Accepted Manuscript Title: Label free targeted detection and quantification of celiac disease immunogenic epitopes by mass spectrometry Author: Hetty C. van den Broeck Jan H.G. Cordewener Merel A. Nessen Antoine H.P. America Ingrid M. van der Meer PII: DOI: Reference:
S0021-9673(15)00343-X http://dx.doi.org/doi:10.1016/j.chroma.2015.02.070 CHROMA 356328
To appear in:
Journal of Chromatography A
Received date: Revised date: Accepted date:
27-11-2014 23-2-2015 26-2-2015
Please cite this article as: H.C. van den Broeck, J.H.G. Cordewener, M.A. Nessen, A.H.P. America, I.M. van der Meer, Label free targeted detection and quantification of celiac disease immunogenic epitopes by mass spectrometry, Journal of Chromatography A (2015), http://dx.doi.org/10.1016/j.chroma.2015.02.070 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
1
Label free targeted detection and quantification of
2
celiac
3
spectrometry
disease
immunogenic
epitopes
by
mass
ip t
4 Hetty C. van den Broecka*, Jan H. G. Cordewenera, Merel A. Nessenb, Antoine H.
6
P. Americaa, Ingrid M. van der Meera
cr
5
8
a
9
b
us
7
Wageningen UR, Plant Research International, PO Box 16, 6700 AA Wageningen, The Netherlands RIKILT Wageningen UR, PO Box 230, 6700 AE Wageningen, The Netherlands
11
*
an
10
Corresponding author: [email protected]; Tel: +31 317 480974; Fax: +31 317 41809
13
M
12 Abstract
ed
14
Celiac disease (CD) is a food-related disease caused by certain gluten peptides
16
containing T-cell stimulating epitopes from wheat, rye, and barley. CD-patients
17
have to maintain a gluten-free diet and are therefore dependent on reliable
18
testing and labeling of gluten-free products. So far, the R5-ELISA is the approved
19
method to detect if food products can be labeled gluten-free. Because the R5-
20
ELISA detects gluten in general, there is a demand for an improved detection
21
method that quantifies specifically CD-epitopes. Therefore, we developed a new
22
method for detection and quantification of CD-epitopes, based on liquid
23
chromatography (LC) coupled to mass spectrometry (MS) in multiple reaction
24
monitoring (MRM) mode. This method enables targeted label free comparative
25
analysis of the gluten proteins present in different wheat varieties and species,
26
and in wheat-based food products. We have tested our method by analyzing
Ac ce
pt
15
1 Page 1 of 49
several wheat varieties that vary in CD-epitope content, as was shown before
28
using immunoblotting and specific monoclonal antibodies. The results showed
29
that a modern bread wheat variety Toronto contained the highest amounts of CD
30
immunogenic peptides compared with the older bread wheat variety Minaret and
31
the tetraploid wheat variety Dibillik Sinde. Our developed method can detect
32
quantitatively and simultaneously multiple specific CD-epitopes in a high
33
throughput manner.
cr
ip t
27
Keywords:
36
-Gliadin, celiac disease, LC-MRM/MS,
T-cell stimulatory epitopes, wheat
Ac ce
pt
ed
M
an
35
us
34
2 Page 2 of 49
37
38
1. Introduction Celiac disease (CD) is a food related disease that results in inflammation of
41
the small intestinal mucosa in genetically predisposed individuals caused by
42
intake of gluten proteins from wheat, rye, and barley [1]. The prevalence of
43
celiac disease is about 0.5 – 2% and is still increasing in Western parts of the
44
world as well as in developing countries [2,3]. CD-patients can present
45
symptoms such as diarrhea, abdominal pain, constipation, weight loss, and
46
dermatitis
47
asymptomatic presentation without gastrointestinal symptoms resulting in a high
48
number of undiagnosed cases [4][2,3]5-7]. For CD-patients still the only cure is
49
to adhere to a strict gluten-free diet. However, gluten proteins are increasingly
50
being applied in all kind of food products because of their interesting features
51
[8]. Therefore, it is very important that food labeling is accurate and reliable. The
52
reduced intake of gluten peptides and proteins containing CD-epitopes by all
53
consumers, including diagnosed and still undiagnosed CD-patients, will reduce
54
symptoms and may reduce the prevalence of CD.
however,
increasingly
more
patients
show
Ac ce
pt
ed
M
an
herpetiformis,
us
cr
ip t
39 40
55
Gluten proteins, i.e. gliadins and glutenins, represent the main part of
56
storage proteins in wheat, are insoluble in water and contain high percentages of
57
the amino acids proline (P) and glutamine (Q). Gliadins form a large protein
58
family in which /-, -, and -gliadins can be distinguished (~30 to ~80 kDa by
59
acidic-PAGE) [9,10]. Glutenins can be subdivided into low-molecular weight
60
glutenin subunits (LMW-GS; ~30 to ~70 kDa by SDS-PAGE) and high-molecular
61
weight
62
[11,12].[13][14]
glutenin
subunits
(HMW-GS;
~80
to
~130
kDa
by
SDS-PAGE)
3 Page 3 of 49
Bread wheat is hexaploid and contains three different genomes (A, B and D)
64
evolved from three different grass species. Therefore, many gluten protein
65
encoding genes are present and copy numbers for -gliadins can range between
66
100 and 150 [15,16]. [17,18]CD symptoms can be caused by many different CD-
67
epitopes present in wheat cultivars and wheat-derived food products. So far, 24
68
epitopes have been defined from wheat gluten proteins that give a T-cell
69
response in CD-patients [19]. These glutamine and proline rich epitopes are
70
derived from all different classes of gliadin and glutenin proteins. The
71
immunogenic epitopes present in -gliadins are Glia-2, Glia-9, and Glia-20 of
72
which the most immunogenic epitopes are Glia-2 and Glia-9 [20,21]. It has
73
been shown by Shan et al. [22] that a large 33-mer from wheat gluten is
74
resistant against human intestinal proteases and therefore can be present in the
75
small intestine. The 33-mer is composed of overlapping immunodominant Glia-2
76
and Glia-9 epitopes that can also be recognized separately and stimulate T-cells
77
in CD [20,21]. This 33-mer can be present in -gliadins that are encoded by the
78
Gli-2 locus (Gli-D2) on the short arms of chromosome 6 of the D-genome in
79
wheat (see Fig. 1) [23-25]. The same holds for the Glia-2 epitope that is
80
present in most -gliadins encoded by genes present at only the D-genome in
81
hexaploid wheat. The -gliadins encoded by the D-genome can also contain
82
derivatives of the 33-mer (a 19-mer and a 26-mer) that contain the Glia-2 and
83
Glia-9 epitopes. The -gliadins encoded by the A-genome in hexaploid and
84
tetraploid wheat contain only a 13-mer containing the Glia-9 epitope. Most -
85
gliadins encoded by both the A- and the D-genome contain the Glia-20 epitope.
86
The -gliadins encoded by the B-genome contain neither the Glia-2/9 epitopes
87
nor the Glia-20 epitope [24,26].
Ac ce
pt
ed
M
an
us
cr
ip t
63
4 Page 4 of 49
Food products can be labeled gluten-free if they contain less than 20 ppm
89
gluten. So far, the only approved test by the Codex Alimentarius to detect the
90
presence of gluten in food products and that can be used to label products
91
‘gluten-free’ is the R5-enzyme-linked immunosorbent assay (R5-ELISA) [27-29].
92
In general, the functionality of ELISA tests depends on the extraction protocol
93
[30-32], reference material for calibration [33], and the detecting antibody. [30-
94
32][33]The latter is one of the limitations of the R5-ELISA that makes use of a
95
monoclonal antibody that detects gluten in general and not specifically CD
96
stimulatory epitopes. The R5-antibody is a-specific and recognizes several, five
97
amino acids long, sequences present in gluten proteins from wheat, rye, and
98
barley, but not from soy, oats, corn, rice, millet, teff, buckwheat, quinoa, and
99
amaranth. This difference in specificity is essential for gluten-free testing.
100
Despite this ability, the short length of the recognition sequence might lead to
101
inaccurate results. Antibodies can recognize sequences as short as five amino
102
acids. This in contrast to the size of the recognition sequence of a CD-epitope
103
recognized by human T-cells in CD, which is at least nine amino acids in length
104
[34]. Therefore, R5-ELISA might result in an overestimation of the amount of
105
true T-cell epitopes present that stimulate CD.
106
[19][20,21][22][20,21][23-25][24,26] Simultaneous detection and quantification
107
of more than one CD-epitope is not possible by immunoblotting using a single
108
blot. By ELISA this is only possible when a multiplexing method with multiple
109
specific
110
chromatography-multiple reaction monitoring mass spectrometry (LC-MRM/MS)
111
method that can detect and quantify multiple CD-peptides in a single short run
112
will have a huge advantage over the existing methods.
Ac ce
pt
ed
M
an
us
cr
ip t
88
antibodies
is
used.
Therefore,
the
development
of
a
liquid
5 Page 5 of 49
LC-MRM/MS is a rapidly emerging method as an alternative to antibody based
114
protein quantification [35-37]. In this approach, the proteins to be quantified are
115
first digested with a specific protease, after which proteotypic peptides are
116
analyzed by MS in MRM mode. Using either a triple quadrupole or a quadrupole-
117
ion trap instrument, peptides are identified and quantified by monitoring several
118
transitions for each peptide. MRM allows sensitive, accurate and reproducible
119
quantification of the peptides and corresponding proteins.So far, several LC-MS
120
detection methods, aiming at the detection of different immunogenic peptides,
121
have been developed and described [38-43]. The developed methods differ in
122
the choice of the selected peptides for quantification and the protease treatments
123
to release the peptides from the gluten proteins.
an
us
cr
ip t
113
Here, we describe the development of an LC-MRM/MS method to quantify
125
individual immunogenic peptides in a gluten protein extract from wheat kernels,
126
using chymotrypsin digestion to release the peptides. We focused on detection of
127
those peptides that have been proven to be the most immunodominant in CD
128
[44-48]. This is in contrast to previously developed LC-MS methods, that focused
129
on the most intensely MS responding peptides. Besides immunodominant
130
peptides, we also incorporated some peptides with amino acid substitutions that
131
make the immunogenic epitopes inactive and are thereby considered safe for
132
CD-patients [24,26]. LC-MS allows detection of these amino acid substitutions,
133
which is not possible by antibody detection as used in ELISA. In addition, we
134
selected peptides that can be used for quantification of the total gliadin protein
135
amount. Furthermore, we show that with this method, wheat varieties,
136
containing different amounts of the epitopes, can be analyzed. This opens the
137
possibilities for a range of future applications such as support for the
138
development of reduced immunogenic wheat lines, the analysis of foods for the
Ac ce
pt
ed
M
124
6 Page 6 of 49
presence of immunogenic CD-epitopes, or the screening of body fluids of
140
allergenic patients for analysis of the presence of selected marker peptides.
Ac ce
pt
ed
M
an
us
cr
ip t
139
7 Page 7 of 49
141
142
2. Materials and methods
143
2.1. Wheat samples
145
ip t
144 Two hexaploid wheat varieties, Minaret (obtained from the Centre for Genetic Resources, CGN, the Netherlands) and Toronto (obtained from Limagrain,
147
Lelystad, The Netherlands), and one tetraploid wheat variety (Dibillik sinde,
148
obtained from CGN) were used for gluten protein extraction followed by LC-MS
149
analysis.
an
us
cr
146
150 2.2. Extraction of gluten proteins
M
151 152
Gluten proteins were extracted from wheat grains according to Van den
154
Broeck et al. [49]. Grains were ground and gluten proteins were extracted from
155
50 mg wheat flour by addition of 0.5 ml of 50% (v/v) aqueous iso-propanol with
156
continuous mixing (MS1 Minishaker, IKA Works, Inc.) at 1000 rpm for 30 min at
157
room temperature, followed by centrifugation at 10,000 rpm for 10 min at room
158
temperature. The residue was re-extracted twice with 50% (v/v) aqueous iso-
159
propanol, 50mM Tris-HCl, pH 7.5 containing 1% (w/v) DTT, for 30 min at 60°C
160
with mixing every 5 to 10 min followed by centrifugation at 10,000 rpm for 10
161
min at room temperature. After addition of each next extraction solution, the
162
residue was resuspended followed by sonication for 10 min in an ultrasonic bath
163
(Branson
164
supernatants were combined and considered the gluten protein extract. The
165
protein content was quantified using the Biorad Protein Assay (Bio-Rad
Ac ce
pt
ed
153
3510,
Branson
Ultrasonics
Corporation).
The
three
obtained
8 Page 8 of 49
166
Laboratories), based on the Bradford dye-binding procedure, according to
167
manufacturer's instruction with BSA as a standard.
168 169
2.3. Digestion conditions
ip t
170 Gluten proteins from the gluten extracts were precipitated with four volumes
172
of ice cold 100% acetone and kept overnight at -20C. After centrifugation for 10
173
min at 14,000 rpm and 4C, the resulting pellet was resuspended in 8 M urea to
174
keep the proteins denaturated. Proteins were reduced in 5 mM DTT for 45 min at
175
60C, followed by alkylation in 15 mM iodoacetamide for 60 min at room
176
temperature in the dark. Aliquots containing 10 g of protein were four fold
177
diluted with 100 mM Tris-HCl (pH 8.0) /10mM CaCl2, to a final concentration of 2
178
M urea. Different experiments were performed to obtain optimal conditions for
179
release of the selected peptides. Proteins were digested with chymotrypsin from
180
bovine pancreas (Roche) and shaken at 25C using three different ratios of
181
protease:protein (1:5, 1:10 and 1:20). The digestion time was varied from 10
182
min up to 22 hours. The proteolytic activity was quenched by addition of
183
trifluoroacetic acid (TFA; 0.5% final concentration and the acidified sample was
184
desalted using solid phase extraction (SPE). Supelco LC-18 1 ml SPE columns
185
were washed with 1 ml 95% acetonitrile (ACN) and equilibrated with 1 ml 2%
186
ACN, 0.1% TFA. Subsequently, the digests were added to the columns and after
187
washing with 1 ml 2% ACN, 0.1% TFA, peptides were eluted with 1 ml 50% ACN,
188
0.1% TFA. Solvents were evaporated using a SpeedVac vacuum concentrator and
189
peptides were dissolved in either 20 µl 0.1 M ammonium formate (AF) or 40 µl
190
5% ACN, 0.1% formic acid (FA).
Ac ce
pt
ed
M
an
us
cr
171
9 Page 9 of 49
191
The formation of peptides P1-P9 was finally studied after both 4 and 22 hours of
192
digestion with chymotrypsin in a 1:5 enzyme to protein ratio.
193
2.4. Marker peptides and preparation of peptide reference solution
194 Nine peptides from gluten proteins (P1-P9) were synthesized and purchased
196
from ProteoGenix (France). An overview of the sequences of the peptides is
197
given in Table 2. For each of the peptides, a stock solution of 10 mM (based on
198
the yield and corrected for purity percentage indicated by the supplier) was
199
prepared in 100% MeOH and stored in aliquots at -20C. The individual stock
200
solutions were mixed to obtain equimolar cocktails of the peptides that were
201
further diluted with MilliQ water to a final concentration of 20 µM in 10% MeOH.
202
Calibration standards were prepared by dilution of the 20 µM cocktail solution in
203
a matrix of tryptic digest of a gluten extract of hexaploid wheat variety Toronto
204
(25 ng/µl 5% ACN, 0.1% FA) or a tryptic digest of bovine serum albumin (Sigma
205
Aldrich) (25 ng/µl 5% ACN, 0.1% FA) .
ed
M
an
us
cr
ip t
195
Calibration standard curves were measured using 20, 50, 100, 200, 500,
207
1000, 2000, 5000, and 10000 fmol per peptide loaded on the reversed phase
208
(RP) TS3 1.0 mm × 150 mm analytical RP column (Waters, Milford, MA, USA).
210 211
Ac ce
209
pt
206
2.5. On-line 2D nano LC-MS/MS configuration for untargeted approach
212
Nanoscale LC separation of complex peptide mixtures was performed using
213
the 2-D nanoAcquity UPLC system on-line coupled to a Synapt HDMS Q-TOF MS
214
instrument (Waters, Milford, MA, USA) as described [50,51]. For each run, 4 μL
215
digest was injected (partial loop method) in 20 mM AF (pH 10) on the first RP
216
column (X-bridge BEH130 C18, 5 μm, 300 μm × 50 mm, Waters). Elution from
10 Page 10 of 49
this column was stepwise under high pH and ultrahigh pressure at 2 μL/min.
218
Samples were eluted in four fractions of 13, 18, 25, and 65% ACN in 20 mM AF
219
(pH 10). Peptide eluates were on-line diluted with an excess of 0.1% FA in water
220
at a flow rate of 20 μL/min to reach a ten-fold dilution before being trapped on a
221
C18 (2G-V/M Symmetry 5 μm) trap column (180 μm × 20 mm). Finally, peptides
222
were
223
75 μm × 150 mm analytical RP column (Waters) at 0.3 μL/min with a gradient
224
using 0.1% FA in water as eluent A and 0.1% FA in ACN as eluent B. The
225
separation was carried out using 5% B for 1 min, 10% B for 2 min, 10–40% B
226
over 62 min and 40–85% B over 9 min. After 6 min of rinsing with 85% B and a
227
linear gradient back to 5% B over 2 min, the column was re-equilibrated at initial
228
conditions. Both the Symmetry trap column temperature and the analytical
229
column temperature were maintained at 55 °C by a built-in column heater. Mass
230
spectrometric analyses were performed in positive mode using ESI with a
231
NanoLockSpray source. Eluates were immediately sprayed into a Q-TOF Synapt
232
G1 (Waters). As lock mass, [Glu1]-fibrinopeptide B (1 pmol/L) was delivered
233
from a syringe pump (Harvard Apparatus, USA) to the reference sprayer of the
234
NanoLockSpray source at a flow rate of 0.2 μL/min. The lock mass channel was
235
sampled every 30 sec. LC-MS data were collected from the Synapt G1 operating
236
in either MS/MS or MSE mode for data-dependent acquisition (DDA) or data-
237
independent acquisition (DIA). For MS/MS, the three most intensive single or
238
multiple charged ions eluting from the column were selected for fragmentation.
239
The eluting peptide ions were detected in the MS survey scan (0.6 sec) from a
240
m/z range of 300 to 1400 and MS/MS scan range from 50 to 2000 m/z. A
241
dynamic exclusion window was set at 60 sec.
from
this
column
and
separated
on
a
HSS
T3
1.7 μm,
Ac ce
pt
ed
M
an
us
cr
eluted
ip t
217
242 11 Page 11 of 49
243
2.6. Triple quadrupole settings for targeted analysis
244 245
Targeted LC-MRM/MS analyses were performed on an Acquity UPLC system
246
coupled to a Xevo TQ-S triple quad instrument (Waters, Milford, MA, USA). For
247
each run, 5 μL and 10 μL sample, corresponding to 250 and 500 ng protein
248
digest,
249
1.0 mm × 150 mm analytical RP column (Waters, Milford, MA, USA) at 60°C with
250
a flow rate of 0.15 mL/min. Peptides were eluted using an ten minute LC-method
251
(starting 1 min after injection): 1% B for 0.5 min, linear gradient of 1–42% B
252
over 5.5 min, linear gradient 42-90% B over 1 min, after 1 min of rinsing with
253
90% B, a linear gradient 90-20% for 1 min and back to 1% B over 1 min. The
254
column was re-equilibrated at initial conditions for 1.9 min, with 0.1% FA in
255
water (LC-MS grade) as eluent A and 0.1% FA in 100% ACN as eluent B.
(partial
loop
method)
and
separated
on
a
TS3
ip t
injected
M
an
us
cr
was
MRM mass spectrometric analyses were performed in positive ion mode. A
257
scheduled MRM method was used with MRM detection window set to 0.5-1.6 min
258
per peptide, the duty cycle was set to automatic and dwell times were minimal 5
259
msec. The transitions and detection windows of the marker peptides are listed in
260
Table 2. Cone voltage was set to 35 V.
262 263 264
pt
Ac ce
261
ed
256
2.7. Data analysis
2.7.1. Untargeted MS data analysis
265
DDA data obtained from the MS/MS analysis with the Synapt G1 were
266
processed by ProteinLynx Global Server software (PLGS version 2.5, Waters)
267
for peak detection. Produced peak lists were used for Mascot search (Matrix
268
Science) to identify peptides. Searches were performed using the Uniprot
12 Page 12 of 49
database (07/2013), green plants. Parameters were: Enzyme selection is
270
chymotrypsin, maximum of three missed cleavages; carbamidomethyl C as fixed
271
modification; deamidated NQ and oxidation M as variable modifications;
272
monoisotopic mass values; protein mass unrestricted; peptide mass tolerance
273
±50 ppm; fragment mass tolerance ± 30 mmu; peptide charges 2+, 3+, and 4+.
275
Raw MSE (DIA) data obtained were processed by Progenesis QI software to visualize chromatograms of peptide resolution per fraction.
cr
274
276 2.7.2. Targeted MS data analysis
us
277
ip t
269
To analyze the targeted MRM data from TQ-S, Skyline software (MacCoss Lab
279
at the University of Washington) was used [52]. The raw data were imported into
280
Skyline and peak detection and quantitation was performed on the given set of
281
transitions. Peak areas were exported as a transition result report and further
282
processed in Excel. Per injection the sum of the area of the multiple transitions
283
per peptide was used as quantitative value. In Skyline the ‘area’ value is the
284
peak
285
subtracted. Calculation of linear regression was performed on the summed area
286
for the dilution series of the reference peptides. We applied an intercept=0
287
setting, as it gave the highest R2 value (between 0.974 and 0.994). For the
288
gluten digests the peak areas were also summed per peptide. The peptide
289
amount was calculated using the slope coefficient obtained from the linear
290
regression of standard curves. Femtomole (fmol) levels were calculated to g
291
amount of injected gliadin protein, using the average Mw of gliadins (32285.5
292
Da).
above
the
background and
therefore
background is already
Ac ce
pt
area
ed
M
an
278
293
13 Page 13 of 49
294
2.8. SDS-PAGE and immunoblotting
295 Equal amounts of gluten proteins were separated on SDS-PAGE gels (10%)
297
as described [53] using a Hoefer SE 260 mighty small II system (GE Healthcare)
298
followed by staining with PageBlue (Fermentas). For immunoblotting, proteins
299
were blotted onto nitrocellulose (0.2 m, Bio-Rad Laboratories), omitting
300
methanol from the blotting buffer, using a Mini Trans-Blot Cell (Bio-Rad
301
Laboratories) at 100V for 1 h. B Blots were stained using a MemCode
302
Reversible Protein Stain Kit for Nitrocellulose Membranes (Fisher Scientific) prior
303
to incubation with monoclonal antibodies (mAbs). Blots were incubated as
304
described [54] using mAbs specific for T-cell stimulatory epitopes Glia-9 and
305
Glia-20 [55-57]. Antibody binding to the blots was visualized by staining for
306
alkaline
307
tetrazolium (NBT) and 5-bromo-4-chloro-3-indolyl phosphate (BCIP) (Sigma-
308
Aldrich). The gluten protein extract of ‘Toronto’ was used on each separate
309
immunoblot as an ‘inter-gel’ control.
an
us
cr
ip t
296
secondary
antibody,
using
Nitro
Blue
pt
ed
M
phosphatase-conjugated
Blots were scanned using a Bio-Rad GS-710 Calibrated Imaging Densitometer
311
(Bio-Rad Laboratories) and saved as TIFF images. Pixel intensities were
312
calculated per lane using Quantity One software (Bio-Rad Laboratories). Relative
313
intensities differed specifically per mAb used, but were normalized to values
314
obtained
Ac ce
310
for
the
‘inter-gel’
control.
14 Page 14 of 49
315
316
3. Results
317 3.1. Identification of gluten proteins in wheat grains by on-line 2D LC-MS/MS
ip t
318 319
In a previous genomics study, the genetic diversity present in -gliadin
321
transcripts and genes has been investigated showing that CD immunogenic
322
epitopes are encoded by the A- and D-genome and not by the B-genome. Fig. 1
323
shows the protein sequence variants of the most frequently found expressed
324
sequence tags (ESTs) originating from -gliadins in wheat, as present in the
325
NCBI-UniGene database. The amino acid sequences derived from the different
326
translated contig sequences have been aligned for the CD-epitope containing
327
region. Sequence variation outside this region is not displayed here. Sequence
328
variants are grouped per genome origin (Gli-A2, Gli-B2 and Gli-D2). The Glia-2
329
and Glia-9 epitopes are marked in yellow and the Glia-20 epitope is marked in
330
green. As mentioned before, these epitopes are encoded by both the A- and D-
331
genome, but not by the B-genome. The large protease resistant 33-mer,
332
containing multiple CD-epitopes, is marked in blue and the 19-mer and 26-mer
333
derivatives, encoded by only the D-genome, are marked in yellow. The -gliadins
334
encoded by the A-genome in hexaploid and tetraploid wheat contain only the 13-
335
mer peptide sequence containing the Glia-9 epitope.
Ac ce
pt
ed
M
an
us
cr
320
336
To study the -gliadins present in wheat varieties at protein level, a
337
previously developed two-step extraction protocol using isopropanol was applied
338
to efficiently extract the gluten proteins [49]. Total gluten concentrations in our
339
extracts were 1.02 g/l (30.45 g/mg flour), 0.72 g/l (21.45 g/mg flour), 15 Page 15 of 49
and 0.80 g/l (23.85 g/mg flour) for ‘Toronto’, Minaret’, and ‘Dibillik sinde’,
341
respectively. For LC-MS/MS analysis of the gluten fraction, proteins were first
342
digested with a specific protease in order to generate peptides with a length that
343
was amenable for proper LC separation and MS detection. Routinely, trypsin is
344
used for enzymatic digestion of complex protein mixtures, which cleaves proteins
345
C-terminal to the basic amino acids arginine (R) and lysine (K). However, due to
346
the low number of these amino acids in gluten proteins, the digestion of the
347
gluten proteins into smaller peptides by trypsin is not feasible. Therefore,
348
chymotrypsin from bovine pancreas was chosen for digestion of the gluten
349
proteins. Chymotrypsin cleaves C-terminal to phenylalanine (F), tyrosine (Y), and
350
tryptophan (W) unless the next residue is a proline (P), and at a lower rate C-
351
terminal to leucine (L), methionine (M), alanine (A), aspartic acid (D), and
352
glutamic acid (E). In Fig. 1, the chymotryptic cleavage sites present in the CD-
353
epitope containing domain of the different -gliadins are indicated in black (F, Y,
354
and W). From this result, we concluded that all -gliadin proteins encoded by the
355
A-genome will, theoretically, yield at least one peptide containing the Glia-9
356
epitope when digested with chymotrypsin (Fig. 1, indicated in yellow). This also
357
holds for -gliadin sequences encoded by the D-genome that will result in
358
different sized peptides (19-mer, 26-mer, and 33-mer) containing the Glia-2/9
359
epitopes (Fig. 1, indicated in yellow) or sequence variants thereof.
Ac ce
pt
ed
M
an
us
cr
ip t
340
360
The use of chymotrypsin as an optimal protease to generate CD-epitope
361
containing peptides was confirmed by digestion of the gluten protein extracts
362
from hexaploid wheat varieties Toronto and Minaret. The generated peptides
363
were separated into four fractions by 2D nano UPLC and on-line analyzed using
364
data-dependent MS/MS mode and data-independent MSE mode. Supplementary
365
Fig. S1 shows the resolution of the detected peptides present in the four different 16 Page 16 of 49
fractions. Supplementary Table S1 summarizes the different gliadin and glutenin
367
proteins identified in the gluten extract of wheat variety Toronto. Whereas the
368
majority of the identified peptides originated from -gliadins, also numerous
369
peptides from -gliadins and LMW-GS have been detected. The - and -gliadins
370
and the LMW-GS were classified into families according to the presence of at
371
least one unique identified (proteotypic) peptide in their sequence. In this way
372
we could discriminate between 24 different -gliadin families originating from
373
Triticum and Aegilops species (Table 1), 15 different -gliadin families, 7 different
374
LMW-GS families, and a single -gliadin and HMW-GS family.
us
cr
ip t
366
The peptide identifications covering the sequence area containing CD-
376
epitopes (displayed in bold) in -gliadins are displayed in Table 1. Different
377
sequence variants of the peptides have been identified and are listed as well.
378
From these identified peptides, we have selected the most frequently occurring
379
peptides as marker peptides for LC-MRM/MS detection, indicated P1 to P9.
ed
M
an
375
Peptides P1 to P4 occurred with peptide P6 in the -gliadin families, whereas
381
the -gliadin peptide containing the 33-mer (P5) was always present in
382
combination with the peptide containing the Glia-20 epitope having a P to S
383
substitution (P7) (Fig. 1; Table 1). Peptide P6 (in combination with P7)
384
containing the Glia-20 epitope occurred in nearly all -gliadin families encoded
385
by the A- and D-genome, but was absent in -gliadin families encoded by the B-
386
genome (Table 1, no. 8, 14, and 16). Peptides from the same region encoded by
387
the B-genome were more difficult to be identified because of the absence of
388
chymotrypsin digestion sites in this region. Instead, three different -gliadin
389
families from the B-genome have been identified (Table 1, no. 8, 14, and 16)
390
with proteotypic peptides outside this region of the gliadin sequence (see
391
Supplementary Table S1). Peptides P8 and P9 appeared in nearly all -gliadin
Ac ce
pt
380
17 Page 17 of 49
392
families, where P9 is the product of a missed cleavage and also contains peptide
393
P8. We reasoned that a targeted LC-MS detection method for the selected
395
peptides would be able to detect the presence of both the different CD-epitopes
396
individually (P1 to P7) and the sum of multiple gliadin proteins (P8 and P9).
ip t
394
397 3.2. Quantitative LC-MRM/MS assay development
cr
398 399
After identification of the different -gliadin peptides in wheat varieties
401
Toronto and Minaret by on-line 2D LC-MS/MS, a targeted MS-based assay was
402
developed for absolute quantification of the selected marker peptides P1-P9
403
(Table 1). In this study a triple quadrupole instrument (Xevo TQ-S) was used to
404
quantify CD-epitopes containing -gliadins in gluten extracts. Non-labelled
405
synthetic marker peptides P1-P9 were used both for the development of the MRM
406
assay and as standards for the generation of calibration curves for absolute
407
quantification of the corresponding -gliadins.Note that P9 has a missed cleavage
408
containing the P8 peptide sequence. Both peptides, P8 and P9, were present in
409
17 of the 24 -gliadin families (Table 1).
Ac ce
pt
ed
M
an
us
400
410
First, the optimal transitions for quantification and confirmation of each of
411
these peptides in a gluten digest were determined. Since the sensitivity of a
412
triple quadrupole MS is critically dependent on ionization conditions and tuning
413
parameters, a mix of all nine synthesized peptides was loaded on a RP column
414
and eluted with a short gradient to optimize cone voltage and collision energy of
415
the TQS. Table 2 summarizes the most optimal parameters achieved for each of
416
the marker peptides, and the transitions used for quantification and confirmation.
417
For most precursor peptides, one m/z value was chosen as the predominant 18 Page 18 of 49
charge state for selection in the first quadrupole. To obtain the most sensitive
419
assay, the collision energy was optimized per precursor to generate maximal
420
fragment ion intensities. In the third quadrupole, the most intense transitions
421
derived from one precursor ion were selected for quantification. In Table 2, the
422
peak retention time measured for each of the marker peptides is shown using the
423
LC conditions as described in section 2.6. (10 min gradient). For each peptide a
424
retention window of 0.5 to 2 min around its peak elution time was used for
425
acquisition in a time-scheduled MRM method. Scheduled MRM instructed the TQS
426
to measure a selected set of transitions during a particular retention time,
427
providing longer dwell times to monitor the target ions and thus improved signal
428
to noise ratios for most of the transitions. A prerequisite in a scheduled MRM
429
analysis is that the chromatography is stable and reproducible within the chosen
430
retention time window.
M
an
us
cr
ip t
418
Since most of the marker peptides contained overlapping sequences, several
432
fragment ions from different marker peptides had identical masses, as listed in
433
Table 2. The specificity for detection was achieved by using unique precursor to
434
product transitions and using scheduled MRM acquisitions around the elution time
435
of the precursor peptide. Supplementary Fig. S2 shows the chromatograms
436
obtained for a mixture of the nine marker peptides and for a ‘Toronto’ gluten
437
protein digestion.
439
pt
Ac ce
438
ed
431
3.3. Calibration curves of peptide standards
440 441
To determine the concentration range at which the nine marker peptides
442
showed a linear MRM response, calibration curves were generated by injection of
443
synthetic peptides in a range of 20-10000 fmol. Peak areas above background 19 Page 19 of 49
were determined by summing the multiple transitions per peptide using Skyline
445
software. Initially, a deviation from linearity was observed for most of the
446
peptides when the injected amount decreased below 100 fmol. Although different
447
glass vials and ‘low binding’ plastic tubes were tested to diminish the loss of
448
peptides when handling these low amounts of peptide, the calibration curves in
449
the lower concentration range remained quite variable. This problem could be
450
solved by preparing peptide dilutions in a matrix of a protein digest that did not
451
contain the marker peptides. When the dilutions of the peptides were prepared in
452
a matrix of tryptic digest from either bovine serum albumin (BSA) or gluten
453
protein, calibration curves showed excellent linearity in a range of four orders of
454
magnitude. The R square correlation values of the nine calibration curves varied
455
between 0.969 and 0.994, as listed in Table 2. In Supplementary Fig. S3, an
456
overview of the calibration curves for all the nine peptides is given.
M
an
us
cr
ip t
444
459
3.4. Digestion conditions
pt
458
ed
457
Prior to quantitative LC-MRM/MS analysis of the nine -gliadin peptides P1-P9
461
in protein digests from gluten extracts (see below), the reproducibility of the
462
digestion procedure with chymotrypsin was determined. Digestion was optimized
463
using gluten extracts from hexaploid wheat variety Toronto and tetraploid wheat
464
variety Dibillik sinde. In contrast to trypsin, which is mostly used in proteomics
465
analysis, chymotrypsin is not an endpoint protease. In addition to its primary
466
cleavage specificity it displays secondary, low affinity activity leading to further
467
breakdown of digested peptides. This means that an optimal digestion time and
468
enzyme to substrate ratio needs to be determined. Therefore, the progress of
469
chymotryptic digestion of gluten proteins dissolved in 2 M urea was followed by
Ac ce
460
20 Page 20 of 49
sampling at different time points of incubation at 25C. The ratio of chymotrypsin
471
to protein was also varied, i.e. 1:20, 1:10, and 1:5. Five peptides (P1-P5)
472
containing Glia-2/9 epitopes had reached a maximum MRM signal already at, or
473
even well-before, 4 h of incubation (Fig. 2A). Longer incubation did not result in
474
a significant decrease of the signals, and indicated that no or limited secondary
475
chymotrypsin activity to peptides P1-P5 took place (Fig. 2B). Peptides P6 and P7,
476
however, showed a much slower release and maximum MRM responses were
477
reached only after 22 h of incubation and an extra addition of chymotrypsin of an
478
enzyme to protein ratio of 1:5 (Fig. 2C and 2D). Finally, formation of all peptides
479
(P1-P9) was studied after 4 and 22 hours of incubation with chymotrypsin in a
480
1:5 enzyme to protein ratio. The results, shown in Supplementary Fig. S5,
481
confirmed the previous findings that P1-P5 reached a maximum response after
482
only 4 hours of digestion, while formation of P6 and P7 reached its maximum
483
after 22 hour of incubation. Also peptide P8, present in most -gliadin families,
484
showed a maximum response after 22 h. The slow formation of P8 is related to
485
the presence of a missed cleavage site, as is shown by the P9 peptide. The signal
486
of peptide P9 decreased over time, reaching a level just above background after
487
22 h when it was fully digested into P8. The reason why peptides P6, P7 (both
488
containing the Glia-20 epitope), and P8were released very slowly during
489
chymotrypsin digestion of -gliadins is not clear. In conclusion, these results
490
indicated that optimal digestion time to detect the CD-epitope containing
491
peptides is 22 hours and this was used for analysis of wheat varieties.
Ac ce
pt
ed
M
an
us
cr
ip t
470
492 493
3.5. Quantification of -gliadins in three wheat varieties by LC-MRM/MS
494
21 Page 21 of 49
After optimization of the digestion conditions, we analyzed protein extracts
496
from three wheat varieties with the developed MRM method: the modern wheat
497
variety Toronto (AABBDD), the landrace variety Minaret (AABBDD), and the
498
tetraploid variety Dibillik sinde (AABB). For LC-MRM/MS analysis two amounts
499
(250 ng and 500 ng) of digested gluten sample were injected in duplicate. Peak
500
areas (minus background) were determined by summing the selected transitions
501
per peptide using Skyline software. The total peak areas were converted to fmol
502
amounts of peptide, using the external calibration curves obtained from a dilution
503
series of the synthetic peptides. The fmol amount of peptide was converted to
504
the correlating amount of -gliadin per microgram digested gluten protein
505
extract using the average Mw of -gliadins (32285.5 Da). This is displayed in
506
Table 3 and Supplementary Fig. S4. Samples from two different digestion times
507
(4 h and 22 h) were analyzed and the 22 h sample was interpreted as completely
508
digested. In Table 3, an overview is given of the quantified peptides in the three
509
wheat varieties. In the tetraploid variety Dibillik sinde, the marker peptide P8
510
and the two peptides P1 and P6 containing the Glia-9 and Glia-20 CD-epitopes
511
were detected. These peptides, P1 and P6, are derived from -gliadins from the
512
A-genome. In the hexaploid varieties ‘Toronto’ and ‘Minaret’ all selected peptides
513
were detected (P9 was only detected in 4 h digests). The standard deviation of
514
the detected peptides was between 3% and 15% for most peptides, indicating a
515
good reproducibility of detection. The amounts of peptides P1 to P5 containing
516
the Glia-2/9 epitopes, could be related to the total ‘load’ of gliadin (P8+P9).
517
This allows comparison of the presence of these CD-epitopes in different wheat
518
varieties and gains insight into the immunogenicity (see Supplementary Fig. S5).
519
Even though peptides P8 and P9 did not cover the total -gliadin content, as will
520
be further discussed in the general discussion, they could be used to indicate the
Ac ce
pt
ed
M
an
us
cr
ip t
495
22 Page 22 of 49
amount of -gliadins present. Results showed that ‘Toronto’ contained the
522
highest amount of -gliadins per total gluten amount and ‘Dibillik sinde’ the
523
lowest amount (Table 3, Supplementary Fig. S5), whereas the total gluten
524
amount for ‘Dibillik sinde’ was higher compared with the total gluten amount for
525
‘Minaret’.
ip t
521
The total amount of Glia-2/9 and Glia-20 epitopes could be compared with
527
results obtained from previous studies where we performed large scale
528
immunoblotting experiments with antibodies against CD-epitopes Glia-9 and
529
Glia-20 (see Supplementary Fig. S6). While equal amounts of total gluten
530
protein extract were used for immunoblotting, the overall signal intensities
531
obtained with mAbs against the Glia-9 (Supplementary Fig. S6B) and Glia-20
532
(Supplementary Fig. S6C) epitopes were different for the three wheat varieties,
533
which indicated that these epitopes were present in different amounts. In Fig. 3
534
an overview is given of the LC-MRM/MS results and shows the sum of P1-P5
535
(Glia-2/9) and P6-P7 (Glia-20) using the average Mw of -gliadins (32285.5
536
Da). Peptide P1, containing the Glia-9 epitope and present in all three varieties,
537
was highly present in both the modern hexaploid variety Toronto and the
538
tetraploid variety Dibillik sinde (Table 3). The total amount of P1, the only
539
peptide present in the tetraploid wheat variety Dibillik sinde encoded by the A-
540
genome, was about half (56%) of the amount present in ‘Toronto’ (Fig. 3A). This
541
result was different from the immunoblotting experiments where the signal in
542
‘Dibillik sinde’ using the Glia-9 mAb (Supplementary Fig. S6B) was about three
543
quarters (75%) compared with the signal in ‘Toronto’, while equal amounts of
544
total gluten protein extract were used for immunoblotting. This was probably
545
caused by a-specific signals by the Glia-9 mAb. With LC-MRM/MS, it was
546
measured that in ‘Minaret’ the total amount of P1 to P5 was about half (52%) of
Ac ce
pt
ed
M
an
us
cr
526
23 Page 23 of 49
547
that in ‘Toronto’ (Fig. 3A). Results from immunoblotting using the Glia-9 mAb
548
(Supplementary Fig. S6B) showed in ‘Minaret’ more than half the signal
549
compared with ‘Toronto’, which was probably also caused by a-specificity of the
550
Glia-9 mAb. The amount of Glia-20 epitope (P6) in ‘Dibillik sinde’ was about one third
552
(32%) of the amount of Glia-20 epitopes (P6 and P7) in ‘Minaret’ and one tenth
553
(13%) of the amount (P6 and P7) in ‘Toronto’ (Fig. 3B). In general, peptides P1
554
to P7 were present in lower amounts in ‘Minaret’ compared with ‘Toronto’ (Table
555
3, Supplementary Fig. S5). The results obtained from immunoblotting using the
556
Glia-20 mAb (Supplementary Fig. S6C) were much more comparable with the
557
LC-MRM/MS results then the results obtained for the Glia-9 mAb, probably
558
because of the higher specificity of the Glia-20 mAb. The number of protein
559
bands reacting with both mAbs was less in ‘Minaret’ and ‘Dibillik sinde’ compared
560
with
‘Toronto’.
Ac ce
pt
ed
M
an
us
cr
ip t
551
24 Page 24 of 49
561
562
4. Discussion
563 The intention of the described study is to develop a peptide targeted
565
detection method that can accurately discriminate between different CD
566
stimulating epitope sequences originating from the multitude of -gliadin
567
proteins expressed in wheat varieties. [25,58-61][60,61][49][60][24]Improved
568
and new detection methods have been under development in the meantime [38-
569
43,62]. However, developing a new approved detection method for immunogenic
570
gluten is very difficult because of the complexity of detecting and quantifying
571
only a subset of ‘harmful’ gluten peptides instead of detecting and quantifying all
572
gluten proteins. We aimed to develop a method detecting and quantifying only
573
harmful epitopes in wheat varieties that stimulate the development of CD. The
574
accuracy of the method will depend on the sensitivity of the mass spectrometer
575
and the optimal digestion conditions of the gluten proteins. It has been shown
576
before by Šalplachta et al. [63] that using trypsin for the digestion of gluten
577
proteins did not result in sufficient amounts of peptides and results in peptides
578
that have too high molecular masses. LC-MS detection methods as developed by
579
Sealey-Voyskner et al. [42] and Prandi et al. [64] make use of the proteases
580
pepsin, trypsin, and chymotrypsin to simulate the human intestinal digestive
581
tract. However, results show that immunodominant peptides are the result of
582
chymotrypsin digestion only. Gastrointestinal digestion may also be different
583
among humans. By using chymotrypsin and optimal digestion conditions this
584
possible difference is excluded.
Ac ce
pt
ed
M
an
us
cr
ip t
564
25 Page 25 of 49
A second success factor for the accuracy of the method is the detection level.
586
The sensitivity and specificity of mass spectrometry are very high and allow
587
detection of peptides containing single amino acid substitutions that can
588
differentiate between immunogenic and non-immunogenic CD-epitope containing
589
peptides. As was shown by Mitea et al. [24], P to S substitutions can cause
590
inactivation of immunogenic CD-peptides. With this knowledge and using MRM
591
mass spectrometry (LC-MRM/MS) it is possible to identify these amino acid
592
substitutions within peptides and search for wheat varieties and species that are
593
low in immunogenic CD-epitopes.
us
cr
ip t
585
It is difficult to compare ELISA read outs with results from LC-MS measuring
595
CD-immunogenic gluten. Both methods are very different in the way of detection
596
and calibration. ELISA uses indirect detection using a mAb and calibrates against
597
a broad range of proteins or peptides depending on whether sandwich or
598
competitive ELISA is being used. LC-MS detects directly the presence of peptides
599
and calibration is against peptide standards. Because of the complexity of gluten
600
proteins, there is not a single peptide present in all gluten proteins and therefore
601
specific peptides of interest should be quantified. Cressey et al. [65] described
602
the use of the method developed by Sealey-Voyksner et al. [42] in which
603
quantification is based on one peptide present in LMW-GS. Because the
604
contribution of LMW-GS to the total gluten protein content varies among
605
hexaploid wheats [66], there should be a certain calculation factor taken into
606
account to obtain the true level of ppm in food products or even in bread wheat
607
flours. The method developed by Prandi et al. [64] described quantification by
608
using a labeled peptide of which the peptide is only encoded by the A-genome in
609
wheat. Because gluten proteins encoded by the A-, B-, and D-genome are not
610
identical, also a certain correcting calculation factor should be included to obtain
Ac ce
pt
ed
M
an
594
26 Page 26 of 49
the correct value for levels of ppm. The authors described the use of three
612
marker peptides for quantification of the gliadin amount. These marker peptides
613
were derived from the N-terminus of -gliadins and were released after pepsin
614
digestion. These three marker peptides, however, are present in only 60% of -
615
gliadins from T. aestivum (Uniprot, 2014). We used two different peptides for
616
total -gliadin quantification (P8 and P9) that are located at the start of the
617
unique domain 1 in -gliadins [67] and are released after chymotrypsin
618
digestion. These peptides are present in about 64% of -gliadins from T.
619
aestivum. This percentage could be increased to 93% when another isoform of
620
these peptides is included (Table 1). This will be necessary when we directly
621
want to quantify peptides present in food products.
an
us
cr
ip t
611
For the developed LC-MS/MRM method, we decided not to use labeled
623
peptides as was argued before by Sealey-Voyksner et al. [42]. The reason was
624
because of increasing costs when many CD-epitope containing peptides need to
625
be quantified. At this moment, more than 24 different CD-epitopes are known.
ed
M
622
Tetraploid wheat, known as pasta wheat, contains the A- and the B-genome
627
and many varieties exist, ranging from old landraces to cultivated lines.
628
Especially old landraces, but also nowadays cultivated pasta wheat, may be
629
mixtures of different genotypes including hexaploid bread wheat [61,68]. Prandi
630
et al. [68] described the detection of contamination of tetraploid wheat batches
631
with hexaploid wheat using the 33-mer as a biomarker. The 33-mer is encoded
632
by -gliadins from the D-genome only. However, not all hexaploid wheat
633
varieties may contain the 33-mer [58]. Here, we describe the use of three
634
additional peptides (P2, P3, and P4) specifically encoded by the D-genome that
635
can be used to detect this type of contamination. One of these peptides (P2,
636
LQLQPFPQPQLPYPQPQPF) was detected by Prandi et al. [40,41] in tetraploid
Ac ce
pt
626
27 Page 27 of 49
637
durum wheat showing that the used durum wheat variety was probably
638
contaminated with hexaploid wheat. It is shown by Mitea et al. (see Fig. 1) [24]
639
that only peptide sequence P1 (LQLQPFPQPQLPY) is encoded by the A-genome
640
and can be present in tetraploid wheat varieties. Immunodominancy of a CD-epitope is probably determined by the speed of
642
release of the gluten peptide containing this epitope and the amount of these
643
peptides present in the small intestine [64]. Gluten peptides shown to be
644
immunodominant (33-mer and derivatives) are indeed released very fast, within
645
one hour, from the gluten extracts in our in vitro digestion. In contrast, the
646
release of the peptide containing the Glia-20 epitope takes much more time, up
647
to 22 hours, and may therefore be less important in stimulating human T-cells.
an
us
cr
ip t
641
Recent research based on detection of CD-epitopes using specific T-cell-,
649
antibody-, and DNA analyses demonstrated that the CD-immunogenicity varies
650
across gluten proteins among and within different wheat varieties and species
651
[25,58-61]. In this study, we analyzed a modern hexaploid wheat variety
652
Toronto, an older hexaploid wheat variety Minaret, and a landrace tetraploid
653
wheat variety Dibillik sinde. Selected wheat varieties have been analyzed before
654
by immunoblotting using mAbs against CD-epitopes Glia-9 and Glia-20 for
655
relative quantification of the amount of these epitopes [60,61]. However, these
656
antibodies do not recognize the complete T-cell specific epitope. The mAb
657
recognition site for the Glia-9 epitope is found in sequences of -gliadins, -
658
gliadins, and -gliadins/D-type LMW-GS obtained from wheat species [49]. The
659
mAb recognition site for the Glia-20 epitope is only found in sequences of -
660
gliadins. In -gliadin sequences, the presence of the recognition site for both
661
mAbs correlates nicely with the presence of the recognition site of the human T-
662
cell [60]. However, both the Glia-9 and Glia-20 mAbs will also recognize
Ac ce
pt
ed
M
648
28 Page 28 of 49
naturally occurring non-epitope sequences that are inactive in T-cell response
664
e.g. because of a P to S substitution present outside the mAb recognition site
665
[24]. Using these mAbs in ELISA or immunoblotting for detecting the Glia-9 and
666
Glia-20 epitopes may therefore result in overestimation of the presence of
667
immuno-active peptides. The development of an LC-MRM/MS method for direct
668
quantitative detection enables us to compare the two different methods and we
669
show that LC-MRM/MS ensures more precise detection and quantification of CD-
670
epitopes. ELISA and LC-MRM/MS could be used in combination where ELISA
671
could be used to identify the presence of (total) gluten proteins and the
672
developed LC-MRM/MS method could be used to specifically identify and quantify
673
the particular immunogenic peptides containing CD-epitopes. The detection and
674
quantification of different and individual immunogenic peptides by LC-MRM/MS
675
will need to be coupled to a different threshold level of exposure levels for CD-
676
patients.
ed
M
an
us
cr
ip t
663
The developed LC-MRM/MS method could also be used in combination with
678
RNA sequencing as described by Salentijn et al. [69]. RNA sequencing is
679
performed on developing wheat seeds to analyze which -gliadin genes are being
680
expressed. This is in contrast to LC-MRM/MS that uses mature seeds for
681
extraction of gluten proteins. The developing stage of the seed is important in
682
determining the amount of mRNA present encoding the different gluten proteins
683
[70,71] and protein composition might change during seed development [72].
684
Furthermore, the relation between mRNA level and protein level is not fully
685
correlated and can differ among genes. In addition, even though Salentijn et al.
686
[69] found novel variants of the CD-epitopes via RNA sequencing, the true
687
presence of these peptide sequence variants in mature seeds can only be shown
688
by LC-MS gluten protein analysis. The results obtained by Salentijn et al. [69] for
Ac ce
pt
677
29 Page 29 of 49
689
the tetraploid ‘Dibillik sinde’ is in accordance with our LC-MRM/MS results. Low
690
expression levels are found for genes encoding the Glia-20 epitope by RNA
691
sequencing from developing seeds and results from LC-MRM/MS show very low
692
levels of corresponding peptides in mature seeds. Combining both techniques
693
may
694
immunogenic
development
of
genetic
markers
for
breeding
of
low
ip t
the
wheat.
Ac ce
pt
ed
M
an
us
cr
enable
30 Page 30 of 49
695
696
5. Conclusions
697 Analysis of hexaploid wheat for the presence of CD-epitopes from -gliadins
699
is more complex than analysis of those present in tetraploid wheat because of
700
the complexity of the D-genome. The targeted LC-MRM/MS method using triple
701
quadrupole mass spectrometry developed in this study, specifically quantifies the
702
amounts of individual immunogenic CD-epitopes from both A- and D-genome -
703
gliadins present in hexaploid wheat. It also quantifies the isoforms present of
704
peptides encoded by the D-genome. LC-MRM/MS technology is useful for
705
identification
706
immunogenic epitopes at low femtomolar levels of detection. It is a fast and
707
sensitive method and detection and quantification are reproducible. The method
708
can be extended to any other CD immunogenic epitope to obtain quantification
709
for all epitopes present in wheat varieties and species. It will enable the selection
710
of varieties or species with low amounts of T-cell stimulatory epitopes that can
711
be used in food products to lower the ‘gluten load’ and thereby prevent
712
development
quantification
of
biomarker
peptides
containing
CD
Ac ce
pt
ed
M
and
an
us
cr
ip t
698
and
prevalence
of
CD.
31 Page 31 of 49
713
714
Acknowledgements
715 This research was partially funded by the European regional development
717
fund and the province of Gelderland and Overijssel (GO EFRO 2007-2013) and
718
the Dutch Ministry of Economic Affairs through the DLO program ‘Plant and
719
Animal for Human Health’ (KB-05-001-019-PRI) and ‘Healthy and safe food in
720
chain perspective’ (KB-15-001-016). The authors thank Noor Bas (Centre for
721
Genetic Resources, CGN, The Netherlands) and Hein de Jong (Limagrain) for
722
providing the seeds, and Elma Salentijn for providing the original figure to adapt
723
Fig.
1.
Ac ce
pt
ed
M
an
us
cr
ip t
716
32 Page 32 of 49
724
725
References
726
728 729
coeliac disease, Nat Rev Immunol. 13 (2013) 294-302. [2]
730 731
L.M. Sollid, B. Jabri, Triggers and drivers of autoimmunity: lessons from N. Reilly, P. Green, Epidemiology and clinical presentations of celiac disease, Semin Immunopathol. 34 (2012) 473-478.
[3]
ip t
[1]
cr
727
K. Mustalahti, C. Catassi, A. Reunanen, E. Fabiani, M. Heier, S. McMillan, L. Murray, M.-H. Metzger, M. Gasparin, E. Bravi, M. Mäki, The prevalence of
733
celiac disease in Europe: Results of a centralized, international mass
734
screening project, Ann Med. 42 (2010) 587-595.
736 737
S.D. Rampertab, N. Pooran, P. Brar, P. Singh, P.H.R. Green, Trends in the
an
[4]
presentation of celiac disease, Am J Med. 119 (2006) 355.e359-355.e314. [5]
J.F. Ludvigsson, A. Rubio-Tapia, C.T. van Dyke, L.J. Melton, III, A.R.
M
735
us
732
738
Zinsmeister, B.D. Lahr, J.A. Murray, Increasing incidence of celiac disease
739
in a North American population, Am J Gastroenterol. 108 (2013) 818-824. [6]
A. Rubio-Tapia, J.F. Ludvigsson, T.L. Brantner, J.A. Murray, J.E. Everhart,
ed
740
The prevalence of celiac disease in the United States, Am J Gastroenterol.
742
107 (2012) 1538-1544.
743
[7]
J.Y. Kang, A.H.Y. Kang, A. Green, K.A. Gwee, K.Y. Ho, Systematic review: worldwide variation in the frequency of coeliac disease and changes over
749
Ac ce
744
pt
741
750
477-482.
745 746
time, Aliment Pharmacol Ther. 38 (2013) 226-245.
[8]
747 748
751 752 753 754
L. Day, in G.O. Phillips, P.A. Williams (Editors), Handbook of food proteins Woodhead Publishing Limited, 2011, pp. 267-288.
[9]
J.H. Woychik, J.A. Boundy, R.J. Dimler, Starch gel electrophoresis of wheat gluten proteins with concentrated urea, Arch Biochem Biophys. 94 (1961)
[10] F. MacRitchie, D. Lafiandra, in S. Damodaran, A. Paraf (Editors), Food proteins and their applications, Dekker, New York, 1997. [11] P.R. Shewry, A.S. Tatham, in P.R. Shewry, R. Casey (Editors), Seed Proteins, Kluwer Academic Publishers, Dordrecht, 1999, pp. 11-36.
33 Page 33 of 49
755
[12] M.C. Gianibelli, O.R. Larroque, F. MacRitchie, C.W. Wrigley, Biochemical,
756
genetic,
and
molecular
characterization
of
wheat
757
component subunits, Cereal Chem. 78 (2001) 635-646.
glutenin
and
its
[13] E.A. Jackson, L.M. Holt, P.I. Payne, Characterisation of high molecular
759
weight gliadin and low-molecular-weight glutenin subunits of wheat
760
endosperm by two-dimensional electrophoresis and the chromosomal
761
localisation of their controlling genes, Theor Appl Genet. 66 (1983) 29-37.
ip t
758
[14] P.I. Payne, L.M. Holt, C.N. Law, Structural and genetic-studies on the high-
763
molecular-weight subunits of wheat glutenin. 1. Allelic variation in subunits
764
amongst varieties of wheat (Triticum aestivum), Theor Appl Genet. 60
765
(1981) 229-236.
us
cr
762
[15] O.D. Anderson, J.C. Litts, F.C. Greene, I. Characterization of ten new
767
wheat alpha-gliadin genomic clones, evidence for limited sequence
768
conservation of flanking DNA, and southern analysis of the gene family,
769
Theor Appl Genet. 95 (1997) 50-58.
an
766
[16] T.W. Okita, V. Cheesbrough, C.D. Reeves, Evolution and heterogeneity of
771
the alpha-/beta-type and gamma-type gliadin DNA sequences, J Biol
772
Chem. 260 (1985) 8203-8213.
M
770
[17] G. Branlard, M. Dardevet, R. Saccomano, F. Lagoutte, J. Gourdon, Genetic
774
diversity of wheat storage proteins and bread wheat quality, Euphytica.
775
119 (2001) 59-67.
pt
777
[18] P.R. Shewry, A.S. Tatham, P.A. Lazzeri, Biotechnology of wheat quality, J Sci Food Agric. 73 (1997) 397-406.
Ac ce
776
ed
773
778
[19] L. Sollid, S.-W. Qiao, R. Anderson, C. Gianfrani, F. Koning, Nomenclature
779
and listing of celiac disease relevant gluten T-cell epitopes restricted by
780
HLA-DQ molecules, Immunogenetics. 64 (2012) 455-460.
781
[20] E.H. Arentz-Hansen, R. Körner, Ø. Molberg, H. Quarsten, W. Vader, Y.M.C.
782
Kooy, K.E.A. Lundin, F. Koning, P. Roepstorff, L.M. Sollid, S.N. McAdam,
783
The intestinal T-cell response to -gliadin in adult celiac disease is focused
784
on a single deamidated glutamine targeted by tissue transglutaminase, J
785
Exp Med. 191 (2000) 603-612.
786
[21] E.H. Arentz-Hansen, S.N. McAdam, Ø. Molberg, B. Fleckenstein, K.E.A.
787
Lundin, T.J.D. Jorgensen, G. Jung, P. Roepstorff, L.M. Sollid, Celiac lesion T
788
cells recognize epitopes that cluster in regions of gliadins rich in proline
789
residues, Gastroenterology. 123 (2002) 803-809. 34 Page 34 of 49
790
[22] L. Shan, Ø. Molberg, I. Parrot, F. Hausch, F. Filiz, G.M. Gray, L.M. Sollid, C.
791
Khosla, Structural basis for gluten intolerance in celiac sprue, Science. 297
792
(2002) 2275-2279. [23] C.L. Marino, N.A. Tuleen, G.E. Hart, J.C. Nelson, M.E. Sorrells, Y.H. Lu, P.
794
Leroy, C.R. Lopes, Molecular genetic maps of the group 6 chromosomes of
795
hexaploid wheat (Triticum aestivum L. em. Thell.), Genome. 39 (1996)
796
359-366.
ip t
793
[24] C. Mitea, E.M.J. Salentijn, P. van Veelen, S.V. Goryunova, I.M. van der
798
Meer, H.C. van den Broeck, J.R. Mujico, V. Monserrat, L.J.W.J. Gilissen,
799
J.W. Drijfhout, L. Dekking, F. Koning, M.J.M. Smulders, A universal
800
approach to eliminate antigenic properties of alpha-gliadin peptides in
801
celiac disease, PLoS ONE. 5 (2010) e15637.
us
cr
797
[25] Ø. Molberg, A.K. Uhlen, T. Jensen, N.S. Flæte, B. Fleckenstein, H. Arentz-
803
Hansen, M. Raki, K.E.A. Lundin, L.M. Sollid, Mapping of gluten T-cell
804
epitopes in the bread wheat ancestors: Implications for celiac disease,
805
Gastroenterology. 128 (2005) 393-401.
M
an
802
[26] T.W.J.M. van Herpen, S. Goryunova, J. van der Schoot, M. Mitreva, E.M.J.
807
Salentijn, O. Vorst, M. Schenk, P. van Veelen, F. Koning, L. van Soest, B.
808
Vosman, D. Bosch, R.J. Hamer, L.J.W.J. Gilissen, M.J.M. Smulders, Alpha-
809
gliadin genes from the A, B, and D genomes of wheat contain different sets
810
of celiac disease epitopes, BMC Genomics. 7 (2006) 1.
811
[27] E. García, M. Llorente, A. Hernando, R. Kieffer, H. Wieser, E. Mendez,
812
Development of a general procedure for complete extraction of gliadins for
813
Ac ce
pt
ed
806
814
heat processed and unheated foods, Eur J Gastroenterol Hepatol. 17 (2005) 529-539.
815
[28] E. Méndez, C. Vela, U. Immer, F.W. Janssen, Report of a collaborative trial
816
to investigate the performance of the R5 enzyme linked immunoassay to
817 818
determine gliadin in gluten-free food, Eur J Gastroenterol Hepatol. 17 (2005) 1053-1063.
819
[29] I. Valdés, E. García, M. Llorente, E. Méndez, Innovative approach to low-
820
level gluten determination in foods using a novel sandwich enzyme-linked
821
immunosorbent assay protocol, Eur J Gastroenterol Hepatol. 15 (2003)
822
465-474.
35 Page 35 of 49
823
[30] B. Geßendorfer, H. Wieser, P. Koehler, Optimisation of a solvent for the
824
complete extraction of prolamins from heated foods, J Cereal Sci. 52
825
(2010) 331-332.
826
[31] P. Kanerva, T. Sontag-Strohm, O. Brinck, H. Salovaara, Improved
827
extraction of prolamins for gluten detection in processed foods, Agr Food
828
Sci. 20 (2011) 206-216. [32] M.C.
Mena,
M.
Lombardía,
A.
Hernando,
E.
Méndez,
J.P.
Albar,
ip t
829
Comprehensive analysis of gluten in processed foods using a new
831
extraction method and a competitive ELISA based on the R5 antibody,
832
Talanta. 91 (2012) 33-40.
cr
830
[33] C. Diaz-Amigo, B. Popping, Accuracy of ELISA detection methods for
834
gluten and reference materials: A realistic assessment, J Agric Food Chem.
835
61 (2013) 5681-5688.
us
833
[34] D. Stepniak, L.W. Vader, Y. Kooy, P.A. Veelen, A. Moustakas, N.A.
837
Papandreou, E. Eliopoulos, J.W. Drijfhout, G.K. Papadopoulos, F. Koning,
838
T-cell recognition of HLA-DQ2-bound gluten peptides can be influenced by
839
an N-terminal proline at p-1, Immunogenetics. 57 (2005) 8-15.
M
an
836
[35] M. Koeberl, D. Clarke, A.L. Lopata, Next generation of food allergen
841
quantification using mass spectrometric systems, J Proteome Res. 13
842
(2014) 3499-3509. [36] C.
Diaz-Amigo,
B.
Popping,
Gluten
and
glutenfree:
Issues
and
pt
843
ed
840
considerations of labeling regulations, detection methods, and assay
845
validation, J AOAC Int. 95 (2012) 337-348.
Ac ce
844 846
[37] R. Haraszi, H. Chassaigne, A. Maquet, F. Ulberth, Analytical methods for
847
detection of gluten in food—Method developments in support of food
848
labeling legislation, J AOAC Int. 94 (2011) 1006-1025.
849
[38] S.B. Altenbach, W.H. Vensel, F.M. DuPont, Integration of transcriptomic
850
and proteomic data from a single wheat cultivar provides new tools for
851
understanding the roles of individual alpha gliadin proteins in flour quality
852
and celiac disease, J Cereal Sci. 52 (2010) 143-151.
853
[39] K.L. Fiedler, S.C. McGrath, J.H. Callahan, M.M. Ross, Characterization of
854
grain-specific peptide markers for the detection of gluten by mass
855
spectrometry, J Agric Food Chem. 62 (2014) 5835-5844.
856 857
[40] B. Prandi, A. Faccini, T. Tedeschi, A. Cammerata, D. Sgrulletta, M. D’Egidio,
G.
Galaverna,
S.
Sforza,
Qualitative
and
quantitative
36 Page 36 of 49
858
determination of peptides related to celiac disease in mixtures derived
859
from different methods of simulated gastrointestinal digestion of wheat
860
products, Anal Bioanal Chem. 406 (2014) 4765-4775. [41] B. Prandi, A. Faccini, T. Tedeschi, G. Galaverna, S. Sforza, LC/MS analysis
862
of proteolytic peptides in wheat extracts for determining the content of the
863
allergen amylase/trypsin inhibitor CM3: Influence of growing area and
864
variety, Food Chem. 140 (2013) 141-146.
ip t
861
[42] J.A. Sealey-Voyksner, C. Khosla, R.D. Voyksner, J.W. Jorgenson, Novel
866
aspects of quantitation of immunogenic wheat gluten peptides by liquid
867
chromatography-mass spectrometry/mass spectrometry, J Chromatogr A.
868
1217 (2010) 4167-4183.
us
cr
865
[43] L. Uvackova, L. Skultety, S. Bekesova, S. McClain, M. Hajduch, MSE based
870
multiplex protein analysis quantified important allergenic proteins and
871
detected relevant peptides carrying known epitopes in wheat grain
872
extracts, J Proteome Res. 12 (2013) 4862-4869.
an
869
[44] A. Camarca, R.P. Anderson, G. Mamone, O. Fierro, A. Facchiano, S.
874
Costantini, D. Zanzi, J. Sidney, S. Auricchio, A. Sette, R. Troncone, C.
875
Gianfrani, Intestinal T cell responses to gluten peptides are largely
876
heterogeneous: Implications for a peptide-based therapy in celiac disease,
877
J Immunol. 182 (2009) 4158-4166. [45] W. Vader, Y. Kooy, P. van Veelen, A. de Ru, D. Harris, W. Benckhuijsen, S.
879
Peña, L. Mearin, J.W. Drijfhout, F. Koning, The gluten response in children
880
with celiac disease is directed toward multiple gliadin and glutenin
881
pt
878
Ac ce
ed
M
873
peptides, Gastroenterology. 122 (2002) 1729-1737.
882
[46] A. Camarca, A. Del Mastro, C. Gianfrani, Repertoire of gluten peptides
883
active in celiac disease patients: Perspectives for translational therapeutic
884 885
applications, Endocr Metab Immune Disord Drug Targets. 12 (2012) 207219.
886
[47] J.A. Tye-Din, J.A. Stewart, J.A. Dromey, T. Beissbarth, D.A. van Heel, A.
887
Tatham, K. Henderson, S.I. Mannering, C. Gianfrani, D.P. Jewell, A.V.S.
888
Hill,
889
quantitative mapping of T cell epitopes in gluten in celiac disease, Sci
890
Transl Med. 2 (2010) 41ra51.
J.
McCluskey,
J.
Rossjohn,
R.P.
Anderson,
Comprehensive,
891
[48] R.P. Anderson, P. Degano, A.J. Godkin, D.P. Jewell, A.V.S. Hill, In vivo
892
antigen challenge in celiac disease identifies a single transglutaminase37 Page 37 of 49
893
modified peptide as the dominant -gliadin T-cell epitope, Nature Med. 6
894
(2000) 337-342.
895
[49] H.C. van den Broeck, A.H.P. America, M.J.M. Smulders, D. Bosch, R.J.
896
Hamer, L.J.W.J. Gilissen, I.M. van der Meer, A modified extraction protocol
897
enables detection and quantification of celiac disease-related gluten
898
proteins from wheat, J Chromatogr B. 877 (2009) 975-982. [50] M. Stapels, K. Fadgen, A reproducible online 2D reversed phase-reversed
900
phase high-low pH method for qualitative and quantitative proteomics,
901
Spectroscopy
902
http://www.spectroscopyonline.com/node/225338.
Special
issue
(2009)
cr
online.
ip t
899
[51] M. Gilar, P. Olivova, A.E. Daly, J.C. Gebler, Two-dimensional separation of
904
peptides using RP-RP-HPLC system with different pH in first and second
905
separation dimensions, J Sep Sci. 28 (2005) 1694-1703.
us
903
[52] B. MacLean, D.M. Tomazela, N. Shulman, M. Chambers, G.L. Finney, B.
907
Frewen, R. Kern, D.L. Tabb, D.C. Liebler, M.J. MacCoss, Skyline: an open
908
source document editor for creating and analyzing targeted proteomics
909
experiments, Bioinformatics. 26 (2010) 966-968.
M
head of bacteriophage T4, Nature. 227 (1970) 680-685. [54] J.H.G. Cordewener, G. Hause, E. Görgen, R. Busink, B. Hause, H.J.M.
913
Dons, A.A.M. Van Lammeren, M.M. Van Lookeren Campagne, P. Pechan,
914
Changes in synthesis and localization of members of the 70-kDa class of
915
heat-shock proteins accompany the induction of embryogenesis in Brassica
916
pt
912
Ac ce
911
[53] U.K. Laemmli, Cleavage of structural proteins during the assembly of the
ed
910
an
906
napus L. microspores, Planta. 196 (1995) 747-755.
917
[55] C. Mitea, Y.M.C. Kooy-Winkelaar, P.A. van Veelen, A. de Ru, J.W. Drijfhout,
918
F. Koning, L. Dekking, Fine specificity of monoclonal antibodies against
919 920
celiac disease-inducing peptides in the gluteome, Am J Clin Nutr. 88 (2008) 1057-1066.
921
[56] E.H.A. Spaenij-Dekking, E.M.C. Kooy-Winkelaar, W.F. Nieuwenhuizen, J.W.
922
Drijfhout, F. Koning, A novel and sensitive method for the detection of T
923
cell stimulatory epitopes of - and -gliadin, Gut. 53 (2004) 1267-1273.
924
[57] C. Mitea, R. Havenaar, J.W. Drijfhout, L. Edens, L. Dekking, F. Koning,
925
Efficient
degradation
of
gluten
by
a
prolyl
endoprotease
in
a
926
gastrointestinal model: implications for coeliac disease, Gut. 57 (2008) 25-
927
32. 38 Page 38 of 49
[58] E.M.J. Salentijn, S. Goryunova, N. Bas, I.M. van der Meer, H.C. van den
929
Broeck, T. Bastien, L.J.W.J. Gilissen, M.J.M. Smulders, Tetraploid and
930
hexaploid wheat varieties reveal large differences in expression of alpha-
931
gliadins from homoeologous Gli-2 loci, BMC Genomics. 10 (2009) 48.
932
[59] L. Spaenij-Dekking, Y. Kooy-Winkelaar, P. van Veelen, J.W. Drijfhout, H.
933
Jonker, L. van Soest, M.J.M. Smulders, D. Bosch, L.J.W.J. Gilissen, F.
934
Koning, Natural variation in toxicity of wheat: potential for selection of
935
nontoxic varieties for celiac disease patients, Gastroenterology. 129 (2005)
936
797-806.
cr
ip t
928
[60] H.C. van den Broeck, H.C. de Jong, E.M.J. Salentijn, L. Dekking, D. Bosch,
938
R.J. Hamer, L.J.W.J. Gilissen, I.M. van der Meer, M.J.M. Smulders,
939
Presence of celiac disease epitopes in modern and old hexaploid wheat
940
varieties: wheat breeding may have contributed to increased prevalence of
941
celiac disease, Theor Appl Genet. 121 (2010) 1527-1539.
an
us
937
[61] H.C. van den Broeck, C. Hongbing, X. Lacaze, J.-C. Dusautoir, L.J.W.J.
943
Gilissen, M.J.M. Smulders, I.M. van der Meer, In search of tetraploid wheat
944
accessions reduced in celiac disease-related gluten epitopes, Mol Biosyst. 6
945
(2010) 2206-2213.
M
942
[62] E. Halbmayr-Jech, E. Hammer, R. Fielder, J. Coutts, A. Rogers, M. Cornish,
947
Characterization of G12 sandwich ELISA, a next-generation immunoassay
948
for gluten toxicity, J AOAC Int. 95 (2012) 372-376.
949
[63] J. Šalplachta, M. Marchetti, J. Chmelík, G. Allmaier, A new approach in
950
proteomics of wheat gluten: combining chymotrypsin cleavage and matrix-
951
Ac ce
pt
ed
946
952
assisted laser desorption/ionization quadrupole ion trap reflectron tandem mass spectrometry, Rapid Commun Mass Sp. 19 (2005) 2725-2728.
953
[64] B. Prandi, M. Bencivenni, A. Faccini, T. Tedeschi, A. Dossena, R. Marchelli,
954
G. Galaverna, S. Sforza, Composition of peptide mixtures derived from
955 956 957
simulated gastrointestinal digestion of prolamins from different wheat varieties, J Cereal Sci. 56 (2012) 223-231.
[65] P. Cressey, P. Grounds, S. Jones, S. Ashmore, D. Saunders, Gluten
958
residues
in
gluten-free
foods
sold
in
Christchurch,
New
Zealand:
959
comparison of LC-MS and ELISA methods, Qual Assur Saf Crop. 5 (2013)
960
207-213.
39 Page 39 of 49
961
[66] H. Wieser, R. Kieffer, Correlations of the amount of gluten protein types to
962
the technological properties of wheat flours determined on a micro-scale, J
963
Cereal Sci. 34 (2001) 19-27.
964
[67] O.D. Anderson, F.C. Greene, The α-gliadin gene family. II. DNA and
965
protein
sequence
variation,
subfamily
structure,
966
pseudogenes, Theor Appl Genet. 95 (1997) 59-65.
and
origins
of
[68] B. Prandi, M. Bencivenni, T. Tedeschi, R. Marchelli, A. Dossena, G.
968
Galaverna, S. Sforza, Common wheat determination in durum wheat
969
samples through LC/MS analysis of gluten peptides, Anal Bioanal Chem.
970
403 (2012) 2909-2914.
cr
ip t
967
[69] E. Salentijn, D. Esselink, S. Goryunova, I. van der Meer, L. Gilissen, M.
972
Smulders, Quantitative and qualitative differences in celiac disease
973
epitopes among durum wheat varieties identified through deep RNA-
974
amplicon sequencing, BMC Genomics. 14 (2013) 905.
an
us
971
[70] K. Kawaura, K. Mochida, Y. Ogihara, Expression profile of two storage-
976
protein gene families in hexaploid wheat revealed by large-scale analysis
977
of expressed sequence tags, Plant Physiol. 139 (2005) 1870-1880.
M
975
[71] P.R. Shewry, R.A.C. Mitchell, P. Tosi, Y. Wan, C. Underwood, A. Lovegrove,
979
J. Freeman, G.A. Toole, E.N.C. Mills, J.L. Ward, An integrated study of
980
grain development of wheat (cv. Hereward), J Cereal Sci. 56 (2012) 21-
981
30. Katagiri,
pt
[72] M.
T.
Masuda,
F.
Tani,
N.
Kitabatake,
Expression
and
983
development of wheat proteins during maturation of wheat kernel and the
984
Ac ce
982
ed
978
985 986
rheological properties of dough prepared from the flour of mature and immature Wheat, Food Sci Technol Res. 17 (2011) 111-120.
40 Page 40 of 49
986
Figure legends
988
Fig. 1.
989
Sequence alignment of the CD-epitope domain of multiple -gliadin proteins from
990
wheat. The deduced amino acid sequences of multiple EST contigs (present in
991
the NCBI-UniGene database) were searched for sequence variants in the CD-
992
epitope containing domain. Homologous ESTs for this sequence domain were
993
grouped and only variants found in more than 5 ESTs (n) are displayed. The
994
variants are grouped per genome origin from which the ESTs are expressed: Gli-
995
A2, Gli-B2, or Gli-D2. The CD-epitopes, known to trigger the adaptive HLA-DQ2+
996
T-cell response (Glia-2/Glia-9, and Glia-20) in CD-patients, are color
997
indicated. In yellow: Glia-2/Glia-9 epitopes. In green: Glia-20 epitope, in
998
blue: 33-mer containing repetitive epitopes. In black: chymotrypsin digestion
999
sites. In red: amino acid (AA) variation in the sequence. In bold underlined:
1000
sequences selected for peptide synthesis. The peptide number code is indicated
1001
in the right-most column. Figure adapted from Mitea et al. [24].
1002
Ac ce
pt
ed
M
an
us
cr
ip t
987
1003
Fig. 2.
1004
Optimization of chymotrypsin digestion conditions for hexaploid wheat variety
1005
Toronto and tetraploid wheat variety Dibillik sinde. A) Release of P1 after
1006
incubation for 10’, 30’, 1 h, 2h, and 4 h at 25C with enzyme:protein ratio of
1007
1:20. The same trend was observed for peptides P2-P5. B) Release of P1 after
1008
incubation for 2 h, 4 h, 8 h, and 18 h at 25C with enzyme:protein ratios of 1:5
1009
and 1:10. The same trend was observed for P2-P5. C) Release of P6 after
1010
incubation for 2 h, 4h, and 8 h at 25C with enzyme:protein ratios of 1:5, 1:10,
41 Page 41 of 49
1011
and 1:20. D) Release of P6 after incubation for 2 h, 4 h, 8 h, 18 h, and 22 h
1012
(extra enzyme addition after 18 h) at 25C with enzyme:protein ratio of 1:5.
1013 Fig. 3.
1015
Total amounts of peptides containing CD-epitopes in mg -gliadin/g gluten (103
1016
ppm) present in three wheat varieties, Toronto (AABBDD), Minaret (AABBDD),
1017
and Dibillik sinde (AABB). Containing A) Glia-2/9 epitopes and B) Glia-20
1018
epitopes.
cr
ip t
1014
us
1019 Supplementary Fig. S1.
1021
Chromatograms, ion intensity maps, of the low pH peptide resolution in the four
1022
subsequent 2D LC-MS fractions of a ‘Minaret’ gluten protein digest by using raw
1023
MSE (DIA) data processed by Progenesis QI software. X-axis represents m/z
1024
value. Y-axis represents retention time in minutes.
1025
ed
M
an
1020
Supplementary Fig. S2.
1027
Chromatograms obtained from LC-MRM/MS showing the resolution of the nine
1028
selected marker peptides. A) Mixture of the nine marker peptides used for
1029
calibration and B) ‘Toronto’ gluten protein digest (22 h). The inlaid images show
1030
an example of the multiple transition signals per peptide as analyzed by Skyline
1031
software. X-axis represents retention time in minutes. Y-axis represents signal
1032
intensity (106).
Ac ce
pt
1026
1033 1034
Supplementary Fig. S3.
42 Page 42 of 49
1056 1057 1058
Highlights
1059
1061
Targeted detection and quantification of multiple celiac disease stimulating
ip t
1060
epitopes.
High-throughput screening for quantification of CD-epitopes in wheat.
1063
Method enables selection of wheat varieties with reduced levels of CD-
us
1065
epitopes.
LC-MRM/MS technology as a standard method for quantifying CD-epitopes.
an
1064
cr
1062
Ac ce
pt
ed
M
1066
44 Page 43 of 49
Ac
ce
pt
ed
M
an
us
cr
i
Figure 1
Page 44 of 49
Ac
ce
pt
ed
M
an
us
cr
i
Figure 2
Page 45 of 49
Ac ce p
te
d
M
an
us
cr
ip t
Figure 3
Page 46 of 49
ip t
Table 1
7
8a
9
10
11*
12
13
14b
15
16
17
18
19c
20
21d
22
23e
24
Q3S4V8_WHEAT
A5JSB3_WHEAT
I3RXV6_AEGTA
Q9M4M4_WHEAT
A5JSA6_WHEAT
A5JSA9_WHEAT
K7X0Q3_WHEAT
I0IT56_WHEAT
Q9M4L8_WHEAT
K7WV47_WHEAT
G9I1T5_AEGUN
Q2QL43_AEGTA
F6M8F7_9POAL
A5JSB7_WHEAT
M4WFD9_AEGSP
K7X1J5_WHEAT
Q2QL52_AEGTA
us
K7XEB5_WHEAT
6
F4YT74_WHEAT
Q0GK30_TRITI
Accession first hit
ce pt
Ac
*
ed
Peptide no. Peptide sequences LQLQPFPQPQPF LQLQPFPQPQLPY P1 LQLQPFPQPQLSY LQLQPFPQPQLPYPQPQPF P2 MQLQPFPQPQLPYPQPQLPYPQPQPF LQLQPFPQPQLPYPQPHLPYPQPQPF P3 LQLQPFPQPQLPYPHPQLPYPQPQPF LQLQPFPQPQLPYPQPQLPYPQPQPF P4 LQLQPFPQPQLPYPQPQLPYPQPQLPYPQPQPF P5 LPQLPYPQPQPFPPQQSYPQPQPQYPQPQQPISQQQAQL RPQQLYPQPQPQY RPQQPYPQPQPQY P6 RPQQPYPQSQPQY P7 RPQQSYPQPQPQY P8 QQQLIPCRDVVL P9 QQILQQQLIPCRDVVL QQQLIPCMDVVL QQILQQQLIPCMDVVL a, b Sequences from Triticum aestivum, probable B-genome c Sequence from Aegilops uniaristata, N-genome d Sequences from Triticum compactum, not full length e Sequences from Aegilops speltoides
-gliadin families
5
J7I026_WHEAT
4
K7WV12_AEGTA
3
I3RXV8_AEGTA
2
M an
1
Q9M4L6_WHEAT
cr
Table 1. Selection of peptide sequences identified from -gliadin families in Triticum and Aegilops species in Mascot search (Uniprot database). Selection of the sequences is based on immunogenic epitopes sequences. Peptide sequences shown include the nine selected peptide sequences used for quantification (P1-P9). Amino acid substitutions in peptide sequences are shown in italics. CD-epitope sequences within the peptides are shown in bold.
Score 46 54 54 47 41 73 60 39 53 33 46 67 59 27 52 51 40
Sequence from Aegilops tauschii, contains Glia-9 epitope on a different peptide
Page 47 of 49
QQILQQQLIPCRDVVL
cr us
14 15 20 21 19 19 29 18 26 16 35 22
6.8 7.0 7.2 6.9 7.3 4.6 4.4 4.4 5.8 5.8 6.3 6.3
6-7.5 6.6-7.5 6.7-8 6.6-7.5 6.7-8 3-5 4-5 4-5 5.5-6 5.5-6 6-6.6 6-6.6
R2 0.987 0.981 0.974 0.969 0.970 0.993 0.991 0.994 0.991 -
Ac
ce pt
MRM schedule (min)
P9
Retention time (min)
QQQLIPCRDVVL
M an
P8
784.9 (2+) 755.1 (3+) 1029.5 (3+) 1032.5 (3+) 978.3 (4+) 813.9 (2+) 808.9 (2+) 539.6 (3+) 734.9 (2+) 490.3 (3+) 976.0 (2+) 651.0 (3+)
Fragment ions m/z 483.3 (b4); 617.3 (y5); 727.4 (b6); 842.4 (y7); 952.5 (b8); 1290.7 (b11) 483.3 (b4); 488.3 (y4); 713.4 (y6); 727.4 (b6); 952.5 (b8); 973.5 (y8) 483.3 (b4); 713.4 (y6); 727.4 (b6); 824.5 (b7); 952.5 (b8); 973.5 (y8) 488.3 (y4); 645.9 (b11); 713.4 (y6); 727.4 (b6); 903.5 (y15); 952.5 (b8) 483.3 (b4); 488.3 (y4); 713.4 (y6); 727.4 (b6); 824.5 (b7); 952.5 (b8); 973.5 (y8) 407.2 (y3); 632.3 (y5); 770.4 (b6); 857.4 (y7); 995.5 (b8); 1220.6 (b10) 510.3 (b4); 770.4 (b6) 407.2 (y3); 510.3 (b4); 770.4 (b6); 847.4 (y7) 858.5 (y7); 971.5 (y8); 1212.7 (y10) 498.3 (b4); 611.4 (b5); 858.5 (y7) 858.5 (y7) 858.5 (y7); 867.5 (b7); 971.5 (y8); 980.6 (b8)
ed
P1 P2 P3 P4 P5 P6 P7
Sequence LQLQPFPQPQLPY LQLQPFPQPQLPYPQPQPF LQLQPFPQPQLPYPQPQLPYPQPQPF LQLQPFPQPQLPYPQPHLPYPQPQPF LQLQPFPQPQLPYPQPQLPYPQPQLPYPQPQPF RPQQPYPQPQPQY RPQQPYPQSQPQY
Precursor m/z (charge state)
Peptide
Table 2. MS conditions of the selected marker peptides.
Collision energy (V)
ip t
Table 2
Page 48 of 49
ip t
Table 3
ed
M an
us
cr
Table 3. Amounts of CD-epitopes present in three wheat varieties: Toronto (AABBDD), Minaret (AABBDD), and Dibillik sinde (AABB). Amounts in 3 mg -gliadin containing CD-epitope/g gluten (10 ppm) at t=22h of peptides from -gliadins containing CD-epitopes, (±SD). peptide sequence Toronto Minaret Dibillik sinde LQLQPFPQPQLPY P1 123.20 (±18.27) 56.14 (±7.07) 122.77 (±19.85) LQLQPFPQPQLPYPQPQPF P2 22.65 (±3.60) 14.93 (±1.23) nd LQLQPFPQPQLPYPQPHLPYPQPQPF P3 12.47 (±0.83) 8.82 (±0.10) nd LQLQPFPQPQLPYPQPQLPYPQPQPF P4 3.34 (±0.17) 2.18 (±0.11) nd LQLQPFPQPQLPYPQPQLPYPQPQLPYPQPQPF P5 56.26 (±4.51) 31.71 (±1.09) nd RPQQPYPQPQPQY P6 133.09 (±8.93) 49.82 (±3.28) 18.69 (±1.38) RPQQPYPQSQPQY P7 13.41 (±1.95) 9.26 (±1.08) nd P8 QQQLIPCRDVVL 162.10 (±21.61) 137.42 (±17.97) 98.75 (±13.06) P9 QQILQQQLIPCRDVVL nd nd nd 217.92 (±27.38) 113.79 (±9.60) 122.77 (±19.85) sum[P1-P5] Glia-2/9 P6 + P7 146.49 (±10.88) 59.08 (±4.36) 18.69 (±1.38) Glia-20 P8 + P9 162.10 (±21.61) 137.42 (±17.97) 98.75 (±13.06) Total -gliadin
Ac
ce pt
nd, not detected
Page 49 of 49
S0021-9673(15)00343-X http://dx.doi.org/doi:10.1016/j.chroma.2015.02.070 CHROMA 356328
To appear in:
Journal of Chromatography A
Received date: Revised date: Accepted date:
27-11-2014 23-2-2015 26-2-2015
Please cite this article as: H.C. van den Broeck, J.H.G. Cordewener, M.A. Nessen, A.H.P. America, I.M. van der Meer, Label free targeted detection and quantification of celiac disease immunogenic epitopes by mass spectrometry, Journal of Chromatography A (2015), http://dx.doi.org/10.1016/j.chroma.2015.02.070 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
1
Label free targeted detection and quantification of
2
celiac
3
spectrometry
disease
immunogenic
epitopes
by
mass
ip t
4 Hetty C. van den Broecka*, Jan H. G. Cordewenera, Merel A. Nessenb, Antoine H.
6
P. Americaa, Ingrid M. van der Meera
cr
5
8
a
9
b
us
7
Wageningen UR, Plant Research International, PO Box 16, 6700 AA Wageningen, The Netherlands RIKILT Wageningen UR, PO Box 230, 6700 AE Wageningen, The Netherlands
11
*
an
10
Corresponding author: [email protected]; Tel: +31 317 480974; Fax: +31 317 41809
13
M
12 Abstract
ed
14
Celiac disease (CD) is a food-related disease caused by certain gluten peptides
16
containing T-cell stimulating epitopes from wheat, rye, and barley. CD-patients
17
have to maintain a gluten-free diet and are therefore dependent on reliable
18
testing and labeling of gluten-free products. So far, the R5-ELISA is the approved
19
method to detect if food products can be labeled gluten-free. Because the R5-
20
ELISA detects gluten in general, there is a demand for an improved detection
21
method that quantifies specifically CD-epitopes. Therefore, we developed a new
22
method for detection and quantification of CD-epitopes, based on liquid
23
chromatography (LC) coupled to mass spectrometry (MS) in multiple reaction
24
monitoring (MRM) mode. This method enables targeted label free comparative
25
analysis of the gluten proteins present in different wheat varieties and species,
26
and in wheat-based food products. We have tested our method by analyzing
Ac ce
pt
15
1 Page 1 of 49
several wheat varieties that vary in CD-epitope content, as was shown before
28
using immunoblotting and specific monoclonal antibodies. The results showed
29
that a modern bread wheat variety Toronto contained the highest amounts of CD
30
immunogenic peptides compared with the older bread wheat variety Minaret and
31
the tetraploid wheat variety Dibillik Sinde. Our developed method can detect
32
quantitatively and simultaneously multiple specific CD-epitopes in a high
33
throughput manner.
cr
ip t
27
Keywords:
36
-Gliadin, celiac disease, LC-MRM/MS,
T-cell stimulatory epitopes, wheat
Ac ce
pt
ed
M
an
35
us
34
2 Page 2 of 49
37
38
1. Introduction Celiac disease (CD) is a food related disease that results in inflammation of
41
the small intestinal mucosa in genetically predisposed individuals caused by
42
intake of gluten proteins from wheat, rye, and barley [1]. The prevalence of
43
celiac disease is about 0.5 – 2% and is still increasing in Western parts of the
44
world as well as in developing countries [2,3]. CD-patients can present
45
symptoms such as diarrhea, abdominal pain, constipation, weight loss, and
46
dermatitis
47
asymptomatic presentation without gastrointestinal symptoms resulting in a high
48
number of undiagnosed cases [4][2,3]5-7]. For CD-patients still the only cure is
49
to adhere to a strict gluten-free diet. However, gluten proteins are increasingly
50
being applied in all kind of food products because of their interesting features
51
[8]. Therefore, it is very important that food labeling is accurate and reliable. The
52
reduced intake of gluten peptides and proteins containing CD-epitopes by all
53
consumers, including diagnosed and still undiagnosed CD-patients, will reduce
54
symptoms and may reduce the prevalence of CD.
however,
increasingly
more
patients
show
Ac ce
pt
ed
M
an
herpetiformis,
us
cr
ip t
39 40
55
Gluten proteins, i.e. gliadins and glutenins, represent the main part of
56
storage proteins in wheat, are insoluble in water and contain high percentages of
57
the amino acids proline (P) and glutamine (Q). Gliadins form a large protein
58
family in which /-, -, and -gliadins can be distinguished (~30 to ~80 kDa by
59
acidic-PAGE) [9,10]. Glutenins can be subdivided into low-molecular weight
60
glutenin subunits (LMW-GS; ~30 to ~70 kDa by SDS-PAGE) and high-molecular
61
weight
62
[11,12].[13][14]
glutenin
subunits
(HMW-GS;
~80
to
~130
kDa
by
SDS-PAGE)
3 Page 3 of 49
Bread wheat is hexaploid and contains three different genomes (A, B and D)
64
evolved from three different grass species. Therefore, many gluten protein
65
encoding genes are present and copy numbers for -gliadins can range between
66
100 and 150 [15,16]. [17,18]CD symptoms can be caused by many different CD-
67
epitopes present in wheat cultivars and wheat-derived food products. So far, 24
68
epitopes have been defined from wheat gluten proteins that give a T-cell
69
response in CD-patients [19]. These glutamine and proline rich epitopes are
70
derived from all different classes of gliadin and glutenin proteins. The
71
immunogenic epitopes present in -gliadins are Glia-2, Glia-9, and Glia-20 of
72
which the most immunogenic epitopes are Glia-2 and Glia-9 [20,21]. It has
73
been shown by Shan et al. [22] that a large 33-mer from wheat gluten is
74
resistant against human intestinal proteases and therefore can be present in the
75
small intestine. The 33-mer is composed of overlapping immunodominant Glia-2
76
and Glia-9 epitopes that can also be recognized separately and stimulate T-cells
77
in CD [20,21]. This 33-mer can be present in -gliadins that are encoded by the
78
Gli-2 locus (Gli-D2) on the short arms of chromosome 6 of the D-genome in
79
wheat (see Fig. 1) [23-25]. The same holds for the Glia-2 epitope that is
80
present in most -gliadins encoded by genes present at only the D-genome in
81
hexaploid wheat. The -gliadins encoded by the D-genome can also contain
82
derivatives of the 33-mer (a 19-mer and a 26-mer) that contain the Glia-2 and
83
Glia-9 epitopes. The -gliadins encoded by the A-genome in hexaploid and
84
tetraploid wheat contain only a 13-mer containing the Glia-9 epitope. Most -
85
gliadins encoded by both the A- and the D-genome contain the Glia-20 epitope.
86
The -gliadins encoded by the B-genome contain neither the Glia-2/9 epitopes
87
nor the Glia-20 epitope [24,26].
Ac ce
pt
ed
M
an
us
cr
ip t
63
4 Page 4 of 49
Food products can be labeled gluten-free if they contain less than 20 ppm
89
gluten. So far, the only approved test by the Codex Alimentarius to detect the
90
presence of gluten in food products and that can be used to label products
91
‘gluten-free’ is the R5-enzyme-linked immunosorbent assay (R5-ELISA) [27-29].
92
In general, the functionality of ELISA tests depends on the extraction protocol
93
[30-32], reference material for calibration [33], and the detecting antibody. [30-
94
32][33]The latter is one of the limitations of the R5-ELISA that makes use of a
95
monoclonal antibody that detects gluten in general and not specifically CD
96
stimulatory epitopes. The R5-antibody is a-specific and recognizes several, five
97
amino acids long, sequences present in gluten proteins from wheat, rye, and
98
barley, but not from soy, oats, corn, rice, millet, teff, buckwheat, quinoa, and
99
amaranth. This difference in specificity is essential for gluten-free testing.
100
Despite this ability, the short length of the recognition sequence might lead to
101
inaccurate results. Antibodies can recognize sequences as short as five amino
102
acids. This in contrast to the size of the recognition sequence of a CD-epitope
103
recognized by human T-cells in CD, which is at least nine amino acids in length
104
[34]. Therefore, R5-ELISA might result in an overestimation of the amount of
105
true T-cell epitopes present that stimulate CD.
106
[19][20,21][22][20,21][23-25][24,26] Simultaneous detection and quantification
107
of more than one CD-epitope is not possible by immunoblotting using a single
108
blot. By ELISA this is only possible when a multiplexing method with multiple
109
specific
110
chromatography-multiple reaction monitoring mass spectrometry (LC-MRM/MS)
111
method that can detect and quantify multiple CD-peptides in a single short run
112
will have a huge advantage over the existing methods.
Ac ce
pt
ed
M
an
us
cr
ip t
88
antibodies
is
used.
Therefore,
the
development
of
a
liquid
5 Page 5 of 49
LC-MRM/MS is a rapidly emerging method as an alternative to antibody based
114
protein quantification [35-37]. In this approach, the proteins to be quantified are
115
first digested with a specific protease, after which proteotypic peptides are
116
analyzed by MS in MRM mode. Using either a triple quadrupole or a quadrupole-
117
ion trap instrument, peptides are identified and quantified by monitoring several
118
transitions for each peptide. MRM allows sensitive, accurate and reproducible
119
quantification of the peptides and corresponding proteins.So far, several LC-MS
120
detection methods, aiming at the detection of different immunogenic peptides,
121
have been developed and described [38-43]. The developed methods differ in
122
the choice of the selected peptides for quantification and the protease treatments
123
to release the peptides from the gluten proteins.
an
us
cr
ip t
113
Here, we describe the development of an LC-MRM/MS method to quantify
125
individual immunogenic peptides in a gluten protein extract from wheat kernels,
126
using chymotrypsin digestion to release the peptides. We focused on detection of
127
those peptides that have been proven to be the most immunodominant in CD
128
[44-48]. This is in contrast to previously developed LC-MS methods, that focused
129
on the most intensely MS responding peptides. Besides immunodominant
130
peptides, we also incorporated some peptides with amino acid substitutions that
131
make the immunogenic epitopes inactive and are thereby considered safe for
132
CD-patients [24,26]. LC-MS allows detection of these amino acid substitutions,
133
which is not possible by antibody detection as used in ELISA. In addition, we
134
selected peptides that can be used for quantification of the total gliadin protein
135
amount. Furthermore, we show that with this method, wheat varieties,
136
containing different amounts of the epitopes, can be analyzed. This opens the
137
possibilities for a range of future applications such as support for the
138
development of reduced immunogenic wheat lines, the analysis of foods for the
Ac ce
pt
ed
M
124
6 Page 6 of 49
presence of immunogenic CD-epitopes, or the screening of body fluids of
140
allergenic patients for analysis of the presence of selected marker peptides.
Ac ce
pt
ed
M
an
us
cr
ip t
139
7 Page 7 of 49
141
142
2. Materials and methods
143
2.1. Wheat samples
145
ip t
144 Two hexaploid wheat varieties, Minaret (obtained from the Centre for Genetic Resources, CGN, the Netherlands) and Toronto (obtained from Limagrain,
147
Lelystad, The Netherlands), and one tetraploid wheat variety (Dibillik sinde,
148
obtained from CGN) were used for gluten protein extraction followed by LC-MS
149
analysis.
an
us
cr
146
150 2.2. Extraction of gluten proteins
M
151 152
Gluten proteins were extracted from wheat grains according to Van den
154
Broeck et al. [49]. Grains were ground and gluten proteins were extracted from
155
50 mg wheat flour by addition of 0.5 ml of 50% (v/v) aqueous iso-propanol with
156
continuous mixing (MS1 Minishaker, IKA Works, Inc.) at 1000 rpm for 30 min at
157
room temperature, followed by centrifugation at 10,000 rpm for 10 min at room
158
temperature. The residue was re-extracted twice with 50% (v/v) aqueous iso-
159
propanol, 50mM Tris-HCl, pH 7.5 containing 1% (w/v) DTT, for 30 min at 60°C
160
with mixing every 5 to 10 min followed by centrifugation at 10,000 rpm for 10
161
min at room temperature. After addition of each next extraction solution, the
162
residue was resuspended followed by sonication for 10 min in an ultrasonic bath
163
(Branson
164
supernatants were combined and considered the gluten protein extract. The
165
protein content was quantified using the Biorad Protein Assay (Bio-Rad
Ac ce
pt
ed
153
3510,
Branson
Ultrasonics
Corporation).
The
three
obtained
8 Page 8 of 49
166
Laboratories), based on the Bradford dye-binding procedure, according to
167
manufacturer's instruction with BSA as a standard.
168 169
2.3. Digestion conditions
ip t
170 Gluten proteins from the gluten extracts were precipitated with four volumes
172
of ice cold 100% acetone and kept overnight at -20C. After centrifugation for 10
173
min at 14,000 rpm and 4C, the resulting pellet was resuspended in 8 M urea to
174
keep the proteins denaturated. Proteins were reduced in 5 mM DTT for 45 min at
175
60C, followed by alkylation in 15 mM iodoacetamide for 60 min at room
176
temperature in the dark. Aliquots containing 10 g of protein were four fold
177
diluted with 100 mM Tris-HCl (pH 8.0) /10mM CaCl2, to a final concentration of 2
178
M urea. Different experiments were performed to obtain optimal conditions for
179
release of the selected peptides. Proteins were digested with chymotrypsin from
180
bovine pancreas (Roche) and shaken at 25C using three different ratios of
181
protease:protein (1:5, 1:10 and 1:20). The digestion time was varied from 10
182
min up to 22 hours. The proteolytic activity was quenched by addition of
183
trifluoroacetic acid (TFA; 0.5% final concentration and the acidified sample was
184
desalted using solid phase extraction (SPE). Supelco LC-18 1 ml SPE columns
185
were washed with 1 ml 95% acetonitrile (ACN) and equilibrated with 1 ml 2%
186
ACN, 0.1% TFA. Subsequently, the digests were added to the columns and after
187
washing with 1 ml 2% ACN, 0.1% TFA, peptides were eluted with 1 ml 50% ACN,
188
0.1% TFA. Solvents were evaporated using a SpeedVac vacuum concentrator and
189
peptides were dissolved in either 20 µl 0.1 M ammonium formate (AF) or 40 µl
190
5% ACN, 0.1% formic acid (FA).
Ac ce
pt
ed
M
an
us
cr
171
9 Page 9 of 49
191
The formation of peptides P1-P9 was finally studied after both 4 and 22 hours of
192
digestion with chymotrypsin in a 1:5 enzyme to protein ratio.
193
2.4. Marker peptides and preparation of peptide reference solution
194 Nine peptides from gluten proteins (P1-P9) were synthesized and purchased
196
from ProteoGenix (France). An overview of the sequences of the peptides is
197
given in Table 2. For each of the peptides, a stock solution of 10 mM (based on
198
the yield and corrected for purity percentage indicated by the supplier) was
199
prepared in 100% MeOH and stored in aliquots at -20C. The individual stock
200
solutions were mixed to obtain equimolar cocktails of the peptides that were
201
further diluted with MilliQ water to a final concentration of 20 µM in 10% MeOH.
202
Calibration standards were prepared by dilution of the 20 µM cocktail solution in
203
a matrix of tryptic digest of a gluten extract of hexaploid wheat variety Toronto
204
(25 ng/µl 5% ACN, 0.1% FA) or a tryptic digest of bovine serum albumin (Sigma
205
Aldrich) (25 ng/µl 5% ACN, 0.1% FA) .
ed
M
an
us
cr
ip t
195
Calibration standard curves were measured using 20, 50, 100, 200, 500,
207
1000, 2000, 5000, and 10000 fmol per peptide loaded on the reversed phase
208
(RP) TS3 1.0 mm × 150 mm analytical RP column (Waters, Milford, MA, USA).
210 211
Ac ce
209
pt
206
2.5. On-line 2D nano LC-MS/MS configuration for untargeted approach
212
Nanoscale LC separation of complex peptide mixtures was performed using
213
the 2-D nanoAcquity UPLC system on-line coupled to a Synapt HDMS Q-TOF MS
214
instrument (Waters, Milford, MA, USA) as described [50,51]. For each run, 4 μL
215
digest was injected (partial loop method) in 20 mM AF (pH 10) on the first RP
216
column (X-bridge BEH130 C18, 5 μm, 300 μm × 50 mm, Waters). Elution from
10 Page 10 of 49
this column was stepwise under high pH and ultrahigh pressure at 2 μL/min.
218
Samples were eluted in four fractions of 13, 18, 25, and 65% ACN in 20 mM AF
219
(pH 10). Peptide eluates were on-line diluted with an excess of 0.1% FA in water
220
at a flow rate of 20 μL/min to reach a ten-fold dilution before being trapped on a
221
C18 (2G-V/M Symmetry 5 μm) trap column (180 μm × 20 mm). Finally, peptides
222
were
223
75 μm × 150 mm analytical RP column (Waters) at 0.3 μL/min with a gradient
224
using 0.1% FA in water as eluent A and 0.1% FA in ACN as eluent B. The
225
separation was carried out using 5% B for 1 min, 10% B for 2 min, 10–40% B
226
over 62 min and 40–85% B over 9 min. After 6 min of rinsing with 85% B and a
227
linear gradient back to 5% B over 2 min, the column was re-equilibrated at initial
228
conditions. Both the Symmetry trap column temperature and the analytical
229
column temperature were maintained at 55 °C by a built-in column heater. Mass
230
spectrometric analyses were performed in positive mode using ESI with a
231
NanoLockSpray source. Eluates were immediately sprayed into a Q-TOF Synapt
232
G1 (Waters). As lock mass, [Glu1]-fibrinopeptide B (1 pmol/L) was delivered
233
from a syringe pump (Harvard Apparatus, USA) to the reference sprayer of the
234
NanoLockSpray source at a flow rate of 0.2 μL/min. The lock mass channel was
235
sampled every 30 sec. LC-MS data were collected from the Synapt G1 operating
236
in either MS/MS or MSE mode for data-dependent acquisition (DDA) or data-
237
independent acquisition (DIA). For MS/MS, the three most intensive single or
238
multiple charged ions eluting from the column were selected for fragmentation.
239
The eluting peptide ions were detected in the MS survey scan (0.6 sec) from a
240
m/z range of 300 to 1400 and MS/MS scan range from 50 to 2000 m/z. A
241
dynamic exclusion window was set at 60 sec.
from
this
column
and
separated
on
a
HSS
T3
1.7 μm,
Ac ce
pt
ed
M
an
us
cr
eluted
ip t
217
242 11 Page 11 of 49
243
2.6. Triple quadrupole settings for targeted analysis
244 245
Targeted LC-MRM/MS analyses were performed on an Acquity UPLC system
246
coupled to a Xevo TQ-S triple quad instrument (Waters, Milford, MA, USA). For
247
each run, 5 μL and 10 μL sample, corresponding to 250 and 500 ng protein
248
digest,
249
1.0 mm × 150 mm analytical RP column (Waters, Milford, MA, USA) at 60°C with
250
a flow rate of 0.15 mL/min. Peptides were eluted using an ten minute LC-method
251
(starting 1 min after injection): 1% B for 0.5 min, linear gradient of 1–42% B
252
over 5.5 min, linear gradient 42-90% B over 1 min, after 1 min of rinsing with
253
90% B, a linear gradient 90-20% for 1 min and back to 1% B over 1 min. The
254
column was re-equilibrated at initial conditions for 1.9 min, with 0.1% FA in
255
water (LC-MS grade) as eluent A and 0.1% FA in 100% ACN as eluent B.
(partial
loop
method)
and
separated
on
a
TS3
ip t
injected
M
an
us
cr
was
MRM mass spectrometric analyses were performed in positive ion mode. A
257
scheduled MRM method was used with MRM detection window set to 0.5-1.6 min
258
per peptide, the duty cycle was set to automatic and dwell times were minimal 5
259
msec. The transitions and detection windows of the marker peptides are listed in
260
Table 2. Cone voltage was set to 35 V.
262 263 264
pt
Ac ce
261
ed
256
2.7. Data analysis
2.7.1. Untargeted MS data analysis
265
DDA data obtained from the MS/MS analysis with the Synapt G1 were
266
processed by ProteinLynx Global Server software (PLGS version 2.5, Waters)
267
for peak detection. Produced peak lists were used for Mascot search (Matrix
268
Science) to identify peptides. Searches were performed using the Uniprot
12 Page 12 of 49
database (07/2013), green plants. Parameters were: Enzyme selection is
270
chymotrypsin, maximum of three missed cleavages; carbamidomethyl C as fixed
271
modification; deamidated NQ and oxidation M as variable modifications;
272
monoisotopic mass values; protein mass unrestricted; peptide mass tolerance
273
±50 ppm; fragment mass tolerance ± 30 mmu; peptide charges 2+, 3+, and 4+.
275
Raw MSE (DIA) data obtained were processed by Progenesis QI software to visualize chromatograms of peptide resolution per fraction.
cr
274
276 2.7.2. Targeted MS data analysis
us
277
ip t
269
To analyze the targeted MRM data from TQ-S, Skyline software (MacCoss Lab
279
at the University of Washington) was used [52]. The raw data were imported into
280
Skyline and peak detection and quantitation was performed on the given set of
281
transitions. Peak areas were exported as a transition result report and further
282
processed in Excel. Per injection the sum of the area of the multiple transitions
283
per peptide was used as quantitative value. In Skyline the ‘area’ value is the
284
peak
285
subtracted. Calculation of linear regression was performed on the summed area
286
for the dilution series of the reference peptides. We applied an intercept=0
287
setting, as it gave the highest R2 value (between 0.974 and 0.994). For the
288
gluten digests the peak areas were also summed per peptide. The peptide
289
amount was calculated using the slope coefficient obtained from the linear
290
regression of standard curves. Femtomole (fmol) levels were calculated to g
291
amount of injected gliadin protein, using the average Mw of gliadins (32285.5
292
Da).
above
the
background and
therefore
background is already
Ac ce
pt
area
ed
M
an
278
293
13 Page 13 of 49
294
2.8. SDS-PAGE and immunoblotting
295 Equal amounts of gluten proteins were separated on SDS-PAGE gels (10%)
297
as described [53] using a Hoefer SE 260 mighty small II system (GE Healthcare)
298
followed by staining with PageBlue (Fermentas). For immunoblotting, proteins
299
were blotted onto nitrocellulose (0.2 m, Bio-Rad Laboratories), omitting
300
methanol from the blotting buffer, using a Mini Trans-Blot Cell (Bio-Rad
301
Laboratories) at 100V for 1 h. B Blots were stained using a MemCode
302
Reversible Protein Stain Kit for Nitrocellulose Membranes (Fisher Scientific) prior
303
to incubation with monoclonal antibodies (mAbs). Blots were incubated as
304
described [54] using mAbs specific for T-cell stimulatory epitopes Glia-9 and
305
Glia-20 [55-57]. Antibody binding to the blots was visualized by staining for
306
alkaline
307
tetrazolium (NBT) and 5-bromo-4-chloro-3-indolyl phosphate (BCIP) (Sigma-
308
Aldrich). The gluten protein extract of ‘Toronto’ was used on each separate
309
immunoblot as an ‘inter-gel’ control.
an
us
cr
ip t
296
secondary
antibody,
using
Nitro
Blue
pt
ed
M
phosphatase-conjugated
Blots were scanned using a Bio-Rad GS-710 Calibrated Imaging Densitometer
311
(Bio-Rad Laboratories) and saved as TIFF images. Pixel intensities were
312
calculated per lane using Quantity One software (Bio-Rad Laboratories). Relative
313
intensities differed specifically per mAb used, but were normalized to values
314
obtained
Ac ce
310
for
the
‘inter-gel’
control.
14 Page 14 of 49
315
316
3. Results
317 3.1. Identification of gluten proteins in wheat grains by on-line 2D LC-MS/MS
ip t
318 319
In a previous genomics study, the genetic diversity present in -gliadin
321
transcripts and genes has been investigated showing that CD immunogenic
322
epitopes are encoded by the A- and D-genome and not by the B-genome. Fig. 1
323
shows the protein sequence variants of the most frequently found expressed
324
sequence tags (ESTs) originating from -gliadins in wheat, as present in the
325
NCBI-UniGene database. The amino acid sequences derived from the different
326
translated contig sequences have been aligned for the CD-epitope containing
327
region. Sequence variation outside this region is not displayed here. Sequence
328
variants are grouped per genome origin (Gli-A2, Gli-B2 and Gli-D2). The Glia-2
329
and Glia-9 epitopes are marked in yellow and the Glia-20 epitope is marked in
330
green. As mentioned before, these epitopes are encoded by both the A- and D-
331
genome, but not by the B-genome. The large protease resistant 33-mer,
332
containing multiple CD-epitopes, is marked in blue and the 19-mer and 26-mer
333
derivatives, encoded by only the D-genome, are marked in yellow. The -gliadins
334
encoded by the A-genome in hexaploid and tetraploid wheat contain only the 13-
335
mer peptide sequence containing the Glia-9 epitope.
Ac ce
pt
ed
M
an
us
cr
320
336
To study the -gliadins present in wheat varieties at protein level, a
337
previously developed two-step extraction protocol using isopropanol was applied
338
to efficiently extract the gluten proteins [49]. Total gluten concentrations in our
339
extracts were 1.02 g/l (30.45 g/mg flour), 0.72 g/l (21.45 g/mg flour), 15 Page 15 of 49
and 0.80 g/l (23.85 g/mg flour) for ‘Toronto’, Minaret’, and ‘Dibillik sinde’,
341
respectively. For LC-MS/MS analysis of the gluten fraction, proteins were first
342
digested with a specific protease in order to generate peptides with a length that
343
was amenable for proper LC separation and MS detection. Routinely, trypsin is
344
used for enzymatic digestion of complex protein mixtures, which cleaves proteins
345
C-terminal to the basic amino acids arginine (R) and lysine (K). However, due to
346
the low number of these amino acids in gluten proteins, the digestion of the
347
gluten proteins into smaller peptides by trypsin is not feasible. Therefore,
348
chymotrypsin from bovine pancreas was chosen for digestion of the gluten
349
proteins. Chymotrypsin cleaves C-terminal to phenylalanine (F), tyrosine (Y), and
350
tryptophan (W) unless the next residue is a proline (P), and at a lower rate C-
351
terminal to leucine (L), methionine (M), alanine (A), aspartic acid (D), and
352
glutamic acid (E). In Fig. 1, the chymotryptic cleavage sites present in the CD-
353
epitope containing domain of the different -gliadins are indicated in black (F, Y,
354
and W). From this result, we concluded that all -gliadin proteins encoded by the
355
A-genome will, theoretically, yield at least one peptide containing the Glia-9
356
epitope when digested with chymotrypsin (Fig. 1, indicated in yellow). This also
357
holds for -gliadin sequences encoded by the D-genome that will result in
358
different sized peptides (19-mer, 26-mer, and 33-mer) containing the Glia-2/9
359
epitopes (Fig. 1, indicated in yellow) or sequence variants thereof.
Ac ce
pt
ed
M
an
us
cr
ip t
340
360
The use of chymotrypsin as an optimal protease to generate CD-epitope
361
containing peptides was confirmed by digestion of the gluten protein extracts
362
from hexaploid wheat varieties Toronto and Minaret. The generated peptides
363
were separated into four fractions by 2D nano UPLC and on-line analyzed using
364
data-dependent MS/MS mode and data-independent MSE mode. Supplementary
365
Fig. S1 shows the resolution of the detected peptides present in the four different 16 Page 16 of 49
fractions. Supplementary Table S1 summarizes the different gliadin and glutenin
367
proteins identified in the gluten extract of wheat variety Toronto. Whereas the
368
majority of the identified peptides originated from -gliadins, also numerous
369
peptides from -gliadins and LMW-GS have been detected. The - and -gliadins
370
and the LMW-GS were classified into families according to the presence of at
371
least one unique identified (proteotypic) peptide in their sequence. In this way
372
we could discriminate between 24 different -gliadin families originating from
373
Triticum and Aegilops species (Table 1), 15 different -gliadin families, 7 different
374
LMW-GS families, and a single -gliadin and HMW-GS family.
us
cr
ip t
366
The peptide identifications covering the sequence area containing CD-
376
epitopes (displayed in bold) in -gliadins are displayed in Table 1. Different
377
sequence variants of the peptides have been identified and are listed as well.
378
From these identified peptides, we have selected the most frequently occurring
379
peptides as marker peptides for LC-MRM/MS detection, indicated P1 to P9.
ed
M
an
375
Peptides P1 to P4 occurred with peptide P6 in the -gliadin families, whereas
381
the -gliadin peptide containing the 33-mer (P5) was always present in
382
combination with the peptide containing the Glia-20 epitope having a P to S
383
substitution (P7) (Fig. 1; Table 1). Peptide P6 (in combination with P7)
384
containing the Glia-20 epitope occurred in nearly all -gliadin families encoded
385
by the A- and D-genome, but was absent in -gliadin families encoded by the B-
386
genome (Table 1, no. 8, 14, and 16). Peptides from the same region encoded by
387
the B-genome were more difficult to be identified because of the absence of
388
chymotrypsin digestion sites in this region. Instead, three different -gliadin
389
families from the B-genome have been identified (Table 1, no. 8, 14, and 16)
390
with proteotypic peptides outside this region of the gliadin sequence (see
391
Supplementary Table S1). Peptides P8 and P9 appeared in nearly all -gliadin
Ac ce
pt
380
17 Page 17 of 49
392
families, where P9 is the product of a missed cleavage and also contains peptide
393
P8. We reasoned that a targeted LC-MS detection method for the selected
395
peptides would be able to detect the presence of both the different CD-epitopes
396
individually (P1 to P7) and the sum of multiple gliadin proteins (P8 and P9).
ip t
394
397 3.2. Quantitative LC-MRM/MS assay development
cr
398 399
After identification of the different -gliadin peptides in wheat varieties
401
Toronto and Minaret by on-line 2D LC-MS/MS, a targeted MS-based assay was
402
developed for absolute quantification of the selected marker peptides P1-P9
403
(Table 1). In this study a triple quadrupole instrument (Xevo TQ-S) was used to
404
quantify CD-epitopes containing -gliadins in gluten extracts. Non-labelled
405
synthetic marker peptides P1-P9 were used both for the development of the MRM
406
assay and as standards for the generation of calibration curves for absolute
407
quantification of the corresponding -gliadins.Note that P9 has a missed cleavage
408
containing the P8 peptide sequence. Both peptides, P8 and P9, were present in
409
17 of the 24 -gliadin families (Table 1).
Ac ce
pt
ed
M
an
us
400
410
First, the optimal transitions for quantification and confirmation of each of
411
these peptides in a gluten digest were determined. Since the sensitivity of a
412
triple quadrupole MS is critically dependent on ionization conditions and tuning
413
parameters, a mix of all nine synthesized peptides was loaded on a RP column
414
and eluted with a short gradient to optimize cone voltage and collision energy of
415
the TQS. Table 2 summarizes the most optimal parameters achieved for each of
416
the marker peptides, and the transitions used for quantification and confirmation.
417
For most precursor peptides, one m/z value was chosen as the predominant 18 Page 18 of 49
charge state for selection in the first quadrupole. To obtain the most sensitive
419
assay, the collision energy was optimized per precursor to generate maximal
420
fragment ion intensities. In the third quadrupole, the most intense transitions
421
derived from one precursor ion were selected for quantification. In Table 2, the
422
peak retention time measured for each of the marker peptides is shown using the
423
LC conditions as described in section 2.6. (10 min gradient). For each peptide a
424
retention window of 0.5 to 2 min around its peak elution time was used for
425
acquisition in a time-scheduled MRM method. Scheduled MRM instructed the TQS
426
to measure a selected set of transitions during a particular retention time,
427
providing longer dwell times to monitor the target ions and thus improved signal
428
to noise ratios for most of the transitions. A prerequisite in a scheduled MRM
429
analysis is that the chromatography is stable and reproducible within the chosen
430
retention time window.
M
an
us
cr
ip t
418
Since most of the marker peptides contained overlapping sequences, several
432
fragment ions from different marker peptides had identical masses, as listed in
433
Table 2. The specificity for detection was achieved by using unique precursor to
434
product transitions and using scheduled MRM acquisitions around the elution time
435
of the precursor peptide. Supplementary Fig. S2 shows the chromatograms
436
obtained for a mixture of the nine marker peptides and for a ‘Toronto’ gluten
437
protein digestion.
439
pt
Ac ce
438
ed
431
3.3. Calibration curves of peptide standards
440 441
To determine the concentration range at which the nine marker peptides
442
showed a linear MRM response, calibration curves were generated by injection of
443
synthetic peptides in a range of 20-10000 fmol. Peak areas above background 19 Page 19 of 49
were determined by summing the multiple transitions per peptide using Skyline
445
software. Initially, a deviation from linearity was observed for most of the
446
peptides when the injected amount decreased below 100 fmol. Although different
447
glass vials and ‘low binding’ plastic tubes were tested to diminish the loss of
448
peptides when handling these low amounts of peptide, the calibration curves in
449
the lower concentration range remained quite variable. This problem could be
450
solved by preparing peptide dilutions in a matrix of a protein digest that did not
451
contain the marker peptides. When the dilutions of the peptides were prepared in
452
a matrix of tryptic digest from either bovine serum albumin (BSA) or gluten
453
protein, calibration curves showed excellent linearity in a range of four orders of
454
magnitude. The R square correlation values of the nine calibration curves varied
455
between 0.969 and 0.994, as listed in Table 2. In Supplementary Fig. S3, an
456
overview of the calibration curves for all the nine peptides is given.
M
an
us
cr
ip t
444
459
3.4. Digestion conditions
pt
458
ed
457
Prior to quantitative LC-MRM/MS analysis of the nine -gliadin peptides P1-P9
461
in protein digests from gluten extracts (see below), the reproducibility of the
462
digestion procedure with chymotrypsin was determined. Digestion was optimized
463
using gluten extracts from hexaploid wheat variety Toronto and tetraploid wheat
464
variety Dibillik sinde. In contrast to trypsin, which is mostly used in proteomics
465
analysis, chymotrypsin is not an endpoint protease. In addition to its primary
466
cleavage specificity it displays secondary, low affinity activity leading to further
467
breakdown of digested peptides. This means that an optimal digestion time and
468
enzyme to substrate ratio needs to be determined. Therefore, the progress of
469
chymotryptic digestion of gluten proteins dissolved in 2 M urea was followed by
Ac ce
460
20 Page 20 of 49
sampling at different time points of incubation at 25C. The ratio of chymotrypsin
471
to protein was also varied, i.e. 1:20, 1:10, and 1:5. Five peptides (P1-P5)
472
containing Glia-2/9 epitopes had reached a maximum MRM signal already at, or
473
even well-before, 4 h of incubation (Fig. 2A). Longer incubation did not result in
474
a significant decrease of the signals, and indicated that no or limited secondary
475
chymotrypsin activity to peptides P1-P5 took place (Fig. 2B). Peptides P6 and P7,
476
however, showed a much slower release and maximum MRM responses were
477
reached only after 22 h of incubation and an extra addition of chymotrypsin of an
478
enzyme to protein ratio of 1:5 (Fig. 2C and 2D). Finally, formation of all peptides
479
(P1-P9) was studied after 4 and 22 hours of incubation with chymotrypsin in a
480
1:5 enzyme to protein ratio. The results, shown in Supplementary Fig. S5,
481
confirmed the previous findings that P1-P5 reached a maximum response after
482
only 4 hours of digestion, while formation of P6 and P7 reached its maximum
483
after 22 hour of incubation. Also peptide P8, present in most -gliadin families,
484
showed a maximum response after 22 h. The slow formation of P8 is related to
485
the presence of a missed cleavage site, as is shown by the P9 peptide. The signal
486
of peptide P9 decreased over time, reaching a level just above background after
487
22 h when it was fully digested into P8. The reason why peptides P6, P7 (both
488
containing the Glia-20 epitope), and P8were released very slowly during
489
chymotrypsin digestion of -gliadins is not clear. In conclusion, these results
490
indicated that optimal digestion time to detect the CD-epitope containing
491
peptides is 22 hours and this was used for analysis of wheat varieties.
Ac ce
pt
ed
M
an
us
cr
ip t
470
492 493
3.5. Quantification of -gliadins in three wheat varieties by LC-MRM/MS
494
21 Page 21 of 49
After optimization of the digestion conditions, we analyzed protein extracts
496
from three wheat varieties with the developed MRM method: the modern wheat
497
variety Toronto (AABBDD), the landrace variety Minaret (AABBDD), and the
498
tetraploid variety Dibillik sinde (AABB). For LC-MRM/MS analysis two amounts
499
(250 ng and 500 ng) of digested gluten sample were injected in duplicate. Peak
500
areas (minus background) were determined by summing the selected transitions
501
per peptide using Skyline software. The total peak areas were converted to fmol
502
amounts of peptide, using the external calibration curves obtained from a dilution
503
series of the synthetic peptides. The fmol amount of peptide was converted to
504
the correlating amount of -gliadin per microgram digested gluten protein
505
extract using the average Mw of -gliadins (32285.5 Da). This is displayed in
506
Table 3 and Supplementary Fig. S4. Samples from two different digestion times
507
(4 h and 22 h) were analyzed and the 22 h sample was interpreted as completely
508
digested. In Table 3, an overview is given of the quantified peptides in the three
509
wheat varieties. In the tetraploid variety Dibillik sinde, the marker peptide P8
510
and the two peptides P1 and P6 containing the Glia-9 and Glia-20 CD-epitopes
511
were detected. These peptides, P1 and P6, are derived from -gliadins from the
512
A-genome. In the hexaploid varieties ‘Toronto’ and ‘Minaret’ all selected peptides
513
were detected (P9 was only detected in 4 h digests). The standard deviation of
514
the detected peptides was between 3% and 15% for most peptides, indicating a
515
good reproducibility of detection. The amounts of peptides P1 to P5 containing
516
the Glia-2/9 epitopes, could be related to the total ‘load’ of gliadin (P8+P9).
517
This allows comparison of the presence of these CD-epitopes in different wheat
518
varieties and gains insight into the immunogenicity (see Supplementary Fig. S5).
519
Even though peptides P8 and P9 did not cover the total -gliadin content, as will
520
be further discussed in the general discussion, they could be used to indicate the
Ac ce
pt
ed
M
an
us
cr
ip t
495
22 Page 22 of 49
amount of -gliadins present. Results showed that ‘Toronto’ contained the
522
highest amount of -gliadins per total gluten amount and ‘Dibillik sinde’ the
523
lowest amount (Table 3, Supplementary Fig. S5), whereas the total gluten
524
amount for ‘Dibillik sinde’ was higher compared with the total gluten amount for
525
‘Minaret’.
ip t
521
The total amount of Glia-2/9 and Glia-20 epitopes could be compared with
527
results obtained from previous studies where we performed large scale
528
immunoblotting experiments with antibodies against CD-epitopes Glia-9 and
529
Glia-20 (see Supplementary Fig. S6). While equal amounts of total gluten
530
protein extract were used for immunoblotting, the overall signal intensities
531
obtained with mAbs against the Glia-9 (Supplementary Fig. S6B) and Glia-20
532
(Supplementary Fig. S6C) epitopes were different for the three wheat varieties,
533
which indicated that these epitopes were present in different amounts. In Fig. 3
534
an overview is given of the LC-MRM/MS results and shows the sum of P1-P5
535
(Glia-2/9) and P6-P7 (Glia-20) using the average Mw of -gliadins (32285.5
536
Da). Peptide P1, containing the Glia-9 epitope and present in all three varieties,
537
was highly present in both the modern hexaploid variety Toronto and the
538
tetraploid variety Dibillik sinde (Table 3). The total amount of P1, the only
539
peptide present in the tetraploid wheat variety Dibillik sinde encoded by the A-
540
genome, was about half (56%) of the amount present in ‘Toronto’ (Fig. 3A). This
541
result was different from the immunoblotting experiments where the signal in
542
‘Dibillik sinde’ using the Glia-9 mAb (Supplementary Fig. S6B) was about three
543
quarters (75%) compared with the signal in ‘Toronto’, while equal amounts of
544
total gluten protein extract were used for immunoblotting. This was probably
545
caused by a-specific signals by the Glia-9 mAb. With LC-MRM/MS, it was
546
measured that in ‘Minaret’ the total amount of P1 to P5 was about half (52%) of
Ac ce
pt
ed
M
an
us
cr
526
23 Page 23 of 49
547
that in ‘Toronto’ (Fig. 3A). Results from immunoblotting using the Glia-9 mAb
548
(Supplementary Fig. S6B) showed in ‘Minaret’ more than half the signal
549
compared with ‘Toronto’, which was probably also caused by a-specificity of the
550
Glia-9 mAb. The amount of Glia-20 epitope (P6) in ‘Dibillik sinde’ was about one third
552
(32%) of the amount of Glia-20 epitopes (P6 and P7) in ‘Minaret’ and one tenth
553
(13%) of the amount (P6 and P7) in ‘Toronto’ (Fig. 3B). In general, peptides P1
554
to P7 were present in lower amounts in ‘Minaret’ compared with ‘Toronto’ (Table
555
3, Supplementary Fig. S5). The results obtained from immunoblotting using the
556
Glia-20 mAb (Supplementary Fig. S6C) were much more comparable with the
557
LC-MRM/MS results then the results obtained for the Glia-9 mAb, probably
558
because of the higher specificity of the Glia-20 mAb. The number of protein
559
bands reacting with both mAbs was less in ‘Minaret’ and ‘Dibillik sinde’ compared
560
with
‘Toronto’.
Ac ce
pt
ed
M
an
us
cr
ip t
551
24 Page 24 of 49
561
562
4. Discussion
563 The intention of the described study is to develop a peptide targeted
565
detection method that can accurately discriminate between different CD
566
stimulating epitope sequences originating from the multitude of -gliadin
567
proteins expressed in wheat varieties. [25,58-61][60,61][49][60][24]Improved
568
and new detection methods have been under development in the meantime [38-
569
43,62]. However, developing a new approved detection method for immunogenic
570
gluten is very difficult because of the complexity of detecting and quantifying
571
only a subset of ‘harmful’ gluten peptides instead of detecting and quantifying all
572
gluten proteins. We aimed to develop a method detecting and quantifying only
573
harmful epitopes in wheat varieties that stimulate the development of CD. The
574
accuracy of the method will depend on the sensitivity of the mass spectrometer
575
and the optimal digestion conditions of the gluten proteins. It has been shown
576
before by Šalplachta et al. [63] that using trypsin for the digestion of gluten
577
proteins did not result in sufficient amounts of peptides and results in peptides
578
that have too high molecular masses. LC-MS detection methods as developed by
579
Sealey-Voyskner et al. [42] and Prandi et al. [64] make use of the proteases
580
pepsin, trypsin, and chymotrypsin to simulate the human intestinal digestive
581
tract. However, results show that immunodominant peptides are the result of
582
chymotrypsin digestion only. Gastrointestinal digestion may also be different
583
among humans. By using chymotrypsin and optimal digestion conditions this
584
possible difference is excluded.
Ac ce
pt
ed
M
an
us
cr
ip t
564
25 Page 25 of 49
A second success factor for the accuracy of the method is the detection level.
586
The sensitivity and specificity of mass spectrometry are very high and allow
587
detection of peptides containing single amino acid substitutions that can
588
differentiate between immunogenic and non-immunogenic CD-epitope containing
589
peptides. As was shown by Mitea et al. [24], P to S substitutions can cause
590
inactivation of immunogenic CD-peptides. With this knowledge and using MRM
591
mass spectrometry (LC-MRM/MS) it is possible to identify these amino acid
592
substitutions within peptides and search for wheat varieties and species that are
593
low in immunogenic CD-epitopes.
us
cr
ip t
585
It is difficult to compare ELISA read outs with results from LC-MS measuring
595
CD-immunogenic gluten. Both methods are very different in the way of detection
596
and calibration. ELISA uses indirect detection using a mAb and calibrates against
597
a broad range of proteins or peptides depending on whether sandwich or
598
competitive ELISA is being used. LC-MS detects directly the presence of peptides
599
and calibration is against peptide standards. Because of the complexity of gluten
600
proteins, there is not a single peptide present in all gluten proteins and therefore
601
specific peptides of interest should be quantified. Cressey et al. [65] described
602
the use of the method developed by Sealey-Voyksner et al. [42] in which
603
quantification is based on one peptide present in LMW-GS. Because the
604
contribution of LMW-GS to the total gluten protein content varies among
605
hexaploid wheats [66], there should be a certain calculation factor taken into
606
account to obtain the true level of ppm in food products or even in bread wheat
607
flours. The method developed by Prandi et al. [64] described quantification by
608
using a labeled peptide of which the peptide is only encoded by the A-genome in
609
wheat. Because gluten proteins encoded by the A-, B-, and D-genome are not
610
identical, also a certain correcting calculation factor should be included to obtain
Ac ce
pt
ed
M
an
594
26 Page 26 of 49
the correct value for levels of ppm. The authors described the use of three
612
marker peptides for quantification of the gliadin amount. These marker peptides
613
were derived from the N-terminus of -gliadins and were released after pepsin
614
digestion. These three marker peptides, however, are present in only 60% of -
615
gliadins from T. aestivum (Uniprot, 2014). We used two different peptides for
616
total -gliadin quantification (P8 and P9) that are located at the start of the
617
unique domain 1 in -gliadins [67] and are released after chymotrypsin
618
digestion. These peptides are present in about 64% of -gliadins from T.
619
aestivum. This percentage could be increased to 93% when another isoform of
620
these peptides is included (Table 1). This will be necessary when we directly
621
want to quantify peptides present in food products.
an
us
cr
ip t
611
For the developed LC-MS/MRM method, we decided not to use labeled
623
peptides as was argued before by Sealey-Voyksner et al. [42]. The reason was
624
because of increasing costs when many CD-epitope containing peptides need to
625
be quantified. At this moment, more than 24 different CD-epitopes are known.
ed
M
622
Tetraploid wheat, known as pasta wheat, contains the A- and the B-genome
627
and many varieties exist, ranging from old landraces to cultivated lines.
628
Especially old landraces, but also nowadays cultivated pasta wheat, may be
629
mixtures of different genotypes including hexaploid bread wheat [61,68]. Prandi
630
et al. [68] described the detection of contamination of tetraploid wheat batches
631
with hexaploid wheat using the 33-mer as a biomarker. The 33-mer is encoded
632
by -gliadins from the D-genome only. However, not all hexaploid wheat
633
varieties may contain the 33-mer [58]. Here, we describe the use of three
634
additional peptides (P2, P3, and P4) specifically encoded by the D-genome that
635
can be used to detect this type of contamination. One of these peptides (P2,
636
LQLQPFPQPQLPYPQPQPF) was detected by Prandi et al. [40,41] in tetraploid
Ac ce
pt
626
27 Page 27 of 49
637
durum wheat showing that the used durum wheat variety was probably
638
contaminated with hexaploid wheat. It is shown by Mitea et al. (see Fig. 1) [24]
639
that only peptide sequence P1 (LQLQPFPQPQLPY) is encoded by the A-genome
640
and can be present in tetraploid wheat varieties. Immunodominancy of a CD-epitope is probably determined by the speed of
642
release of the gluten peptide containing this epitope and the amount of these
643
peptides present in the small intestine [64]. Gluten peptides shown to be
644
immunodominant (33-mer and derivatives) are indeed released very fast, within
645
one hour, from the gluten extracts in our in vitro digestion. In contrast, the
646
release of the peptide containing the Glia-20 epitope takes much more time, up
647
to 22 hours, and may therefore be less important in stimulating human T-cells.
an
us
cr
ip t
641
Recent research based on detection of CD-epitopes using specific T-cell-,
649
antibody-, and DNA analyses demonstrated that the CD-immunogenicity varies
650
across gluten proteins among and within different wheat varieties and species
651
[25,58-61]. In this study, we analyzed a modern hexaploid wheat variety
652
Toronto, an older hexaploid wheat variety Minaret, and a landrace tetraploid
653
wheat variety Dibillik sinde. Selected wheat varieties have been analyzed before
654
by immunoblotting using mAbs against CD-epitopes Glia-9 and Glia-20 for
655
relative quantification of the amount of these epitopes [60,61]. However, these
656
antibodies do not recognize the complete T-cell specific epitope. The mAb
657
recognition site for the Glia-9 epitope is found in sequences of -gliadins, -
658
gliadins, and -gliadins/D-type LMW-GS obtained from wheat species [49]. The
659
mAb recognition site for the Glia-20 epitope is only found in sequences of -
660
gliadins. In -gliadin sequences, the presence of the recognition site for both
661
mAbs correlates nicely with the presence of the recognition site of the human T-
662
cell [60]. However, both the Glia-9 and Glia-20 mAbs will also recognize
Ac ce
pt
ed
M
648
28 Page 28 of 49
naturally occurring non-epitope sequences that are inactive in T-cell response
664
e.g. because of a P to S substitution present outside the mAb recognition site
665
[24]. Using these mAbs in ELISA or immunoblotting for detecting the Glia-9 and
666
Glia-20 epitopes may therefore result in overestimation of the presence of
667
immuno-active peptides. The development of an LC-MRM/MS method for direct
668
quantitative detection enables us to compare the two different methods and we
669
show that LC-MRM/MS ensures more precise detection and quantification of CD-
670
epitopes. ELISA and LC-MRM/MS could be used in combination where ELISA
671
could be used to identify the presence of (total) gluten proteins and the
672
developed LC-MRM/MS method could be used to specifically identify and quantify
673
the particular immunogenic peptides containing CD-epitopes. The detection and
674
quantification of different and individual immunogenic peptides by LC-MRM/MS
675
will need to be coupled to a different threshold level of exposure levels for CD-
676
patients.
ed
M
an
us
cr
ip t
663
The developed LC-MRM/MS method could also be used in combination with
678
RNA sequencing as described by Salentijn et al. [69]. RNA sequencing is
679
performed on developing wheat seeds to analyze which -gliadin genes are being
680
expressed. This is in contrast to LC-MRM/MS that uses mature seeds for
681
extraction of gluten proteins. The developing stage of the seed is important in
682
determining the amount of mRNA present encoding the different gluten proteins
683
[70,71] and protein composition might change during seed development [72].
684
Furthermore, the relation between mRNA level and protein level is not fully
685
correlated and can differ among genes. In addition, even though Salentijn et al.
686
[69] found novel variants of the CD-epitopes via RNA sequencing, the true
687
presence of these peptide sequence variants in mature seeds can only be shown
688
by LC-MS gluten protein analysis. The results obtained by Salentijn et al. [69] for
Ac ce
pt
677
29 Page 29 of 49
689
the tetraploid ‘Dibillik sinde’ is in accordance with our LC-MRM/MS results. Low
690
expression levels are found for genes encoding the Glia-20 epitope by RNA
691
sequencing from developing seeds and results from LC-MRM/MS show very low
692
levels of corresponding peptides in mature seeds. Combining both techniques
693
may
694
immunogenic
development
of
genetic
markers
for
breeding
of
low
ip t
the
wheat.
Ac ce
pt
ed
M
an
us
cr
enable
30 Page 30 of 49
695
696
5. Conclusions
697 Analysis of hexaploid wheat for the presence of CD-epitopes from -gliadins
699
is more complex than analysis of those present in tetraploid wheat because of
700
the complexity of the D-genome. The targeted LC-MRM/MS method using triple
701
quadrupole mass spectrometry developed in this study, specifically quantifies the
702
amounts of individual immunogenic CD-epitopes from both A- and D-genome -
703
gliadins present in hexaploid wheat. It also quantifies the isoforms present of
704
peptides encoded by the D-genome. LC-MRM/MS technology is useful for
705
identification
706
immunogenic epitopes at low femtomolar levels of detection. It is a fast and
707
sensitive method and detection and quantification are reproducible. The method
708
can be extended to any other CD immunogenic epitope to obtain quantification
709
for all epitopes present in wheat varieties and species. It will enable the selection
710
of varieties or species with low amounts of T-cell stimulatory epitopes that can
711
be used in food products to lower the ‘gluten load’ and thereby prevent
712
development
quantification
of
biomarker
peptides
containing
CD
Ac ce
pt
ed
M
and
an
us
cr
ip t
698
and
prevalence
of
CD.
31 Page 31 of 49
713
714
Acknowledgements
715 This research was partially funded by the European regional development
717
fund and the province of Gelderland and Overijssel (GO EFRO 2007-2013) and
718
the Dutch Ministry of Economic Affairs through the DLO program ‘Plant and
719
Animal for Human Health’ (KB-05-001-019-PRI) and ‘Healthy and safe food in
720
chain perspective’ (KB-15-001-016). The authors thank Noor Bas (Centre for
721
Genetic Resources, CGN, The Netherlands) and Hein de Jong (Limagrain) for
722
providing the seeds, and Elma Salentijn for providing the original figure to adapt
723
Fig.
1.
Ac ce
pt
ed
M
an
us
cr
ip t
716
32 Page 32 of 49
724
725
References
726
728 729
coeliac disease, Nat Rev Immunol. 13 (2013) 294-302. [2]
730 731
L.M. Sollid, B. Jabri, Triggers and drivers of autoimmunity: lessons from N. Reilly, P. Green, Epidemiology and clinical presentations of celiac disease, Semin Immunopathol. 34 (2012) 473-478.
[3]
ip t
[1]
cr
727
K. Mustalahti, C. Catassi, A. Reunanen, E. Fabiani, M. Heier, S. McMillan, L. Murray, M.-H. Metzger, M. Gasparin, E. Bravi, M. Mäki, The prevalence of
733
celiac disease in Europe: Results of a centralized, international mass
734
screening project, Ann Med. 42 (2010) 587-595.
736 737
S.D. Rampertab, N. Pooran, P. Brar, P. Singh, P.H.R. Green, Trends in the
an
[4]
presentation of celiac disease, Am J Med. 119 (2006) 355.e359-355.e314. [5]
J.F. Ludvigsson, A. Rubio-Tapia, C.T. van Dyke, L.J. Melton, III, A.R.
M
735
us
732
738
Zinsmeister, B.D. Lahr, J.A. Murray, Increasing incidence of celiac disease
739
in a North American population, Am J Gastroenterol. 108 (2013) 818-824. [6]
A. Rubio-Tapia, J.F. Ludvigsson, T.L. Brantner, J.A. Murray, J.E. Everhart,
ed
740
The prevalence of celiac disease in the United States, Am J Gastroenterol.
742
107 (2012) 1538-1544.
743
[7]
J.Y. Kang, A.H.Y. Kang, A. Green, K.A. Gwee, K.Y. Ho, Systematic review: worldwide variation in the frequency of coeliac disease and changes over
749
Ac ce
744
pt
741
750
477-482.
745 746
time, Aliment Pharmacol Ther. 38 (2013) 226-245.
[8]
747 748
751 752 753 754
L. Day, in G.O. Phillips, P.A. Williams (Editors), Handbook of food proteins Woodhead Publishing Limited, 2011, pp. 267-288.
[9]
J.H. Woychik, J.A. Boundy, R.J. Dimler, Starch gel electrophoresis of wheat gluten proteins with concentrated urea, Arch Biochem Biophys. 94 (1961)
[10] F. MacRitchie, D. Lafiandra, in S. Damodaran, A. Paraf (Editors), Food proteins and their applications, Dekker, New York, 1997. [11] P.R. Shewry, A.S. Tatham, in P.R. Shewry, R. Casey (Editors), Seed Proteins, Kluwer Academic Publishers, Dordrecht, 1999, pp. 11-36.
33 Page 33 of 49
755
[12] M.C. Gianibelli, O.R. Larroque, F. MacRitchie, C.W. Wrigley, Biochemical,
756
genetic,
and
molecular
characterization
of
wheat
757
component subunits, Cereal Chem. 78 (2001) 635-646.
glutenin
and
its
[13] E.A. Jackson, L.M. Holt, P.I. Payne, Characterisation of high molecular
759
weight gliadin and low-molecular-weight glutenin subunits of wheat
760
endosperm by two-dimensional electrophoresis and the chromosomal
761
localisation of their controlling genes, Theor Appl Genet. 66 (1983) 29-37.
ip t
758
[14] P.I. Payne, L.M. Holt, C.N. Law, Structural and genetic-studies on the high-
763
molecular-weight subunits of wheat glutenin. 1. Allelic variation in subunits
764
amongst varieties of wheat (Triticum aestivum), Theor Appl Genet. 60
765
(1981) 229-236.
us
cr
762
[15] O.D. Anderson, J.C. Litts, F.C. Greene, I. Characterization of ten new
767
wheat alpha-gliadin genomic clones, evidence for limited sequence
768
conservation of flanking DNA, and southern analysis of the gene family,
769
Theor Appl Genet. 95 (1997) 50-58.
an
766
[16] T.W. Okita, V. Cheesbrough, C.D. Reeves, Evolution and heterogeneity of
771
the alpha-/beta-type and gamma-type gliadin DNA sequences, J Biol
772
Chem. 260 (1985) 8203-8213.
M
770
[17] G. Branlard, M. Dardevet, R. Saccomano, F. Lagoutte, J. Gourdon, Genetic
774
diversity of wheat storage proteins and bread wheat quality, Euphytica.
775
119 (2001) 59-67.
pt
777
[18] P.R. Shewry, A.S. Tatham, P.A. Lazzeri, Biotechnology of wheat quality, J Sci Food Agric. 73 (1997) 397-406.
Ac ce
776
ed
773
778
[19] L. Sollid, S.-W. Qiao, R. Anderson, C. Gianfrani, F. Koning, Nomenclature
779
and listing of celiac disease relevant gluten T-cell epitopes restricted by
780
HLA-DQ molecules, Immunogenetics. 64 (2012) 455-460.
781
[20] E.H. Arentz-Hansen, R. Körner, Ø. Molberg, H. Quarsten, W. Vader, Y.M.C.
782
Kooy, K.E.A. Lundin, F. Koning, P. Roepstorff, L.M. Sollid, S.N. McAdam,
783
The intestinal T-cell response to -gliadin in adult celiac disease is focused
784
on a single deamidated glutamine targeted by tissue transglutaminase, J
785
Exp Med. 191 (2000) 603-612.
786
[21] E.H. Arentz-Hansen, S.N. McAdam, Ø. Molberg, B. Fleckenstein, K.E.A.
787
Lundin, T.J.D. Jorgensen, G. Jung, P. Roepstorff, L.M. Sollid, Celiac lesion T
788
cells recognize epitopes that cluster in regions of gliadins rich in proline
789
residues, Gastroenterology. 123 (2002) 803-809. 34 Page 34 of 49
790
[22] L. Shan, Ø. Molberg, I. Parrot, F. Hausch, F. Filiz, G.M. Gray, L.M. Sollid, C.
791
Khosla, Structural basis for gluten intolerance in celiac sprue, Science. 297
792
(2002) 2275-2279. [23] C.L. Marino, N.A. Tuleen, G.E. Hart, J.C. Nelson, M.E. Sorrells, Y.H. Lu, P.
794
Leroy, C.R. Lopes, Molecular genetic maps of the group 6 chromosomes of
795
hexaploid wheat (Triticum aestivum L. em. Thell.), Genome. 39 (1996)
796
359-366.
ip t
793
[24] C. Mitea, E.M.J. Salentijn, P. van Veelen, S.V. Goryunova, I.M. van der
798
Meer, H.C. van den Broeck, J.R. Mujico, V. Monserrat, L.J.W.J. Gilissen,
799
J.W. Drijfhout, L. Dekking, F. Koning, M.J.M. Smulders, A universal
800
approach to eliminate antigenic properties of alpha-gliadin peptides in
801
celiac disease, PLoS ONE. 5 (2010) e15637.
us
cr
797
[25] Ø. Molberg, A.K. Uhlen, T. Jensen, N.S. Flæte, B. Fleckenstein, H. Arentz-
803
Hansen, M. Raki, K.E.A. Lundin, L.M. Sollid, Mapping of gluten T-cell
804
epitopes in the bread wheat ancestors: Implications for celiac disease,
805
Gastroenterology. 128 (2005) 393-401.
M
an
802
[26] T.W.J.M. van Herpen, S. Goryunova, J. van der Schoot, M. Mitreva, E.M.J.
807
Salentijn, O. Vorst, M. Schenk, P. van Veelen, F. Koning, L. van Soest, B.
808
Vosman, D. Bosch, R.J. Hamer, L.J.W.J. Gilissen, M.J.M. Smulders, Alpha-
809
gliadin genes from the A, B, and D genomes of wheat contain different sets
810
of celiac disease epitopes, BMC Genomics. 7 (2006) 1.
811
[27] E. García, M. Llorente, A. Hernando, R. Kieffer, H. Wieser, E. Mendez,
812
Development of a general procedure for complete extraction of gliadins for
813
Ac ce
pt
ed
806
814
heat processed and unheated foods, Eur J Gastroenterol Hepatol. 17 (2005) 529-539.
815
[28] E. Méndez, C. Vela, U. Immer, F.W. Janssen, Report of a collaborative trial
816
to investigate the performance of the R5 enzyme linked immunoassay to
817 818
determine gliadin in gluten-free food, Eur J Gastroenterol Hepatol. 17 (2005) 1053-1063.
819
[29] I. Valdés, E. García, M. Llorente, E. Méndez, Innovative approach to low-
820
level gluten determination in foods using a novel sandwich enzyme-linked
821
immunosorbent assay protocol, Eur J Gastroenterol Hepatol. 15 (2003)
822
465-474.
35 Page 35 of 49
823
[30] B. Geßendorfer, H. Wieser, P. Koehler, Optimisation of a solvent for the
824
complete extraction of prolamins from heated foods, J Cereal Sci. 52
825
(2010) 331-332.
826
[31] P. Kanerva, T. Sontag-Strohm, O. Brinck, H. Salovaara, Improved
827
extraction of prolamins for gluten detection in processed foods, Agr Food
828
Sci. 20 (2011) 206-216. [32] M.C.
Mena,
M.
Lombardía,
A.
Hernando,
E.
Méndez,
J.P.
Albar,
ip t
829
Comprehensive analysis of gluten in processed foods using a new
831
extraction method and a competitive ELISA based on the R5 antibody,
832
Talanta. 91 (2012) 33-40.
cr
830
[33] C. Diaz-Amigo, B. Popping, Accuracy of ELISA detection methods for
834
gluten and reference materials: A realistic assessment, J Agric Food Chem.
835
61 (2013) 5681-5688.
us
833
[34] D. Stepniak, L.W. Vader, Y. Kooy, P.A. Veelen, A. Moustakas, N.A.
837
Papandreou, E. Eliopoulos, J.W. Drijfhout, G.K. Papadopoulos, F. Koning,
838
T-cell recognition of HLA-DQ2-bound gluten peptides can be influenced by
839
an N-terminal proline at p-1, Immunogenetics. 57 (2005) 8-15.
M
an
836
[35] M. Koeberl, D. Clarke, A.L. Lopata, Next generation of food allergen
841
quantification using mass spectrometric systems, J Proteome Res. 13
842
(2014) 3499-3509. [36] C.
Diaz-Amigo,
B.
Popping,
Gluten
and
glutenfree:
Issues
and
pt
843
ed
840
considerations of labeling regulations, detection methods, and assay
845
validation, J AOAC Int. 95 (2012) 337-348.
Ac ce
844 846
[37] R. Haraszi, H. Chassaigne, A. Maquet, F. Ulberth, Analytical methods for
847
detection of gluten in food—Method developments in support of food
848
labeling legislation, J AOAC Int. 94 (2011) 1006-1025.
849
[38] S.B. Altenbach, W.H. Vensel, F.M. DuPont, Integration of transcriptomic
850
and proteomic data from a single wheat cultivar provides new tools for
851
understanding the roles of individual alpha gliadin proteins in flour quality
852
and celiac disease, J Cereal Sci. 52 (2010) 143-151.
853
[39] K.L. Fiedler, S.C. McGrath, J.H. Callahan, M.M. Ross, Characterization of
854
grain-specific peptide markers for the detection of gluten by mass
855
spectrometry, J Agric Food Chem. 62 (2014) 5835-5844.
856 857
[40] B. Prandi, A. Faccini, T. Tedeschi, A. Cammerata, D. Sgrulletta, M. D’Egidio,
G.
Galaverna,
S.
Sforza,
Qualitative
and
quantitative
36 Page 36 of 49
858
determination of peptides related to celiac disease in mixtures derived
859
from different methods of simulated gastrointestinal digestion of wheat
860
products, Anal Bioanal Chem. 406 (2014) 4765-4775. [41] B. Prandi, A. Faccini, T. Tedeschi, G. Galaverna, S. Sforza, LC/MS analysis
862
of proteolytic peptides in wheat extracts for determining the content of the
863
allergen amylase/trypsin inhibitor CM3: Influence of growing area and
864
variety, Food Chem. 140 (2013) 141-146.
ip t
861
[42] J.A. Sealey-Voyksner, C. Khosla, R.D. Voyksner, J.W. Jorgenson, Novel
866
aspects of quantitation of immunogenic wheat gluten peptides by liquid
867
chromatography-mass spectrometry/mass spectrometry, J Chromatogr A.
868
1217 (2010) 4167-4183.
us
cr
865
[43] L. Uvackova, L. Skultety, S. Bekesova, S. McClain, M. Hajduch, MSE based
870
multiplex protein analysis quantified important allergenic proteins and
871
detected relevant peptides carrying known epitopes in wheat grain
872
extracts, J Proteome Res. 12 (2013) 4862-4869.
an
869
[44] A. Camarca, R.P. Anderson, G. Mamone, O. Fierro, A. Facchiano, S.
874
Costantini, D. Zanzi, J. Sidney, S. Auricchio, A. Sette, R. Troncone, C.
875
Gianfrani, Intestinal T cell responses to gluten peptides are largely
876
heterogeneous: Implications for a peptide-based therapy in celiac disease,
877
J Immunol. 182 (2009) 4158-4166. [45] W. Vader, Y. Kooy, P. van Veelen, A. de Ru, D. Harris, W. Benckhuijsen, S.
879
Peña, L. Mearin, J.W. Drijfhout, F. Koning, The gluten response in children
880
with celiac disease is directed toward multiple gliadin and glutenin
881
pt
878
Ac ce
ed
M
873
peptides, Gastroenterology. 122 (2002) 1729-1737.
882
[46] A. Camarca, A. Del Mastro, C. Gianfrani, Repertoire of gluten peptides
883
active in celiac disease patients: Perspectives for translational therapeutic
884 885
applications, Endocr Metab Immune Disord Drug Targets. 12 (2012) 207219.
886
[47] J.A. Tye-Din, J.A. Stewart, J.A. Dromey, T. Beissbarth, D.A. van Heel, A.
887
Tatham, K. Henderson, S.I. Mannering, C. Gianfrani, D.P. Jewell, A.V.S.
888
Hill,
889
quantitative mapping of T cell epitopes in gluten in celiac disease, Sci
890
Transl Med. 2 (2010) 41ra51.
J.
McCluskey,
J.
Rossjohn,
R.P.
Anderson,
Comprehensive,
891
[48] R.P. Anderson, P. Degano, A.J. Godkin, D.P. Jewell, A.V.S. Hill, In vivo
892
antigen challenge in celiac disease identifies a single transglutaminase37 Page 37 of 49
893
modified peptide as the dominant -gliadin T-cell epitope, Nature Med. 6
894
(2000) 337-342.
895
[49] H.C. van den Broeck, A.H.P. America, M.J.M. Smulders, D. Bosch, R.J.
896
Hamer, L.J.W.J. Gilissen, I.M. van der Meer, A modified extraction protocol
897
enables detection and quantification of celiac disease-related gluten
898
proteins from wheat, J Chromatogr B. 877 (2009) 975-982. [50] M. Stapels, K. Fadgen, A reproducible online 2D reversed phase-reversed
900
phase high-low pH method for qualitative and quantitative proteomics,
901
Spectroscopy
902
http://www.spectroscopyonline.com/node/225338.
Special
issue
(2009)
cr
online.
ip t
899
[51] M. Gilar, P. Olivova, A.E. Daly, J.C. Gebler, Two-dimensional separation of
904
peptides using RP-RP-HPLC system with different pH in first and second
905
separation dimensions, J Sep Sci. 28 (2005) 1694-1703.
us
903
[52] B. MacLean, D.M. Tomazela, N. Shulman, M. Chambers, G.L. Finney, B.
907
Frewen, R. Kern, D.L. Tabb, D.C. Liebler, M.J. MacCoss, Skyline: an open
908
source document editor for creating and analyzing targeted proteomics
909
experiments, Bioinformatics. 26 (2010) 966-968.
M
head of bacteriophage T4, Nature. 227 (1970) 680-685. [54] J.H.G. Cordewener, G. Hause, E. Görgen, R. Busink, B. Hause, H.J.M.
913
Dons, A.A.M. Van Lammeren, M.M. Van Lookeren Campagne, P. Pechan,
914
Changes in synthesis and localization of members of the 70-kDa class of
915
heat-shock proteins accompany the induction of embryogenesis in Brassica
916
pt
912
Ac ce
911
[53] U.K. Laemmli, Cleavage of structural proteins during the assembly of the
ed
910
an
906
napus L. microspores, Planta. 196 (1995) 747-755.
917
[55] C. Mitea, Y.M.C. Kooy-Winkelaar, P.A. van Veelen, A. de Ru, J.W. Drijfhout,
918
F. Koning, L. Dekking, Fine specificity of monoclonal antibodies against
919 920
celiac disease-inducing peptides in the gluteome, Am J Clin Nutr. 88 (2008) 1057-1066.
921
[56] E.H.A. Spaenij-Dekking, E.M.C. Kooy-Winkelaar, W.F. Nieuwenhuizen, J.W.
922
Drijfhout, F. Koning, A novel and sensitive method for the detection of T
923
cell stimulatory epitopes of - and -gliadin, Gut. 53 (2004) 1267-1273.
924
[57] C. Mitea, R. Havenaar, J.W. Drijfhout, L. Edens, L. Dekking, F. Koning,
925
Efficient
degradation
of
gluten
by
a
prolyl
endoprotease
in
a
926
gastrointestinal model: implications for coeliac disease, Gut. 57 (2008) 25-
927
32. 38 Page 38 of 49
[58] E.M.J. Salentijn, S. Goryunova, N. Bas, I.M. van der Meer, H.C. van den
929
Broeck, T. Bastien, L.J.W.J. Gilissen, M.J.M. Smulders, Tetraploid and
930
hexaploid wheat varieties reveal large differences in expression of alpha-
931
gliadins from homoeologous Gli-2 loci, BMC Genomics. 10 (2009) 48.
932
[59] L. Spaenij-Dekking, Y. Kooy-Winkelaar, P. van Veelen, J.W. Drijfhout, H.
933
Jonker, L. van Soest, M.J.M. Smulders, D. Bosch, L.J.W.J. Gilissen, F.
934
Koning, Natural variation in toxicity of wheat: potential for selection of
935
nontoxic varieties for celiac disease patients, Gastroenterology. 129 (2005)
936
797-806.
cr
ip t
928
[60] H.C. van den Broeck, H.C. de Jong, E.M.J. Salentijn, L. Dekking, D. Bosch,
938
R.J. Hamer, L.J.W.J. Gilissen, I.M. van der Meer, M.J.M. Smulders,
939
Presence of celiac disease epitopes in modern and old hexaploid wheat
940
varieties: wheat breeding may have contributed to increased prevalence of
941
celiac disease, Theor Appl Genet. 121 (2010) 1527-1539.
an
us
937
[61] H.C. van den Broeck, C. Hongbing, X. Lacaze, J.-C. Dusautoir, L.J.W.J.
943
Gilissen, M.J.M. Smulders, I.M. van der Meer, In search of tetraploid wheat
944
accessions reduced in celiac disease-related gluten epitopes, Mol Biosyst. 6
945
(2010) 2206-2213.
M
942
[62] E. Halbmayr-Jech, E. Hammer, R. Fielder, J. Coutts, A. Rogers, M. Cornish,
947
Characterization of G12 sandwich ELISA, a next-generation immunoassay
948
for gluten toxicity, J AOAC Int. 95 (2012) 372-376.
949
[63] J. Šalplachta, M. Marchetti, J. Chmelík, G. Allmaier, A new approach in
950
proteomics of wheat gluten: combining chymotrypsin cleavage and matrix-
951
Ac ce
pt
ed
946
952
assisted laser desorption/ionization quadrupole ion trap reflectron tandem mass spectrometry, Rapid Commun Mass Sp. 19 (2005) 2725-2728.
953
[64] B. Prandi, M. Bencivenni, A. Faccini, T. Tedeschi, A. Dossena, R. Marchelli,
954
G. Galaverna, S. Sforza, Composition of peptide mixtures derived from
955 956 957
simulated gastrointestinal digestion of prolamins from different wheat varieties, J Cereal Sci. 56 (2012) 223-231.
[65] P. Cressey, P. Grounds, S. Jones, S. Ashmore, D. Saunders, Gluten
958
residues
in
gluten-free
foods
sold
in
Christchurch,
New
Zealand:
959
comparison of LC-MS and ELISA methods, Qual Assur Saf Crop. 5 (2013)
960
207-213.
39 Page 39 of 49
961
[66] H. Wieser, R. Kieffer, Correlations of the amount of gluten protein types to
962
the technological properties of wheat flours determined on a micro-scale, J
963
Cereal Sci. 34 (2001) 19-27.
964
[67] O.D. Anderson, F.C. Greene, The α-gliadin gene family. II. DNA and
965
protein
sequence
variation,
subfamily
structure,
966
pseudogenes, Theor Appl Genet. 95 (1997) 59-65.
and
origins
of
[68] B. Prandi, M. Bencivenni, T. Tedeschi, R. Marchelli, A. Dossena, G.
968
Galaverna, S. Sforza, Common wheat determination in durum wheat
969
samples through LC/MS analysis of gluten peptides, Anal Bioanal Chem.
970
403 (2012) 2909-2914.
cr
ip t
967
[69] E. Salentijn, D. Esselink, S. Goryunova, I. van der Meer, L. Gilissen, M.
972
Smulders, Quantitative and qualitative differences in celiac disease
973
epitopes among durum wheat varieties identified through deep RNA-
974
amplicon sequencing, BMC Genomics. 14 (2013) 905.
an
us
971
[70] K. Kawaura, K. Mochida, Y. Ogihara, Expression profile of two storage-
976
protein gene families in hexaploid wheat revealed by large-scale analysis
977
of expressed sequence tags, Plant Physiol. 139 (2005) 1870-1880.
M
975
[71] P.R. Shewry, R.A.C. Mitchell, P. Tosi, Y. Wan, C. Underwood, A. Lovegrove,
979
J. Freeman, G.A. Toole, E.N.C. Mills, J.L. Ward, An integrated study of
980
grain development of wheat (cv. Hereward), J Cereal Sci. 56 (2012) 21-
981
30. Katagiri,
pt
[72] M.
T.
Masuda,
F.
Tani,
N.
Kitabatake,
Expression
and
983
development of wheat proteins during maturation of wheat kernel and the
984
Ac ce
982
ed
978
985 986
rheological properties of dough prepared from the flour of mature and immature Wheat, Food Sci Technol Res. 17 (2011) 111-120.
40 Page 40 of 49
986
Figure legends
988
Fig. 1.
989
Sequence alignment of the CD-epitope domain of multiple -gliadin proteins from
990
wheat. The deduced amino acid sequences of multiple EST contigs (present in
991
the NCBI-UniGene database) were searched for sequence variants in the CD-
992
epitope containing domain. Homologous ESTs for this sequence domain were
993
grouped and only variants found in more than 5 ESTs (n) are displayed. The
994
variants are grouped per genome origin from which the ESTs are expressed: Gli-
995
A2, Gli-B2, or Gli-D2. The CD-epitopes, known to trigger the adaptive HLA-DQ2+
996
T-cell response (Glia-2/Glia-9, and Glia-20) in CD-patients, are color
997
indicated. In yellow: Glia-2/Glia-9 epitopes. In green: Glia-20 epitope, in
998
blue: 33-mer containing repetitive epitopes. In black: chymotrypsin digestion
999
sites. In red: amino acid (AA) variation in the sequence. In bold underlined:
1000
sequences selected for peptide synthesis. The peptide number code is indicated
1001
in the right-most column. Figure adapted from Mitea et al. [24].
1002
Ac ce
pt
ed
M
an
us
cr
ip t
987
1003
Fig. 2.
1004
Optimization of chymotrypsin digestion conditions for hexaploid wheat variety
1005
Toronto and tetraploid wheat variety Dibillik sinde. A) Release of P1 after
1006
incubation for 10’, 30’, 1 h, 2h, and 4 h at 25C with enzyme:protein ratio of
1007
1:20. The same trend was observed for peptides P2-P5. B) Release of P1 after
1008
incubation for 2 h, 4 h, 8 h, and 18 h at 25C with enzyme:protein ratios of 1:5
1009
and 1:10. The same trend was observed for P2-P5. C) Release of P6 after
1010
incubation for 2 h, 4h, and 8 h at 25C with enzyme:protein ratios of 1:5, 1:10,
41 Page 41 of 49
1011
and 1:20. D) Release of P6 after incubation for 2 h, 4 h, 8 h, 18 h, and 22 h
1012
(extra enzyme addition after 18 h) at 25C with enzyme:protein ratio of 1:5.
1013 Fig. 3.
1015
Total amounts of peptides containing CD-epitopes in mg -gliadin/g gluten (103
1016
ppm) present in three wheat varieties, Toronto (AABBDD), Minaret (AABBDD),
1017
and Dibillik sinde (AABB). Containing A) Glia-2/9 epitopes and B) Glia-20
1018
epitopes.
cr
ip t
1014
us
1019 Supplementary Fig. S1.
1021
Chromatograms, ion intensity maps, of the low pH peptide resolution in the four
1022
subsequent 2D LC-MS fractions of a ‘Minaret’ gluten protein digest by using raw
1023
MSE (DIA) data processed by Progenesis QI software. X-axis represents m/z
1024
value. Y-axis represents retention time in minutes.
1025
ed
M
an
1020
Supplementary Fig. S2.
1027
Chromatograms obtained from LC-MRM/MS showing the resolution of the nine
1028
selected marker peptides. A) Mixture of the nine marker peptides used for
1029
calibration and B) ‘Toronto’ gluten protein digest (22 h). The inlaid images show
1030
an example of the multiple transition signals per peptide as analyzed by Skyline
1031
software. X-axis represents retention time in minutes. Y-axis represents signal
1032
intensity (106).
Ac ce
pt
1026
1033 1034
Supplementary Fig. S3.
42 Page 42 of 49
1056 1057 1058
Highlights
1059
1061
Targeted detection and quantification of multiple celiac disease stimulating
ip t
1060
epitopes.
High-throughput screening for quantification of CD-epitopes in wheat.
1063
Method enables selection of wheat varieties with reduced levels of CD-
us
1065
epitopes.
LC-MRM/MS technology as a standard method for quantifying CD-epitopes.
an
1064
cr
1062
Ac ce
pt
ed
M
1066
44 Page 43 of 49
Ac
ce
pt
ed
M
an
us
cr
i
Figure 1
Page 44 of 49
Ac
ce
pt
ed
M
an
us
cr
i
Figure 2
Page 45 of 49
Ac ce p
te
d
M
an
us
cr
ip t
Figure 3
Page 46 of 49
ip t
Table 1
7
8a
9
10
11*
12
13
14b
15
16
17
18
19c
20
21d
22
23e
24
Q3S4V8_WHEAT
A5JSB3_WHEAT
I3RXV6_AEGTA
Q9M4M4_WHEAT
A5JSA6_WHEAT
A5JSA9_WHEAT
K7X0Q3_WHEAT
I0IT56_WHEAT
Q9M4L8_WHEAT
K7WV47_WHEAT
G9I1T5_AEGUN
Q2QL43_AEGTA
F6M8F7_9POAL
A5JSB7_WHEAT
M4WFD9_AEGSP
K7X1J5_WHEAT
Q2QL52_AEGTA
us
K7XEB5_WHEAT
6
F4YT74_WHEAT
Q0GK30_TRITI
Accession first hit
ce pt
Ac
*
ed
Peptide no. Peptide sequences LQLQPFPQPQPF LQLQPFPQPQLPY P1 LQLQPFPQPQLSY LQLQPFPQPQLPYPQPQPF P2 MQLQPFPQPQLPYPQPQLPYPQPQPF LQLQPFPQPQLPYPQPHLPYPQPQPF P3 LQLQPFPQPQLPYPHPQLPYPQPQPF LQLQPFPQPQLPYPQPQLPYPQPQPF P4 LQLQPFPQPQLPYPQPQLPYPQPQLPYPQPQPF P5 LPQLPYPQPQPFPPQQSYPQPQPQYPQPQQPISQQQAQL RPQQLYPQPQPQY RPQQPYPQPQPQY P6 RPQQPYPQSQPQY P7 RPQQSYPQPQPQY P8 QQQLIPCRDVVL P9 QQILQQQLIPCRDVVL QQQLIPCMDVVL QQILQQQLIPCMDVVL a, b Sequences from Triticum aestivum, probable B-genome c Sequence from Aegilops uniaristata, N-genome d Sequences from Triticum compactum, not full length e Sequences from Aegilops speltoides
-gliadin families
5
J7I026_WHEAT
4
K7WV12_AEGTA
3
I3RXV8_AEGTA
2
M an
1
Q9M4L6_WHEAT
cr
Table 1. Selection of peptide sequences identified from -gliadin families in Triticum and Aegilops species in Mascot search (Uniprot database). Selection of the sequences is based on immunogenic epitopes sequences. Peptide sequences shown include the nine selected peptide sequences used for quantification (P1-P9). Amino acid substitutions in peptide sequences are shown in italics. CD-epitope sequences within the peptides are shown in bold.
Score 46 54 54 47 41 73 60 39 53 33 46 67 59 27 52 51 40
Sequence from Aegilops tauschii, contains Glia-9 epitope on a different peptide
Page 47 of 49
QQILQQQLIPCRDVVL
cr us
14 15 20 21 19 19 29 18 26 16 35 22
6.8 7.0 7.2 6.9 7.3 4.6 4.4 4.4 5.8 5.8 6.3 6.3
6-7.5 6.6-7.5 6.7-8 6.6-7.5 6.7-8 3-5 4-5 4-5 5.5-6 5.5-6 6-6.6 6-6.6
R2 0.987 0.981 0.974 0.969 0.970 0.993 0.991 0.994 0.991 -
Ac
ce pt
MRM schedule (min)
P9
Retention time (min)
QQQLIPCRDVVL
M an
P8
784.9 (2+) 755.1 (3+) 1029.5 (3+) 1032.5 (3+) 978.3 (4+) 813.9 (2+) 808.9 (2+) 539.6 (3+) 734.9 (2+) 490.3 (3+) 976.0 (2+) 651.0 (3+)
Fragment ions m/z 483.3 (b4); 617.3 (y5); 727.4 (b6); 842.4 (y7); 952.5 (b8); 1290.7 (b11) 483.3 (b4); 488.3 (y4); 713.4 (y6); 727.4 (b6); 952.5 (b8); 973.5 (y8) 483.3 (b4); 713.4 (y6); 727.4 (b6); 824.5 (b7); 952.5 (b8); 973.5 (y8) 488.3 (y4); 645.9 (b11); 713.4 (y6); 727.4 (b6); 903.5 (y15); 952.5 (b8) 483.3 (b4); 488.3 (y4); 713.4 (y6); 727.4 (b6); 824.5 (b7); 952.5 (b8); 973.5 (y8) 407.2 (y3); 632.3 (y5); 770.4 (b6); 857.4 (y7); 995.5 (b8); 1220.6 (b10) 510.3 (b4); 770.4 (b6) 407.2 (y3); 510.3 (b4); 770.4 (b6); 847.4 (y7) 858.5 (y7); 971.5 (y8); 1212.7 (y10) 498.3 (b4); 611.4 (b5); 858.5 (y7) 858.5 (y7) 858.5 (y7); 867.5 (b7); 971.5 (y8); 980.6 (b8)
ed
P1 P2 P3 P4 P5 P6 P7
Sequence LQLQPFPQPQLPY LQLQPFPQPQLPYPQPQPF LQLQPFPQPQLPYPQPQLPYPQPQPF LQLQPFPQPQLPYPQPHLPYPQPQPF LQLQPFPQPQLPYPQPQLPYPQPQLPYPQPQPF RPQQPYPQPQPQY RPQQPYPQSQPQY
Precursor m/z (charge state)
Peptide
Table 2. MS conditions of the selected marker peptides.
Collision energy (V)
ip t
Table 2
Page 48 of 49
ip t
Table 3
ed
M an
us
cr
Table 3. Amounts of CD-epitopes present in three wheat varieties: Toronto (AABBDD), Minaret (AABBDD), and Dibillik sinde (AABB). Amounts in 3 mg -gliadin containing CD-epitope/g gluten (10 ppm) at t=22h of peptides from -gliadins containing CD-epitopes, (±SD). peptide sequence Toronto Minaret Dibillik sinde LQLQPFPQPQLPY P1 123.20 (±18.27) 56.14 (±7.07) 122.77 (±19.85) LQLQPFPQPQLPYPQPQPF P2 22.65 (±3.60) 14.93 (±1.23) nd LQLQPFPQPQLPYPQPHLPYPQPQPF P3 12.47 (±0.83) 8.82 (±0.10) nd LQLQPFPQPQLPYPQPQLPYPQPQPF P4 3.34 (±0.17) 2.18 (±0.11) nd LQLQPFPQPQLPYPQPQLPYPQPQLPYPQPQPF P5 56.26 (±4.51) 31.71 (±1.09) nd RPQQPYPQPQPQY P6 133.09 (±8.93) 49.82 (±3.28) 18.69 (±1.38) RPQQPYPQSQPQY P7 13.41 (±1.95) 9.26 (±1.08) nd P8 QQQLIPCRDVVL 162.10 (±21.61) 137.42 (±17.97) 98.75 (±13.06) P9 QQILQQQLIPCRDVVL nd nd nd 217.92 (±27.38) 113.79 (±9.60) 122.77 (±19.85) sum[P1-P5] Glia-2/9 P6 + P7 146.49 (±10.88) 59.08 (±4.36) 18.69 (±1.38) Glia-20 P8 + P9 162.10 (±21.61) 137.42 (±17.97) 98.75 (±13.06) Total -gliadin
Ac
ce pt
nd, not detected
Page 49 of 49
