Transgenic expression of bean alpha-amylase inhibitor in peas results in altered structure and immunogenicity.

Abstract

The development of modern gene technologies allows for the expression of recombinant proteins in non-native hosts. Diversity in translational and post-translational modification pathways between species could potentially lead to discrete changes in the molecular architecture of the expressed protein and subsequent cellular function and antigenicity. Here, we show that transgenic expression of a plant protein (alpha-amylase inhibitor-1 from the common bean (Phaseolus vulgaris L. cv. Tendergreen)) in a non-native host (transgenic pea (Pisum sativum L.)) led to the synthesis of a structurally modified form of this inhibitor. Employing models of inflammation, we demonstrated in mice that consumption of the modified alphaAI and not the native form predisposed to antigen-specific CD4+ Th2-type inflammation. Furthermore, consumption of the modified alphaAI concurrently with other heterogeneous proteins promoted immunological cross priming, which then elicited specific immunoreactivity of these proteins. Thus, transgenic expression of non-native proteins in plants may lead to the synthesis of structural variants possessing altered immunogenicity.

Full text

Page 1

Transgenic Expression of Bean r-Amylase Inhibitor in Peas
Results in Altered Structure and Immunogenicity
VANESSA E. PRESCOTT,†PETER M. CAMPBELL,§ANDREW MOORE,|
JOERG MATTES,†MARC E. ROTHENBERG,‡PAUL S. FOSTER,†
T. J. V. HIGGINS,|AND SIMON P. HOGAN*,†
Division of Molecular Bioscience, The John Curtin School of Medical Research, Australian National
University, Canberra, ACT, Australia, Division of Allergy and Immunology, Department of Pediatrics,
Cincinnati Children’s Hospital Medical Center, University of Cincinnati College of Medicine,
Cincinnati, Ohio 45229, and Divisions of Entomology and Plant Industry, Commonwealth Scientific
and Industrial Research Organization, Canberra, ACT, Australia
The development of modern gene technologies allows for the expression of recombinant proteins in
non-native hosts. Diversity in translational and post-translational modification pathways between
species could potentially lead to discrete changes in the molecular architecture of the expressed
protein and subsequent cellular function and antigenicity. Here, we show that transgenic expression
of a plant protein (R-amylase inhibitor-1 from the common bean (Phaseolus vulgaris L. cv.
Tendergreen)) in a non-native host (transgenic pea (Pisum sativum L.)) led to the synthesis of a
structurally modified form of this inhibitor. Employing models of inflammation, we demonstrated in
mice that consumption of the modified RAI and not the native form predisposed to antigen-specific
CD4+Th2-type inflammation. Furthermore, consumption of the modified RAI concurrently with other
heterogeneous proteins promoted immunological cross priming, which then elicited specific immuno-
reactivity of these proteins. Thus, transgenic expression of non-native proteins in plants may lead to
the synthesis of structural variants possessing altered immunogenicity.
KEYWORDS:
photometry
r-Amylase inhibitor; transgenic plant; animal model; Th2 inflammation; mass spectro-
INTRODUCTION
Genetically modified (GM) plants are designed to enhance
agronomic productivity or product quality and are being
increasingly employed in both agricultural and livestock indus-
tries (1, 2). Recently, peas (Pisum satiVum L.) expressing a gene
for R-amylase inhibitor-1 (RAI) from the common bean
(Phaseolus vulgaris L. cv. Tendergreen) were generated to
protect the seeds from damage by inhibiting the R-amylase
enzyme in old world bruchids (pea, cowpea, and azuki bean
weevils) and are currently undergoing risk assessments (3-6).
The present study was initiated to (1) characterize the
proteolytic processing and glycopeptide structures of RAI when
transgenically expressed in peas (pea-RAI) and (2) evaluate in
an in vivo model system the immunological consequence of
oral consumption of pea-RAI. We demonstrate that expression
of RAI in pea leads to a structurally modified form of this
inhibitor. Employing experimental models, we show that the
structural modification can lead to altered antigenicity. These
investigations reveal that expression of proteins in non-native
hosts can lead to the synthesis of a protein variant with altered
immunogenicity.
MATERIALS AND METHODS
Nontransgenic and Transgenic Plants. Seed meal was obtained
from nontransgenic peas, genetically modified peas expressing bean
R-amylase inhibitor-1 (RAI) (5), genetically modified narrow leaf lupin
(Lupinus angustifolius L.) expressing sunflower seed albumin protein
(SSA) in the seeds (SSA-lupin) (7), and from nontransgenic Pinto bean.
Seeds were ground into fine flour in liquid N2using a mortar and pestle.
This seed meal was then suspended in PBS (0.166 g meal/mL),
homogenized, sieved through a 70 µm mesh, and stored at -70 °C. In
some experiments, seed meal homogenates were cooked at 100 °C for
30 min before administration to mice (indicated in text).
Purification of SSA from Transgenic Lupin and rAI from
Common Beans and from Transgenic Peas. RAI was purified from
the common beans (Pinto and Tendergreen) and transgenic peas and
SSA from genetically modified narrow leafed lupin (SSA-lupin) as
previously described (7, 8). Purified proteins were analyzed by sodium
dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE, 15-
* Author to whom correspondence should be addressed [telephone (513)
636-6620; fax (513) 636-3310; e-mail simon.hogan@cchmc.org].
†Australian National University.
§Division of Entomology, Commonwealth Scientific and Industrial
Research Organization.
|Division of Plant Industry, Commonwealth Scientific and Industrial
Research Organization.
‡University of Cincinnati College of Medicine.
J. Agric. Food Chem. 2005, 53, 9023−9030
9023
10.1021/jf050594v CCC: $30.25© 2005 American Chemical Society
Published on Web 10/15/2005

Page 2

25% gradient, 1 mm thick, mini-gel format) and MALDI-TOF mass
spectrometry.
Western Immunoblot Analysis. RAI polypeptide composition was
determined in protein extracts from common bean and transgenic peas
as previously described (3). Protein was extracted from seeds with 0.5
M NaCl, 1 mM EDTA, and 0.1 M N-tris(hydroxymethyl)methylamino-
ethanesulfonic acid at pH 7.8. Aliquots of reduced protein (20 µg by
Bradford assay) were fractionated by SDS-PAGE and electroblotted
onto nitrocellulose membrane. RAI polypeptides were detected with
an RAI antiserum from rabbit and goat anti-rabbit IgG conjugated to
alkaline phosphatase (3). The concentration of RAI in transgenic peas
was determined as 4% of total protein as previously described (3).
Structural Analysis of Purified rAI from the Pinto and Ten-
dergreen Beans and from Transgenic Peas. Purified RAI from the
common beans, Pinto and Tendergreen, and from transgenic peas were
analyzed by matrix-assisted laser desorption/ionization-time-of-flight-
mass spectrometry (MALDI-TOF-MS). The proteins were dissolved
in water (approximately 1 µg/µL), and then 1 µL was mixed with 1
µL of matrix solution (saturated sinapinic acid in 50% acetonitrile/
0.1% trifluoroacetic acid) on the sample plate of a Voyager Elite
MALDI-TOF mass spectrometer (Perseptive Biosystems) and allowed
to dry. Spectra were collected in linear mode with myoglobin used for
close external calibration (Sigma, Cat. No. M-1882, 16952.6 [M +
H]+, 8476.8 [M + 2H]2+).
Mice and Intragastric Administration of Seed Meal from
Nontransgenic and Transgenic Plants. BALB/c mice were obtained
from specific pathogen-free facilities at the Australian National
University. Mice were intragastrically administered 250 µL of seed
meal suspension (∼100 mg/mL) containing either transgenic peas,
nontransgenic peas, SSA-lupin, or Pinto bean twice a week for 4 weeks.
In some experiments, serum was taken from the mice at the start of
the third and fifth weeks during feeding. The serum antibody titers
were determined as previously described (9).
Mice and Delayed Type Hypersensitivity Responses. BALB/c
mice were administered seed meal as described above. Seven days
following the final intra-gastric challenge, mice were subcutaneously
injected with 25 µL of the appropriate antigen [Tendergreen-RAI, pea
R-AI, or lupin SSA (1 mg/mL in PBS)] into the footpad. The positive
control [(+) control] is mice immunized by i.p. injection of 200 µL
containing 50 µg of Tendergreen-RAI dissolved in PBS with Alum
(1 mg/mL) and subsequently receiving 25 µL of purified Tendergreen-
RAI (1 mg/mL PBS). The negative control [(-) control] is mice
immunized by i.p. injection of 200 µL containing 50 µg of Tendergreen-
RAI dissolved in PBS with Alum (1 mg/mL) and subsequently receiving
25 µL of PBS. DTH responses were assessed by measuring the specific
increase in footpad thickness using a digmatic calliper (Mitutoyo,
Kawasaki, Japan) 24 h following the challenge. Serum was collected
on day 14, and antibody titers were determined as previously described
(9).
Murine Model of CD4+Th2 Cell-Mediated Inflammation.
BALB/c WT mice were administered seed meal as indicated in the
text. Seven and nine days following the final intra-gastric challenge,
mice were anesthetized with an intravenous injection of 100 µL of
Saffan solution (1:4 diluted in PBS). Mice were intubated with a 22
gauge catheter needle, through which purified RAI from Tendergreen
bean or transgenic pea (1 mg/mL PBS), or vehicle control (PBS), was
instilled. Airway responsiveness (AHR), mucus production, and
eosinophilia were measured 24 h following the final intra-tracheal
challenge. AHR to methacholine was assessed in conscious, unrestrained
mice by barometric plethysmography, using apparatus and software
supplied by Buxco (Troy, NY) as previously described (9). This system
yields a dimensionless parameter known as enhanced pause (Penh),
reflecting changes in waveform of the pressure signal from the
plethysmography chamber combined with a timing comparison of early
and late expiration, which can be used to empirically monitor airway
function. Measurements were performed as previously described (9).
Lung tissue representing the central (bronchi-bronchiole) and peripheral
(alveoli) airways was fixed, processed, and stained with Alcian Blue-
PAS for enumeration of mucin-secreting cells or Charbol’s chromo-
trope-Haematoxylin for identification of eosinophils as previously
described (9).
Intragastric Administration of Purified rAI and OVA. Mice were
administered 200 µL of affinity purified Tendergreen- or transgenic
pea-RAI (5 µg) with ovalbumin (OVA, 1 mg/mL) in a PBS suspension
three times a week for 2 weeks. One week following feeding, the mice
were intubated with a 22 gauge catheter needle, through which 25 µL
of OVA (1 mg/mL PBS), or vehicle control (PBS), was instilled and
the CD4+Th2-inflammation indices determined as described above.
Serum was taken from the mice 1 day after the final intra-tracheal
challenge, and serum antibody titers were determined as described (9).
Antigen Specific CD4+T-Cell Response. Peribronchial lymph
nodes (PBLN) were subjected to pea-RAI or RCD3/RCD28 stimulation
as previously described (9). In brief, 5 × 105PBLN cells/mL were
cultured with RAI (50 µg/mL) or RCD3 (5 µg/mL)/RCD28 (1 µg/mL)
for 96 h. IL-4, IL-5, IFNγ levels were determined in supernatants from
stimulated PBLN homogenates by using the OptEIA Mouse IL-4, IL-
5, and IFNγ kits (Pharmingen).
Statistical Analysis. The significance of differences between
experimental groups was analyzed using Student’s unpaired t-test.
Values are reported as the mean ( SEM. Differences in means were
considered significant if p < 0.05.
RESULTS
MALDI-TOF-MS Analysis of rAI. To assess the conse-
quences of transgenic expression of the bean RAI in peas, we
initially performed a structural analysis of the transgenically
expressed protein (pea-RAI). Pea-RAI was compared by Western
blot analysis and MALDI-TOF-MS with natively expressed RAI
from the common beans, cvs. Pinto (Pinto-RAI) and Tender-
green (Tendergreen-RAI) (collectively termed bean-RAI). Previ-
ous studies have shown that bean-RAI is synthesized as a pre-
pro-RAI polypeptide that is cleaved following Asn77to form
two peptide chains (R and ?), both of which are glycosylated
and have one or more amino acid residue(s) removed from their
C-termini (8). This post-translational processing results in major
forms of the R and ? chains with masses of 11 646 and 17 319,
respectively, and minor forms containing alternative glycans
(10-12). Western immunoblot analysis of Tendergreen-RAI and
pea-RAI revealed immunoreactive bands in the 11 000-18 000
mass range consistent with the reported structure (10-13).
Detailed comparison of Tendergreen-RAI with pea-RAI revealed
differences in the banding profile, suggesting possible differ-
ences in the molecular structure of natively and transgenically
expressed RaI (Figure 1A).
To better resolve the differences between pea-RAI and bean-
RAI, affinity purified RAI was analyzed by MALDI-TOF-MS
(Figure 1B). The mass spectra of Tendergreen-RAI and Pinto-
RAI closely matched a previously published spectrum (10) of
a bean-RAI (Phaseolus Vulgaris L. cv. Greensleeves) confirming
that both Tendergreen- and Pinto-RAI possess similar well-
characterized post-translational modifications and very similar
relative abundance of minor processing variants (10, 11).
Alignment of our spectra with the previously published data
(10) allowed identification of peaks in the pea-, Tendergreen-,
and Pinto-RAI spectra. The major form of the R-chain (11 646
Da) of bean-RAI contains residues 1-76 by cleavage of the
pro-protein following Asn77, removal of Asn77, and the addition
of sugar residues (Man6GlcNAc2at Asn12and Man9GlcNAc2
at Asn65). Minor forms of the R-chain of bean-RAI differed by
having one to three fewer mannose residues resulting in a series
of peaks in the MALDI-TOF spectrum that differ by 162 mass
units. In contrast, less heavily glycosylated forms dominated
for the R-chain of pea-RAI. In particular, an R-chain with two
fewer mannose residues (11 322 Da) was the most abundant
for pea-RAI but the least abundant for Tendergreen-RAI (Figure
1C(i)). A further difference in the pea-RAI spectrum was a series
of minor peaks differing from the main R-chain peaks by either
9024
J. Agric. Food Chem., Vol. 53, No. 23, 2005 Prescott et al.

Page 3

+98 or -64 mass units, indicating another modification of some
of the pea-RAI R-chains (Figure 1C(i)).
The major form of the ?-chain of Greensleeves-RAI (16527
Da) contains residues 78-216 by cleavage of the pro-protein
following Asn77, the removal of the seven C-terminal residues
following Asn216, and the addition of sugar residues (Man3-
GlcNAc2Xyl1at Asn140) (10-13). The ?-chain region of the
Tendergreen-RAI spectrum closely aligned with that of Green-
sleeves-RAI (Figure 1C). The ?-chain region of the Pinto-RAI
spectrum also closely resembled that of Greensleeves-RAI
except that both major and minor peaks of Pinto-RAI were
shifted by approximately +104 mass units. This mass discrep-
ancy is consistent with five amino acid residue differences
between the ?-chains of Tendergreen-RAI and Pinto-RAI as
predicted by gene sequence comparison (see Supporting Infor-
mation Figure 1). Further, there are also three predicted residue
differences between the Tendergreen-RAI and Pinto-RAI R-chains
that result in a difference of +1 mass unit, which would not be
Figure 1. Western immunoblot and MALDI-TOF-MS analysis of common bean-derived-RAIs and RAI from transgenic peas. (A) Western blot analysis
of RAI protein in extracts of transgenic peas and the Tendergreen variety of common bean. The masses of standard proteins are indicated. (B) Aligned
MALDI-TOF mass spectra of purified RAI from transgenic pea and the common beans, Tendergreen and Pinto. (C) Detail from the spectra in panel B
showing the regions of the R-chain (i) and the ?-chain (ii).
Bean R-Amylase Inhibitor and ImmunogenicityJ. Agric. Food Chem., Vol. 53, No. 23, 2005
9025

Page 4

detected by our methods. These sequence differences are
consistent with previous reports of RAI polymorphisms among
bean cultivars (12, 13). The pea-RAI spectrum showed major
peaks corresponding to the two major and minor forms of the
?-chain found in Tendergreen-RAI; however, the pea-RAI
spectrum also showed a number of other peaks (Figure 1C(ii)).
DNA sequencing of the transgene in pea and comparison with
the published sequence (14) confirmed that the nucleotide
sequences were identical, establishing that the observed further
forms of the pea-RAI are related by variations in post-
translational modifications including glycosylation (Figure
1C(ii)).
Analysis of the spectra of pea- and bean-RAI also revealed
several other differences. First, a number of peaks at ∼8-9000
and 5824 mass units and below were observed in the bean-RAI
spectrum, which are consistent with a previously reported protein
that copurifies with bean-RAI (10) and doubly charged ((MH2)2+)
forms of the R-chain, respectively. Further, a peak at 4223 mass
units was detected in the pea-RAI spectrum, which has not been
previously reported. While this peak is barely detected in the
bean-RAI spectrum presented here, the peak was observed in a
number of other bean-RAI preparations (results not shown). The
mass of this peak is consistent with the first 39 residues of the
?-chain, which could be obtained by cleavage following an Asn
residue, the same protease specificity that provides the reported
processing of RAI at Asn77. Consistent with this hypothesis, a
small peak was detected in some preparations at about 12 304
mass units that could correspond to the remainder of the ?-chain.
While pea-RAI has not yet been characterized as thoroughly
as the bean-RAI, it is clear that the transgenic expression of
the bean RAI gene in the pea led to differences of glycosylation
and possibly other differences in both the R- and the ?-chains.
Immunological Consequence of Oral Consumption of
Beans. Peas are used as a feed component in the livestock
industry and also in human diets. Generally, dietary protein
antigens undergo gastric digestion leading to the formation of
nonimmunogenic peptides and the induction of a state of specific
immunological unresponsiveness termed oral tolerance (15, 16).
However, the demonstration of structural differences between
the transgenic RAI in pea and the natively expressed bean forms
raised the concern that the tolerance mechanism may be
perturbed, possibly leading to enhanced immunoreactivity.
The induction of oral tolerance results in the failure of the
immune system to elicit an active immune response to subse-
quent exposure to the same antigen in the skin (delayed type
hypersensitivity [DTH] response) or lung (CD4+T-helper [Th2]
cell-mediated inflammation). To examine potential differences
in immunological responsiveness following oral consumption,
mice were fed Pinto bean, which expresses a native form of
RAI and subsequently received purified Tendergreen-RAI in the
skin and lung. Most varieties of common beans such as Red
Kidney or Tendergreen contain high levels of phytohemagglu-
tinin (PHA), an anti-nutritional factor that induces dietary
toxicity in rodents and birds. We therefore used the Pinto variety
that contains very low levels of PHA (17, 18) as the appropriate
control for oral exposure. Oral consumption of native uncooked
Pinto bean seed flour followed by intra-tracheal (i.t.) challenge
with Tendergreen-RAI or phosphate buffered saline (PBS) failed
to induce an RAI-specific IgG1antibody response (Figure 2A).
Similarly, sub-cutaneous (s.c.) challenge of the footpad or i.t.
challenge of Pinto bean-fed mice with Tendergreen-RAI also
failed to promote a DTH response (results not shown) or a
pulmonary Th2-inflammatory response [pulmonary eosinophilia,
mucus hypersecretion, and enhanced AHR to a bronchocon-
strictive agents], respectively (Figure 2B-D). While the level
of AHR in the Pinto bean-fed RAI-challenged mice was higher
than PBS-challenged mice, the level of responsiveness is not
significantly different from that of naı ¨ve mice i.t. challenged
with Tendergreen-RAI (Figure 2D). As a positive control, mice
were sensitized by intra-peritoneal (i.p.) injection and subse-
quently challenged via the airways with bean-derived RAI to
induce immunological responsiveness (Figure 2A-D). Col-
lectively, these data showed that oral consumption of the native
bean form of RAI followed by respiratory exposure to bean-
RAI did not promote immunological responsiveness or inflam-
mation.
Immunological Consequence of Oral Consumption of
Transgenic Peas. To determine whether oral consumption of
the transgenic RAI (from pea) elicited an immunological
response, mice were orally administered transgenic pea seed
meal and RAI; serum antibody titers and DTH responses were
examined. Interestingly, in mice that were fed transgenic pea,
but not nontransgenic pea, RAI-specific IgG1was detected at 2
weeks and at significant levels after 4 weeks of oral exposure
(Figure 3A). Consistent with the antibody findings, mice fed
nontransgenic pea seed meal did not develop DTH responses
following footpad challenge with purified pea-RAI (Figure 3B).
In contrast, mice fed transgenic pea seed meal exhibited a
significant DTH response as compared to the nontransgenic pea
exposed group when purified pea-RAI was injected into the
footpad (Figure 3B). As a control for any general effect of
genetic modification, we repeated the experiment with material
from two other genetically modified plants, lupin (Lupinus
angustifolius L.) expressing sunflower seed albumin (SSA)
[transgenic lupin] (9) and chickpeas (Cicer arietinum L.)
expressing bean derived RAI. Mice were orally administered
lupin or transgenic lupin or chickpea or transgenic chickpea
seed meal and subsequently footpad challenged with SSA or
RAI and DTH responses were examined. In contrast to trans-
genic pea, mice fed transgenic lupin or chickpea did not develop
Figure 2. Experimental consumption of bean (cv. Pinto) seed meal does
not predispose to inflammation. (A) RAI-specific IgG1in serum and (B)
mucus-secreting cell numbers and (C) eosinophil levels in lung tissue
from Pinto bean-fed mice i.t. challenged with PBS or Tendergreen-RAI.
(D) AHR in Pinto bean-fed mice i.t. challenged with PBS or Tendergreen-
RAI. Data are expressed as the (A−D and F) mean ± SEM and (E)
mean O.D. of the serum dilution 1/10 ± SEM from 4 to 6 mice per group
from duplicate experiments. (A−D) * p < 0.05 as compared to Pinto bean-
fed i.t. RAI.
9026
J. Agric. Food Chem., Vol. 53, No. 23, 2005Prescott et al.

Page 5

DTH responses following footpad challenge with the transgeni-
cally expressed and purified SSA or RAI protein (Figure 3B;
results not shown). Thus, consumption of transgenic pea
containing RAI promoted RAI-specific immunological respon-
siveness.
To characterize the type of immune response elicited against
pea-RAI following oral consumption of transgenic pea, we
employed a well-characterized murine model of CD4+Th2cell-
mediated inflammation (19). Mice were orally administered
transgenic pea seed meal and subsequently i.t. challenged with
purified pea-RAI, and key features of Th2-inflammation [pul-
monary eosinophilia, mucus hypersecretion, and AHR] were
examined. I.t. challenge of nontransgenic pea-fed mice with
purified pea-RAI failed to induce features of Th2-inflammation
(Figure 4A-G). Furthermore, airways responsiveness to the
cholinergic spasmogen, methacholine, was not induced in these
Figure 3. Experimental consumption of transgenic pea seed meal predisposed to antigen-specific IgG1and DTH responses. (A) Antigen-specific IgG1
and (B) DTH responses in pea nontransgenic and pea transgenic-fed mice. Data are expressed as the (F) mean ± SEM and (E) mean O.D. of the serum
dilution 1/10 ± SEM from 4 to 6 mice per group from duplicate experiments. (A−C) * p < 0.05 as compared to nontransgenic pea or transgenic lupin
fed mice i.t. RAI.
Figure 4. Consumption of transgenic pea seed meal predisposed to CD4+Th2-type inflammatory response. Eosinophil accumulation in bronchoaveolar
lavage fluid (BAL) (A), tissue (B), and mucus-secreting cell numbers (C) in lung tissue from nontransgenic and transgenic pea-fed mice i.t. challenged
with RAI purified from pea. (D−G) Representative photomicrographs of eosinophil accumulation in lung of (D) nontransgenic and (E) pea transgenic-fed
mice and mucus-secreting cell numbers in lung tissue of (F) nontransgenic and (G) pea transgenic-fed mice i.t. challenged with RAI from pea. (H)
Airways hyperresponsiveness (AHR) in nontransgenic and pea transgenic-fed mice i.t. challenged with RAI from pea. Data are expressed as the mean
± SEM from 3 to 6 mice per group from duplicate experiments. Statistical significance of differences (p < 0.05) was determined using Student’s unpaired
t-test. (D−G) ×400 magnification.
Bean R-Amylase Inhibitor and ImmunogenicityJ. Agric. Food Chem., Vol. 53, No. 23, 2005
9027

Page 6

mice (Figure 4H). However, instillation of pea-RAI into the
lungs of mice fed transgenic pea induced key features of Th2-
type inflammation including pulmonary eosinophilia, mucus
hypersecretion, and AHR (Figure 4A-H).
Pulmonary eosinophilia, mucus hypersecretion, and AHR are
critically linked to the effector function of the Th2cytokines
(20). To examine whether consumption of transgenic pea
promoted a RAI-specific CD4+Th2-type T-cell response, CD4+
T-cells in peribronchial lymph node (PBLN) cultures from mice
fed nontransgenic pea or transgenic pea seeds challenged with
pea-RAI were stimulated with pea-RAI and cytokine profiles
determined. Stimulation of CD4+T-cells in peribronchial lymph
node (PBLN) cultures from nontransgenic pea-fed mice chal-
lenged with pea-RAI did not elicit Th2(interleukin (IL)-4 and
IL-5)- or Th1-type (gamma interferon, IFNγ) cytokine produc-
tion in response to pea-RAI stimulation (Figure 5A-C). By
contrast, stimulation of PBLN cultures with pea-RAI from i.t.
challenged mice fed transgenic pea resulted in the significant
production of Th2 cytokines (Figure 5A-C). Thus, oral
exposure of mice to transgenic pea, but not nontransgenic seed
meal, predisposed to systemic immunological responsiveness
characterized by a Th2-type immune profile.
Pea-rAI Promotes Immune Responses to Other Oral
Antigens. Previous investigations have demonstrated that vari-
ous plant-derived proteins such as tomatine possess immuno-
modulatory activity and potentiate and polarize immune re-
sponses (21-23). We have demonstrated that consumption of
transgenic pea in the presence of a large number of potential
dietary antigens in the gastrointestinal tract induces an active
systemic Th2-immune response against pea-RAI. In light of these
findings, we were next interested in determining whether
consumed pea-RAI possessed immunomodulatory activity for
Th2immune responses and could sensitize mice to heteroge-
neous nongenetically modified food antigens. Thus, we intra-
gastrically (i.g.) administered purified Tendergreen- or pea-RAI
with the well-characterized dietary antigen, chicken egg white
protein OVA, or OVA alone and subsequently i.t. challenged
mice with OVA. I.g. administration of OVA alone did not
systemically sensitize mice to OVA (Figure 6A). Further,
subsequent OVA challenge in the airways did not promote Th2-
inflammation (mucus hypersecretion, pulmonary eosinophilia,
or AHR). Similarly, i.g. administration of bean-RAI and OVA
did not systemically sensitize mice or predispose to Th2-
inflammatory processes. However, consumption of pea-RAI and
OVA promoted a strong OVA-specific Th2-type antibody
response (Figure 6A) and predisposed mice to OVA-induced
Th2-inflammation (Figure 6B-D). To support this observation,
we examined serum levels of antigen-specific IgG1against pea
seed proteins (pea globulins, lectin, and vicilin-4) in transgenic
pea and nontransgenic pea-fed mice. Interestingly, levels of
antigen-specific IgG1against pea globulins, lectin, and vicilin-4
in serum of transgenic pea fed mice were significantly higher
than those of nontransgenic pea-fed mice, suggesting a height-
ened immune responsiveness to dietary proteins due to pea-
RAI (Figure 7). Thus, these studies demonstrate that modified
RAI possesses immunodulatory activity and that consumption
Figure 5. Consumption of transgenic pea seed meal predisposed to CD4+
T-cell derived Th2-type cytokine production. IL-4 (A), IL-5 (B), and IFNγ
(C) levels in supernatants from RCD3/RCD28 or pea-RAI or media alone
stimulated PBLN cells from nontransgenic and transgenic pea-fed mice
i.t. challenged with RAI from pea. Data are expressed as the mean ±
SEM from 3 to 6 mice per group from duplicate experiments. Statistical
significance of differences (p < 0.05) was determined using Student’s
unpaired t-test.
Figure 6. Intra-gastric administration of RAI from pea induces cross-
priming of heterogeneous food antigens. OVA-specific IgG1levels (A)
and the Th2-inflammation phenotype (mucus hypersecretion) (B), pulmo-
nary eosinophilia (C), and airways hyperreactivity (D) in mice that were
fed (i.g. challenged) ovalbumin (OVA) alone (the control) or in combination
with natively expressed Tendergreen bean-RAI or transgenically expressed
(pea) RAI and subsequently intra-tracheal challenged with purified OVA.
Data are expressed as the mean ± SEM from 4 to 6 mice per group. *
p < 0.05 as compared to OVA and bean RAI/OVA.
Figure 7. RAI from pea induces cross-priming of pea proteins. Pea
globulin-, vicilin-4, and lectin-specific IgG1levels in serum from mice that
were intragastrically administered 250 µL (∼100 mg/mL) of either
nontransgenic or transgenic pea seed meal twice a week for 4 weeks.
Data are expressed as mean ± SEM from 4 to 5 mice per group. * p <
0.05 as compared to nontransgenic pea.
9028
J. Agric. Food Chem., Vol. 53, No. 23, 2005 Prescott et al.

Page 7

of the modified RAI concurrently with heterogeneous proteins
can promote immunological cross priming, which predisposes
to specific immunoreactivity to these proteins.
DISCUSSION
Recently, peas expressing a gene for RAI from the common
bean were generated for protection against field and storage pests
(3-6). Characterization of RAI by structural analysis has
demonstrated that transgenic expression of this protein in peas
led to the synthesis of a modified form of RAI. Further, we
show that the modified form of RAI possessed altered antigenic
properties and consumption of this protein by mice predisposed
to RAI-specific CD4+Th2-type inflammation and elicited
immunoreactivity to concurrently consumed heterogeneous food
antigens.
Bean-RAI undergoes significant post-translational modifica-
tion including variable glycosylation and proteolytic processing
leading to the synthesis of a mature functional protein (8, 11).
We demonstrate that differences in glycosylation and/or other
modifications of the pea-RAI lead to altered antigenicity.
Consistent with our observations, investigators have previously
demonstrated that differential glycosylation of subunits of a
cereal R-amylase-inhibitor family (unrelated to legume RAIs)
enhances IgE-binding capacity (24). Moreover, glycosylated
cereal RAI subunits have been shown to possess significantly
enhanced IgE-binding affinity when compared to the unglyco-
sylated forms (24). These cereal proteins possess identical amino
acid sequences and only differ in their carbohydrate moieties,
indicating that glycosylation can confer IgE-binding capacity
and Th2-inflammation. In particular, recent investigations have
demonstrated that glycan side chains linked to high mannose-
type N-glycans on plant-derived glycoproteins can confer
immunogenicity and are IgE binding determinants (25, 26).
Moreover, R(1,3)-fucose and ?(1,2)-xylose linkage to high
mannose-type N-glycans (Man5GlcNAc2-Man9GlcNAc2) pro-
mote immunogenicity and IgE binding. The ?-chain of pea-
RAI possesses ?(1,2)-xylose linked high mannose-type N-gly-
cans, and other complex glycoforms and the R-chain may
possess an as yet undefined glycoform variant, and it remains
to be determined how these modifications alter pea-RAI
immunogenicity.
Functional and structural properties of pea-RAI may con-
tribute to its ability to circumvent immune tolerance and elicit
inflammatory responses. Bean-RAI is a potent inhibitor of
human R-amylase activity and can induce gastrointestinal
dysfunction (27). Comparison of bean- and pea-derived RAI
activity revealed no difference in enzymatic activity between
the two proteins (results not shown). Furthermore, we examined
the gastrointestinal tract of pea and transgenic pea-fed mice and
observed no histological abnormalities to the gastrointestinal
tissue in either group (results not shown). Bean-RAI is also a
heat-stable protein and partially resistant to proteolytic degrada-
tion (28, 29). Extensive boiling (100 °C for 20 min), while
significantly reducing R-amylase inhibitory activity, failed to
alter the ability of the transgenic pea to prime for Th2-
inflammation when challenged in the lung [results not shown:
see Supporting Information Figure 2]. These findings are
consistent with previous demonstrations that cooking of plant
material such as lentils and peanuts does not diminish the
allergenic potential of certain proteins (30, 31). Furthermore,
these studies suggest that the altered immunogenicity of RAI is
unrelated to its properties as an amylase inhibitor.
We demonstrate that the immune response elicited against
pea-RAI following oral consumption of transgenic pea is
characterized by CD4+Th2 cell-mediated inflammation, in
particular, the presence of IL-4 and IL-5. To examine whether
the immune response was dependent on IL-5 and eosinophils,
we employed IL-5 and eotaxin-deficient mice. IL-5/eotaxin-
deficient mice were i.g. administered nontransgenic and trans-
genic seed meal and subsequently i.t. challenged with purified
RAI. We show that i.t. challenge of transgenic pea fed IL-5/
eotaxin-deficient mice induced Th2-inflammation that was
significantly elevated over nontransgenic fed mice (32). These
investigations suggest that the immune response elicited against
pea-RAI following oral consumption of transgenic pea is not
dependent on IL-5 and eosinophils.
In this study, we have demonstrated that transgenic expression
of RAI in a pea can lead to the synthesis of a modified form of
the protein with altered antigenic properties. Furthermore, we
show that concomitant exposure of the gastrointestinal tract to
modified RAI and heterogeneous food antigens cross primes
and elicits immunogenicity. Currently, we do not know the
frequency at which alterations in structure and immunogenicity
of transgenically expressed proteins occur or whether this is
unique to transgenically expressed RAI. These investigations,
however, demonstrate that transgenic expression of non-native
proteins in plants may lead to the synthesis of structural variants
with altered immunogenicity.
ABBREVIATIONS USED
RAI, R-amylase inhibitor-1; pea (Pisum satiVum L.), trans-
genic pea; Phaseolus Vulgaris L. cv. Tendergreen, Pisum
satiVum L. expressing R-amylase inhibitor-1 from the common
bean; MALDI-TOF-MS, matrix-assisted laser desorption/ioniza-
tion-time-of-flight-mass spectrometry.
ACKNOWLEDGMENT
We thank Aulikki Koskinen and Anne Prins for excellent
technical assistance and David Tremethick, Ian Young, and
Klaus Matthaei for their helpful discussions and preparation of
the manuscript. GenBank accession number for common bean
cv. Pinto is AY603476.
Supporting Information Available: Amino acid sequence
of RA1 from common bean and consumption of pea seed meal
predisposed to Th2-type inflammation. This material is available
free of charge via the Internet at http://pubs.acs.org.
LITERATURE CITED
(1) Daniell, H.; Streatfield, S. J.; Wycoff, K. Medical molecular
farming: production of antibodies, biophmaceuticals and edible
vaccines in plants. Trends Plant Sci. 2001, 6, 219-226.
(2) Brandt, P. Overview of the current status of genetically modified
plants in Europe as compared to the USA. J. Plant. Physiol.
2003, 160, 735-742.
(3) Schroeder, H. E.; Gollasch, S.; Moore, A. E.; Tabe, L. M.; Craig,
S.; Hardie, D. C.; Chrispeels, M. J.; Spencer, D.; Higgins, T. J.
V. Bean R-amylase inhibitor confers resistance to the pea weevil
(Bruchus pisorum) in transgenic peas (Pisum satiVum L.). Plant
Physiol. 1995, 107, 1233-1239.
(4) Ishimoto, M. Bruchid resistance of transgenic azuki bean
expressing seed R-amylase inhibitor of common bean. Entomol.
Exp. Appl. 1996, 309-315.
(5) Shade, R. E.; Schroeder, H. E.; Pueyo, J. J.; Tabe, L. M.;
Murdock, L. L.; Higgins, T. J. V.; Chrispeels, M. J. Transgenic
pea seeds expressing the R-amylase inhibitor of the common
bean are resistant to bruchid beetles. Biotechnology 1994, 12,
793-796.
Bean R-Amylase Inhibitor and ImmunogenicityJ. Agric. Food Chem., Vol. 53, No. 23, 2005
9029

Page 8

(6) Morton, R. L.; Schroeder, H. E.; Bateman, K. S.; Chrispeels,
M. J.; Armstrong, E.; Higgins, T. J. V. Bean R-amylase inhibitor
1 in transgenic peas (Pisum satiVum) provides complete protec-
tion from pea weevil (Bruchus pisorum) under field conditions.
Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 3820-3825.
(7) Molvig, L.; Tabe, L. M.; Eggum, B. O.; Moore, A. E.; Carig,
S.; Spencer, D.; Higgins, T. J. V. Enhanced methionine levels
and increase nutritive value of seeds of transgenic lupins (Lupins
angustifolius L.) expressing a sunflower seed albumin gene. Proc.
Natl. Acad. Sci. U.S.A. 1997, 94, 8393-8398.
(8) Moreno, J.; Chrispeels, M. J. A lectin gene encodes the
R-amylase inhibitor of the common bean. Proc. Natl. Acad. Sci.
U.S.A. 1989, 86, 7885-7889.
(9) Smart, V.; Foster, P. S.; Rothenberg, M. E.; Higgins, T. J.;
Hogan, S. P. A plant-based allergy vaccine suppresses experi-
mental asthma via an IFN-gamma and CD4+CD45RBlow T cell-
dependent mechanism. J. Immunol. 2003, 171, 2116-2126.
(10) Martin Young, N.; Thibault, P.; Watson, D. C.; Chrispeels, M.
J. Post-translational processing of two R-amylase inhibitors and
an arcelin from the common bean, Phaseolus Vulgaris. FEBS
Lett. 1999, 446, 203-206.
(11) Yamaguchi, H.; Funaoka, H.; Iwamoto, H. Structures of sugar
chains of the subunits of an alpha-amylase inhibitor from
Phaseolus Vulgaris white kidney beans. J. Biochem. (Tokyo)
1992, 111, 388-395.
(12) Sawada, S.; Takeda, Y.; Tashiro, M. primary structures of alpha-
and beta-subunits of alpha-amylase inhibitors from seeds of three
cultivars of phaseolus beans. J. Protein Chem. 2002, 21, 9-17.
(13) Ishimoto, M.; Suzuki, K.; Iwanaga, M.; Kikuchi, F.; Kitamura,
I. Variation of seed R-amylase inhibitors in the common bean.
Theor. Appl. Genet. 1995, 90, 425-429.
(14) Hoffman, L. M.; Ma, Y.; Barker, R. F. Molecular cloning of
Phaseolus vulgaris lectin mRNA and use of cDNA as a probe
to estimate lectin transcript levels in various tissues. Nucleic
Acids Res. 1982, 10, 7819-7828.
(15) Mowat, A. M. I. Anatomical basis of tolerance and immunity to
intestinal anitgens. Nat. ReV. Immunol. 2003, 3, 331-341.
(16) Michael, J. G. The role of digestive enzymes in orally induced
immune tolerance. Immunol. InVest. 1989, 18, 1049-1054.
(17) Pusztai, A.; G., G.; Stewart, J. C. A new type of Phaseolus
Vulgaris (cv. Pinto III) seed lectin: isolation and characterization.
Biochim. Biophys. Acta 1981, 671, 146-154.
(18) Banwell, J. G.; Abramowsky, C. R.; Weber, F.; Howard, R.;
Boldt, D. H. Phytohemagglutinin-induced diarrheal disease. Dig.
Dis. Sci. 1984, 29, 921-929.
(19) Wills-Karp, M. Murine models of asthma in understanding
immune dysregulation in human asthma. Immunopharmacology
2000, 48, 263-268.
(20) Wills-Karp, M. Immunologic basis of antigen-induced airway
hyperresponsiveness. Annu. ReV. Immunol. 1999, 17, 255-
281.
(21) Rajananthanan, P.; Attard, G. S.; Sheikh, N. A.; Morrow, W. J.
Novel aggregate structure adjuvants modulate lymphocyte
proliferation and Th1 and Th2 cytokine profiles in ovalbumin
immunized mice. Vaccine 1999, 18, 140-152.
(22) Rajananthanan, P.; Attard, G. S.; Sheikh, N. A.; Morrow, W. J.
Evaluation of novel aggregate structures as adjuvants: composi-
tion, toxicity studies and humoral responses. Vaccine 1999, 17,
715-730.
(23) Morrow, W. J. W.; Wang, Y. W.; Sheikh, N. A. Immunobiology
of the tomatine adjuvant. Vaccine 2004, 22, 2380-2384.
(24) Sanchez-Monge, R.; Gomez, L.; Barber, D.; Lopez-OTIN, C.;
Armentia, A.; Salcedo, G. Wheat and barley allergens associated
with bakers asthma. Biochem. J. 1992, 281, 401-405.
(25) Faye, L.; Gomord, V.; Fitchette-Laine, A. C.; Chrispeels, M. J.
Affinity purification of antibodies specific for Asn-linked glycans
containing a1-3 fucose or b1-2 xylose. Anal. Biochem. 1993,
209, 104-108.
(26) Van Ree, R.; Cabanes-Macheteau, M.; Akkerdaas, J.; Milazzo,
J.; Loutelier-Bourhis, C.; Rayon, C.; Villalba, M.; Koppelman,
S.; Aalberse, R.; Rodriguez, R.; Faye, L.; Lerouge, P. b(1,2)-
xylose and a(1,3)-fucose residues have a string contribution in
IgE binding to plant glycoallergens. J. Biol. Chem. 2000, 275,
11451-11458.
(27) Layer, P.; Carlson, G. L.; DiMagno, E. P. Partially purified
amylase inhibitor reduces starch digestion in vitro and inactivates
intraduodenal amylase in humans. Gastoenterology 1985, 88,
1895-1902.
(28) Marshall, J. J.; Lauda, C. M. Purification and properties of
phaseolamin, an inhibitor of R-amylase, from the kidney bean,
Phaseolus Vulgaris. J. Biol. Chem. 1975, 250, 8030-8037.
(29) Yoshikawa, H.; Kotaru, M.; Tanaka, C.; Ikeuchi, T.; Kawabata,
M. Characterization of Kintoki bean (Phaseolus Vulgaris) alpha-
amylase inhibitor: inhibitory activities against human salivary
and procine pancreatic alpha-amylases and activity changes by
prteolytic digestion. J. Nutr. Sci. Vitaminol. 1999, 45, 797-802.
(30) Ibanez Sandin, D.; MArtinez San Ireneo, M.; Maranon Lizana,
F.; Fernandez-Caldas, E.; Alonso Lebrero, E.; Laso Borrego, T.
Specific IgE determinations to crude and boiled lentil (Lens
culinaris) extracts in lentil-sensitive children and controls. Allergy
1999, 54, 1209-1214.
(31) Beyer, K.; Morrow, E.; Li, X.; Bardina, L.; Bannon, G. A.; Burks,
A. W.; Sampson, H. Effects of cooking methods on peanut
allergenicity. J. Allergy Clin. Immunol. 2001, 107, 1077-1081.
(32) Prescott, V.; Foster, P. S.; Matthaei, K. I.; Hogan, S. P.
Mechanistic analysis of food allergy-induced skin and pulmonary
inflammation. J. Allergy Clin. Immunol. 2005, 115(2 Suppl), S1-
360.
Received for review March 16, 2005. Revised manuscript received
August 26, 2005. Accepted September 6, 2005. This work was supported
in part by National Health Medical Research Council (Australia)
Program Grant 224207.
JF050594V
9030
J. Agric. Food Chem., Vol. 53, No. 23, 2005Prescott et al.