Báo cáo khoa học: Activity of the plant peptide aglycin in mammalian systems potx

Activity of the plant peptide aglycin in mammalian
systems
Xin-Peng Dun
1
, Jian-He Wang
1
, Lei Chen
1
, Jie Lu
1
, Fa-Fang Li
1
, Yan-Ying Zhao
1
, Ella Cederlund
2
,
Galina Bryzgalova
3
, Suad Efendic
3
, Hans Jo
¨
rnvall
2
, Zheng-Wang Chen
1,2
and Tomas Bergman
2
1 School of Life Science and Technology, Huazhong University of Science and Technology, Wuhan, China
2 Department of Medical Biochemistry and Biophysics, Karolinska Institutet, Stockholm, Sweden

3 Department of Molecular Medicine and Surgery, Karolinska University Hospital, Stockholm, Sweden
Polypeptide hormones have long been recognized as
important regulatory molecules in animals and the
human. Since the discovery of secretin in 1902 [1], and
insulin in 1921 [2,3], polypeptides have been assigned
signaling functions in the regulation of physiological
processes, and several peptides have been used as
drugs in specific diseases. Discovery of polypeptide sig-
nals in plant defense, growth, and development shows
the presence of peptide signaling also in plants [4,5]. It
has been reported that plant peptides may be found
in animals through alimentary absorption or through
coexistence as homologous counterparts in animals,
sharing common structures [6,7].
In the present study, we have isolated a bioactive pep-
tide, aglycin, from porcine intestine and found it to be
identical to a segment of the hormone-like plant poly-
peptide albumin 1 B precursor (PA1B, chain b) from
pea seeds (Pisum sativum) [8]. PA1B, chain b, is involved
in plant signal transduction to regulate growth and dif-
ferentiation and is increasingly expressed during seed
development (SwissProt entry P62927). In total, six iso-
forms of this polypeptide have been described, revealing
sequence homology (PA1A–F, SwissProt entries
P62926–62931, respectively). In relation to the PA1B
chain b sequence, the other five isoforms are 75–94%
identical (BLAST search at ).
Keywords
aglycin; albumin 1 B precursor; blood
glucose; mice; voltage-dependent anion-

selective channel protein 1
Correspondence
T. Bergman, Medical Biochemistry and
Biophysics, Karolinska Institutet, SE-171 77
Stockholm, Sweden
Fax: +46 8 337 462
Tel: +46 8 524 87780
E-mail:
Z W. Chen, School of Life Science and
Technology, Huazhong University of Science
and Technology, Wuhan 430074, China
Fax ⁄ Tel: +86 27 8779 2027
E-mail:
(Received 3 October 2006, revised 23
November 2006, accepted 29 November
2006)
doi:10.1111/j.1742-4658.2006.05619.x
A 37 residue peptide, aglycin, has been purified from porcine intestine. The
sequence is identical to that of residues 27–63 of plant albumin 1 B precur-
sor (PA1B, chain b) from pea seeds. Aglycin resists in vitro proteolysis by
pepsin, trypsin and Glu-C protease, compatible with its intestinal occur-
rence and an exogenous origin from plant food. When subcutaneously
injected into mice (at 10 lgÆg
)1
body weight), aglycin has a hyperglycemic
effect resulting in a doubling of the blood glucose level within 60 min.
Using surface plasmon resonance biosensor technology, an aglycin binding
protein with an apparent molecular mass of 34 kDa was detected in mem-
brane protein extracts from porcine and mice pancreas. The polypeptide
was purified by affinity chromatography and identified through peptide

mass fingerprinting as the voltage-dependent anion-selective channel pro-
tein 1. The results indicate that aglycin has the potential to interfere with
mammalian physiology.
Abbreviations
CTIP, concentrate of thermostable intestinal polypeptides; HRP, horseradish peroxidase; PA1B, albumin 1 B precursor; VDAC-1, voltage-
dependent anion-selective channel protein 1.
FEBS Journal 274 (2007) 751–759 ª 2007 The Authors Journal compilation ª 2007 FEBS 751
We now show that aglycin interferes with mamma-
lian physiology as revealed by an increase of blood
glucose concentration in mice upon subcutaneous
injection. Furthermore, a protein purified by binding
to aglycin in porcine and mice pancreas membrane
protein extracts is identified as the voltage-dependent
anion-selective channel protein 1 (VDAC-1) [9]. Hence
aglycin apparently has several effects in mammalian
systems.
Results
Aglycin was purified as outlined in Fig. 1, monitoring
glucose-induced insulin release from pancreatic b-cells
[10]. Fraction 4 from Sephadex G-25 fine chromato-
graphy was active. Ion-exchange chromatography of
this fraction revealed an active component eluting at
0.1 m ammonium bicarbonate. This fraction was
lyophilized and further purified by reversed phase
high performance liquid chromatography (RP-HPLC)
(Fig. 1). An average yield of 6 lg aglycin was obtained
with 2.0 mg lyophilized peptides from 0.1 m ammo-
nium bicarbonate fraction, corresponding to about
3 lg pure peptide from 1 kg tissue (wet). The molecu-
lar mass determined by matrix-assisted laser desorp-

tion ⁄ ionization time-of-flight (MALDI-TOF) mass
spectrometry was 3742.3 Da. Edman degradation
revealed the amino acid sequence ASCNGVCSPFEM
PPCGSSACRCIPVGLVVGYCRHPSG (37 residues).
Database searches with this sequence revealed that it is
identical to residues 27–63 of the polypeptide PA1B
from pea seeds [8].
After exposure of aglycin to common proteases,
samples were analyzed by RP-HPLC. The results show
that pepsin, trypsin and Glu-C protease do not affect
aglycin to any appreciable extent because essentially
no proteolytic products were observed after 12 h of
incubation (Fig. 2). Mass measurements revealed only
the intact material, both before and after the proteo-
lytic treatment (data not shown).
Aglycin subcutaneously injected into mice at a dose
of 10 lgÆg
)1
(n ¼ 20) was found to enhance the blood
glucose concentration with a peak value at 60 min cor-
responding to a doubling of the glucose concentration
in relation to the saline group (Fig. 3). When the pea
albumin isoform PA1F, chain b (purified from pea
seeds, 91% identical sequence to that of aglycin) was
tested for the effect on blood glucose, a similar hyper-
glycemic pattern was observed (data not shown). Ana-
lysis by electrospray mass spectrometry of the PA1F
isoform directly before testing its influence on blood
glucose concentration revealed a mass value within
Fig. 1. Purification of aglycin from pig intestine. Details of the purifi-

cation scheme and chromatographic steps are given in the text.
Fig. 2. Aglycin stability against proteolytic cleavage. After treatment
with pepsin, trypsin or Glu-C protease, the reaction mixtures were
analyzed by RP-HPLC (chromatogram after pepsin treatment
shown). No significant hydrolysis products were observed for any
of these enzymes after 12 h of incubation.
Aglycin activity in mammals X P. Dun et al.
752 FEBS Journal 274 (2007) 751–759 ª 2007 The Authors Journal compilation ª 2007 FEBS
0.02 Da of that for the fully oxidized peptide, showing
disulfide bridges to be intact in the biologically active
peptide.
Aglycin immobilized onto an surface plasmon reson-
ance biosensor chip (coupling efficiency checked by
atomic force microscopy) was tested for interaction
with protein components in the microsomal membrane
fraction of homogenized porcine tissues. Significant
change of the refractive index (increase by 7 · 10
)4
)
was detected with the pancreatic extract only, showing
that one or several binding proteins are abundantly
present in this tissue preparation (Fig. 4).
Using affinity chromatography with immobilized
aglycin, a binding protein was recovered from porcine
pancreas membrane extract. ELISA at various condi-
tions confirmed the interaction between aglycin and
the purified protein (Fig. 5). The binding protein was
also purified from mice pancreas extract using the
same protocol and tested in ELISA with the same
monoclonal antibody (data not shown). SDS ⁄ PAGE

of the fraction from affinity chromatography revealed
a single band at 34 kDa for both the porcine and mice
preparations (Fig. 6). The aglycin binding protein
was identified by peptide mass fingerprinting using
MALDI-TOF mass spectrometry after tryptic in-gel
Fig. 3. The effect of aglycin on blood glucose concentration in nor-
mal mice. Means ± SEM are shown. The probability of random dif-
ference between saline and aglycin groups is < 0.001 (n ¼ 20).
Fig. 4. Surface plasmon resonance measurements reveal the exist-
ence of a binding protein in membrane extracts tested. P, pan-
creas; L, liver; K, kidney; M, muscle. Upward arrows indicate
beginning of injection of the extract, downward arrows indicate
beginning of washing. An aglycin interacting protein is present in
the pancreatic extract.
Fig. 5. ELISA of the purified protein from porcine pancreas mem-
brane extract reveals that it binds to aglycin. (A) Wells coated with
the purified protein followed by addition of aglycin, and then mono-
clonal antibody against aglycin. (B) Wells coated with the purified
protein followed by monoclonal antibody but without previous addi-
tion of aglycin. (C) After blocking with gelatin, aglycin was added,
followed by monoclonal antibody. (D) Wells coated with aglycin fol-
lowed by the monoclonal antibody.
Fig. 6. SDS ⁄ PAGE of the pancreatic aglycin binding protein purified
by affinity chromatography. Lane A, protein molecular mass mark-
ers; lane B, the aglycin binding protein isolated from mice; lane C,
the aglycin binding protein isolated from pig. Staining was with
Coomassie blue and numbers indicate molecular masses in kDa.
X P. Dun et al. Aglycin activity in mammals
FEBS Journal 274 (2007) 751–759 ª 2007 The Authors Journal compilation ª 2007 FEBS 753
digestion of the material from mice pancreas as the

282 residue voltage-dependent anion-selective channel
protein 1 (VDAC-1, Sus scrofa SwissProt Q9MZ16,
theoretical mass 30.6 kDa [9,11]) at a sequence cover-
age of 60%.
Discussion
We have isolated the thermostable plant peptide agly-
cin from pig intestine and investigated its possible
interactions and activities in mammalian systems. The
results reveal that it has a clear physiological effect in
raising the blood glucose concentration in mice about
two-fold upon subcutaneous injection. Furthermore,
specific binding to the ion channel protein VDAC-1
was detected in the membrane protein extracts from
porcine and mice pancreas.
Aglycin is a single-chain, 37 residue polypeptide,
containing six half-cystine residues at positions 3, 7,
15, 20, 22 and 32, an N-terminal alanine and a C-ter-
minal glycine. A search for the aglycin amino acid
sequence in the SwissProt database revealed that it is
identical to residues 27–63 of the plant polypeptide
PA1B from pea seeds, first reported in 1986 [8]. PA1B
was later characterized as the pea counterpart of a
4 kDa hormone-like peptide in soybean [12] associated
with plant cell proliferation and differentiation [13].
Interestingly, insulin and insulin-like growth factor I
and II from mammals are able to compete with the
4 kDa peptide in binding to the receptor-like protein
basic 7S globulin isolated from soybean [12,14,15].
Due to the similarity with animal insulin in binding
the basic 7S globulin and stimulating its protein kinase

activity [16], the 4 kDa soybean peptide was initially
designated leginsulin [12] but this name was later aban-
doned to avoid any confusion with insulin [13].
We have now purified PA1B, chain b, from a porcine
intestinal extract of thermostable polypeptides and
detected novel activities ⁄ interactions. Therefore, we
believe that PA1B, chain b, deserves a descriptive name
and suggest aglycin to emphasize the first and the last
residue in the amino acid sequence. Considering the
sequences of all the six known isoforms of the peptide,
all but one (PA1C, chain b, N-terminal residue Ile),
starts with an Ala residue and ends with a Gly residue
which makes the name even more appropriate.
Structurally aglycin belongs to the cystine-knot pep-
tide family that has been found in several sources
(plants, fungi, animal venoms, insects) [17]. The mem-
bers reveal diverse biological activities and are com-
monly ion channel blockers and toxins as well as
enzyme inhibitors [17]. The cystine-knot structural
motif consists of a ring-like structure formed by two
disulfide bonds and their connection held together by a
third disulfide bond. This motif is invariably associated
with a nearby b-sheet structure and the overall design
appears highly efficient for structure stabilization.
Aglycin has furthermore been described as an entomo-
toxin because of its highly toxic activity against cereal
weevils (Sitophilus spp.) [18]. A high-affinity binding
site in the insect gut has been detected and character-
ized, but the identity of the corresponding target pro-
tein and the mechanisms involved are unknown [18].

There are reports describing homologous counter-
parts to plant peptides in the animal kingdom with
identical or very similar sequences [6,7]. We now find
that aglycin significantly resists hydrolysis by trypsin,
pepsin and Glu-C protease in vitro, and it is conceiv-
able that aglycin isolated in this study is of exogenous
origin from plant food sources. This conclusion is sup-
ported by results from studies on soluble proteins pre-
sent in ileal digests from pigs on pea diets where
albumin PA1B was found totally resistant to gastric
and small intestine digestion [19]. However, it cannot
yet be excluded that aglycin, or a structural homolog
of aglycin, exists in animals and thus represent a
cross-kingdom bioactive peptide family. Interestingly,
similarity searches using the Swiss Institute of Bio-
informatics (SIB) BLAST network service (http://
www.expasy.org/cgi-bin/blast.pl) identified a mouse
protein segment with 60% identity (15 ⁄ 25, 72% posi-
tives) to the aglycin sequence (TrEMBL Q9D7N2),
and a human protein segment with 58% identity
(10 ⁄ 17, 58% positives) to the aglycin sequence (TrEM-
BL Q76B61). In these alignments, the four residue
sequence PCGS (aglycin residues 14–17) was common
to both protein segments. The presence of aglycin-like
sequences in proteins from mouse and man indicates
that the plant peptide aglycin may have structurally
related counterparts in mammals generated by frag-
mentation of larger precursor proteins.
Subcutaneous injection of aglycin at a dose of
10 lgÆg

)1
body weight increases the blood glucose con-
centration in normal mice (Kunming type) about two-
fold. Another aglycin isoform (PA1F, chain b) was also
tested employing C57BL ⁄ 6 mice ( n ¼ 6) with similar
results. In other words, aglycin is bioactive in a mam-
malian system represented by the mice with statistically
significant effects on the blood glucose level. However,
it should be emphasized that even though aglycin, like
insulin, influences the concentration of glucose in
blood, the effect is opposite that of insulin, increasing
rather than decreasing the blood glucose concentration.
VDAC-1 was purified and identified as a specific
aglycin binding protein. A high-affinity protein binding
site for aglycin in the gut of cereal weevils has been
Aglycin activity in mammals X P. Dun et al.
754 FEBS Journal 274 (2007) 751–759 ª 2007 The Authors Journal compilation ª 2007 FEBS
described but without identification of the correspond-
ing protein target [18]. It is therefore likely that the tar-
get for aglycin binding is an ion channel protein of the
VDAC type. Gressent et al. [18] points out that the
binding activity was found in the microsomal fraction,
as we also did in this study, and of the two well-known
activities for members of the cystine-knot peptide
family, ion channel toxicity and enzyme inhibition [20],
the latter has so far not been demonstrated for aglycin
[18]. Furthermore, ion channel blockers have with few
exceptions been described only for venoms originating
from the animal kingdom which makes the present
finding that the plant peptide aglycin binds to VDAC-1

in porcine pancreas even more interesting, in particular
because all cystine-knot plant peptides for which the
target is known are enzyme inhibitors [17].
Aglycin increases blood glucose concentration in
mice (see above). It is tempting to suggest that the
mechanism of enhancement involves binding to
VDAC-1. The VDAC-1 protein is involved in energy
metabolism of cells, and is mainly distributed to the
outer mitochondrial membrane where it controls energy
homeostasis by transport of ATP and ADP [21]. How-
ever, VDAC-1 was now purified from pancreatic cell
membranes, not from mitochondrial membranes, which
implicates a novel function of VDAC-1 (i.e., interaction
with aglycin) facilitated by its distribution to the cell
membranes of pancreatic b-, a- and pp-cells that pro-
duce and secret insulin, glucagon and pancreatic pep-
tide, respectively. It has been reported that VDAC-1
has been found also in other types of secretory cell
membranes such as those of B-lymphocytes and mem-
branes associated with cell secretion [22,23]. Further-
more, VDAC-1 was recently identified as a NADH–
ferricyanide reductase in the plasma membrane [24].
Members of the VDAC protein family are found in
both animals and plants. VDAC-1 (SwissProt entry
Q9MZ16), now identified as a binding partner to agly-
cin, forms channels through both the outer mitochond-
rial membrane and the plasma membrane. This allows
diffusion of small hydrophilic molecules. Despite its
name, VDAC-1 is permeable to both anions and cati-
ons depending on the actual membrane potential. To

speculate, the mechanism behind the increase of blood
glucose concentration could potentially involve the flux
of calcium ions through the plasma membrane of pan-
creatic b-cells. Because aglycin, upon binding to
VDAC-1, probably blocks the channel function and
consequently slows down or stops the transport of pos-
itive ions such as calcium into the b-cell, this will lead
to low levels of cellular calcium that potentially can
affect insulin secretion resulting in lower than normal
insulin exocytosis and elevated blood glucose levels.
In conclusion, aglycin is a plant albumin fragment
which binds to VDAC-1 in membrane protein extracts
from porcine and mice pancreas. It can also increase
blood glucose concentration when injected subcutane-
ously into mice. Therefore, aglycin represents a plant
peptide with physiological effects in mammalian sys-
tems.
Experimental procedures
Purification and identification of aglycin
The starting material for peptide purification was a concen-
trate of thermostable intestinal polypeptides (CTIP) from
porcine gut [10,25]. Bioactivity during the purification pro-
cess was monitored as the effect on glucose-induced insulin
release from isolated pancreatic b-cells (rat) [10]. A 30 g
quantity of CTIP was dissolved in 0.24 L water [containing
0.5% (v ⁄ v) thiodiglycol], and 1.08 L isopropanol was added
to the clear solution. After vigorous stirring, a precipi-
tate (the first fraction) was removed by centrifugation
(7000 g, 25 min, Yingtai instrument GL21MC, rotor
GL21MC30110, Changsha, China). To the supernatant,

additionally 1.32 L isopropanol (precooled to )20 °C) was
added. After 24 h at )20 °C, a precipitate was collected by
suction filtration. This precipitate of crude peptides (11.5 g,
the second fraction) was dissolved in 500 mL 0.2 m acetic
acid and chromatographed on a Sephadex G-25 fine (Phar-
macia, Uppsala, Sweden) column (10 · 90 cm) in 0.2 m
acetic acid. After lyophilization, fraction 4 (1.2 g dry mater-
ial) was found to be bioactive and was partly soluble in
120 mL 0.01 m ammonium bicarbonate, pH 8.0. Insoluble
material was removed by centrifugation (12 000 g, 25 min,
Yingtai instrument GL21MC, rotor GL21MC30107).
The supernatant was chromatographed on an Express
Ion-Exchange C (Whatman, Maidstone, UK) column
(2.5 · 60 cm) by stepwise elution with 0.01, 0.02, 0.05, 0.1
and 0.2 m ammonium bicarbonate, pH 8.0. The polypeptide
fractions eluted with 0.1 m ammonium bicarbonate exhib-
ited activity and were lyophilized (90.5 mg dry material).
An aliquot (2.0 mg) of this material was subjected to
RP-HPLC using an Agilent 1100 system (Agilent Technol-
ogies, Wilmington, DE, USA) fitted with a Zorbax C
18
col-
umn (4.6 · 150 mm, 5 l m particles). Eluent A was 0.1%
trifluoroacetic acid in water, and eluent B, 0.1% trifluoro-
acetic acid in acetonitrile. A linear gradient of 10–60%
eluent B in 50 min (1 mLÆmin
)1
) was employed. Eluted
components absorbing at 214 nm were collected, lyophilized
and analyzed for bioactivity.

After establishing the amino acid sequence and hence
identity of aglycin (below), the peptide was purified from
pea seeds to acquire sufficient amounts for further charac-
terization and analysis. A similar protocol was then
employed and the homogeneity of the preparations was
X P. Dun et al. Aglycin activity in mammals
FEBS Journal 274 (2007) 751–759 ª 2007 The Authors Journal compilation ª 2007 FEBS 755
checked by coelution of plant and animal material in
RP-HPLC and by electrophoresis [26].
Structural analysis of aglycin
Molecular masses of components recovered from RP-HPLC
were determined using MALDI-TOF mass spectrometry in
an Applied Biosystems (Foster City, CA, USA) Voyager
4307 instrument, using a-cyano-4-hydroxycinnamic acid at
10 mgÆmL
)1
70% acetonitrile, 0.1% trifluoroacetic acid as
matrix. The PA1F, chain b, preparation was analyzed by
electrospray mass spectrometry directly before biological test-
ing using a Waters (Manchester, UK) Q-TOF Ultima instru-
ment fitted with a PicoTip nanospray emitter (New Objective,
Woburn, MA, USA). Edman degradation was carried out
without prior reduction and alkylation of the sample in an
Applied Biosystems Procise HT instrument. Cysteine residues
were indirectly identified by the presence of gaps in the other-
wise clearly interpretable sequence. Computer searches of
peptide sequences were performed in the SwissProt and
TrEMBL databases. Protein concentrations were determined
using the Bio-Rad (Hercules, CA, USA) protein assay.
Stability against proteolytic cleavage

Stability of aglycin towards pepsin (Calbiochem, San Diego,
CA, USA), trypsin (Promega, Madison, WI, USA) and
Glu-C protease (Roche Diagnostics, Basel, Switzerland) was
tested. For pepsin, 2 lLat1lgÆlL
)1
was added to 20 lg
aglycin dissolved in 18 lL water (adjusted to pH 2.0 with
1 m HCl); for trypsin, 2 lLat1lgÆlL
)1
was added to 20 lg
aglycin dissolved in 18 lL 1% ammonium bicarbonate.
Both reaction mixtures were incubated at 37 °C for 12 h.
For Glu-C protease, 2 lLat1lgÆlL
)1
was added to 20 lg
aglycin dissolved in 18 l L 1% ammonium acetate, followed
by incubation at 25 °C for 12 h. The reaction mixtures
were analyzed by RP-HPLC on a Zorbax C
18
column
(4.6 · 150 mm, 5 lm particles). HPLC conditions were elu-
ent A, 0.1% trifluoroacetic acid in water; eluent B, 0.1% tri-
fluoroacetic acid in acetonitrile; flow rate, 1 mLÆmin
)1
;
gradient 10–60% B, 0–50 min; and detection at 214 nm.
Effect on blood glucose in mice
Normal mice (Kunming, 18–20 g, n ¼ 40) were obtained
from the standard animal center of China Medical College
(Beijing, China). The mice were fasted for 8 h and the ini-

tial blood glucose concentration was determined with Accu-
Chek Advantage blood glucose monitor (Roche Diagnos-
tics). The mice were then divided into an equal number of
animals receiving saline and aglycin, respectively. For ani-
mals in the saline group, 100 lL 0.9% NaCl was injected
subcutaneously, and in the aglycin group, the aglycin pep-
tide at 10 lgÆg
)1
body weight (in 0.9% NaCl at 2 lgÆlL
)1
).
After injection, the blood glucose concentration was meas-
ured at time points 20, 40, 60 and 80 min. Blood for deter-
mination of glucose concentration was taken from the tail
of each animal. Blood glucose values are given as means of
data collected from 20 animals ± SEM. A statistical com-
parison between the groups was performed with the Stu-
dent t-test. P < 0.05 was considered significant. Animal
experiments were designed and carried out according to the
directive 86 ⁄ 609 ⁄ EEC to minimize pain and discomfort.
The effect on blood glucose was further tested using one
of the other polypeptide isoforms, PA1F, chain b, for
which the sequence is 91% identical to that of PA1B,
chain b (three amino acid replacements: Ser17Thr, Val29Ile
and His34Asn). The experimental conditions were the same
(except that C57BL ⁄ 6 mice were used and aglycin was
injected into six animals).
Detection of an aglycin binding protein
Tissue extracts
Fresh porcine pancreas, liver, kidney and muscle (500 g

each) were collected from a local slaughter house and
immediately washed with 0.25 m sucrose (precooled to
4 °C). After removal of connective tissue and fat, the
material was cut into small pieces and washed with buffer
A (50 mm Hepes, pH 7.6, containing 1 mm phenyl-
methanesulfonyl fluoride, 1 mm dithiothreitol, 1 mm
EDTA, 0.2 mgÆmL
)1
soybean trypsin inhibitor, 2 lgÆmL
)1
aprotinin, 5 lgÆmL
)1
leupeptin and 1 mgÆmL
)1
bacitracin),
disintegrated in a JJ-2 homogenizer (GuoHua Instrument
Co., Wuhan, China) with two volumes of buffer A contain-
ing 0.25 m sucrose, centrifuged first at 600 g for 10 min at
4 °C, and then the supernatants again at 12 000 g for
15 min at 4 °C (Yingtai instrument GL21MC, rotor
GL21MC30107). The second supernatants were finally cen-
trifuged at 200 000 g for 60 min at 4 °C (rotor TLA-100,
Beckman Coulter, Fullerton, CA, USA) to obtain a micro-
somal membrane pellet, washed once with buffer A, dis-
solved to a final protein concentration of approximately
10 mgÆmL
)1
with buffer A containing 1% Triton X-100,
stirred for 45 min at 4 °C, then centrifuged again at
200 000 g for 45 min at 4 °C (rotor TLA-100, Beckman

Coulter, Fullerton, CA, USA) [27]. The clear supernatant
was used for surface plasmon resonance measurements to
detect binding proteins and for subsequent isolation by
affinity chromatography (below). Pancreas from 20 mice
was similarly processed and the extract also used for isola-
tion and characterization of aglycin binding proteins.
Biosensor analysis
Spreeta biosensor and software (American TI Corp., Attle-
boro, MA, USA) was used to detect aglycin binding pro-
teins. Aglycin was immobilized onto the surface plasmon
Aglycin activity in mammals X P. Dun et al.
756 FEBS Journal 274 (2007) 751–759 ª 2007 The Authors Journal compilation ª 2007 FEBS
resonance sensor chip according to the manufacturer’s pro-
tocol. The efficiency of immobilization was evaluated by
atomic force microscopy using a NanoIIIa instrument
(Digital Instrument Company, Santa Barbara, CA, USA).
Extracts of membrane proteins prepared from pancreas,
liver, kidney and muscle (above) were diluted with Hepes-
buffered saline (HBS: 10 mm Hepes, pH 7.4, containing
0.15 m NaCl) to final protein concentration 200 lgÆmL
)1
.
Aliquots were injected over the sensor chip surface at a
flow rate of 20 lLÆmin
)1
for 2–3 min at 25 °C. After each
injection, the sensor chip was thoroughly washed with HBS
containing 0.05% Triton X-100 and equilibrated with HBS.
Binding interactions were continuously monitored and plot-
ted as refractive index versus time and displayed in a sen-

sorgram [28].
Affinity purification
Aglycin (5 mg) was coupled to 1 mL CNBr-activated
Sepharose 4B (Amersham Pharmacia Biotech, Uppsala,
Sweden) according to the manufacturer’s protocol. The
affinity resin was transferred to a column and equili-
brated with buffer B (50 mm Hepes, pH 7.6, containing
0.1% Triton X-100) at 4 °C. The pancreatic membrane
protein extract (porcine or mice) was diluted three-fold
with buffer A and applied at 0.5 mLÆmin
)1
(4 °C). After
adsorption, the column was washed with buffer C (buf-
fer B containing 1 mm phenylmethanesulfonyl fluoride
and 1 mm dithiothreitol), followed by thorough washing
with buffer C containing 1 m NaCl, at 40 mLÆh
)1
. For
elution, monitoring was at 254 nm, with first buffer D
(50 mm acetate, pH 5.0, containing 1 m NaCl and 0.1%
Triton X-100, 1 mm phenylmethanesulfonyl fluoride,
1mm dithiothreitol), and then buffer D containing 1.5 m
urea at 20 mLÆh
)1
. Eluted fractions (2 mL) were collected
in tubes containing 1 mL 0.5 m Tris ⁄ HCl, pH 8.25, and
pooled according to the peak patterns. After immediate
dialysis against 10 mm Hepes buffer, 0.1% Triton X-100,
pH 7.6 [27], 2 mL aliquots were taken for interaction
studies between pancreatic proteins and aglycin by

ELISA (below). Remaining parts of the fractions were
lyophilized for protein characterization.
ELISA measurements
The interaction between porcine pancreas proteins and
aglycin was studied in an ELISA array. Briefly, 96 well pol-
yvinylchloride plates were coated with 50 lL porcine pan-
creas protein fraction from affinity chromatography
(10 lgÆmL
)1
in 50 mm Na
2
CO
3
⁄ NaHCO
3
, pH 9.6) and
incubated at 4 °C overnight. The wells were washed with
20 mm NaCl ⁄ P
i
containing 0.1% gelatin, then blocked with
20 mm NaCl ⁄ P
i
containing 1% gelatin for 1 h at 37 °C.
The wells were washed three times with NaCl ⁄ P
i
-T (20 mm
NaCl ⁄ P
i
, 0.1% Tween-20) containing 0.1% gelatin, before
incubation overnight at 4 °C with aglycin, 50 lLofa

50 lgÆmL
)1
solution in NaCl ⁄ P
i
-T. The wells were washed
with NaCl ⁄ P
i
-T, followed by addition of an antiaglycin
monoclonal antibody (prepared in our laboratory as des-
cribed [29]) at a 1 : 20 000 dilution of 2 mgÆmL
)1
in
NaCl ⁄ P
i
-T and incubation 1 h at 37 °C. After removal of
nonbinding antibodies with NaCl ⁄ P
i
-T (five times, 3 min
each), horseradish peroxidase (HRP) labeled rabbit anti-
mouse IgG secondary antibody was applied and incubated
for 1 h at 37 °C. Non-adsorbed IgG-HRP complex was
thoroughly removed by washing with NaCl ⁄ P
i
-T. Bound
HRP was monitored by addition of o-phenylenediamine
and detection at 492 nm.
Identification of an aglycin binding protein
Gel electrophoresis
SDS ⁄ PAGE of fractions from the affinity purification was
carried out in 0.75 mm 12% slab gels (Bio-Rad) [30]. Sam-

ples were dissolved in 5% SDS containing 20 mm dithio-
threitol and incubated for 12 h at room temperature. The
electrophoresis was conducted in the presence of 0.1% SDS
and 20 mm dithiothreitol. Rabbit phosphorylase b
(97.4 kDa), bovine serum albumin (66.2 kDa), rabbit actin
(43.0 kDa), bovine carbonic anhydrase (31.0 kDa) and
trypsin inhibitor (20.1 kDa) were used as molecular mass
standards.
In-gel digestion and peptide mass fingerprinting
After separation by SDS ⁄ PAGE, the gel was stained with
Coomassie blue and the single band detected from the
aglycin binding fraction was excised and cut into small
pieces (1 mm
2
). The pieces were placed in a 0.65 mL silic-
onized tube, washed twice with 250 lL 100 mm ammo-
nium bicarbonate, vortexed in 250 lL 50% (v ⁄ v)
acetonitrile ⁄ 100 mm ammonium bicarbonate for 10 min,
and dehydrated in 150 lL neat acetonitrile until the gel
turned opaque. It was then dried in a Speed Vac (Globule
Medical Instrument, Ramsey, MN, USA) for 20 min and
subsequently reswelled in trypsin solution (three-fold the
gel volume; 300 lL 100 mm ammonium bicarbonate con-
taining 3 lg trypsinÆmL
)1
, Calbiochem, San Diego, CA,
USA) for 10 min. After addition of 100 lL 100 mm
ammonium bicarbonate, digestion was carried out at 37 °C
overnight. The solution was then transferred to a 0.65 mL
siliconized tube and the gel pieces were extracted twice

under sonication (10 min) with 50 lL 50% (v ⁄ v) acetonit-
rile containing 5% trifluoroacetic acid. The digest and
extracts were combined and concentrated under vacuum.
Peptide mass fingerprints were determined by MALDI-
TOF mass spectrometry in a Tof Spec instrument (Micro-
mass, Manchester, UK) [31,32] and submitted to data-
base searches using the mascot software (http://www.
matrixscience.com).
X P. Dun et al. Aglycin activity in mammals
FEBS Journal 274 (2007) 751–759 ª 2007 The Authors Journal compilation ª 2007 FEBS 757
Acknowledgements
This work was supported by grants from the National
Science Foundation of China (30370647 and
30470823), the Chinese 863 Program (2002AA214061),
the Swedish Research Council (03X-3532, 629-2002-
8654 and 621-2003-3616), the Swedish Cancer Society
(4159), the Wallenberg Consortium North (WCN), the
Juvenile Diabetes Foundation (JDFI-4-99-647), the
European Commission (LSHC-CT-2003–503297), and
Karolinska Institutet.
References
1 Bayliss WM & Starling EH (1902) On the causation of
the so-called ‘peripheral reflex secretion’ of the pancreas.
Proc Roy Soc Lond 69, 352–353.
2 Banting FG, Best CH, Collip JB, Campbell WR &
Fletcher AA (1922) Pancreatic extracts in the treatment
of diabetes mellitus. Preliminary report. Can Med Assoc
J 12, 141–146.
3 Macleod JJR (1922) The source of insulin. A study of
the effect produced on blood sugar by extracts of the

pancreas and principal islets of fishes. J Metab Res 2,
149–172.
4 Ryan CA, Pearce G, Scheer J & Moura DS (2002) Poly-
peptide Hormones. Plant Cell Suppl. 251–264.
5 Lindsey K (2001) Plant peptide hormones: the long and
the short of it. Current Biology 11, R741–R743.
6 Oliveira AEA, Machado OLT, Gomes VM, Neto JX,
Pereira AC, Vieira JGH, Fernandez KVS & Xavier-
Filho J (1999) Jack bean seed coat contains a protein
with complete sequence homology to bovine insulin.
Protein Peptide Lett 6, 15–22.
7 Maryani MM, Morse MV, Bradley G, Irving HR,
Cahill DM & Gehring CA (2003) In situ localization
associates biologically active plant natriuretic peptide
immuno-analogues with conductive tissue and stomata.
J Exp Bot 54, 1553–1564.
8 Higgins TJ, Chandler PM, Randall PJ, Spencer D,
Beach LR, Blagrove RJ, Kortt AA & Inglis AS (1986)
Gene structure, protein structure, and regulation of the
synthesis of a sulfur-rich protein in pea seeds. J Biol
Chem 261, 11124–11130.
9 Strausberg RL, Feingold EA, Grouse LH, Derge JG,
Klausner RD, Collins FS, Wagner L, Shenmen CM,
Schuler GD, Altschul SF et al. (2002) Generation and
initial analysis of more than 15 000 full-length human
and mouse cDNA sequences. Proc Natl Acad Sci USA
99, 16899–16903.
10 Chen Z-W, Agerberth B, Gell K, Andersson M, Mutt
V, O
¨

stenson CG, Efendic S & Jo
¨
rnvall H (1988) Isola-
tion and characterization of porcine diazepam-binding
inhibitor, a polypeptide not only of cerebral occurrence
but also common in intestinal tissues and with effects
on regulation of insulin release. Eur J Biochem 174,
239–245.
11 Yu WH, Wolfgang W & Forte M (1995) Subcellular
localization of human voltage-dependent anion channel
isoforms. J Biol Chem 270, 13998–14006.
12 Watanabe Y, Barbashov SF, Komatsu S, Hemmings
AM, Miyagi M, Tsunasawa S & Hirano H (1994) A
peptide that stimulates phosphorylation of the plant
insulin-binding protein. Isolation, primary structure and
cDNA cloning. Eur J Biochem 224, 167–172.
13 Yamazaki T, Takaoka M, Katoh E, Hanada K, Sakita
M, Sakata K, Nishiuchi Y & Hirano H (2003) A possible
physiological function and the tertiary structure of a
4-kDa peptide in legumes. Eur J Biochem 270, 1269–1276.
14 Komatsu S & Hirano H (1991) Plant basic 7S globulin-
like proteins have insulin and insulin-like growth factor
binding activity. FEBS Lett 294, 210–212.
15 Hanada K & Hirano H (2004) Interaction of a 43-kDa
receptor-like protein with a 4-kDa hormone-like peptide
in soybean. Biochemistry 43, 12105–12112.
16 Hanada K, Nishiuchi Y & Hirano H (2003) Amino acid
residues on the surface of soybean 4-kDa peptide
involved in the interaction with its binding protein. Eur
J Biochem 270, 2583–2592.

17 Norton RS & Pallaghy PK (1998) The cystine knot
structure of ion channel toxins and related polypeptides.
Toxicon 36, 1573–1583.
18 Gressent F, Rahioui I & Rahbe Y (2003) Characteriza-
tion of a high-affinity binding site for the pea albumin
1b entomotoxin in the weevil Sitophilus. Eur J Biochem
270, 2429–2435.
19 Le Gall M, Quillien L, Gue
´
guen J, Rogniaux H & Se
`
ve
B (2005) Identification of dietary and endogenous ileal
protein losses in pigs by immunoblotting and mass spec-
trometry. J Nutr 135, 1215–1222.
20 Craik DJ, Daly NL & Waine C (2001) The cystine knot
motif in toxins and implications for drug design. Toxi-
con 39, 43–60.
21 Blachly-Dyson E, Zambronicz EB, Yu WH, Adams V,
McCabe ER, Adelman J, Colombini M & Forte M
(1993) Cloning and functional expression in yeast of
two human isoforms of the outer mitochondrial mem-
brane channel, the voltage-dependent anion channel.
J Biol Chem 268, 1835–1841.
22 Thinnes FP, Gotz H, Kayser H, Benz R, Schmidt WE,
Kratzin HD & Hilschmann N (1989) Identification of
human porins. I. Purification of a porin from human
B-lymphocytes (Porin 31HL) and the topochemical
proof of its expression on the plasmalemma of the
progenitor cell. Biol Chem Hoppe Seyler 370, 1253–

1264.
23 Buettner R, Papoutsoglou G, Scemes E, Spray DC &
Dermietzel R (2000) Evidence for secretory pathway
localization of a voltage-dependent anion channel iso-
form. Proc Natl Acad Sci USA 97, 3201–3206.
Aglycin activity in mammals X P. Dun et al.
758 FEBS Journal 274 (2007) 751–759 ª 2007 The Authors Journal compilation ª 2007 FEBS
24 Lawen A, Ly JD, Lane DJ, Zarschler K, Messina A &
DePinto VC (2005) Voltage-dependent anion-selective
channel 1 (VDAC1): a mitochondrial protein rediscov-
ered as a novel enzyme in the plasma membrane. Int J
Biochem Cell Biol 37, 277–282.
25 Lee JY, Boman A, Sun CX, Andersson M, Jo
¨
rnvall
H, Mutt V & Boman HG (1989) Antibacterial
peptides from pig intestine: isolation of a mam-
malian cecropin. Proc Natl Acad Sci USA 86,
9159–9162.
26 Schagger H & von-Jagow G (1987) Tricine-sodium
dodecyl sulfate-polyacrylamide gel electrophoresis for
the separation of proteins in the range from 1 to 100
kDa. Anal Biochem 166, 368–379.
27 Fujita-Yamaguchi Y, Choi S, Sakamoto Y & Itakura K
(1983) Purification of insulin receptor with full binding
activity. J Biol Chem 258, 5045–5049.
28 Bender V, Ali M, Amon M, Diefenbach E & Manolios
N (2004) T cell antigen receptor peptide–lipid mem-
brane interactions using surface plasmon resonance.
J Biol Chem 279, 54002–54007.

29 Wang J-H, Dun X-P, Qu L-N, Zhao Y-Y, Yang T-B &
Chen Z-W (2005) Preparation and identification of
monoclonal antibodies against pea albumin 1b (PA1b).
Hybridoma 24, 197–200.
30 Laemmli UK (1970) Cleavage of structural proteins
during the assembly of the head of bacteriophage T4.
Nature 227, 680–685.
31 Jeno P, Mini T, Moes S, Hintermann E & Horst M
(1995) Internal sequences from proteins digested in
polyacrylamide gels. Anal Biochem 224, 75–82.
32 Medzihradszky KF, Leffler H, Baldwin MA & Burlin-
game AL (2001) Protein identification by in-gel digestion,
high-performance liquid chromatography, and mass spec-
trometry: peptide analysis by complementary ionization
techniques. J Am Soc Mass Spectrom 12, 215–221.
X P. Dun et al. Aglycin activity in mammals
FEBS Journal 274 (2007) 751–759 ª 2007 The Authors Journal compilation ª 2007 FEBS 759