b-glycosidase catalyzed by a double-displace-
ment mechanism.36 It is well known that various b-
glucuronidases prepared from liver, snail, and Esche-
richia coli catalyze transglycosylation reactions in the
presence of various glucuronides as donors and vari-
ous acceptors.37 A rat liver b-glucuronidase catalyzes
the transfer of GlcA residues from phenyl b-GlcA to
various acceptor sugars through b-(1?3) linkages.38
Similarly, a bovine liver b-glucuronidase produces b-
GlcA-(1?3)-b-GlcA-PNP and its (1?2) regioisomer,
when incubated with PNP b-GlcA alone.39 Since
rAnGlcAase produces b-GlcA-(1?6)-Gal as a major
transglycosylation product when incubated with PNP
b-GlcA and Gal, we note that fungal b-glucuronidases
seem to markedly differ from the b-glucuronidases of
other organisms in the specificity of their transglycosy-
lation reactions.
Transglycosylation reactions of rAnGlcAase were
performed in reaction mixtures containing various
acceptor compounds. After incubation for 3 h, consider-
able amounts of transglycosylation products, possibly
formed by the transfer of GlcA residues from PNP
b-GlcA onto the monosaccharides, were observed by
paper chromatography (data not shown). The formation
of transglycosylation products was also confirmed color-
imetrically by measuring the reduced content of free
GlcA compared with that of the control assay
(Fig. 3D). Glc, Gal, Xyl, GlcNAc, and GalNAc were
good acceptors, better than Man, L-Fuc, L-Rha, D-
Ara, L-Ara, Rib, and sorbitol. It seems that rAnGlcAase
prefers saccharides with stereochemistry similar to Gal
at C-5 and C-3 positions as acceptors. The transfer of
a considerable amount of GlcA to Xyl also suggests that
rAnGlcAase can recognize OH groups in positions other
than C-6.
In order to identify the transglycosylation products,
we performed a large-scale reaction. MALDI-TOFMS
of the main transglycosylation product from Gal as
acceptor gave a signal at m/z 379.4, which matches the
mass value (379.2) of the disaccharide GlcAÁGal as a so-

1200
T. Konishi et al. / Carbohydrate Research 343 (2008) 1191–1201
B
Acceptor: Gal
A
Acceptor: Gal
100
75
50
25
0
0
6
12
18
24
Reaction time (h)
40
30
20
10
0
C
D
0.03
0.02
0.01
*
-0.5
0
0.5
1
[S]
0
0.2 0.4 0.6 0.8 1.0
Acceptor Gal (M)
100
80
60
40
20
0
Figure 3. Transglycosylation reaction of rAnGlcAase. (A) The transglycosylation reaction was examined using 20 mM PNP b-GlcA and 0.5 M Gal
for various incubation times. The amount of liberated p-nitrophenol (closed circles) and GlcA (open triangles) was determined. The amount of the
transglycosylation product (open circles) was calculated by subtraction of the amount of free GlcA from that of free p-nitrophenol. The initial
amount of PNP b-GlcA (0.4 lmol) is indicated as 100%. The data shown are the averages with SE of triplicate assays. (B) The transglycosylation
products catalyzed by rAnGlcAase were separated on paper chromatography. Standard GlcA, PNP b-GlcA, Gal, b-GlcA-(1?6)-Gal (S1), b-GlcA-
(1?3)-Gal (S2), and reaction products at various incubation times (from 0 to 25 h) are shown. The white broken line represents the origin of
chromatogram. The white arrow and arrowhead indicate the main and minor transglycosylation products, respectively. The asterisk indicates the
migration position of GlcA. (C) Kinetic parameters of the transglycosylation reaction were determined for rAnGlcAase. Reaction mixtures
containing various concentrations of Gal as acceptor and 20 mM of PNP b-GlcA as donor were incubated at 37 °C for 20 min. The inset is the
Hanes–Woolf plot of the mean SE of three independent experiments, based on which the kinetic parameters were calculated by the least squares
method. (D) The specificity of rAnGlcAase transferring GlcA residues onto monosaccharides was examined. Reactions were performed with 20 mM
PNP b-GlcA and 0.5 M of each monosaccharide at 37 °C for 3 h. Open bar and hatched bar indicate the amount of p-nitrophenol and GlcA,
respectively. The closed bar represents the difference of the amount between p-nitrophenol and GlcA, which presumably corresponds to the amount
of transglycosylation products. The initial amount of PNP b-GlcA (0.4 lmol) is indicated as 100%. Data are averages SE of triplicate assays.
3.7. Conclusion
radation of the sugar moieties of AGPs under the con-
certed action of glycosidases and glycanases secreted
by the fungi.
We have identified two representatives of a novel type of
b-glucuronidase in A. niger and N. crassa, categorized
them as family 79 glycoside hydrolases, and established
the actions of the recombinant enzymes on b-GlcA and
4-Me-b-GlcA residues of AGPs. The physiological func-
tion of AnGlcAase and NcGlcAase is likely the specific
hydrolysis of uronosyl residues of AGPs, leading to deg-
Supplementary data
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.carres.
2008.03.004.

T. Konishi et al. / Carbohydrate Research 343 (2008) 1191–1201
1201
19. Aspinall, G. O.; Hirst, E. L.; Ramstad, E. J. Chem. Soc.
1958, 593–601.
References
20. Tsumuraya, Y.; Hashimoto, Y.; Yamamoto, S.; Shibuya,
N. Carbohydr. Res. 1984, 134, 215–228.
1. Fincher, G. B.; Stone, B. A.; Clarke, A. E. Annu. Rev.
Plant Physiol. 1983, 34, 47–70.
21. Dubois, M.; Gilles, K. A.; Hamilton, J. K.; Rebers, P. A.;
Smith, F. Anal. Chem. 1956, 28, 350–356.
22. Galambos, J. T. Anal. Biochem. 1967, 19, 119–132.
23. Bradford, M. M. Anal. Biochem. 1976, 72, 248–254.
24. Laemmli, U. K. Nature 1970, 227, 680–685.
25. Milner, Y.; Avigad, G. Carbohydr. Res. 1967, 4, 359–
361.
26. Dygert, S.; Li, L. H.; Florida, D.; Thoma, J. A. Anal.
Biochem. 1965, 13, 367–374.
27. Nelson, N. J. Biol. Chem. 1944, 153, 375–380.
28. Somogyi, N. J. Biol. Chem. 1952, 195, 19–23.
29. Ishikawa, M.; Kuroyama, H.; Takeuchi, Y.; Tsumuraya,
Y. Planta 2000, 210, 782–791.
30. Bendtsen, J. D.; Nielsen, H.; von Heijne, G.; Brunak, S. J.
Mol. Biol. 2004, 340, 783–795.
31. Dempsey, L. A.; Plummer, T. B.; Coombes, S. L.; Platt, J.
L. Glycobiology 2000, 10, 467–475.
32. Nasser, N. J.; Nevo, E.; Shafat, I.; Ilan, N.; Vlodavsky, I.;
Avivi, A. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 15161–
15166.
33. Podyma-Inoue, K. A.; Yokote, H.; Sakaguchi, K.; Ikuta,
M.; Yanagishita, M. J. Biol. Chem. 2002, 277, 32459–
32465.
34. Raemaekers, R. J. M.; de Muro, L.; Gatehouse, J. A.;
Fordham-Skelton, A. P. Eur. J. Biochem. 1999, 265, 394–
403.
35. Gemmill, T. R.; Trimble, R. B. Biochim. Biophys. Acta
1999, 1426, 227–237.
36. Rye, C. S.; Withers, S. G. Curr. Opin. Chem. Biol. 2000, 4,
573–580.
37. Fishman, W. H.; Green, S. J. Biol. Chem. 1957, 225, 435–
452.
38. Niemann, R.; Buddecke, E. Biochim. Biophys. Acta 1979,
567, 196–206.
39. Langlois, V.; Parisot, J.; Bonnet, V.; Rabiller, C. Tetra-
hedron: Asymmetry 2002, 13, 2369–2373.
2. Nothnagel, E. A. Int. Rev. Cytol. 1997, 174, 195–291.
3. Majewska-Sawka, A.; Nothnagel, E. A. Plant Physiol.
2000, 122, 3–9.
4. Shi, H.; Kim, Y.; Guo, Y.; Stevenson, B.; Zhu, J.-K. Plant
Cell 2003, 15, 19–32.
5. Tsumuraya, Y.; Ogura, K.; Hashimoto, Y.; Mukoyama,
H.; Yamamoto, S. Plant Physiol. 1988, 86, 155–160.
6. Sekimata, M.; Ogura, K.; Tsumuraya, Y.; Hashimoto, Y.;
Yamamoto, S. Plant Physiol. 1989, 90, 567–574.
7. Kotake, T.; Dina, S.; Konishi, T.; Kaneko, S.; Igarashi,
K.; Samejima, M.; Watanabe, Y.; Kimura, K.; Tsumu-
raya, Y. Plant Physiol. 2005, 138, 1563–1576.
8. Hata, K.; Tanaka, M.; Tsumuraya, Y.; Hashimoto, Y.
Plant Physiol. 1992, 100, 388–396.
9. Kotake, T.; Tsuchiya, K.; Aohara, T.; Konishi, T.;
Kaneko, S.; Igarashi, K.; Samejima, M.; Tsumuraya, Y.
J. Exp. Bot. 2006, 57, 2353–2362.
´ˇ
10. Tsumuraya, Y.; Mochizuki, N.; Hashimoto, Y.; Kovac, P.
J. Biol. Chem. 1990, 265, 7207–7215.
11. Ichinose, H.; Yoshida, M.; Kotake, T.; Kuno, A.; Igar-
ashi, K.; Tsumuraya, Y.; Samejima, M.; Hirabayashi, J.;
Kobayashi, H.; Kaneko, S. J. Biol. Chem. 2005, 280,
25820–25829.
12. Okemoto, K.; Uekita, T.; Tsumuraya, Y.; Hashimoto, Y.;
Kasama, T. Carbohydr. Res. 2003, 338, 219–230.
13. Kotake, T.; Kaneko, S.; Kubomoto, A.; Haque, M. A.;
Kobayashi, H.; Tsumuraya, Y. Biochem. J. 2004, 377,
749–755.
14. Kuroyama, H.; Tsutsui, N.; Hashimoto, Y.; Tsumuraya,
Y. Carbohydr. Res. 2001, 333, 27–39.
15. Haque, M. A.; Kotake, T.; Tsumuraya, Y. Biosci.
Biotechnol. Biochem. 2005, 69, 2170–2177.
16. Henrissat, B. Biochem. J. 1991, 280, 309–316.
17. Henrissat, B.; Bairoch, A. Biochem. J. 1993, 293, 781–788.
18. Sasaki, K.; Taura, F.; Shoyama, Y.; Morimoto, S. J. Biol.
Chem. 2000, 275, 27466–27472.