Báo cáo khoa học: Comparative analysis of carbohydrate-binding properties of two tandem repeat-type Jacalin-related lectins, Castanea crenata agglutinin and Cycas revoluta leaf lectin docx

Comparative analysis of carbohydrate-binding properties
of two tandem repeat-type Jacalin-related lectins,
Castanea crenata agglutinin and Cycas revoluta leaf lectin
Sachiko Nakamura
1
, Fumio Yagi
2
, Kiichiro Totani
3
, Yukishige Ito
3
and Jun Hirabayashi
1
1 Glycostructure Analysis Team, Research Center for Glycoscience, National Institute of Advanced Industrial Science and Technology, Japan
2 Department of Biochemical Science and Technology, Faculty of Agriculture, Kagoshima University, Japan
3 RIKEN (The Institute of Physical and Chemical Research), Saitama, Japan
Lectins are carbohydrate-binding proteins distributed
in all of the organisms characterized so far. A large
number of plant lectins have been isolated and charac-
terized [1,2]. Van Damme et al. classified them into
seven families based on their molecular structures and
carbohydrate-binding specificities [3]. Members belong-
ing to each family share some common properties.
Compared with legume and monocot lectin families,
jacalin-related lectins (JRLs) were originally assumed
to form a relatively small family [4]. During the last
decade, however, many new members belonging to this
family were discovered, and some novel features
became evident. As a result, JRLs can now be classi-
fied into two subgroups in terms of their carbohydrate
binding specificities, i.e. galactose-binding-type JRLs

(gJRLs) and mannose-binding-type JRLs (mJRLs) [5].
Keywords
carbohydrate binding specificity; frontal
affinity chromatography; Jacalin-related
lectin family; lectin
Correspondence
J. Hirabayashi, Glycostructure Analysis
Team, Research Center for Glycoscience,
National Institute of Advanced Industrial
Science and Technology, AIST Tsukuba
Central 2, 1-1-1, Umezono, Tsukuba,
Ibaraki 305-8568, Japan
Fax: +81 29 861 3125
Tel: +81 29 861 3124
E-mail:
(Received 27 December 2004, revised 23
February 2005, accepted 4 April 2005)
doi:10.1111/j.1742-4658.2005.04698.x
Lectins belonging to the jacalin-related lectin family are distributed widely
in the plant kingdom. Recently, two mannose-specific lectins having tan-
dem repeat-type structures were discovered in Castanea crenata (angio-
sperm) and Cycas revoluta (gymnosperm). The occurrence of such similar
molecules in taxonomically less related plants suggests their importance in
the plant body. To obtain clues to understand their physiological roles, we
performed detailed analysis of their sugar-binding specificity. For this
purpose, we compared the dissociation constants (K
d
)ofCastanea crenata
agglutinin (CCA) and Cycas revoluta leaf lectin (CRLL) by using 102
pyridylaminated and 13 p-nitrophenyl oligosaccharides with a recently

developed automated system for frontal affinity chromatography. As a
result, we found that the basic carbohydrate-binding properties of CCA
and CRLL were similar, but differed in their preference for larger N-linked
glycans (e.g. Man7–9 glycans). While the affinity of CCA decreased with
an increase in the number of extended a1–2 mannose residues, CRLL
could recognize these Man7–9 glycans with much enhanced affinity. Nota-
bly, both lectins also preserved considerable affinity for mono-antennary,
complex type N-linked glycans, though the specificity was much broader
for CCA. The information obtained here should be helpful for understand-
ing their functions in vivo as well as for development of useful probes for
animal cells. This is the first systematic approach to elucidate the fine spe-
cificities of plant lectins by means of high-throughput, automated frontal
affinity chromatography.
Abbreviations
CCA, Castanea crenata agglutinin; CRLL, lectin from leaves of Cycas revoluta; CRD, carbohydrate-recognition domain; FAC, frontal affinity
chromatography; gJRLs, galactose-binding-type Jacalin-related lectins; JRLs, Jacalin-related lectins; mJRLs, mannose-binding-type Jacalin-
related lectins; M2M2M3Mb,Mana1–2Mana1–2Mana1–3Manb; M2M3M6Mb,Mana1–2Mana1–3Mana1–6Manb; MTX, methotrexate;
M3GN2, Man3GlcNAc2; PA, pyridylaminated; pNP, p-nitrophenyl.
2784 FEBS Journal 272 (2005) 2784–2799 ª 2005 FEBS
gJRLs are represented by Jacalin, known as a useful
probe for IgA [6], and also by Maclura pomifera agglu-
tinin [4] and Morus nigra agglutinin (Morniga G) [3].
All of these lectins, which come from Moraceae plants,
have common structural features: they are composed of
four protomers consisting of a short b-chain (2 kDa)
and a long a-chain (13 kDa) as a result of proteolytic
cleavage of a precursor polypeptide [3,7]. On the other
hand, the mJRL subgroup comprises many more mem-
bers, such as Calsepa isolated from Calystegia sepium
[8], Artocarpin from Artocarpus integrifolia [9], Heltuba

from Helianthus tuberosus [10], BanLec from Musa acu-
minate [11], Morniga M from M. nigra [7], Orysata
from Oryza sativa [12,13] and PAL from Phlebodium
aureum [14]. Although these mJRLs were proved to
form a b-prism fold I structure consisting of one carbo-
hydrate-recognition domain (CRD) similar to Jacalin
[15–17], no proteolytic modification of mJRLs occurs.
There is a view that the lack of proteolysis of them
may result in preservation of their mannose ⁄ glucose
specificity [18]. Although all of the mJRLs mentioned
above are of the single CRD type (15–16 kDa), new
members of much larger molecular size (% 33 kDa)
were discovered recently from Castanea crenata (angio-
sperm) and Cycas revoluta (gymnosperm) [19,20].
Sequence analysis of these large mJRLs, i.e. C. crenata
agglutinin (CCA) and C. revoluta leave lectin (CRLL)
revealed that they form a tandem-repeat structure
composed of two jacalin-type CRDs [21,22]. Therefore,
mJRLs are now thought to be distributed more widely
in the plant kingdom with more structural diversity
than had ever been previously thought.
Both CCA and CRLL showed a similar extent of
homology (30–40% in amino acid identity) to other
JRLs. Although the complete amino acid sequence has
not yet been determined for CRLL, a high extent of
intramolecular homology, i.e. between N-terminal and
C-terminal CRDs, is also evident for both CCA and
CRLL (> 35%) [19,21]. Since Greek key motif 3, a
region assumed to form a carbohydrate-binding site, is
highly conserved in both N-terminal and C-terminal

CRDs in CCA and CRLL, all of these CRDs are
likely to maintain a sugar-binding function. Thus, dis-
tribution of a similar type of molecule in taxonomically
unrelated plants raises a basic question about the bio-
logical significance of tandem repeat-type mJRLs. In
this context, there are some lines of evidence that cer-
tain mJRLs are induced by treatment with methyl
jasmonate or by salt stress [12,23]. This suggests that
mJRLs have some defensive roles in the plant body.
However, there is no clear evidence for this hypothesis
or no such report for CCA and CRLL. To understand
their functions in plants, it is essential to elucidate
their biochemical properties in terms of carbohydrate-
binding specificities.
For this purpose, we recently developed an automa-
ted frontal affinity chromatography (FAC) system,
which enables us to analyse lectin–oligosaccharide
interactions in a high-throughput manner [24]. Signifi-
cant advantages of FAC include high sensitivity and
reproducibility. In addition, the method is well suited
for determination of dissociation constants (K
d
s) for
relatively low-affinity binding (e.g. K
d
>10
)6
m), thus
making it practically advantageous for analysis of
lectin–oligosaccharide interactions. FAC was originally

developed by Kasai et al. [25], reinforced by Hirabaya-
shi et al. [26], and proved to be an effective alternative
successfully applied to comparative analysis of animal
lectins with a set of fluorescently labelled glycans [27].
Other investigators also used an FAC system equipped
with an MS detector for analysis of mushroom lectins,
and demonstrated its efficiency [28,29].
In this present study we applied two tandem repeat-
type mJRLs, CCA and CRLL, to this automated FAC
system. As a result, both conserved and divergent fea-
tures of these taxonomically unrelated mJRLs became
evident.
Results
Evaluation of the lectin columns
CCA and CRLL were purified by affinity chromatog-
raphy on asialofetuin and mannose–agarose columns,
respectively, as previously described [19,20]. The thus
purified proteins were immobilized on NHS-activated
Sepharose 4FF, and the resulting resins were packed
into miniature columns (inner diameter, 2 mm; length,
10 mm; bed volume, 31.4 lL) as described under
Experimental procedures. The amounts of immobilized
CCA and CRLL were determined to be 2.0 and
1.1 mgÆmL
)1
gel, respectively. For evaluation of the
prepared columns, it was necessary to determine the
effective ligand content (B
t
) based on the so-called

‘concentration-dependence analysis’ [26,27]. For this
purpose, however, none of the commercially available
p-nitrophenyl (pNP) derivatives of simple saccharides
tested, i.e. Man-a, Man-b, Glc-a, Gal-a, Gal-b,
GalNAc-a, GalNAc-b, Fuc-a, Galb1–4Glc-b, Galb1–
4GlcNAc-b, Galb1–3GalNAc-a, Glca1–4Glc-a or
(Glca1–4)
5
-a, showed any significant affinity for these
columns. Since these lectins are known to show high
affinity for the mannotriose structure, Mana1–
3(Mana1–6)Man [19,20], we tested methotrexate
(MTX)-derivatized Man
3
GlcNAc
2
(M3GN2-MTX,
Fig. 1A) previously synthesized successfully [30] to see
S. Nakamura et al. Oligosaccharide specificity of Jacalin-related lectins
FEBS Journal 272 (2005) 2784–2799 ª 2005 FEBS 2785
if it would be appropriate for the above concentration-
dependence analysis. M3GN2-MTX showed strong
retardation when it was applied to these lectin columns
at the concentration of 5 lm, whereas it showed no
significant binding to a BSA–agarose column
(2.6 mgÆmL
)1
, data not shown). Hence, concentration-
dependence analysis was performed with M3GN2-
MTX at various concentrations ranging from 5 to

20 lm (Fig. 1B,C). As a result, B
t
and K
d
values were
determined to be 1.49 nmol and 1.2 · 10
)5
m, respect-
ively, for the CCA column, and 0.81 nmol and
2.0 · 10
)6
m, respectively, for the CRLL column.
Based on these data, availability of the CCA and
CRLL columns was calculated to be 39 and 40%,
respectively. Specifications obtained for these columns
are summarized in Table 1.
As regards the detection limit of low-affinity bind-
ing, such as those for the pNP-sugars described above,
we found approximately 2 lL of experimental error in
the V–V
0
value in the present FAC system considering
the data collection interval (1 s) and the flow rate
(0.125 mLÆmin
)1
). Under such conditions, low-affinity
saccharides having K
d
values > 7.5 · 10
)4

m and
> 4.1 · 10
)4
m for the above CCA and CRLL col-
umns, respectively, cannot be precisely characterized.
On the other hand, the maximum V–V
0
value measur-
able in the present system is at least 120 lL, which
corresponds to K
d
values of 1.3 · 10
)5
m and
6.8 · 10
)6
m for the CCA and CRLL columns,
respectively. Thus, a dynamic range of 60-fold is
achieved by using these columns with no change in lig-
and concentrations.
Fig. 1. Determination of Bt values. (A) Struc-
tural formula of MTX-derivatized Man3Glc-
NAc2 (M3GN2-MTX), which was used for
concentration-dependence analysis. For det-
ermination of B
t
values for the immobilized
CCA (B) and CRLL (C), M3GN2-MTX was
diluted to various concentrations (5–20 l
M)

and applied to each column. The solid lines
and dotted lines indicate elution profiles of
M3GN2-MTX and control sugar (pNP-Lac-
tose), respectively (left). Woolf–Hofstee-type
plots were made by using V–V
0
values
(right). For details, see text.
Table 1. Specifications of CCA- and CRLL-immobilized columns used in this study.
Lectin name Origin Immobilized (mg ⁄ mL) Bt (nmol) Availability (%) r
2a
M3GN2-MTX K
d
(M)
CCA Castanea crenata 2.0 1.49 39 1.00 1.2 · 10
)5
CRLL Cycas revoluta 1.1 0.81 40 0.97 2.0 · 10
)6
a
Reliability of lines obtained as a result of Woolf–Hofstee-type plot in each concentration-dependence analysis.
Oligosaccharide specificity of Jacalin-related lectins S. Nakamura et al.
2786 FEBS Journal 272 (2005) 2784–2799 ª 2005 FEBS
Overall features of oligosaccharide specificities
of CCA and CRLL
In order to briefly profile oligosaccharide specificities
of CCA and CRLL from a global viewpoint, we pre-
pared a panel of 102 pyridylaminated (PA) glycans
including 55 N-linked glycans and 38 glycolipid-type
glycans (Fig. 2). For the determination of K
d

, retarda-
tion of the elution front relative to that of PA-lactose,
i.e. V–V
0
, was measured for each analyte solution dilu-
ted to either 2.5 or 5.0 nm. The amount of glycan
required for determination of reliable K
d
value was
< 3 pmol, which is much smaller (< 10
)3
) compared
with other methods. Since the concentrations are much
lower than K
d
values assumed for the present lectin
columns as described (> % 10
)5
m),K
d
values could
be calculated according to Eqn (2) by using observed
V–V
0
values in a manner independent of [A]
0
. For the
sake of comparison, bar graph representation was
made in terms of affinity constant (K
a

) in Fig. 3. Obvi-
ously, both CCA and CRLL showed affinity for a
wide range of N-linked glycans, but not at all for
glycolipid-type glycans, which lack mannose. This
observation is reasonable, because these lectins were
characterized as mannose-specific lectins. Thus, global
features of CCA and CRLL are apparently similar,
but detailed features are different as described below.
Comparison of fine specificities
High-mannose-type N-glycans
Chromatograms obtained for N-linked type glycans
are shown in Fig. 4. In the case of CCA (Fig. 4A), the
strongest affinity was observed for 005 (designated
M5, K
d
¼ 1.0 · 10
)5
m), followed by 007 (M6,
1.2 · 10
)5
m) and 003 (M3, 1.4 · 10
)5
m). On the
other hand, CCA showed no affinity toward 001 (M2).
These results indicate that the Mana1–3Manb struc-
ture is essential for CCA binding and that the removal
of a1–3Man abolished the affinity for CCA. Because
CCA could recognize 004 (M4, 2.6 · 10
)5
m), which

lacks the depicted Mana1–3Manb structure, it proved
to have significant affinity to Mana1–3Mana, too. In
contrast, removal of a1–6Man from M3 (compare 002
and 003) had rather a small effect on affinity (78% rel-
ative to 003). This observation agreed with pervious
analysis by means of haemagglutination inhibition
assay [20,22], isothermal titration caloriemetry and
enzyme-linked lectin assay (unpublished data) toward
simple saccharides. The binding affinity toward CCA
was much reduced, when the nonreducing terminal
Man of the Mana1–3Manb structure was modified
by a1–2Man (008–014, corresponding to M7–9). The
tendency was confirmed by comparison of M6 isomers,
006 and 007; i.e. their affinities relative to that of 005
were 36% and 83%, respectively.
CRLL also showed significant affinity for relatively
small high-mannose type glycans, except for 001 and
004 (Figs 3 and 4B). CRLL showed moderate affinity
for an M2 saccharide, 002 (K
d
¼ 4.7 · 10
)5
m), but
not at all for its isomer, 001. Moreover, CRLL did not
bind to 004, which lacks the Mana1–3Manb structure.
These results indicate that the core Mana1–3Manb
structure forms an essential unit for CRLL recogni-
tion. Considering the inability of CRLL to bind to
004, the terminal Mana1–3Mana cannot substitute for
the core Man a1–3Manb structure, unlike the case for

CCA. However, CRLL showed somewhat (30%)
higher affinity for 005 (K
d
¼ 3.0 · 10
)5
m) than for
003 (3.9 · 10
)5
m). By comparison between 002 (M2,
4.7 · 10
)5
m) and 003 (3.9 · 10
)5
m), it is clear that
addition of the Mana1–6Manb branch had only a
small effect, if any, on CRLL–glycan interaction.
CRLL also showed a significant difference in affinity
for M6 isomers (006 and 007). When compared with
their parental molecule, M5 (005, K
d
¼ 3.0 · 10
)5
m),
the addition of a1–2Man to the Mana1–3Manb
branch resulted in almost complete loss of affinity;
whereas that to the Mana1–3Mana1–6Manb branch
still enhanced the affinity (2.4 · 10
)5
m). The tendency
observed here is essentially the same as that observed

for CCA. However, the effect of the a1–2Man addition
makes a clear contrast between the two lectins. There-
fore, the addition of a1–2Man to the core Mana1–
3Manb had rather a destructive effect on CRLL
recognition.
The most distinguishing feature of CRLL is its highly
enhanced affinity for relatively large high-mannose-type
glycans, i.e. M8–9. As a matter of fact, CRLL showed
the strongest affinity for 013 (M8, K
d
¼ 1.1 · 10
)5
m)
and 014 (M9, 9.5 · 10
)6
m), whereas CCA could not
bind to these large saccharides at all. Binding ability of
CRLL to large high-mannose-type glycans was consis-
tent with the results obtained by haemagglutination
inhibition assay using glycopeptides [19]. Since these
saccharides share two common structures, i.e. Mana1–
2Mana1–2Mana1–3Manb (M2M2M3Mb) and Mana1–
2Mana1–3Mana1–6Manb (M2M3M6Mb), coincidence
of these extended structural units may contribute to the
observed high affinity in CRLL. In this regard, CRLL
showed 1.9-times stronger affinity for an M7 saccharide,
009, containing the M2M2M3Mb unit (K
d
¼ 4.1 ·
10

)5
m) than for its M7 isomer, 010, containing the
M2M3M6Mb unit (7.9 · 10
)5
m). Similar results were
observed for M8 isomers, 012 (K
d
¼ 3.0 · 10
)5
m) con-
taining the M2M2M3Mb unit and 011 (4.9 · 10
)5
m)
S. Nakamura et al. Oligosaccharide specificity of Jacalin-related lectins
FEBS Journal 272 (2005) 2784–2799 ª 2005 FEBS 2787
Fig. 2. Schematic representation of oligosaccharide structures. Note that the reducing terminal is pyridylaminated for FAC analysis. Symbols
used to represent pyranose rings of monosaccharides are shown in the box at the bottom of the figure. Anomeric carbon, i.e. position 1, is
placed at the right side, and 2, 3, 4… are placed clockwise. Thin and thick bars represent a and b linkage, respectively.
Oligosaccharide specificity of Jacalin-related lectins S. Nakamura et al.
2788 FEBS Journal 272 (2005) 2784–2799 ª 2005 FEBS
containing the M2M3M6Mb unit. In contrast, the
increase in affinity was relatively small, when a1–2Man
was added to the M6M6Mb unit; i.e. compare M8 iso-
mers, 011 (4.9 · 10
)5
m) and 012 (3.0 · 10
)5
m) with
their parental M7 isomers, 010 (7.9 · 10
)5

m) and 090
(4.1 · 10
)5
m), respectively. Since the increase was
approximately 1.5-fold in these cases, the addition of
the terminal a1–2Man to the M6M6Mb unit still
made some contribution to affinity enhancement. In this
context, the strongest contribution of a1–2Man was
observed when it was added to the M2M3Mb unit:
compare 010 with 013, and 011 with 014 (affinity
Fig. 3. Bar graph representation of affinity
constants (K
a
) of CCA (left) and CRLL (right)
toward N-linked glycans. The small Arabic
figures in the centre correspond to sugar
numbers indicated in Fig. 2; whereas large
Roman figures on the left side of graphs
represent types of glycans: high-mannose-
type (I), agalacto-type (II), galactosylated-
type (III) and sialylated-type (IV) N-linked
glycans, glycolipid-type glycans (V), and
others (VI).
S. Nakamura et al. Oligosaccharide specificity of Jacalin-related lectins
FEBS Journal 272 (2005) 2784–2799 ª 2005 FEBS 2789
Fig. 4. Elution profiles of N-linked glycans obtained with CCA (A) and CRLL (B) columns. Chromatograms of N-linked glycans are shown in
the order of sugar numbers together with retardation volumes (upper, V–V
0
, lL) and dissociation constants (lower, K
d

, M). For the sake of
convenience, the elution pattern of each saccharide is overlaid with that of PA-lactose, which has no affinity for either lectin, i.e. the negative
control.
Oligosaccharide specificity of Jacalin-related lectins S. Nakamura et al.
2790 FEBS Journal 272 (2005) 2784–2799 ª 2005 FEBS
enhanced by 7.2 and 5.2 times, respectively). In the
remaining case of the a1–2Man addition, i.e. to the
M3M6Mb unit, the effect was rather intermediate
(approximately 3.5 times), when the affinity of 013 and
014 were compared with that of their parent saccha-
rides, 009 and 012, respectively. These results indicate
that the presence of nonreducing end mannose in the
M2M2M3Mb unit plays a dominant role in the strong
Fig. 4. (Continued).
S. Nakamura et al. Oligosaccharide specificity of Jacalin-related lectins
FEBS Journal 272 (2005) 2784–2799 ª 2005 FEBS 2791
interaction between CRLL and large high-mannose-
type glycans, whereas M2M3M6Mb and M2M6M6Mb
contribute to lesser extents (i.e. M2M3M6Mb >
M2M6M6Mb). From a practical viewpoint, glycans
having both M2M2M3Mb and M2M3M6Mb units
show the highest affinity for CRLL.
Complex-type glycans
In the present study, both CCA and CRLL were found
to bind to several complex-type N-linked glycans, too.
As these lectins were previously characterized as man-
nose-specific lectins [19,20], possible reasons for this
discrepancy should be given. In the case of CCA, it
showed significant affinity for 101 (K
d

¼ 2.2 · 10
)5
m),
201 (2.8 · 10
)5
m), 301 (1.6 · 10
)5
m), and 401
(2.3 · 10
)5
m); whereas it showed lower affinity for
their position isomers, i.e. 102 (3.5 · 10
)5
m), 302
(3.7 · 10
)5
m), and 402 (5.5 · 10
)5
m) relative to 101,
301 and 401, respectively. This observation indicates
that CCA prefers mono-antennary, complex-type
N-linked glycans, which have a1–6 branch, or nonsub-
stituted a1–3 Man. Such a feature is consistent with
the idea that Mana1–3Manb forms the core recogni-
tion unit. On the other hand, we also found that CCA
showed significant affinity for bi-antennary glycans,
i.e. 103 (K
d
¼ 2.9 · 10
)5

m), 202 (5.2 · 10
)5
m), 304
(3.4 · 10
)5
m), 307 (3.7 · 10
)5
m), 403 (5.0 · 10
)5
m),
404 (5.4 · 10
)5
m), and 405 (5.3 · 10
)5
m). Their
affinities for CCA were similar (3–5 · 10
)5
m), but
somewhat (40–80%) reduced in comparison with those
for the respective parental mono-antennary glycans,
i.e. 101, 201, 301, and 401. These results indicate that
modification of a1–3Man is never detrimental. This
may explain why CCA could bind to asialofetuin-
agarose [20], because this glycoprotein contains bi-
antennary, complex-type glycans [31]. On the other
hand, tri-antennary glycans showed no affinity for
CCA. Therefore, C4-OH group of a1–3Man is essen-
tial for CCA recognition.
Similar to CCA, CRLL also showed significant
binding to mono-antennary, complex-type N-linked

glycans having the a1–6 branch, i.e. 101 (K
d
¼
3.9 · 10
)5
m), 201 (2.7 · 10
)5
m), 301 (4.1 · 10
)5
m),
and 401 (2.4 · 10
)5
m). Unlike CCA, however, CRLL
did not show affinity at all for their position isomers
(102, 302, and 402). Moreover, CRLL had no affinity
for bi-antennary glycans (e.g. 103, 104, 202, 203, 303,
304, 307, 308, 403, 404, 405, and 406). Neither tri-
antennary nor tetra-antennary N-linked glycans were
targets for CRLL, either. Therefore, CRLL is stricter
than CCA in that the former never permits substitu-
tion of a1–3Man in complex-type glycans.
In the present study, the influence of a1–6 (core)
fucosylation on lectin–glycan interaction could also be
examined by comparison between 003 and 015, 101
and 201, 103 and 202, 104 and 203, 301 and 401, 302
and 402, 304 and 403, and 307 and 405 (Figs 3 and 4).
In the case of CCA, a1–6 (core) fucosylation slightly
reduced the affinity; for example, compare 101 (K
d
¼

2.2 · 10
)5
m) with 201 (2.8 · 10
)5
m). With CRLL,
however, this type of fucosylation enhanced the affinity
by 1.2–1.7 times; compare 101 (K
d
¼ 3.9 · 10
)5
m)
with 201 (2.7 · 10
)5
m). The effect of a1–6 fucosyla-
tion was significant, but not very drastic in both cases.
So, the presence of a1–6Fuc does not have any essen-
tial role for the recognition of these mJRLs.
Oligosaccharides 501–505 represent sialylated gly-
cans. Among them, bi-antennary glycans (501, 502,
and 503) were recognized by CCA to some degree,
whereas none of them showed significant affinity for
CRLL. In the case of CCA, it is clear that the effect
of sialylation of bi-antennary glycans was different
between mono-sialylated isomers i.e. 501 (K
d
¼
1.1 · 10
)4
m) and 502 (4.5 · 10
)5

m). By comparison
with nonsialylated glycan, 307 (3.7 · 10
)5
m), the inhi-
bitory effect of sialylation was more drastic on the
a1–3 branch (affinity reduced to 34%) than on the
a1–6 branch (to 82%). Again, this result is consistent
with the above observation that Mana1–3Manb is an
essential unit for CCA recognition.
Summary of FAC analysis of CCA and CRLL
The sugar-binding properties of CCA described above
may be summarized as follows: (a) CCA shows affinity
for mannose-containing N-linked glycans, but not for
glycolipid-type glycans; (b) it binds to relatively small
high-mannose-type glycans, which contain a nonsubsti-
tuted a1–3Man residue in the tri-mannosyl core struc-
ture; (c) the binding is greatly diminished by the
addition of a1–2Man residue(s) to the a1–3Man; (d)
CCA can also bind to mono-, bi-antennary N-liked
glycans with varied affinities, but not to tri- and tetra-
antennary ones.
On the other hand, CRLL has the following fea-
tures: (a) similar to CCA, CRLL shows affinity only
for N-linked glycans but not for glycolipid-type ones;
(b) unlike CCA, however, CRLL binds to relatively
large high-mannose-type glycans with much increased
affinity; (c) a1–2Man extension in the M2M2M3Mb
unit makes the strongest contribution to such high
affinity for the large high-mannose type glycans; (d)
affinity enhancement is also supported by a1–2Man

extension in the other units M2M3M6Mb and
M2M6M6Mb; (e) CRLL can bind only to a1–6
Oligosaccharide specificity of Jacalin-related lectins S. Nakamura et al.
2792 FEBS Journal 272 (2005) 2784–2799 ª 2005 FEBS
branched mono-antennary, complex-type N-linked gly-
cans, but never to a1–3 branched mono-, bi-, tri- and
tetra-antennary glycans.
Discussion
By using our recently developed FAC system, we ana-
lysed detailed sugar-binding specificities of CCA and
CRLL. The advantage of FAC in sensitivity, economy
and reliability enabled us to reveal fine specificity of
these lectins. Based on the results, both conserved and
divergent properties were reveled for these two
mJRLs. For relatively small high-mannose-type gly-
cans, CCA and CRLL showed similar sugar-binding
profiles: they bound to Mana1–3Manb as a main
recognition unit, and the affinity was reduced by the
addition of an a1–2Man residue to the Mana1–3Manb
structure. However, according to the X-ray crystallo-
graphic study on Heltuba, another mJRL, in complex
with a1–3 mannobiose, no direct hydrogen bond was
found at the O2 atom of mannose [5]. Thus, substi-
tution of 2-OH with a1–2Man seems to have no
destructive effect on interactions. On the other hand,
a modelling study on Artocarpin, another mJRL, with
ManNAc also revealed that severe steric hindrance
occurs between the N-acetyl group of ManNAc and
the residues in the loops b1-b2 and b11-b12, which
construct the primary-binding site [32]. At a glance,

the amino acid sequences around the region critical
for sugar-binding function are tightly conserved
among mJRLs (Fig. 5). Therefore, the reason why the
affinity was diminished by the addition of a1–2Man
residue(s) to the core Mana1–3Manb unit can prob-
ably be attributed to significant steric hindrance by
the residue.
As regards the relatively large high-mannose-type
glycans (M7–M9), which showed much enhanced affin-
ity only for CRLL, considerable steric hindrance will
Fig. 5. Sequence alignment of mJRLs. Individual CRDs of CCA (N- and C-domains) and partial amino acid sequences reported for CRLL [19]
are aligned with other representative mJRLs. Residues conserved in all mJRLs are indicated in bold letters. Secondary structure elements
(b-sheet) and a1–3 mannobiose-binding residues reported for Heltuba are indicated by arrows and filled circles, respectively [5].
S. Nakamura et al. Oligosaccharide specificity of Jacalin-related lectins
FEBS Journal 272 (2005) 2784–2799 ª 2005 FEBS 2793
occur, if they interact in the same manner as the
mannobiose discussed above. Therefore, CRLL may
create an alternative binding mode specifically for such
large high-mannose-type glycans. In Heltuba, three
loops (b1–b2, b7–b8 and b11–b12) construct the sugar-
binding site, and another loop ( b5–b6) affects protein–
sugar interactions. Among these loops, CRLL, but not
CCA, has two large deletions corresponding to b5 and
loop b7–b8 (Fig. 5, T Haraguchi & F Yagi, unpub-
lished data). These deletions should affect the overall
conformation of these loops and thus leave space,
which results in a change in the binding mode. In this
context, the interaction between CRLL and large high-
mannose-type glycans was drastically enhanced by the
coexistence of M2M2M3Mb and M2M3M6Mb units.

This observation implies multiple binding via either
inter- or intramolecular interactions. In fact, an X-ray
crystallographic study demonstrated that octamer
assembly enables Heltuba to cross-link with mannose
moieties belonging to different antennae of a single
N-glycan [5]. Because CRLL exists as a monomer [19],
such an effect may be compensated by its tandem-repeat
type architecture. At present we don’t know whether
such assumed cross-linkage actually occurs, and whether
CRLL is composed of two homologous but distinct
domains in terms of sugar-binding properties or not. To
clarify these points, it will be necessary to analyse sugar-
binding profiles of individual domains for both lectins.
In addition to having affinity for high-mannose-
type glycans, CCA showed significant affinity for
mono- and bi-antennary, complex-type N-linked gly-
cans. In other words, the addition of b1–2GlcNAc to
Mana1–3Manb is acceptable for CCA. This fact sug-
gests that steric hindrance is more significant when a
‘double axial’ linkage is formed, e.g. by C1-a-OH of
nonreducing terminal mannose and C2-OH of redu-
cing terminal mannose in Mana1–2Man, than in the
case of an ‘equatorial–axial’ linkage, e.g. when Glc-
NAcb1–2Man is formed. On the other hand, because
CRLL had no affinity for a1–3 branched (a1–3Man
substituted), mono-antennary N-linked glycans, it
permits neither kind of linkage. In contrast, a1–6
branched (a1–6Man substituted), mono-antennary
N-linked glycans showed affinity for both CCA and
CRLL. In the Artocarpin–mannotriose complex, it

was revealed that a1–6Man was exposed to solvent
and had no significant interaction with the protein
[32]. Therefore, the observed affinity of CCA and
CRLL for a series of a1–6 branched, mono-antennary
N-linked glycans is consistent with the assumption
that nonmodified Mana1–3Manb is the basic recogni-
tion unit and a1–6Man has essentially no interference
with the recognition.
There are various lines of evidence indicating that
some mJRLs are specifically induced by treatment with
jasmonate [23], by salt stress [12] or by microbial (e.g.
Xanthomonas campestris) infection [13]. Thus, one
possible role for them is protection against environ-
mental stress and foreign enemies. In this study, we
demonstrated that CCA and CRLL bound to high-
mannose-type glycans, which are widely found in
fungi, insects and animals, while at the same time each
of them showed respective characteristic sugar-binding
properties. Therefore, it is possible that they are
involved in nonself recognition targeting of pathogens
or herbivorous animals in respective contexts.
Although the relationship between plant lectins and
host glycans remains to be elucidated, the wealth of
information obtained in the present study may give us
clues to understand better the physiological functions
of lectins in the plant body.
From an evolutionary viewpoint, it is of interest to
argue the issue of origin of sugar-binding specificities
of JRLs, in the particular context of biological signi-
ficance of tandem-repeat type structures. At the

moment, however, there have been only few data of
fine oligosaccharide specificities of JRLs for systematic
comparison. To this approach, more comprehensive
analysis of other JRLs is in progress.
Experimental procedures
Materials
NHS-activated Sepharose was purchased from Amersham
Pharmacia Biotech (Little Chalfont, Bucks, UK). All
chemical reagents used in this study were of analytical
grade.
Oligosaccharides
p-Nitrophenyl glycosides of Gal-a, Gal-b, GalNAc-a,
Galb1–3GalNAc-a, Man-b, and Galb1–4GlcNAc-b, were
from Sigma (St Louis, MO, USA); and Glc-a was from
Calbiochem (San Diego, CA, USA). Other pNP-glycosides
(GalNAc-b-, Galb1–4Glc-b-, Man-a-, Fuc-a-, Glca1–4Glc-
a-, and (Glca1–4)
5
-a -pNP) were obtained from Funakoshi
Co. (Tokyo, Japan).
Pyridylaminated oligosaccharides used in this study were
listed in Fig. 2. PA-N-linked glycans numbered 001–014,
103, 105, 107, 108, 307, 313, 314, 323, 405, 410, 418, 419,
420, and 503 were from Takara Bio Inc. (Kyoto, Japan);
and the others were from Seikagaku Co. (Tokyo, Japan).
Glycolipid-type glycans numbered 701–703, 705–713, 715–
721, 724, 726, and 728–731 were obtained from Takara
Bio. The sources of nonlabelled glycans were as follow:
Oligosaccharide specificity of Jacalin-related lectins S. Nakamura et al.
2794 FEBS Journal 272 (2005) 2784–2799 ª 2005 FEBS

727 from Funakoshi Co., 733, 734, and 908 from Dextra
Laboratories, Ltd. (Reading, UK); 725, 909, and 910 from
Calbiochem; 906 and 907 from Seikagaku Co. Oligo-lacto-
samines 901, 902, 903, and 905 and milk oligosaccharides
722, 723, 732 , and 735–739 were generous gifts from K i.
Yoshida (Seikagaku Co.) and from T. Urashima (Obihiro
University of Agriculture and Veterinary Medicine, Obi-
hiro, Japan), respectively. The nonlabelled glycans were
pyridylaminated with GlycoTag (Takara Bio) before use.
Synthesis of MTX-derivatized Man
3
GlcNAc
2
Methotrexate-derivatized Man
3
GlcNAc
2
was synthesized
as previously described [30]. Man
3
GlcNAc
2
(21.8 mg,
0.0239 mmol) was dissolved in saturated aqueous
NH
4
HCO
3
(2 mL) and stirred at 40 °C for 14 h, and then
the mixture was concentrated and coevaporated with H

2
O
in vacuo. The residue was dissolved in dioxane/H
2
O(1:1;
2 mL), after which NaHCO
3
(4.8 mg, 0.057 mmol) and
FmocGlyCl (9.1 mg, 0.029 mmol) were added at 0 °C.
After the mixture had been stirred at room temperature for
1 h, NaHCO
3
(2.4 mg, 0.029 mmol) and FmocGlyCl
(4.5 mg, 0.014 mmol) were added to it at 0 °C. After hav-
ing been stirred at room temperature for 1 h, the mixture
was purified by gel filtration (Sephadex LH20,
H
2
O ⁄ MeOH, 1 : 3) to give FmocGly-possessing saccharide.
The FmocGly-possessing saccharide was dissolved in di-
methylformamide (1 mL), and piperidine (0.2 mL) was
added at 0 °C. After having been stirred at room tempera-
ture for 2 h, the mixture was concentrated in vacuo. The
residue was purified by gel filtration (Sephadex G15, H
2
0)
to give 21.5 mg (93%) of Man
3
GlcNAc
2

NHGly.
To a cold ()18 °C) solution of Man
3
GlcNAc
2
NHGly
(10.7 mg, 0.0111 mmol) and MTX (a
t
Bu; 8.5 mg,
0.017 mmol) in dimethylformamide (1 mL) was added
2-(1H-benzotriazole-1-yl)-1,1,3,3-tetramethyluronium tetra-
fluoroborate (TBTU) (6.2 mg 0.019 mmol). The mixture
was stirred at 0 °C for 19 h, at room temperature for 4 h,
and then concentrated in vacuo. The residue was purified by
preparative thin-layer chromatography (PTLC) (H
2
O ⁄
i
PrOH, 1 : 3) to give the coupled product, which was then
dissolved in CF
3
COOH at 0 °C. After having been stirred
at room temperature for 2 h, the mixture was concentrated
and coevaporated with toluene in vacuo. The residue was
purified by PTLC (H
2
O ⁄ MeOH, 1 : 5) to give 11.2 mg
(72%) of Man
3
GlcNAc

2
-MTX: TLC, R
f
0.63 (H
2
O ⁄
i
PrOH,
1 : 2);
1
H NMR (400 MHz, D
2
O) d 1.90 (s, 3 H), 2.06 (s,
3 H), 2.03–2.50 (m, 4 H), 3.13 (m, 3 H), 3.50–3.96 (m,
34 H), 4.06 (s, 1 H), 4.25 (s, 1 H), 4.36 (m, 1 H), 4.59 (d,
J ¼ 6.8 Hz, 1 H), 4.60–4.85 (m, 1 H), 4.90 (s, 1 H), 5.09 (s,
1 H), 6.79 (m, 2 H), 7.64 (m, 2 H), 8.52 (m, 1 H);
13
C
NMR (125 MHz, D
2
O) d 22.55, 22.98, 28.38, 32.88, 39.42,
43.32, 44.06, 51.32, 54.15, 54.95, 55.54, 55.66, 60.50, 60.64,
61.62, 61.81, 63.18, 66.49, 67.46, 67.55, 70.56, 70.68, 70.81,
71.01, 71.08, 72.61, 73.13, 73.34, 74.10, 74.86, 75.06, 76.83,
79.20, 79.46, 80.27, 80.74, 81.16, 100.27, 100.01, 101.93,
103.16, 113.15 · 2, 129.40, 129.52 · 2, 130.04, 130.10,
149.89, 150.02, 152.61, 163.86, 169.83, 173.13, 175.24 · 2,
175.31, 176.89 178.97 L; MS (MALDI-TOF) calcd for
C

56
H
82
N
2
O
30
Na (M + Na)
+
m ⁄ z 1425.5, found 1425.5.
A
B
Fig. 6. Scheme of the FAC procedure. (A) Analyte solution, the ini-
tial concentration of which is represented by [A]
0
, is continuously
applied to a lectin-immobilized column. The elution front of an ana-
lyte that does not interact with the immobilized lectin (Analyte I) is
observed immediately, whereas that of an analyte that specifically
interacts with the immobilized lectin (Analyte II) is retarded depend-
ing on the strength of the interaction. (B) Curve I (solid line) indi-
cates the elution profile of Analyte I, showing no interaction with
the immobilized lectin; and Curve II (dotted line), that of Analyte II,
showing significant affinity for the immobilized lectin. The elution
volumes of Analyte I and Analyte II are indicated as V
0
and V,
respectively. The basic equation of FAC is shown as Eqn (1). For
details about the principle of FAC, see Experimental procedures.
S. Nakamura et al. Oligosaccharide specificity of Jacalin-related lectins

FEBS Journal 272 (2005) 2784–2799 ª 2005 FEBS 2795
Purification of CCA and CRLL
CCA was purified from Castanea crenata cotyledons as pre-
viously described [20]. Briefly, 100 g of cotyledons were
homogenized in 50 mm phosphate buffer pH 7.0, contain-
ing 0.9% (w ⁄ v) NaCl (NaCl ⁄ P
i
), 10 mm 2-mercaptoethanol,
5mm EDTA, and 5% (w ⁄ w) polyvinylpolypyrrolidone.
After the extract had been stirred at 4 °C for 1 h, it was
applied onto an asialo-fetuin Sepharose 4B column
(10 mgÆmL
)1
) equilibrated with NaCl ⁄ P
i
. After the column
had been washed with the same buffer, the lectin fraction
was eluted with NaCl ⁄ P
i
containing 0.2 md-mannose. The
active fraction was further applied to a Mono Q column
equilibrated with 50 mm Tris ⁄ HCl pH 8.0. Elution was car-
ried out with a linear gradient of NaCl from 0 to 0.5 m.
The lectin fraction was concentrated by salting with ammo-
nium sulfate, and then applied on a Superose 12 column.
The mobile phase was NaCl ⁄ P
i
and the flow rate was kept
at 0.6 mLÆmin
)1

.
Cycas revoluta leaf lectin (CRLL) was purified from
leaves of Japanese cycad as previously described [19].
Briefly, leaf powder was extracted with 10 vols 6.67 mm
sodium phosphate buffer pH 7.2, containing 0.8% (w ⁄ v)
NaCl, and the homogenate was filtered through two layers
of gauze and centrifuged. After salting out by ammonium
sulfate, the precipitate was dissolved in the same buffer as
used for the extraction. The supernatant was loaded onto
a column of mannose-agarose (2 lmolÆ mL
)1
) previously
equilibrated with 50 mm Tris ⁄ HClpH 7.5. The column was
washed with the same buffer, and the bound fraction was
eluted with 0.2 md-mannose.
Preparation of lectin columns
Thus purified CCA and CRLL were dissolved in 10 mm
phosphate buffer pH 7.4, containing 0.7% NaCl and cou-
pled to NHS-activated Sepharose according to the manu-
facturer’s instructions. After deactivation of excess NHS
groups by 1 m monoethanolamine followed by extensive
washing, the lectin-Sepharose was suspended in 10 mm
Tris ⁄ HCl pH 7.4, containing 1 mm CaCl
2
and 1 mm
MnCl
2
. The slurry was packed into a capsule-type minia-
ture column (inner diameter, 2 mm; length, 10 mm; bed
volume, 31.4 lL). The amount of immobilized protein was

determined by measuring the amount of uncoupled protein
in the above wash fraction by the method of Bradford [33].
The resulting lectin-columns were slotted into a stainless
holder, and connected to the FAC-1 machine, which had
been specially designed and manufactured by Shimadzu Co
[24].
Fig. 7. Scheme of the developed system for
automated FAC (FAC-1). (A) The FAC-1
machine is equipped with two pumps, an
autosampler, a column oven, and two minia-
ture columns connected in parallel. One of
the pumps is used exclusively for analysis,
and the other is for regeneration of the
columns. Analyte solutions are applied via
the autosampler by pump A, and elution of
analytes from the lectin-immobilized column
is monitored by UV or fluorescence detec-
tors. In the figure, the bold line indicates the
pathway when analysis is done with column
1. (B) Outline of the dual column switching
system. The 2 columns are used in ‘one
after the other’ fashion to reduce by half the
total analysis time.
Oligosaccharide specificity of Jacalin-related lectins S. Nakamura et al.
2796 FEBS Journal 272 (2005) 2784–2799 ª 2005 FEBS
Principle of FAC
The principle of FAC was previously described by Hira-
bayashi et al. [26]. Briefly, an excess volume (0.6–1.0 mL)
of dilute PA-oligosaccharide solution is continuously
applied to a lectin-immobilized column (Fig. 6A). When the

amount of the applied analyte molecules exceeds the ability
of the column, leakage occurs and the concentration of
oligosaccharides in the elution finally reaches a plateau,
where the concentration is equal to that of the initial solu-
tion (Fig. 6B). In the case of oligosaccharides having some
affinity for the immobilized lectin (Fig. 6A, right), they
interact with and are retained by the immobilized lectin,
and the elution is observed as in curve II (Fig. 6B). The
volume of the elution front (V) of each oligosaccharide is
calculated essentially as described previously [26]. Retarda-
tion of the elution front relative to that of an appropriate
standard oligosaccharide (Fig. 6A, left, and Fig. 6B, curve
I), i.e. V–V
0
, is then determined. K
d
values for dissociation
of lectin and oligosaccharide are obtained from V–V
0
and
B
t
, according to the basic equation of FAC, Eqn (1), where
B
t
is the effective ligand content (expressed in mol), and
[A]
0
is the initial concentration of PA-oligosaccharide.
Equation (1) can be simplified to Eqn (2), where [A]

0
(e.g. < 10
)8
m) is negligibly small compared with K
d
(e.g. > 10
)6
m).
K
d
¼ B
t
=ðV À V
0
ÞÀ½A
0
ð1Þ
K
d
¼ B
t
=ðV À V
0
Þ; if K
d
)½A
0
ð2Þ
It is often favourable to discuss the matter of lectin–oligo-
saccharide interactions in terms of affinity constant (K

a
)
instead of K
d
. The two equilibrium constants are in the fol-
lowing relationship:
K
d
¼ 1=K
a
ð3Þ
FAC procedure
Frontal affinity chromatography was performed by using a
recently developed machine for automated FAC (FAC-1)
[24]. A scheme for the total procedure is illustrated in
Fig. 7. Briefly, the system consists of an FAC-1 machine
(Shimadzu) equipped with two isocratic pumps, an auto-
sampler, and a couple of miniature columns (2.0 · 10 mm,
31.4 lL) connected in parallel to either a fluorescence detec-
tor (Shimadzu, RF10AXL) or UV detector (Shimadzu,
SPD-10A VP) and a PC workstation loaded with ‘LCsolu-
tion’ software (Fig. 7A). By use of the parallel column sys-
tem, the time for analysis can be cut almost in half; i.e.
during analysis with one column, the other column is being
washed (Fig. 7B).
Lectin columns were equilibrated with 10 mm Tris ⁄ HCl
pH 7.4, containing 0.8% NaCl (TBS). The flow rate and
the column temperature were kept at 0.125 mLÆmin
)1
and

25 °C, respectively. After equilibrium of the miniature col-
umns had been obtained, either PA- (2.5 nm) or pNP-oligo-
saccharides (5.0 lm) dissolved in TBS were successively
injected into a pair of lectin-columns by the auto-sampling
system. Elution of PA-oligosaccharides was monitored by
fluorescence (excitation and emission wavelengths of 310
and 380 nm, respectively), whereas that of pNP-glycosides
was detected by UV (280 nm).
Evaluation of lectin columns
For the determination of effective ligand content, B
t
, concen-
tration-dependence analysis and subsequent Woolf–
Hofstee-type plot were performed as described previously
[26]. Here, various concentrations ([A]
0
) of MTX-derivatized
Man
3
GlcNAc
2
dissolved in TBS were applied to the minia-
ture column, and the elution was monitored by absorbance
at 304 nm. Woolf–Hofstee-type plots, i.e. (V-V
0
) vs.
(V-V
0
)[A]
0

, were made to determine B
t
and K
d
values from
the intercept and the slope, respectively, of the fitted curve.
Acknowledgements
This work was supported by NEDO (New Energy and
Industrial Technology Organization) under the METI
(The Ministry of Economy, Trade, and Industry,
Japan).
References
1 Iglesias JL, Halina L & Sharon N (1982) Purification
and properties of a d-galactose ⁄ N-acetyl-d-galactosa-
mine-specific lectin from Erythrina cristagalli. Eur J Bio-
chem 123, 247–252.
2 Shibuya N, Goldstein IJ, Van Damme EJM & Peumans
WJ (1988) Binding properties of a mannose-specific lec-
tin from the snowdrop (Galanthus nivalis) bulb. J Biol
Chem 263, 728–734.
3 Van Damme EJM, Peumans WJ, Barre A & Rouge
´
P
(1998) Plant lectins: a composite of several distinct
families of structurally and evolutionary related proteins
with diverse biological roles. Crit Rev Plant Sci 17, 575–
692.
4 Young NM, Johnston RA & Watson DC (1991) The
amino acid sequences of jacalin and the Maclura pomi-
fera agglutinin. FEBS Lett 282, 382–384.

5 Bourne Y, Zamboni V, Barre A, Peumans WJ, Van
Damme EJM & Rouge P (1999) Helianthus tuberosus
lectin reveals a widespread scaffold for mannose-binding
lectins. Structure 7, 1473–1482.
6 Loomes LM, Stewart WW, Mazengera RL, Senior
BW & Kerr MA (1991) Purification and characteriza-
tion of human immunoglobulin IgA1 and IgA2
isotypes from serum. J Immunol Methods 141, 209–
218.
S. Nakamura et al. Oligosaccharide specificity of Jacalin-related lectins
FEBS Journal 272 (2005) 2784–2799 ª 2005 FEBS 2797
7 Van Damme EJM, Hause B, Hu J, Barre A, Rouge P,
Proost P & Peumans WJ (2002) Two distinct jacalin-
related lectins with a different specificity and subcellular
location are major vegetative storage proteins in the bark
of the black mulberry tree. Plant Physiol 130, 757–769.
8 Peumans WJ, Winter HC, Bemer V, Van Leuven F,
Goldstein IJ, Truffa-Bachi P & Van Damme EJM
(1997) Isolation of a novel plant lectin with an unusual
specificity from Calystegia sepium. Glycoconj J 14, 259–
265.
9 Misquith S, Rani PG & Surolia A (1994) Carbohydrate
binding specificity of the B-cell maturation mitogen
from Artocarpus integrifolia seeds. J Biol Chem 269,
30393–30401.
10 Van Damme EJM, Barre A, Mazard AM, Verhaert P,
Horman A, Debray H, Rouge P & Peumans WJ (1999)
Characterization and molecular cloning of the lectin
from Helianthus tuberosus. Eur J Biochem 259, 135–142.
11 Koshte VL, van Dijk W, van der Stelt ME & Aalberse

RC (1990) Isolation and characterization of BanLec-I, a
mannoside-binding lectin from Musa paradisiac
(banana). Biochem J 272, 721–726.
12 Zhang W, Peumans WJ, Barre A, Astoul CH, Rovira P,
Rouge P, Proost P, Truffa-Bachi P, Jalali AA & Van
Damme EJM (2000) Isolation and characterization of a
jacalin-related mannose-binding lectin from salt-stressed
rice (Oryza sativa) plants. Planta 210, 970–978.
13 Hirano K, Teraoka T, Yamanaka H, Harashima A,
Kunisaki A, Takahashi H & Hosokawa D (2000) Novel
mannose-binding rice lectin composed of some isolectins
and its relation to a stress-inducible salT gene. Plant
Cell Physiol 41, 258–267.
14 Tateno H, Winter HC, Petryniak J & Goldstein IJ
(2003) Purification, characterization, molecular cloning,
and expression of novel members of jacalin-related lec-
tins from rhizomes of the true fern Phlebodium aureum
(L) J. Smith (Polypodiaceae). J Biol Chem 278, 10891–
10899.
15 Sankaranarayanan R, Sekar K, Banerjee R, Sharma V,
Surolia A & Vijayan M (1996) A novel mode of
carbohydrate recognition in jacalin, a Moraceae plant
lectin with a beta-prism fold. Nat Struct Biol 3, 596–
603.
16 Bourne Y, Roig-Zamboni V, Barre A, Peumans WJ,
Astoul CH, Van Damme EJM & Rouge P (2004) The
crystal structure of the Calystegia sepium agglutinin
reveals a novel quaternary arrangement of lectin sub-
units with a beta-prism fold. J Biol Chem 279, 527–
533.

17 Pratap JV, Jeyaprakash AA, Rani PG, Sekar K, Surolia
A & Vijayan M (2002) Crystal structures of artocarpin,
a Moraceae lectin with mannose specificity, and its com-
plex with methyl-alpha-D-mannose: implications to the
generation of carbohydrate specificity. J Mol Biol 317,
237–247.
18 Peumans WJ, Van Damme EJM, Barre A & Rouge P
(2001) Classification of plant lectins in families of struc-
turally and evolutionary related proteins. Adv Exp Med
Biol 491, 27–54.
19 Yagi F, Iwaya T, Haraguchi T & Goldstein IJ (2002)
The lectin from leaves of Japanese cycad, Cycas revoluta
Thunb. (gymnosperm), is a member of the jacalin-
related family. Eur J Biochem 269, 4335–4341.
20 Nomura K, Ashida H, Uemura N, Kushibe S, Ozaki T
& Yoshida M (1998) Purification and characterization
of a mannose ⁄ glucose-specific lectin from Castanea
crenata. Phytochemistry 49, 667–673.
21 Nomura K, Nakamura S, Fujitake M & Nakanishi T
(2000) Complete amino acid sequence of Japanese chest-
nut agglutinin. Biochem Biophys Res Commun 276,
23–28.
22 Nakamura S, Ikegami A, Matsumura Y, Nakanishi T &
Nomura K (2002) Molecular cloning and expression of
the mannose ⁄ glucose specific lectin from Castanea cre-
nata cotyledons. J Biochem (Tokyo) 131 (2), 241–246.
23 Peumans WJ, Zhang W, Barre A, Houles Astoul C,
Balint-Kurti PJ, Rovira P, Rouge P, May GD, Van
Leuven F, Truffa-Bachi P & Van Damme EJM (2000)
Fruit-specific lectins from banana and plantain. Planta

211, 546–554.
24 Hirabayashi J (2004) Lectin-based structural glycomics:
glycoproteomics and glycan profiling. Glycoconj J 21,
35–40.
25 Kasai K, Oda Y, Nishikawa M & Ishii S (1986) Frontal
affinity chromatography: theory, for its application to
studies on specific interaction of biomolecules. J Chro-
matogr A 379, 33–47.
26 Hirabayashi J, Arata Y & Kasai K (2003) Frontal affi-
nity chromatography as a tool for elucidation of sugar
recognition properties of lectins. Methods Enzymol 362,
352–368.
27 Hirabayashi J, Hashidate T, Arata Y, Nishi N, Nakam-
ura T, Hirashima M, Urashima T, Oka T, Futai M,
Muller WE, Yagi F & Kasai K (2002) Oligosaccharide
specificity of galectins: a search by frontal affinity chro-
matography. Biochem Biophys Acta 1572 , 232–254.
28 Zhang B, Palcic MM, Mo H, Goldstein IJ & Hindsgaul
O (2001) Rapid determination of the binding affinity
and specificity of the mushroom Polyporus squamosus
lectin using frontal affinity chromatography coupled to
electrospray mass spectrometry. Glycobiology 11, 141–
147.
29 Rempel BP, Winter HC, Goldstein IJ & Hindsgaul O
(2002) Characterization of the recognition of blood
group B trisaccharide derivatives by the lectin from
Marasmius oreades using frontal affinity chromatogra-
phy-mass spectrometry. Glycoconj J 19, 175–180.
30 Totani K, Matsuo I & Ito Y (2004) Tight binding
ligand approach to oligosaccharide-grafted protein.

Bioorg Med Chem Lett 14, 2285–2289.
2798 FEBS Journal 272 (2005) 2784–2799 ª 2005 FEBS
Oligosaccharide specificity of Jacalin-related lectins S. Nakamura et al.
31 Green ED, Adelt G, Baenziger JU, Wilson S &
Van Halbeek H (1988) The asparagine-linked oligosac-
charides on bovine fetuin. Structural analysis of
N-glycanase-released oligosaccharides by 500-mega-
hertz 1H NMR spectroscopy. J Biol Chem 263,
18253–18268.
32 Jeyaprakash AA, Srivastav A, Surolia A & Vijayan M
(2004) Structural basis for the carbohydrate specificities
of artocarpin: variation in the length of a loop as a
strategy for generating ligand specificity. J Mol Biol
338, 757–770.
33 Bradford MM (1976) A rapid and sensitive method for
the quantitation of microgram quantities of protein util-
izing the principle of protein-dye binding. Anal Biochem
72, 248–254.
FEBS Journal 272 (2005) 2784–2799 ª 2005 FEBS 2799
S. Nakamura et al. Oligosaccharide specificity of Jacalin-related lectins