Food Chemistry 246 (2018) 156–163

Contents lists available at ScienceDirect

Food Chemistry
journal homepage: www.elsevier.com/locate/foodchem

Highly sensitive detection of gluten-containing cereals in food samples by T

real-time Loop-mediated isothermal AMPlification (qLAMP) and real-time
polymerase chain reaction (qPCR)
⁎
Alejandro Garrido-Maestu , Sarah Azinheiro, Pablo Fuciños, Joana Carvalho, Marta Prado
Department of Life Sciences, Nano4Food – Food Quality and Safety Research Group, International Iberian Nanotechnology Laboratory, Av. Mestre José Veiga s/n, 4715-
330 Braga, Portugal

A R T I C L E I N F O A B S T R A C T

Keywords: The treatment of gluten-related disorders is based on a lifelong, and strict, gluten-free diet. Thus, reliable and
Real-time Loop-mediated isothermal sensitive methods are required to detect the presence of gluten contamination. Traditional techniques rely on the
AMPlification (qLAMP) detection of these proteins based on specific antibodies, but recent approaches go for an indirect route detecting
Real-time PCR (qPCR) the DNA that indicates the presence of cereals with gluten content. In the current study two different DNA
α2-gliadin
amplification techniques, real-time PCR (qPCR) and real-time Loop-mediated isothermal AMPlification
Gluten
(qLAMP), were evaluated for their capability to detect and quantify gluten. Different detection strategies, based
Gluten-free food
on these DNA amplification techniques, were tested. Even though good specificity results were obtained with the
different approaches, overall qPCR proved more sensitive than qLAMP. This is the first study reporting a qLAMP
based-method for the detection of gluten-containing cereals, along with its evaluation in comparison with qPCR.

1. Introduction to the Codex Alimentarius and the European Union regulation, only
those products with a gluten content below 20 ppm or 20 mg/kg can be
Celiac disease (CD) is an autoimmune disorder associated with a labeled as “gluten-free” If the content does not exceed 100 mg/kg the
permanent intolerance to gluten proteins of wheat, barley, and rye. CD product can be labeled as “very low gluten” (EFSA, 2014; European
is characterized by an inflammatory response that leads to atrophy of Commision, 2009; WHO, 2010).
the intestinal villi and malabsorption. Reactions to gluten are not lim- The most frequently used methods for gluten detection in food
ited to CD, and a spectrum of gluten-related disorders have been de- products are immunological based, due to their ease of use and rapid
scribed, including wheat allergy (WA), a food allergy that is im- analytical procedure. More than 15 immunochemical methods for de-
munoglobulin E (IgE) mediated (Mittag et al., 2004), and more tecting gluten are marketed commercially, mainly ELISA kits and
recently: non-celiac gluten sensitivity (NCGS) (Sapone et al., 2012). Lateral Flow Devices (LFD), which have been recently reviewed (Bruins
Although different conditions, the treatment of gluten-related disorders Slot, van der Fels-Klerx, Bremer, & Hamer, 2015). However, there is still
is based on a gluten-free diet (GFD). Besides the people needing to a strong need for improvement due to the difficulties related to the full
follow a GFD for any of the conditions mentioned above, a new segment extraction of gluten proteins from food matrixes, the choice of epitopes
of consumers has arisen who consume gluten-free products as a lifestyle to be used as targets for detection methods, and the lack of certified
choice (Foschia, Horstmann, Arendt, & Zannini, 2016). As a result, in reference materials (CRMs) (Bruins Slot et al., 2015; Diaz-Amigo &
the last 15 years, the global market of gluten-free products has in- Popping, 2012; Haraszi, Chassaigne, Maquet, & Franz, 2011).
creased exponentially approaching 2.5 billion $ in global sales in 2010 As for allergens, although the detection of the offending protein is
(Sapone et al., 2012), and with an even higher growth since then, desirable, this is not always feasible due to a lack of well-characterized
reaching up to an estimated 6.2 billion $ by 2018 (Wieser, Koehler, & compounds, or lack of sensitive enough detection techniques (Prado
Konitzer, 2014). et al., 2016). DNA-based assays have proven to be useful tools by tar-
Gluten is a complex mixture of storage proteins from grains, which geting DNA as a marker for the presence of the offending ingredient
are ingredients in a wide range of foods (Miranda-Castro, de-Los- (Carmen Diaz-Amigo & Popping, 2013; M. Prado et al., 2016), both for
Santos-Álvarez, Miranda-Ordieres, & Lobo-Castañón, 2016). According the detection of undeclared ingredients in food products, and for the

⁎
Corresponding author at: International Iberian Nanotechnology Laboratory, Av. Mestre José Veiga s/n, 4715-330 Braga, Portugal.
E-mail address: alejandro.garrido@inl.int (A. Garrido-Maestu).

https://doi.org/10.1016/j.foodchem.2017.11.005
Received 5 May 2017; Received in revised form 8 September 2017; Accepted 2 November 2017
Available online 06 November 2017
0308-8146/ © 2017 Elsevier Ltd. All rights reserved.
A. Garrido-Maestu et al. Food Chemistry 246 (2018) 156–163

evaluation of cleaning procedures for industrial equipment (Stephan & Table 1

Vieths, 2004). The use of DNA targets is interesting for methods re- Result comparison among qPCR and qLAMP in food samples.
quiring a high sensitivity since DNA is not affected by the variability of
Type of sample Number of Result
the phenotype, they present a high thermal stability, and DNA ampli- samples
fication methods, such as PCR and qPCR, allow the amplification of the qPCR qLAMP
number of copies of the initial DNA target. Moreover, they have the
Probe Sybr In-house Commercial
advantage of allowing the detection of individual species (Mujico,
Green I
Lombardía, Mena, Méndez, & Albar, 2011), or the simultaneous de-
tection of a collective allergen group (Marta Prado, Boix, & von Holst, Bread 1 + + + +
2012) depending on the designed primers and the chosen sequence, Wheat flour 2 + + + +
which gives an additional flexibility to this approach. Rye flour 1 + + + +
Spelt flour 1 + + + +
DNA amplification approaches based on polymerase chain reaction Rice flour* 1 + + + +
(PCR) (Dahinden, von Büren, & Lüthy, 2001) or more recently q PCR Barley grain 2 +/− + − −
(Mujico et al., 2011; Sandberg, Lundberg, Ferm, & Malmheden Yman, Rye grain 1 +/− + − −
2003; Zeltner, Glomb, & Maede, 2009) have been described for the Pasta 1 + + + +
Couscous 1 + + + +
detection of gluten-containing cereals.
Corn flour 1 − − − −
In recent years, an isothermal amplification technique named Loop- Gluten-free 1 − − − −
mediated isothermal AMPlification (LAMP) has been developed. LAMP cookies
allows the specific detection of DNA, likewise PCR/qPCR, but targeting Rice** 1 − − − −−
six to eight different sequences and being performed at a constant Corn 2 − − − −
Popcorn 1 − − − −
temperature. It has additional advantages over the former techniques,
being one of the most attractive the possibility of naked-eye result * In the package the producer indicated that may contain traces of gluten.
observation due to the formation of insoluble by-products (magnesium ** Without rinsing with water previous to DNA extraction, resulted positive. “+/−”
pyrophosphate) whenever the reaction is positive (presence of target indicate that one of the replicates within the sample failed to amplify.
sequence) (Mori, Kitao, Tomita, & Notomi, 2004; Notomi et al., 2000).
In this regard recent studies have also included the use of intercalating 2.2.2. Grains
dyes for naked-eye observation and/or for fluorescence tracking in real- When the samples consisted of grains, these were previously frozen
time assays (D’Agostino, Diez-Valcarce, Robles, Losilla-Garcia, & Cook, in liquid nitrogen and homogenized using a mortar and pestle until a
2015; D’Agostino et al., 2016; Hara-Kudo, Yoshino, Kojima, & Ikedo, fine powder was obtained. Regarding the rice sample, the grain was
2005; Ma et al., 2010; Tomita, Mori, Kanda, & Notomi, 2008; Zhang, washed with regular tap water, to eliminate traces of other cereals
Brown, & González-Escalona, 2011). which may contaminate the samples during the processing in the fa-
Surprisingly, to our knowledge, this relatively new technique has cility. Then, the DNA was extracted using the NucleoSpin Food and
not been previously applied for the detection gluten-containing cereals. Plant kits (Macherey-Nagel, Düren, Germany).
Thus, the goal of the current study was to evaluate the performance of a
LAMP based method for the specific detection of gluten-containing 2.3. Gene selection
cereals in food samples, and its comparison against qPCR.
The use of α2-gliadin gene as target for qPCR has been previously
reported to provide a good sensitivity for the detection of gluten-con-
2. Materials and methods taining cereals in foods (Martín-Fernández et al., 2015). Therefore, this
gene was selected for the design of the new methods. Sequences from
2.1. Samples common wheat, rye and barley: Triticum aestivum (AJ133612), Secale
cereale (JX843487) and Hordeum vulgare (HQ875781) were obtained
A total of 17 different samples, including grains and foods, were from GeneBank and aligned with CLC Sequence Viewer (C L C Bio-
purchased from local suppliers, and analyzed with the qPCR and Qiagen, 2016). The consensus sequence was used for primer design.
qLAMP approaches, described below, to evaluate the specificity of the
different protocols. A detailed list of the samples analyzed is provided 2.4. Primer design for qPCR and qLAMP
in Table 1.
In order to determine the real potential of each technique for The design of primers for qPCR was performed with Primer3 free
quantification/detection purposes, a binary model was prepared as online software (Untergasser et al., 2007). Along with the primers, a
described by Martín-Fernández, Costa, Oliveira, López-Ruiz, and Mafra hydrolysis probe was also designed. Regarding qLAMP, the design was
(2015). To this end, wheat flour was mixed with gluten-free corn flour performed with the free software Primer Explorer V4 (https://
in the following proportions: 50% (500000 mg/kg), 10% (100000 mg/ primerexplorer.jp/e/index.html). The specificity of all primers was
kg), 5% (50000 mg/kg), 1% (10000 mg/kg), 0.5% (5000 mg/kg), 0.1% verified in silico with BLAST® (Basic Local Alignment Search Tool,
(1000 mg/kg), 0.05% (500 mg/kg), 0.01% (100 mg/kg), 0.005 (50 mg/ https://blast.ncbi.nlm.nih.gov/Blast.cgi?CMD = Web&PAGE_TYPE=
kg), 0.002% (20 mg/kg) and 0.001% (10 mg/kg). Each sample was BlastHome), and later on with real food samples. For comparison pur-
processed as described in the Section 2.2.1 and analyzed in triplicate by poses, the use of an intercalating dye was also tested by qPCR. To this
the different amplification approaches (qPCR and qLAMP). end, the F3/B3 primers designed for qLAMP were selected, as seen in
Table 2.

2.2. DNA extraction 2.5. qPCR

2.2.1. Food For comparison purposes, two different qPCR approaches were as-
DNA extraction and purification from food samples was accom- sessed. The first one using the primers F3/B3 from the qLAMP design,
plished with the NucleoSpin Food kit (Macherey-Nagel, Düren, and SYBR® Green I as intercalating dye. The second approach consisted
Germany) with the modifications described by Martín-Fernández et al. on a newly designed set of primers and included a hydrolysis probe. In
(2015). both sets of experiments, the samples were analyzed in duplicate, and

157
A. Garrido-Maestu et al. Food Chemistry 246 (2018) 156–163

Table 2 2.6.1. In-house master mix

Primer list for qPCR and qLAMP. The reaction mixture contained: 2 μL of 10X Isothermal
Amplification Buffer (New England BioLabs, Inc., Ipswich, Mass., USA),
Detection Sequence 5′-3′
method 0.35 mM dNTP mix (Thermo Fisher Scientific Inc., Waltham, MA, USA),
8 U Bst 2.0 WarmStart® DNA Polymerase (New England BioLabs, Inc.,
qLAMP a2g-F3* GCCAAGCCATCCACAATGT Ipswich, Mass., USA), and 1 μL of 1:500 dilution of Midori Green
a2g-B3* CAACTGGAGCAATGGTGC
(Nippon Genetics Europe GmbH, Düren, Germany) for fluorescence
a2g-FIP GGAAGGAGCCCTGGCCTGAT-
CAACCGTTGAGCCAGGTCT detection. The remaining volume was completed with DNAse, RNAse
a2g-BIP CAGGGCTCTGTCCAGCCTCA- free water.
TGCAGGTAGCGTCTCTAGC

qPCR a2g-F CAGGGCTCTGTCCAGCCTCAAC 2.6.2. Commercial master mix

a2g-R TGCCAACTGGAGCAATGGTGCAA When the analyses were performed with the commercial master
FAM−
a2g-P CGCTAGAGA/ZEN/CGCTACCTGCAATGTGC- mix, the primers, DMSO, water and sample (in the concentrations de-
IABkFQ
tailed above) were added to 10 μL of GspSSD Isothermal Master Mix
* F3/B3: outer primers, also used for qPCR with Sybr Green I. FIP/BIP: inner primers.
(OptiGene Ltd., Horsham, UK).
“ZEN” is an internal quencher proprietary from IDT
2.7. Evaluation of the limit of detection with pure DNA
the reaction volume was fixed at 20 μL and sample volume at 2 μL. The
experiments were performed in a StepOne Plus™ Real-Time PCR system An amplification curve was created with tenfold dilutions of wheat
(Applied Biosystems™, Foster City, CA, USA) with StepOne™ Software flour DNA extract ranging from 150 to 0.0015 ng/μL in Tris-EDTA 1X
v2.2.2. (TE1X, 100 mM Tris-HCl, 1 mM EDTA, pH 7.5). DNA extracts were
amplified in duplicates for each dilution, by both qPCR and qLAMP
methods. DNA extract was obtained from wheat flour as described in
2.5.1. Detection using intercalating dye (F3/B3-qPCR) Section 2.2.1.
The F3/B3 primers from the qLAMP method were added at a final
concentration of 200 nM together with 10 μL of SsoAdvanced™ 2.8. Mathematical modeling
Universal SYBR® Green Supermix (Bio-Rad Laboratories, Inc., USA).
The thermal profile was: Hot Start at 98 °C for 3 min, followed by 55 The newly developed qLAMP method was evaluated using a regular
cycles of dissociation at 95 °C for 30 s and annealing-extension at 65 °C Real-Time PCR system. The main objective was to get more information
for 45 s. A melt curve analysis was performed at the end of the qPCR about the process in real-time, and allow a comparison among the
run to verify the specificity of the reaction. different chemistries, the use of enhancer/supplements, comparison
with qPCR, and evaluation of the performance criteria. However, the
software in qPCR systems is designed to perform data analysis of qPCR
2.5.2. Detection using hydrolysis probe
results, which are based on cycle of quantification (Cq), values calcu-
In this approach, primers and probe were added at a final con-
lated by determining the point at which the fluorescence produced in
centration of 100 nM and 200 nM respectively, with 10 μL Maxima
each sample reaches the chosen threshold limit, and it is inversely re-
Probe/ROX qPCR Master Mix (Thermo Fisher Scientific Inc., Waltham,
lated to the starting copy number of the target sequence. However,
MA, USA). The thermal profile consisted of 2 min at 50 °C for Uracil-
qLAMP is an isothermal reaction, which implies that all biochemical
DNA Glycosylase (UDG) treatment, followed by 10 min at 95 °C hot
events take place in parallel and the total time of the reaction is mea-
start polymerase activation, and 55 cycles of dissociation at 95 °C for
sured rather than the number of cycles. Therefore not following cycles
30 s and annealing-extension at 65 °C for 45 s.
as qPCR, results are expressed as Time-to-threshold (Tt) which is the
measure of the reaction time at which the fluorescence of each sample
2.6. qLAMP exceeds the threshold. In order to avoid infra- or overestimation errors
in the determination of the Tt values, that would lead to false negatives
To obtain an optimal performance of the newly designed qLAMP or false positives, the Tt values were calculated by mathematical
method, two parameters were optimized, the amplification temperature modeling the qLAMP kinetics using a sigmoid equation (reparametrized
and the enhancers of the reaction (betaine and DMSO). The final re- Gompertz) as described by Garrido-Maestu, Fuciños, Azinheiro,
action volume was 20 μL, and sample volume 2 μL. For the steps of the Carvalho, and Prado (2017).
optimization process, the primer concentration was fixed at 1600 nM
for FIP/BIP and 200 nM for F3/B3, and the rest of the reaction com- 3. Results
ponents were kept as indicated below. Regarding optimization of the
temperature, it was evaluated ranging from 60 °C to 65 °C for 60 min. 3.1. Optimization of the qLAMP assay
Regarding the supplements, betaine was added to the reactions at 0.6,
0.8, and 1.0 M final concentrations, while DMSO was tested at 5% and The optimization of the temperature of the qLAMP assay showed
7.5%. Additionally, a commercial master mix, as well as one in-house that the optimal amplification temperature was 65 °C, as shown in
prepared, were tested for comparison purposes. All these tests were Fig. 1a where an estimated Tt reduction of 28 min was obtained respect
performed in duplicate. to the reaction performed at 60 °C. Regarding the supplements, better
Reactions for food sample analyses were performed in duplicate, results were obtained with DMSO rather than with betaine, being the
and the reactions contained: 5% DMSO; primer concentration was best results obtained with a 5% concentration, as depicted in Fig. 1b
1600 nM for FIP/BIP and 200 nM for F3/B3; amplification was ac- where a Tt reduction of about 15 min was obtained respect to the re-
complished at 65 °C for 90 min, with a post-amplification melting action performed with 1 M Betaine.
analysis consisting in heating up to 98 °C for 15 s, cooling down to 80 °C
for 1 min, and then raising the temperature again up to 95 °C for 15 s, 3.2. Specificity of new primers designed for qPCR and qLAMP
acquiring fluorescence every 0.3 °C. The experiments were run in the
same thermocycler used for qPCR, where each cycle corresponds to The specificity of all newly designed primers was satisfactorily
30 s, when the fluorescence was acquired. confirmed by BLAST. In addition to this, a total of 17 samples were

158
A. Garrido-Maestu et al. Food Chemistry 246 (2018) 156–163

Fig. 1. a) qLAMP temperature optimization. b) qLAMP supplement optimization. Each cycle corresponds to 30 s.

analyzed, 6 without gluten-containing cereals and 11 with gluten-con- detectable concentration. It was observed that the qPCR method with
taining cereals, which were all correctly identified with the newly de- the hydrolysis probe presented the lowest LoD, reaching 0.0015 ng/μL,
signed qPCR assays (probe and intercalating dye), while the qLAMP followed by the qPCR with the intercalating dye, which reliably de-
tests failed to detect gluten-containing cereals in the samples consisting tected down to 0.15 ng/μL, as shown in Fig. 2a and b. Regarding the
of grains (barley and rye). Also, two rice samples (flour and grains), qLAMP approaches, the in-house mixture reached reliably 0.15 ng/μL
which should not contain gluten, obtained a positive result regardless while with lower concentrations some replicates failed to amplify, and
the DNA amplification method applied (Table 1). the determined Tt values were not statistically significative (Table S1);
in this regard, similar results were obtained with the commercial master
mix as shown in Fig. 2c–d.
3.3. Pure DNA LoD

Pure DNA obtained from wheat flour was selected as the reference
material, and was ten-fold serially diluted to assess the lowest

159
A. Garrido-Maestu et al. Food Chemistry 246 (2018) 156–163

Fig. 2. LoD with pure DNA. a) Probe qPCR, b) SYBR Green qPCR c) qLAMP results obtained with in-house master mix. d) qLAMP results obtained with commercial master mix. Gray lines
represent the 95% confidence bands for the linear estimation.

3.4. Detection of gluten-containing cereals in food samples and evaluation Tovoli, & De Giorgio, 2013)
of the quantification potential in a binary model LAMP based methods are becoming increasingly applied for point-
of-care and limited-resource settings allowing in-situ analysis, due to the
Whenever qPCR approaches were selected, whether using hydro- possibility of performing nucleic acid amplification without the need of
lysis probe or intercalating dye, even the samples with the lowest expensive laboratory equipment such as thermocyclers, a high toler-
percentage of wheat flour were detected (0.001%) (Table S2), although ance for inhibitors and the possibility of detecting results with the
for quantification purposes the lower limit of the linear range should be naked eye (Oscorbin, Belousova, Zakabunin, Boyarskikh, & Filipenko,
set at 0.005% for the hydrolysis probe and 0.01% for the intercalating 2015). For their application in such conditions, LAMP based methods
dye (Fig. 3a and b). need to be thoroughly optimized and evaluated, since there are several
Regarding the qLAMP methods, the lowest detectable concentration factors that might influence their speed and analytical sensitivity, in-
was 0.005% when the commercial master mix was selected, while with cluding: the annealing events taking place that might compete with
the in-house master mix the limit of detection was 0.1% (Table S2). each other, and influence as well the speed of isothermal reactions, and
Regarding quantification, the lower limits were set at 0.1% for the the influence of the template’s innate secondary structures due to the
commercial master mix and 1% for the in-house mixture. At con- absence of multiple denaturation steps (Khorosheva, Karymov, Selck, &
centrations lower than those, the determined Tt values had high con- Ismagilov, 2016). In the absence of specific guidelines for optimization
fidence intervals and, despite being statistically significative (Table S2), as for qPCR, sensitivity of newly developed LAMP methods are typically
including them in the quantification range increased the deviation from evaluated by using 10-fold serial template dilutions that are then
linearity (Fig. 3c and d). compared to the sensitivities of a PCR method. The use of real-time
These values were set with the concentrations from which all re- measurement of product amplification for optimizing analysis condi-
plicates amplified, as whenever one of the replicates failed to amplify, tions, provides quantitative data that allows comparison among the
the sample was considered negative. effect of the different parameters, and gives a quantification potential to
the method. However, in order to use regular Real-Time PCR systems, a
mathematical model was proposed to calculate the corresponding Tt
4. Discussion values.
The optimization of the qLAMP methods focused on the temperature
The application of DNA methods for the detection of gluten-con- and the supplementation of the reactions with either betaine or DMSO.
taining cereals, was initially intended for confirmation of protein assays The estimated optimal temperature was 65 °C. Concerning the optimi-
(Miranda-Castro et al., 2016; Sandberg et al., 2003). However, in recent zation focused on the supplements added to the reaction, our results
years they have been gaining interest as fast and reliable alternatives, agree with those of Wang et al., who reported an increased performance
particularly since the development of qPCR due to its advantages in of the qLAMP reactions whenever DMSO was added rather than be-
terms of specificity, possibility of quantification, and rapidity respect to taine. However, our optimal concentration was obtained with 5% in-
conventional PCR (Martín-Fernández et al., 2015; Sandberg et al., stead with 7.5% as reported by them (Wang, Brewster, Paul, &
2003; Zeltner et al., 2009). Moreover, DNA methods seem an inter- Tomasula, 2015).
esting alternative taking into account that some of the gluten-related The first step in the evaluation of the newly developed DNA
disorders are not actually exclusively trigger by gluten (Volta, Caio,

160
A. Garrido-Maestu et al. Food Chemistry 246 (2018) 156–163

Fig. 3. Flour mix comparison, only the linear range was represented. a) results obtained with qPCR and probe b) results obtained with qPCR and SYBR Green c) results obtained with
qLAMP in-house mixture of the quantification range d) results obtained with qLAMP commercial master mix of the quantification range. Gray lines represent the 95% confidence bands
for the linear estimation.

amplification approaches (qPCR and qLAMP) consisted in the con- plant/seed samples (Demeke, Malabanan, Holigroski, & Eng, 2017;
firmation of the specificity of the new primers and probes. In this sense, Hasan et al., 2012; Xin & Chen, 2012). Although, not the scope of the
an in silico analysis performed with BLAST yielded good results for all present study, these results suggest the need for improving DNA ex-
the oligonucleotides. When the specificity analyses were performed in traction protocols from certain matrixes, and study the effect in dif-
vitro, with the corresponding protocols, to confirm the BLAST results, a ferent amplification protocols.
false positive was obtained for the rice flour. Nevertheless, these results In terms of speed, the qLAMP methods proved to be faster than both
were later considered as correct because the producer of the flour in- qPCRs, being completed the analyses in about 90 min against the
dicated that it might contain traces of gluten. Indeed, rice grain rinsed 110 min and 140 min for the qPCR-probe and the qPCR-SYBR respec-
with tap water, previous to DNA extraction, did not give a positive tively. This is an advantage of qLAMP approaches over the qPCR
amplification result, meanwhile if the rinsing step was omitted a false whenever fast results are needed, being possible to reduce it further if a
positive result was obtained. These results were consistent within both turbidity or colorimetric product detection is implemented (Arunrut
techniques in their different approaches, and highlight the sensitivity of et al., 2016; Goto, Honda, Ogura, Nomoto, & Hanaki, 2009; Mori et al.,
both. It is important to bear in mind the importance of identifying cross- 2004).
contamination during the processing or storage of raw materials or the The following step consisted in the determination of the LoD (lowest
final product, as this particular case exemplifies, and to identify how detectable analyte) which was assessed with pure wheat flour DNA. The
and when this cross-contamination occurred, since low allergen con- best results were obtained with the qPCR-probe method, followed by
centration may trigger allergenic reactions in sensitized individuals the F3/B3-qPCR and then both qLAMP approaches, proving that qPCR
(Prado et al., 2016). Sensitive enough methods are therefore required, provides higher sensitivity for the detection of gluten-containing cer-
and DNA methods involving DNA amplification have shown to provide eals.
the required sensitivity. Finally, due to the absence of certified reference materials (CRMs)
Even though good sensitivity results were obtained, gluten-con- for wheat, a binary mixture model containing various proportions of
taining grain samples (two barleys and one rye) were not detected by wheat and corn flour was built as described by Martín-Fernández at al.,
qLAMP. Nevertheless, this lack of amplification seemed related to the (Martín-Fernández et al., 2015). Following this approach, similar re-
extraction protocol rather than to the sensitivity of the technique. In sults as those obtained for the LoD with pure wheat DNA were obtained.
this sense, it is worth mentioning that even though positive results were In this sense, the lowest quantification limits were obtained with qPCR
obtained with qPCR, a delay in the Cq values was also observed. reaching 0.005% and 0.01% when using the hydrolysis probe and
Moreover, in those cases, replicates of the same samples produced in- SYBR® Green I respectively. It was even possible to detect gluten-con-
consistent amplification. This, in combination with the fact that taining cereals in lower proportion mixtures (0.002% and 0.001%),
whenever flour instead whole grain was used for the analyses, gluten- although those samples fell out of the linear range and were not con-
containing cereal was always correctly detected, which emphasizes the sidered for quantification purposes (Fig. 3a and b). The obtained qPCR
hypothesis that it is an issue related to the type of sample rather than to results are in agreement with those reported by Martín-Fernández et al.
the performance of the DNA amplification techniques. Indeed, previous (2015).
studies have already reported difficulties in the extraction of DNA from Regarding the qLAMP methods, the limit of detection and

161
A. Garrido-Maestu et al. Food Chemistry 246 (2018) 156–163

quantification were 0.005% and 0.1% when the commercial master mix Diaz-Amigo, C., & Popping, B. (2012). Gluten and gluten-free: Issues and considerations
was selected, while with the in-house master mix the obtained values of labeling regulations, detection methods, and assay validation. Journal of AOAC
International, 95(2), 337–348. http://dx.doi.org/10.5740/jaoacint.SGE_Diaz-Amigo.
were ten times higher (0.1% and 1%). Within the qLAMP assays, more Diaz-Amigo, C., & Popping, B. (2013). Real-time PCR analysis of food allergens and
uniform results among replicates were achieved with the commercial gluten. In David Rodríguez-Lázaro (Ed.). Real-time PCR in food science: Current tech-
master mix. This may be related to the higher number of pipetting steps nology and applications (pp. 239–252). Burgos: Caister Academic Press.
EFSA (2014). Scientific Opinion on the evaluation of allergenic foods and food ingredients
required for the in-house method (Bst pol, dNTP, buffer, DMSO, pri- for labelling purposes. EFSA Journal, 12(11), 1–277. http://dx.doi.org/10.2903/j.
mers, and water) compared to the commercial master mix (mix, DMSO, efsa.2014.NNNN.
primers, and water). Also, the commercial mix contains inorganic European Commision (2009). Commission regulation (EC) No. 41/2009 of 20 January
2009 concerning the composition and labeling of foodstuffs suitable for people in-
pyrophosphatase thus no pyrophosphate is accumulated, making the tolerant to gluten. Official Journal of the European Union, 52(41), 3–5.
solutions clear even in positive reactions and so avoiding any possible Foschia, M., Horstmann, S., Arendt, E. K., & Zannini, E. (2016). Nutritional therapy
interference associated with the increase in turbidity in the fluorescence -Facing the gap between coeliac disease and gluten-free food. International Journal of
Food Microbiology, 239(2016), 113–124. http://dx.doi.org/10.1016/j.ijfoodmicro.
detection; resulting overall in an enhancement of both limits, detection
2016.06.014.
and quantification. Garrido-Maestu, A., Fuciños, P., Azinheiro, S., Carvalho, J., & Prado, M. (2017).
Systematic Loop-mediated isothermal amplification assays for rapid detection and
5. Conclusions characterization of Salmonella spp., Enteritidis and Typhimurium in food samples.
Food Control, 80, 297–306. http://dx.doi.org/10.1016/j.foodcont.2017.05.011.
Goto, M., Honda, E., Ogura, A., Nomoto, A., & Hanaki, K. I. (2009). Colorimetric detection
In the current study four different approaches based on two dif- of loop-mediated isothermal amplification reaction by using hydroxy naphthol blue.
ferent DNA amplification techniques were successfully developed, BioTechniques, 46(3), 167–172. http://dx.doi.org/10.2144/000113072.
Hara-Kudo, Y., Yoshino, M., Kojima, T., & Ikedo, M. (2005). Loop-mediated isothermal
tested and compared, being this the first study to describe the devel- amplification for the rapid detection of Salmonella. FEMS Microbiology Letters, 253(1),
opment of a specific method for the detection of gluten based on 155–161.
qLAMP. All protocols behave equally well in terms of specificity. Haraszi, R., Chassaigne, H., Maquet, A., & Franz, Ulberth. (2011). Analytical methods for
detection of gluten in food-method developments in support of food labeling legis-
Regarding the sensitivity, the best results were obtained by qPCR. It is lation. Journal of AOA C International, 94(4), 1006–1025.
worth to highlight that, even though the qLAMP approaches were not Hasan, S., Prakash, J., Vashishtha, A., Sharma, A., Srivastava, K., Sagar, F., ... Mishra, S.
suitable for seeds, they provided the fastest results for the rest of the (2012). Optimization of DNA extraction from seeds and leaf tissues of
Chrysanthemum (Chrysanthemum indicum) for polymerase chain reaction.
samples. Bioinformation, 8(5), 225–228. http://dx.doi.org/10.6026/97320630008225.
Khorosheva, E. M., Karymov, M. A., Selck, D. A., & Ismagilov, R. F. (2016). Lack of
Acknowledgements correlation between reaction speed and analytical sensitivity in isothermal amplifi-
cation reveals the value of digital methods for optimization: Validation using digital
real-time RT-LAMP. Nucleic Acids Research, 44(2), e10. http://dx.doi.org/10.1093/
This work was supported by a Marie Curie COFUND Action (Project nar/gkv877.
No: 600375. NanoTRAINforGrowth – INL Fellowship programme in Ma, X. Jun, Shu, Y., Nie, K. Long, Qin, M., Wang, D. Yan, Gao, R. Bao, ... Dong, X. Ping
nanotechnologies for biomedical, environment and food applications), (2010). Visual detection of pandemic influenza A H1N1 Virus 2009 by reverse-
transcription loop-mediated isothermal amplification with hydroxynaphthol blue
by the project Nanotechnology Based Functional Solutions (NORTE-01- dye. Journal of Virological Methods, 167(2), 214–217. http://dx.doi.org/10.1016/j.
0145-FEDER-000019), supported by Norte Portugal Regional jviromet.2010.03.027.
Operational Programme (NORTE2020), under the PORTUGAL 2020 Martín-Fernández, B., Costa, J., Oliveira, M. B. P. P., López-Ruiz, B., & Mafra, I. (2015).
Screening new gene markers for gluten detection in foods. Food Control, 56, 57–63.
Partnership Agreement, through the European Regional Development http://dx.doi.org/10.1016/j.foodcont.2015.02.047.
Fund (ERDF), and by NANOIMMUNOTECH S.L. in the framework of the Miranda-Castro, R., de-Los-Santos-Álvarez, N., Miranda-Ordieres, A. J., & Lobo-Castañón,
project Smart Factory for Safe Foods (SF4SF) supported by FEDER M. J. (2016). Harnessing aptamers to overcome challenges in gluten detection.
Biosensors, 6(2), http://dx.doi.org/10.3390/bios6020016.
Innterconecta 2015 (EXP-00082964/ITC-20151195). Mittag, D., Niggemann, B., Sander, I., Reese, I., Fiedler, E. M., Worm, M., ... Reese, G.
(2004). Immunoglobulin E-reactivity of wheat-allergic subjects (baker’s asthma, food
Appendix A. Supplementary data allergy, wheat-dependent, exercise-induced anaphylaxis) to wheat protein fractions
with different solubility and digestibility. Molecular Nutrition and Food Research,
48(5), 380–389. http://dx.doi.org/10.1002/mnfr.200400016.
Supplementary data associated with this article can be found, in the Mori, Y., Kitao, M., Tomita, N., & Notomi, T. (2004). Real-time turbidimetry of LAMP
online version, at http://dx.doi.org/10.1016/j.foodchem.2017.11.005. reaction for quantifying template DNA. Journal of Biochemical and Biophysical
Methods, 59(2), 145–157. http://dx.doi.org/10.1016/j.jbbm.2003.12.005.
Mujico, J. R., Lombardía, M., Mena, M. C., Méndez, E., & Albar, J. P. (2011). A highly
References sensitive real-time PCR system for quantification of wheat contamination in gluten-
free food for celiac patients. Food Chemistry, 128(3), 795–801. http://dx.doi.org/10.
Arunrut, N., Kampeera, J., Sirithammajak, S., Sanguanrut, P., Proespraiwong, P., 1016/j.foodchem.2011.03.061.
Suebsing, R., & Kiatpathomchai, W. (2016). Sensitive visual detection of AHPND Notomi, T., Okayama, H., Masubuchi, H., Yonekawa, T., Watanabe, K., Amino, N., &
bacteria using loop-mediated isothermal amplification combined with DNA-functio- Hase, T. (2000). Loop-mediated isothermal amplification of DNA. Nucleic Acids
nalized gold nanoparticles as probes. PLoS One, 11(3), e0151769. Research, 28(12), e63. http://dx.doi.org/10.1093/nar/28.12.e63.
Bio-Qiagen, C. L. C. (2016). CLC sequence viewer. Denmark: Aarhus. Oscorbin, I. P., Belousova, E. A., Zakabunin, A. I., Boyarskikh, U. A., & Filipenko, M. L.
Bruins Slot, I. D., van der Fels-Klerx, H. J., Bremer, M. G. E. G., & Hamer, R. J. (2015). (2015). Comparison of fluorescent intercalating dyes for quantitative loop-mediated
Immunochemical detection methods for gluten in food products: Where do we go isothermal amplification (qLAMP). BioTechniques, 58(June), 308–317. http://dx.doi.
from here? Critical Reviews in Food Science and Nutrition, 8398(March), http://dx.doi. org/10.2144/000114432.
org/10.1080/10408398.2013.847817. Prado, M., Boix, A., & von Holst, C. (2012). Novel approach for the simultaneous de-
D’Agostino, M., Diez-Valcarce, M., Robles, S., Losilla-Garcia, B., & Cook, N. (2015). A tection of DNA from different fish species based on a nuclear target: Quantification
loop-mediated isothermal amplification-based method for analysing animal feed for potential. Analytical and Bioanalytical Chemistry, 403, 3041–3050. http://dx.doi.org/
the presence of Salmonella. Food Analytical Methods, 8(10), 2409–2416. http://dx.doi. 10.1007/s00216-012-5852-7.
org/10.1007/s12161-015-0148-0. Prado, M., Ortea, I., Vial, S., Rivas, J., Calo-Mata, P., & Barros-Velázquez, J. (2016).
D’Agostino, M., Robles, S., Hansen, F., Ntafis, V., Ikonomopoulos, J., Kokkinos, P., & Advanced DNA- and protein-based methods for the detection and investigation of
Cook, N. (2016). Validation of a loop-mediated amplification/iso 6579-based method food allergens. Critical Reviews in Food Science and Nutrition, 56(16), 2511–2542.
for analysing soya meal for the presence of Salmonella enterica. Food Analytical http://dx.doi.org/10.1080/10408398.2013.873767.
Methods, 9(11), 2979–2985. http://dx.doi.org/10.1007/s12161-016-0602-7. Sandberg, M., Lundberg, L., Ferm, M., & Malmheden Yman, I. (2003). Real Time PCR for
Dahinden, I., von Büren, M., & Lüthy, J. (2001). A quantitative competitive PCR system to the detection and discrimination of cereal contamination in gluten-free foods.
detect contamination of wheat, barley or rye in gluten-free food for coeliac patients. European Food Research and Technology, 217(4), 344–349. http://dx.doi.org/10.
European Food Research and Technology, 212(2), 228–233. http://dx.doi.org/10. 1007/s00217-003-0758-4.
1007/s002170000252. Sapone, A., Bai, J. C., Ciacci, C., Dolinsek, J., Green, P. H. R., Hadjivassiliou, M., ...
Demeke, T., Malabanan, J., Holigroski, M., & Eng, M. (2017). Effect of source of DNA on Fasano, A. (2012). Spectrum of gluten-related disorders: Consensus on new nomen-
the quantitative analysis of genetically engineered traits using digital PCR and real- clature and classification. BMC Medicine, 10(1), 13. http://dx.doi.org/10.1186/1741-
time PCR. Journal of AOAC International, 100(2), 492–498. http://dx.doi.org/10. 7015-10-13.
5740/jaoacint.16-0284 492-498(7). Stephan, O., & Vieths, S. (2004). Development of a real-time PCR and a sandwich ELISA
for detection of potentially allergenic trace amounts of peanut (Arachis hypogaea) in

162
A. Garrido-Maestu et al. Food Chemistry 246 (2018) 156–163

processed foods. Journal of Agricultural and Food Chemistry, 52(12), 3754–3760. food hygiene.
http://dx.doi.org/10.1021/jf035178u. Wieser, H., Koehler, P., & Konitzer, K. (2014). Celiac disease and gluten: Multidisciplinary
Tomita, N., Mori, Y., Kanda, H., & Notomi, T. (2008). Loop-mediated isothermal ampli- challenges and opportunities. In P. Koehler, W. Herbert, & K. Konitzer (Eds.). London:
fication (LAMP) of gene sequences and simple visual detection of products. Nature Academic Press Elsevier.
Protocols, 3(5), 877–882. Xin, Z., & Chen, J. (2012). A high throughput DNA extraction method with high yield and
Untergasser, A., Nijveen, H., Rao, X., Bisseling, T., Geurts, R., & Leunissen, J. A. M. quality. Plant Methods, 8(1), 26. http://dx.doi.org/10.1186/1746-4811-8-26.
(2007). Primer3Plus, an enhanced web interface to Primer3. Nucleic Acids Research, Zeltner, D., Glomb, M. A., & Maede, D. (2009). Real-time PCR systems for the detection of
35(Suppl. 2), W71–W74. http://dx.doi.org/10.1093/nar/gkm306. the gluten-containing cereals wheat, spelt, kamut, rye, barley and oat. European Food
Volta, U., Caio, G., Tovoli, F., & De Giorgio, R. (2013). Non-celiac gluten sensitivity: Research and Technology, 228(3), 321–330. http://dx.doi.org/10.1007/s00217-008-
Questions still to be answered despite increasing awareness. Cellular and Molecular 0937-4.
Immunology, 10(10), 383–392. http://dx.doi.org/10.1038/cmi.2013.28. Zhang, G., Brown, E. W., & González-Escalona, N. (2011). Comparison of real-time PCR,
Wang, D., Brewster, J., Paul, M., & Tomasula, P. (2015). Two methods for increased reverse transcriptase real-time PCR, loop-mediated isothermal amplification, and the
specificity and sensitivity in loop-mediated isothermal amplification. Molecules, FDA conventional microbiological method for the detection of Salmonella spp. in
20(4), 6048–6059. http://dx.doi.org/10.3390/molecules20046048. produce. Applied and Environmental Microbiology, 77(18), 6495–6501. http://dx.doi.
WHO and FAO. (2010). Joint FAO/WHO food standards programme Codex committee on org/10.1128/AEM.00520-11.

163