Development (2020-21)

Wageningen Academic
World Mycotoxin Journal, 2022; 15 (1): 3-25 P u b l i s h e r s
Developments in mycotoxin analysis: an update for 2020-2021
S.A. Tittlemier1*, B. Cramer2, C. Dall’Asta3, M.C. DeRosa4, V.M.T. Lattanzio5, R. Malone6, C. Maragos7, M. Stranska8
and M.W. Sumarah9
1Canadian Grain Commission, Grain Research Laboratory, 1404-303 Main St, Winnipeg, MB, R3C 3G8, Canada;
2Westfälische Wilhelms-Universität Münster, Institute of Food Chemistry, Corrensstr. 45, 48149 Münster, Germany; 3Università
di Parma, Department of Food and Drug, Viale delle Scienze 27/A, 43124 Parma, Italy; 4Department of Chemistry, Carleton
University, Ottawa, ON, K1S 5B6, Canada; 5National Research Council of Italy, Institute of Sciences of Food Production,
via Amendola 122/O, 70126 Bari, Italy; 6Trilogy Analytical Laboratory, 870 Vossbrink Dr, Washington, MO 63090, USA;
7United States Department of Agriculture, ARS National Center for Agricultural Utilization Research, Peoria, IL 61604,
USA; 8Department of Food Analysis and Nutrition, Faculty of Food and Biochemical Technology, University of Chemistry
and Technology, Technicka 5, Prague, 166 28, Prague, Czech Republic; 9Agriculture and Agri-Food Canada, London Research
and Development Centre, 1391 Sandford Street, London, ON, N5V 4T3, Canada; [email protected]
Received: 8 November 2021 / Accepted: 4 February 2022

© 2022 Wageningen Academic Publishers
OPEN ACCESS REVIEW ARTICLE

Abstract
This review summarises developments published in the period from mid-2020 to mid-2021 on the analysis of a
number of diverse matrices for mycotoxins. Notable developments in all aspects of mycotoxin analysis, from sampling
and quality assurance/quality control of analytical results, to the various detection and quantitation technologies
ranging from single mycotoxin biosensors to comprehensive instrumental methods are presented and discussed.
The summary and discussion of this past year’s developments in detection and quantitation technology covers
chromatography with targeted or non-targeted high resolution mass spectrometry, tandem mass spectrometry,
detection other than mass spectrometry, biosensors, as well as assays using alternatives to antibodies. This critical
review aims to briefly present the most important recent developments and trends in mycotoxin determination, as
well as to address limitations of the presented methodologies.
Keywords: sampling, multi-mycotoxin analysis, quality control, multiplex, biosensor, chromatography, non-targeted
analysis, aptamer, ELISA, molecularly imprinted polymer, high resolution mass spectrometry, tandem mass spectrometry
1. Introduction We have continued with organising the review to cover

sampling (section 2), quality control of mycotoxin
This article is the latest instalment in a series of annual analyses (section 3), chromatography with tandem mass
reviews highlighting analytical method developments for spectrometry (MS/MS; section 4), chromatography with
mycotoxin determination, continuing from the previous targeted high resolution mass spectrometry (HRMS;
review covering the mid-2019 to mid-2020 period section 5), chromatography with non-targeted HRMS
(Tittlemier et al., 2021). As with the previous reviews in this (section 6), chromatography with non-mass spectrometric
series, our primary purpose is to raise awareness of notable (MS) detection (section 7), multiplex biosensors (section 8),
developments and advances in the analysis of mycotoxins single mycotoxin or single mycotoxin family biosensors
and to highlight where improvements can help the wider (section 9), and assays using antibody analogues (section 10).
‘mycotoxin community’ of analysts, researchers, regulators,
and those relying on analytical data. This review is not While many valuable advances were published over the
meant to be an exhaustive list of publications on mycotoxin past year, contributing authors from most sections of
analytical methods. the review noted the continued absence of important
details in many publications. In particular, the absence
ISSN 1875-0710 print, ISSN 1875-0796 online, DOI 10.3920/WMJ2021.27523

S.A. Tittlemier et al.
of specific information on sampling and processing in individual kernels. These data reinforce the constitution
equipment and procedures used, method sensitivity heterogeneity of aflatoxins and fumonisins in bulk maize
reported on a sample mass basis, and the application of (i.e. differences in amount of mycotoxin between kernels),
novel detection technologies to ‘real’ matrices continued in as there were low frequencies of high concentrations in
many publications. Authors, reviewers, and editors need to single kernels. Even though the selection of kernels for
ensure that this information is communicated thoroughly analysis from the bulk samples was biased towards those
and clearly, in addition to ensuring reported results are kernels displaying risk factors associated with the presence
based on sound science. In order to promote this valuable of mycotoxins, only 1 and 5% of kernels contained aflatoxins
communication approach, the World Mycotoxin Journal is ≥20 μg/kg and fumonisins ≥2 mg/kg, respectively. The
planning to update its instructions to authors and reviewers aim of the biased selection was to produce a set of test
to minimise the publication of articles lacking important materials to facilitate development of the screening
information. prediction model. Interestingly, estimates of aflatoxin and
fumonisin concentrations in bulk maize made from the
2. Sampling targeted individual kernel results showed no bias when
compared to the concentrations obtained from analysis of
Research on sampling and sample processing published a 250 g sub-sample taken from the bulk samples. The lack
over the past year include characterisation of mycotoxin of apparent bias is most likely reflecting very high kernel-
distributions among maize kernels (Chavez et al., 2022) to-kernel differences, that overshadowed any effect from
and within fields (Yi et al., 2021), an enhanced analysis kernel selection bias.
of existing data sets on mycotoxin spatial distribution
in bulk volumes (Kerry et al., 2021), and descriptions of Yi et al. (2021) sampled two fields of wheat and rice and
sampling techniques for blood (Vidal et al., 2021), grain performed a spatial analysis of deoxynivalenol (DON)
dust (Limay-Rios and Schaafsma, 2021), and for pooling and aflatoxins, respectively. The fields were segmented
samples for mycotoxin screening (Cheng et al., 2020). A very into three sections (middle, peripheral, and ‘other’ – the
welcome retrospective analysis was also published on the span between the middle and the periphery of the fields).
disappointing decrease in researchers’ consideration of The authors reported a ‘filtration effect’ by the peripheral
proper sampling and processing of maize (Kumphanda section, with more dust, fungal spores, and mycotoxins
et al., 2021). on grain collected from this section. The reported results
emphasise the need to consider spatial variation within a
Kumphanda et al. (2021) reviewed papers on the analysis field when sampling plant material for mycotoxin analysis.
of mycotoxins in maize published in 1991 through 2020. Unfortunately, no methodological details were given on the
In particular, they focused on the laboratory sample mass fungal analyses, and the description of sampling was also
(amount of maize comminuted) and test portion mass severely lacking. There was no information provided on
(amount of ground maize extracted). The results of their the mass of grain collected for samples; no information on
review were disheartening. Only 50% of publications the tools and procedures used for comminution of whole
explicitly stated the laboratory sample mass used, greater grain, nor the particle sizes obtained; and no information
than 67% of publications did not specify the type of grinder on how the comminuted grain was sub-sampled. The lack
and sieve size used to comminute grain, and laboratory of methodological details prevent full consideration of the
sample and test portion masses both showed a statistically results. However, the purported conclusions could be used
significant decline over the three decades reviewed. The to develop more robust field studies on the situational
authors proposed that high solvent costs, increased factors that impact field-sampling for mycotoxin analysis.
time needed for sample preparation and processing, and
increased resources needed to dispose of excess sample, all Kerry et al. (2021) re-visited spatial analyses of DON and
contributed to the decrease in laboratory sample and test ochratoxin A (OTA) in a truckload of wheat and fumonisins
portion sizes at the expense of increasing total variance in a maize pile. Because of the heterogeneity of OTA
of test results. In addition to the journal editors and food and fumonisins in the grain studied, the highly skewed
safety bodies challenged to uphold the quality and reporting concentrations obtained through previous testing were
of sampling processes by the authors, funding bodies and transformed to normal scores to properly investigate the
reviewers also need to hold researchers accountable for spatial correlation of the mycotoxins within the wheat
using sound sampling and sample preparation and for truckload and maize pile. Different from the previous
clearly providing this necessary information in publications. analysis (Rivas Casado et al., 2009), Kerry et al. found that
OTA showed clustering within the truckload. The authors
As part of their work on improving high throughput UV- also suggested that understanding spatial variability of
vis-near-infrared (NIR) screening of maize kernels for bulk volumes could improve accuracy and reduce the cost
contamination with aflatoxin and fumonisins, Chavez et al. of sampling plans, but unfortunately did not discuss how
(2022) generated a substantial data set on concentrations consistent (or not) the spatial distribution of DON, OTA,
4 World Mycotoxin Journal 15 (1)

Developments in mycotoxin analysis 2020-2021
and fumonisins would be amongst different truckloads, reduction in numbers of analyses gained by pooling was
piles, and other bulk volumes of grain. offset by increased cost in supplies due to the complexity
of pooling strategies. Considering two-dimensional 48-well
Other research published over the last year investigated (6 rows × 8 columns) and 96-well (8 rows x 12 columns)
sampling techniques. Vidal et al. (2021) reported on ELISA plates, four pooling strategies were examined: (1)
‘volumetric absorptive microsampling’ (VAMS) as an one-dimensional ‘by rows’ (pooling of 8 or 12 samples), (2)
alternative to dried blood spot sampling. With VAMS a one-dimensional ‘by columns’ (pooling of 6 or 8 samples),
fixed volume of blood is absorbed by a polymeric absorbent (3) two-dimensional (pooling of samples by considering
tip, which is subsequently dried and extracted. This blood results from all relevant columns and rows), plus (4) Shifted
sampling technique avoids the challenges with spreading Transversal Design (‘layers’ of sample combinations are
and associated heterogeneity issues when hole punch constructed and varied to minimise the number of times
sampling dried blood spots. The authors provided details any two samples appear together in a pool). Cheng et al.
on the absorbing and extraction parameters of VAMS used concluded that one-dimensional and Shifted Transversal
in their method, and detected similar concentrations of Design pooling minimise reagent costs, taken as the relative
a number of mycotoxins present in whole blood [OTA, number of tests needed to identify positive individual
ochratoxin α (OTα), zearalenone (ZEN), aflatoxin B 1 samples as compared to the scenario of no pooling, by 70
(AFB1)] using VAMS and an existing validated liquid-liquid and 80%, respectively, but only when the expected frequency
extraction method. Stability assessment showed absorbed of analyte prevalence in samples is below 21% (for 48-
whole blood samples were stable after 1 and 3 weeks of well plates) and <13-21% (for 96-well plates). In these two
storage, suggesting VAMS is a robust sampling method that pooling strategies, pipetting costs increased up to a factor
can be a useful tool when analysis of whole blood samples of 4.3-fold. The authors did not assign monetary value to
cannot take place immediately after sampling. costs of reagents, other consumables, and personnel time
in order to estimate total savings as laboratories may value
Limay-Rios and Schaafsma (2021) reported on research these expenditures differently.
that measured a number of mycotoxins in paired whole
grain winter wheat and grain dust samples collected during 3. Quality control of mycotoxin analyses
unloading of combines during harvest and unloading of
grain storage bins into trucks. The grain flow was manually To guarantee a uniform application of regulations for
sampled using a cup, and dust was sampled using a modified mycotoxins, harmonisation and verification of standardised
commercially available vacuum cleaner during the entire methods for analysis of mycotoxins is of the uppermost
period of grain sampling near the point the grain flow was importance. From mid-2020 to mid-2021, results of two
deposited. Concentrations of 25 out of the 26 mycotoxins collaborative studies focused on validation of methods
detected were greater in dust than in whole grain (by in support of EU regulations as requested by the M/520
factors of 3 to 62), with only deoxynivalenol-3-glucoside standardisation mandate of the European Commission,
(DON-3G) at higher concentrations in the whole grain, were published. The first study of De Girolamo et al.
indicating dust may be a good matrix for screening for the (2020) addressed the fourth item of the standardisation
presence of many mycotoxins. The authors reported strong mandate (EC, 2013), aiming at optimising and validating
correlations between dust and grain concentrations of OTA an analytical method for the simultaneous determination of
(r2=0.976) and DON (r2=0.949), and discussed how models nivalenol (NIV), DON, 3-acetyl-deoxynivalenol (3-ADON),
developed from their data can be used to predict grain 15-acetyl-deoxynivalenol (15-ADON), T-2 toxin (T-2),
mycotoxin concentrations in some cases but that variability HT-2 toxin (HT-2), and ZEN in cereals and cereal products
in the dust/grain relationship limits the application of the by liquid chromatography with tandem mass spectrometry
model to more stringent specifications, such as 1000 μg/kg (LC-MS/MS). Sample was extracted using acetonitrile/
DON in wheat. The low concentrations and detection water 84/16 (v/v), and the extract was analysed after solid
frequencies observed precludes use of the model to predict phase extraction (SPE; using Oasis® HLB; Waters, Milford,
concentrations in grain for the other mycotoxins included MA, USA) clean-up. Data from the collaborative study,
in the study. as well as feedback from participants, confirmed that the
chromatographic separation of 3-ADON and 15-ADON
Cheng et al. (2020) examined pooling strategies to remains one of the most challenging steps of the method.
reduce the cost of mycotoxin screening by 48- and 96- At the same time, lower (71-78%) recoveries were observed
well enzyme-linked immunosorbent assay (ELISA) tests for NIV. These were still deemed ‘acceptable’ by the authors;
using simulations incorporating the maize single kernel unfortunately, the benchmark for ‘acceptable’ was not
aflatoxin and fumonisin data from Chavez et al. (2022). discussed. Despite this, the results of the collaborative
The authors modelled outcomes of combining extracts study met the required method performance criteria with
of multiple samples and analysing the composite extracts HorRat values <2.0 and have been recently adopted as a
to quickly rule out negative samples and assessed if the CEN standard (EN17280:2019).
World Mycotoxin Journal 15 (1) 5

The second study of Tangni et al. (2021) was focused on calculating a mean value from participants’ results for
determination of CIT in food by LC-MS/MS, as a response generating the assigned value, without any additional
to the eleventh priority of the M/520 standardisation measurements (i.e. of a certified reference material).
mandate. Within this validation study, in-house prepared They point to the fact that this approach suffers from the
matrix reference materials of red yeast rice, wheat flour, rather high risk of biased results when the consensus of
and Ginkgo biloba leaves were used, which is in line with participants’ results is low (Çetinkaya and Çetinkaya 2021).
current trends preferring the matrix reference materials The topic of non-conformity of PT results and influence
with incurred mycotoxins, over the reference materials on the assigned value was previously discussed in study of
prepared using fortification (Tittlemier et al., 2021). From Zachariasova et al. (2014), where bimodal distribution of
all of the tested matrices, Ginkgo biloba was reported to be results for DON and ZEN in cereals was observed. In this
problematic by some participants (in particular, incorrect particular case it was caused by the detection technique
CIT ion rations, as well as influencing of the signal of the used by the laboratories, and overestimation of results
13C-labelled internal standard and native CIT were observed obtained by immunoassays cross-reacting with DON
in this matrix), indicating that increased attention should and ZEN ‘masked’ conjugates and matrix interferences,
be paid when analysing Ginkgo biloba-based products. when compared to results of liquid chromatography mass
Nevertheless, evidenced by HorRat values <2.0, the results spectrometry methods (Zachariasova et al., 2014). This
of the collaborative trial demonstrated that the applied demonstrates that the design of the PT is essential, and
analytical method could be standardised for all of the new contributions to this topic may be inspirational, and
matrices, at levels considered by the European Commission lead to improvements in quality control of generated
for setting regulatory limits (Tangni et al. 2021). results and improved performance standards in analytical
laboratories. This study also emphasises, even though there
Çetinkaya and Çetinkaya (2021) published a paper is an increased cost, one of the best ways of assigned value
introducing a new proficiency testing (PT) scheme suggested generation is by utilising isotopically labelled mycotoxins
by the experienced PT organiser, Turkish National Food as internal standards, which are able to effectively manage
Reference Laboratory. This institution has eight years of the matrix effects and achieve a satisfactory precision
experience with organising PTs focused on determination of (Tittlemier et al., 2021).
AFs and OTA, as the most frequent mycotoxins in Turkish
export products (dried fruits, pistachio, hazelnuts, and The last paragraph of this sub-chapter is devoted to
others). Instead of the generally used PT scheme utilising quality assurance in conjugated mycotoxins analysis, as
the Horwitz-Thompson model to obtain the standard the gradually increasing area of interest of the mycotoxin
deviation for z-score calculation (Thompson 2000), the analysts’ community. The biggest obstacle in accurate
newly suggested scheme introduced the possibility of analysis and quantification of mycotoxin conjugates is the
calculation a ‘robust standard deviation’ by the Q/Hampel limited commercial availability of analytical standards. To
method (ISO, 2015). The results from previous rounds of date, only some mycotoxin conjugates can be purchased in
PTs obtained between the years 2013 and 2020 were utilised sufficient purity. In addition to well-known DON-3G from
to recalculate and compare results from both evaluation Sigma Aldrich (St. Louis, MO, USA), conjugates of ZEN and
principles. Contrary to the Horwitz-Thompson model, its metabolites as e.g. zearalenone-14-glucoside (ZEN-14G),
where the relative standard deviation (RSD) is set to 22% zearalenone-14-sulfate (ZEN-14S), zearalenone-14,16-
for concentrations below 120 µg/kg (this value has been set disulfate (ZENdiS), α-zearalenol-14-glucoside (α-ZEL-
empirically based on observations from many collaborative 14G), α-zearalenol-14-sulfate (α-ZEL-14S), β-zearalenol-
trials of many parameters over an extended time period), 14-glucoside (β-ZEL-14G), β-zearalenol-14-sulfate
the new model calculates the standard deviation from a (β-ZEL-14S), and zearalanone-14-glucoside (ZAN-14G)
more homogenous data set (results of specific mycotoxins can be purchased from ASCA GmbH (Berlin, Germany).
from specific matrices). The authors claimed use of the Q/ These eight conjugates were incorporated into the multi-
Hampel model would provide more realistic results, but analyte method for determination of mycotoxins and their
did not elaborate on this claim. The recalculated ‘robust metabolites in beer (Rausch et al., 2021).
standard deviation’ was 17%. The results of the study showed
that reducing RSD from 22 to 17% narrowed the limit values It is also worth mentioning biotransformation by plants
and satisfactory range of z-score by about 30%. Despite the or microorganisms as another option of gaining these
fact that a more strict RSD limits the success rate of the rare and rather expensive mycotoxin conjugates. Recently,
participants and is more challenging for the laboratories, Peters et al. (2021) introduced an easily implementable
it is beneficial in terms of improved quality assurance of microbiological method for production of ZEN glycosides.
the mycotoxins analysis (Çetinkaya and Çetinkaya 2021). This was achieved by Cunninghamella fungal strains in
sulphate-depleted media, where the lack of sulphur steered
The authors of the study also mention that nowadays, many the reaction towards the preferred abundant production
PT organisers frequently use the low-cost approach of of ZEN-14G and zearalenone-16-glucoside (ZEN-16G).

Interestingly, when the developed approach was applied The same group was also involved in a validation study
for other mycotoxins (namely DON, AFB 1, FB1, T-2, offering relevant insight into the parameters to control for
and OTA, no obvious biotransformation was observed accurate quantification of multiple contaminants (Sulyok
at the conditions previously used for ZEN (Peters et al., et al., 2020). In addition to the value of the data provided,
2021) indicating that different microorganisms have the paper is a blueprint for further validation studies of
specific metabolic pathways for detoxification of different multiclass protocols. In particular, the authors underlined
mycotoxins. For checking the concentrations of standards the importance of a thorough evaluation of matrix effects,
of conjugated mycotoxins, measurement of absorbance which can vary a lot among chemical classes, and may
and calculation of concentrations via the molar absorption affect final results.
coefficients, as described by Scheibenzuber et al. (2021)
is highly recommended. Scheibenzuber et al. (2021) also As already said, several papers have addressed the
describe a method for determining molar absorption occurrence of mycotoxins in unusual food matrices, or have
coefficients of previously uncharacterised conjugated enlarged the number of mycotoxins potentially detected in
mycotoxins. usual food matrices. In this context, an interesting example
is the multi-toxin determination of regulated, emerging and
4. Chromatography with MS/MS modified mycotoxins in milk proposed by González-Jartín et
al. (2021). The authors validated a method for the accurate
As already described in the last updates (Tittlemier et al., quantification of 40 analytes in milk, using a QuEChERS-
2020, 2021), the trend towards the development of multi- based extraction. Although aflatoxin M1 (AFM1) is the
toxin methods is still rising, as attested by the number of only mycotoxin regulated in milk so far, protocols able
reports published over the past year. The growing interest to provide accurate data about the potential occurrence
of policy-makers around multiple chemicals in food, has of other mycotoxins are needed for a more informed risk
supported the expansion of multi-analyte protocols, able assessment. In particular, the authors reported on the
to detect different groups of contaminants in the same frequent occurrence of enniatins and beauvericin (BEA)
food matrix (i.e. mycotoxins and pesticides in vegetables, in milk, although at low concentrations. It should be noted
or mycotoxins and veterinary drugs in food from animal that, in spite of the large range of chemically diverse analytes
origin). and the complex matrix, no internal standard was used.
This might affect the overall analytical performance of
While the application of LC-MS/MS methods to mycotoxins the method, especially when a large number of samples is
in regulated matrices is almost routinely performed, and screened over time.
therefore less novelty can be expected in that field, several
protocols have been developed for unusual matrices, thus As for infant food monitoring, Braun et al. (2021) developed
offering a valuable contribution in the assessment of a multi-toxin strategy based on two complementary
emerging risks. Two papers published over this year can methods, covering 46 mycotoxins overall in infant formulae
be regarded as the real breakthrough in the use of LC-MS/ and meals. In particular, the authors adopted two LC-MS/
MS for mycotoxin analysis. MS methods targeting two sets of mycotoxins, the former
covering the most common Aspergillus, Penicillium, and
Steiner et al. (2020) presented the first multiclass Fusarium mycotoxins, and the latter considering Alternaria
quantitative method for the determination of more than mycotoxins. Since the study focused on infant food, the
1,200 biotoxins and other contaminants in complex feed. need to achieve a very high sensitivity did not allow for
The efforts in the optimisation of the chromatographic any compromise in terms of chromatographic separation
separation together with the appropriate MS/MS conditions, and matrix effect, thus requiring the use of two separate
have allowed the accurate quantification of compounds with protocols of analysis. Due to the lack of certified reference
a wide range of physico-chemical properties. All the quality materials and the large variety of food components used
parameters fall within the EU validation guidelines (EC, for infant food formulation, no matrix-matched calibration
2006). Validation data provided have confirmed matrix was used. Satisfactory validation parameters were obtained
effects as the more critical control point of the entire using a thorough standard calibration followed by spiking
procedure. Although the inter-laboratory transferability experiments, although 12 analytes were beyond the
of such protocol could be cumbersome due to the number satisfactory recovery range 50-150%; likely due to strong
of parameters to keep under control, it provides a strong matrix effects. The sensitivity range was good, with limits of
proof of concept about the potential of the most advanced detection (LODs) ranging between 0.1 µg/kg for aflatoxicol,
LC-MS/MS instrumentation for multiclass analysis. In a metabolite of AFB1, and 12 µg/kg for tenuazonic acid
consideration of increasing interest in combined effects (TeA). According to the data collected, trace levels of 17
of multiple toxicants (EC, 2020), this study provides a mycotoxins were found in samples from the market (n=59),
first reliable tool for the assessment of co-exposure from among them the previously unreported sterigmatocystin
multiple contaminants in animal feed. and aflatoxicol.

Although beer has been often targeted for mycotoxin (76-110%), RSD (18%), and LOQs (1-40 ng/l) with the only
analysis, the method proposed by Rausch et al. (2021) exception of alternariol monomethyl ether (AME), which
provided a very comprehensive overview of potentially required a matrix-matched calibration curve. In terms of
occurring parent and modified forms. Beer being a complex sample storage and preparation, the authors reported on
matrix, the extraction and clean up procedures reported the possible microbial degradation of analytes over time,
in the literature are commonly SPE, immunoaffinity, or and therefore the need to perform the analysis no longer
QuEChERS based. On the contrary, the authors proposed than 96 hours after sample collection. The method was
a solvent-precipitation strategy, using acetonitrile as applied for the analysis of freshwaters from ponds across
precipitation agent. Such simple clean up, although time- Flanders, Belgium. Enniatins and BEA, together with AME,
and cost-convenient, led to a strong matrix effect for many ZEN and the more potent oestrogenic compound α-ZEL
analytes (-67±319%), compensated for by using zearalenone- were frequently found in water.
dimethylether-D6 as an isotopic-labelled internal standard.
The protocol returned validation parameters compliant A novel approach in terms of extraction has been
with the EU guidelines (EC, 2006). In particular, sensitivity reported by Gbashi et al. (2020). The authors proposed
[limits of quantitation (LOQs) 0.04-75 µg/l] and recovery the use of pressurised hot water for the extraction of 15
(79-100%) were highly satisfactory for all the analytes. different mycotoxins from cereals. By means of advanced
chemometrics, results obtained by conventional extraction
Rämö et al. (2021) applied LC-MS/MS for the accurate and pressurised hot water extraction were compared,
quantification of Fusarium toxins in onions. It is actually finding a good consistency. The application of green
known that Fusarium species may cause a disease called extraction strategies for multiclass analysis is relevant,
Fusarium basal rot in onion, but the possible occurrence of as it may improve the environmental sustainability of the
mycotoxins in the edible tissues is still understudied. While protocols and may offer an alternative for those areas where
a dilute-and-shoot approach was applied for fumonisins and solvent storage may be an issue.
BEA with a minor matrix effect, a concentration step was
required for the quantification of moniliformin (MON). Taken altogether, these examples reflected the large variety
13C-FB was used as an isotope-labelled internal standard. of food matrices analysed so far by LC-MS/MS for multiple
1
The method was able to detect the main fumonisins (FB1, mycotoxins. It is clear that the scientific community is more
FB2, and FB3) together with MON and BEA with good and more interested in capturing mycotoxins co-occurrence
recoveries (67-122%) and sensitivity (LOQs: 2.5-10 µg/kg). in major crops, as well as in understanding the potential
accumulation of unregulated mycotoxins in unregulated
Fruit juice and purées are a food category of interest for the commodities. A few papers also addressed environmental
development of multi-analyte methods, in consideration matrices, such as soil and water. For accurate quantification,
of the reported increasing fungal infections due to climate targeted LC-MS/MS is still the benchmark thanks to the
change. Guo et al. (2021) presented a LC-MS/MS protocol large affordability and the robustness of the technique.
for the quantification of 15 major mycotoxins produced by However, a thorough validation workflow, and in particular
Aspergillus, Penicillium and Alternaria species in orange, a careful assessment of the matrix effect, is still somehow
grape, and apple juices. The extraction procedure was based missing due to the lack of certified reference materials and
on the use of QuEChERS, and allowed for satisfactory calibrants. In terms of sample preparation, many reports
recoveries (74-110%) and sensitivity (LODs: 0.05-0.1 µg/l). are based on QuEChERS approach, on account of their ease
Matrix-matched calibration was used to compensate for in storage and handling. A growing number of papers are
matrix effects. The study demonstrated frequent occurrence focused on the application of green strategies, with a general
of mycotoxins in juices from the market, thus confirming trend in decreasing the use of solvents. The use of isotope-
the need for further studies and proper monitoring. labelled standards is growing, although still not routine.
A very innovative application of the multi-toxin approach Besides the breakthrough protocols proposed by Steiner
was reported by Goessens et al. (2021), who developed a et al. (2020) and Sulyok et al. (2020), a vast majority of
method for the determination of 20 mycotoxin in freshwater. less sophisticated, but still efficient, methods are still in
The authors started from the observation that mycotoxins use, offering a good compromise between the affordability
can leach out of plant tissues and be transported through of the analytical system and the quality of the analytical
runoff water into ponds and water basins. In view of the performance.
increasing number of extreme climate events, the potential
role of mycotoxins in ecotoxicology has to be considered. 5. Chromatography with targeted HRMS
Due to the very low expected amounts, the authors applied
an SPE clean-up and concentration step, coupled with the High resolution mass spectrometry (HRMS) has found
use of isotope-labelled internal standards. Satisfactory data its place in many different research areas, being a core
were collected for matrix effects, extraction recoveries analytical technique for many omics applications. For

targeted analysis, HRMS is becoming more and more possible. However, even modern MS/MS systems are still
popular, as well. The improved selectivity due to high not sensitive enough to measure the baseline exposure to
resolution in combination with the fast generation of many mycotoxins. Thus, a combination of MS/MS with
product ion spectra can reduce the number of ambiguous online SPE can also be a valuable tool, as described by
results and streamline peak detection. However, regarding Schmidt et al. (2021).
the limits of detection and quantitation (LOD, LOQ), a gap
between HRMS and MS/MS remains for most analytes. To Another HRMS-based biomonitoring method for the
overcome this gap and broaden applicability for mycotoxin detection of citrinin (CIT) and dihydrocitrinone has
trace analysis, different analyte enrichment strategies been published by Narváez et al. (2021). In this paper, an
such as solid phase extraction have been implemented in interesting chromatography is applied which seems to elute
workflows. Unfortunately, these additional steps always analytes in a ‘wave’ of methanol from a reversed phase
cause a strong increase of manual laboratory work and Luna Omega series column (Phenomenex, Torrance, CA,
costs, thus reducing possible benefits of HRMS systems. USA). Water and methanol containing 5 mM ammonium
formate and 0.1% formic acid were applied as solvents
Online SPE as an alternative to classical SPE allows users and the analytes were eluted between 4.75 and 5.00 min.
to omit additional lab work by installing a re-usable SPE The gradient started with 0% methanol and increased
column into a valve unit placed into the ultra-high pressure to 95% methanol within 1 min which was held for 0.5
liquid chromatograph (UHPLC) flow path ahead of the min before the methanol content dropped to 75% within
chromatographic column. When in load position, the SPE 2.5 min and 60% in the subsequent minute. At the end,
is not connected with the chromatographic column, but the column was equilibrated with 0% methanol for 1.5 min
sample extracts are loaded onto the column by means of prior to the next injection. The effect of this fast increase
a standard UHPLC autosampler. Additionally, washing high elution power that afterwards dropped during the
of the SPE column with appropriate solvents or solvent whole chromatographic run is not further discussed by
combinations is possible. Switching the valve unit to elute the authors. Whether some HILIC-type interactions are
mode places the online SPE column into the main flow taking place during the run is not clear, but the gradient
path and the concentrated, purified extract is reversely itself surely requires special emphasis to the performance
eluted from the SPE onto the chromatographic column for of the UHPLC system and transferability to other UHPLC-
separation. Materials of different selectivities are available systems will probably be difficult.
for online SPE, including most standard SPE phases but
specific turbulent flow SPE columns can also be used. In the last decade, ion mobility spectrometry devices based
Turboflow means that columns with narrow diameter on travelling wave (Waters), drift tube ion mobility (Agilent,
and large particle sizes are applied, with the idea, that Santa Clara, CA, USA) and trapped ion mobility (Bruker,
large (macro) molecules, due to their slow diffusion, Billerica, MA, USA) became available as parts of enhanced
are less adsorbed to the particles compared to small ToF-HRMS Systems, providing an extra dimension of
analyte molecules such as mycotoxins. As a consequence, separation and selectivity. The core idea of all of these
additional purification and removal of macromolecules is technologies is that compounds can be separated based
achieved. Ndaw et al. (2021) investigated the applicability on their collisional cross section (CCS). While the idea
of the cyclone P (ThermoFisher Scientific, Waltham, of CCS values is rather old and, for instance, has been
MA, USA) turboflow online SPE cartridge, for the implemented in screening devices for chemical warfare
detection of mycotoxins and mycotoxin biomarkers in agents, its application and meaningfulness for compound
urine. Subsequently, they compared offline SPE with two identification in combination with mass spectrometry
different cartridges as an alternative manual enrichment is currently under discussion. Especially when coupled
and purification technique with the system. The authors with HRMS instruments, CCS data can improve selectivity
observed that the highest sensitivity could generally be and, depending on the system and matrix, lower the
achieved with online SPE, but retention of DON was not LOQ and LOD for mycotoxin detection. Following their
possible with the applied material. Thus, for optimised 2018 publication on the determination of CCS values of
sensitivity, the authors had to split the method into an mycotoxins for travelling wave systems, Righetti et al. (2020)
online-SPE method for AFM1, FB1, ochratoxins, HT-2 now provided data on the reproducibility of CCS values
and ZEN and a second single analyte method using offline among different travelling wave instruments from the same
SPE on an Oasis HLB column for DON detection. Phase vendor. In the study three instruments from two different
2 metabolites of mycotoxins were only indirectly detected platforms were compared, resulting in 96.4% of the CCS
after enzymatic cleavage. Using this sample enrichment, values differing only by up to 2% between laboratories while
LOD and LOQ values comparable to current MS/MS for single labs, an average RSD as low as 0.15% was reached.
dilute and shoot approaches for most analytes could be Thus, CCS seems to be a very reproducibly measurable
obtained, demonstrating that the application of HRMS molecular property. Additionally, characteristic differences
instruments for human biomonitoring of mycotoxins is between CCS of isomers can be used for identification and

probably reduce the requirements for UHPLC separation. The European Commission (EC) recently introduced new
Currently a bottleneck of the applied instruments in some rules limiting 12 ‘major’ ergot alkaloids in certain food
circumstances seems to be the low overall sensitivity – products for infants and young children (EC, 2021). A timely
for a large number of analytes, the lowest concentration study by Uhlig et al. (2021) applied non-targeted HRMS
where a CCS value could be assigned was around 10 ng/ml analysis with post-acquisition diagnostic fragment filtering
in solvent. For most instruments, when the ion mobility (DFF) to unravel the complexity of ergot alkaloids and
separation mechanism is activated, there is a reduction in indole diterpenoids in sclerotia (Uhlig et al., 2021). Analysis
signal intensity that can affect CCS value characterisation. was performed on samples obtained from wild grasses, rye
In the analysis of samples, this reduction in signal intensity and oats from Norway or Canada. MS/MS datasets were
can be more than offset by the alleviation of the negative screened for the ergot alkaloid specific product ions of
impact from interfering compounds. In the future it will m/z 208.0748 and 223.1230 and 130.0651 and 146.0600 for
be interesting to see the sensitivity of other ion mobility- indole diterpenoids. Using this technique they were able to
quadrupole-time of flight (Q-ToF) instruments and to what identify 67 ergot alkaloids and 5 indole diterpenoids from
extend the additional selectivity can help to identify and the grain samples. Interestingly, the 6 major ergot alkaloids
quantify mycotoxins in complex matrices. and their epimers made up more than 50% of the total ergot
alkaloids detected in samples from Canada and Norway.
6. Chromatography with non-targeted HRMS These data support monitoring for the 12 EC ergot alkaloids
in food and feed as being sufficient to manage risk from a
LC-HRMS for targeted analysis has become more food safety perspective. One additional benefit that NTA
mainstream as reviewed above with the increased provided was insight into chemotaxonomic markers that
availability of Q-ToF and Orbitrap instruments and the were specific to certain species of Claviceps.
added benefits that they provide. Complementary non-
target methods such as data-independent acquisition HRMS with DDA was used for the analysis of mycotoxins
(DIA), data-dependent acquisition (DDA) and all ion from Alternaria-inoculated tomatoes to assess potential
fragmentation (AIF) have also greatly increased in many risks associated with these emerging mycotoxins (Zhang et
fields of science including metabolomics and the study of al., 2021a). The authors noted that conjugation with sulphate
environmental contaminants. However, a similar increase is a common mechanism for Alternaria compounds.
in non-targeted analysis (NTA) has not been observed In order to screen for these sulphated metabolites, data
for the investigation of mycotoxins. Steiner et al. (2021) were processed using DFF set for identification of the
provided a useful commentary on this subject in their neutral loss of 79.9568 corresponding to the loss of the
‘trend’ article on the ‘Challenges and future directions in sulphate. Twenty-four sulphated Alternaria metabolites
LC-MS-based multiclass method development for the were identified from fungal cultures, these data were then
quantification of food contaminants’. They identified a used to build a custom database of Alternaria mycotoxins
number of strengths of non-targeted HRMS analysis, and sulphated metabolites. Screening of Alternaria
including the ease of adopting existing targeted HRMS inoculated tomatoes revealed that two of these sulphated
methods for NTA and the enhanced sensitivity for ultra- compounds [alternariol-sulphate (AOH-sulphate) and
trace analysis. Additionally, retrospective analysis of AME-sulphated] were present, indicating the potential
non-targeted data enables the possibility of identifying risk of these Alternaria metabolites in food. The authors
additional metabolites and transformation products of were also able to retrospectively screen for and confirm
important mycotoxins that would otherwise be missed. that there were no glucose conjugated metabolites from
They also identified two important impediments to wider Alternaria in the tomatoes. This work demonstrated the
uptake of this technology by the mycotoxin community. power of combining HRMS for target analysis and NTA
The first is that the linear working range of these but could have been improved by expanding the analysis
instruments needs to be expanded to meet the needs of to more food products including samples that were not
routine analysis. The second is the cost of the instruments inoculated.
makes them prohibitive for many labs. The key take away
message is that in order for multiclass analyte screening A metabolomics based approach was used to analyse
methods to increase beyond 1000 target compounds, for regulated and emerging mycotoxins in bulk milk
HRMS and NTA will be necessary. Since the previous samples from cows fed maize silage (Rocchetti et al.,
review in mid-2020, there have been a few significant 2021). Retrospective screening identified 46 mycotoxins
papers describing advances in NTA with HRMS for within milk samples that were previously placed into one
mycotoxin determination. These papers describe analysis of five groups based on the level of fungal contamination
of ergot alkaloids from grain, Alternaria metabolites in in the silage the cows were eating. Multivariate statistical
tomatoes, and mycotoxins in milk. analysis of the data resulted in clustering of the milk samples
from cows fed the higher mycotoxin contaminated silage.
Additionally, orthogonal projections to latent structures

discriminant analysis and k-means clustering revealed four Kumar et al. (2020) undertook a similar study and developed
discriminant marker compounds that were driving the an FLD method for analysing and validating aflatoxins in
separation of the milk samples. These four compounds could animal feeds using UHPLC without derivatisation. This
potentially be linked back to specific fungal contaminants technique used 80% methanol extraction with a water-based
within the feed. This work demonstrates the value of using immunoaffinity column clean-up. The method was validated
non-targeted HRMS analysis combined with retrospective using a variety of feed matrices. The technique is useful in
screening for assessing mycotoxin contamination. aflatoxin detection as it minimises co-extractives and matrix
meddling. This UHPLC-FLD method proved a simple process
Although NTA with HRMS for mycotoxin determination for aflatoxin analysis in an extensive range of animal feed,
has not experienced the uptake by the community as it has with high sensitivity and cost-effectiveness. For instance, in
in other fields, these papers clearly demonstrate its value. pigeon pea husk feed, an LOQ of 0.5 μg/kg for each aflatoxin,
NTA for screening will likely continue to gain traction as with recoveries of AFB1, AFB2, AFG1, and AFG2 of 71.5,
databases improve and software packages for data process 75.6, 82.4, and 78.2% respectively, were obtained (Kumar et
are better equipped to handle these large datasets. The al., 2020). The development of methods for complex feed
ability to retrospectively screen for >1000 compounds of matrices with FLD continues to be important as MS-based
interest is not possible for most lab doing targeted analysis methods with complex matrices may experience matrix
due to the cost of standards and their limited commercial specific signal suppression or enhancement.
availability. Alternatively, as noted by Steiner et al. (2021)
the cost of these instruments will be a major factor for A separate multi-mode mechanism for detecting ZEN, a
labs, as well as the additional knowledge required by the highly potent oestrogenic mycotoxin in rice, was recently
analyst due to the complexity of the analysis. There is a lot published (Xu et al., 2020). A hydrophilic polyhedral
of discussion in other fields about how NTA will support oligomeric silsesquioxane-containing monolithic column
regulation, it is expected that this will be important for was prepared and coupled to high-performance liquid
mycotoxin determination, as is evident from the papers chromatography (HPLC) in order to detect ZEN using FLD
discussed here. (Xu et al., 2020). The functional and structural evaluation
of the column revealed that this technique provided good
7. Chromatography with non-MS detection selective recognition and detection of ZEN; cross-reactivity
assessment showed no recognition of AFB1 and minimal
Various analytical methods for mycotoxin quantification recognition of α-ZEL (3.3% of ZEN). Recoveries of ZEN
with non-MS detection were developed from the middle from rice fortified at 0.2-10 μg/kg ranged from 99-101%
of 2020 to the middle of 2021. While a vast number of with RSDs of 2-3%. The authors reported an LOD of 0.02
analytical investigations have been undertaken, further ng/ml; unfortunately only on the basis of sample extract and
research on non-MS mycotoxin detection is important to not sample mass basis. Based on the sample preparation
support laboratories that do not have the financial resources information provided, the reported LOD was calculated
or expertise to utilise MS detection methodologies. as 0.15 μg/kg on a sample mass basis.
One of the recent advances in mycotoxin analysis was the Another method of determining mycotoxins T-2 and HT-2
development of an advanced pre-treatment technique for in wheat grain using HPLC-FLD with 1-anthroylnitrile
the analysis of aflatoxins in spices (Saito et al., 2020). This derivatization was recently developed and validated
FLD method uses dispersive solid phase extraction (dSPE) (Laposha et al., 2020). The method met EU method
together with solid-phase fluorescence derivatization and performance criteria (EC, 2006) with regards to accuracy,
immunoaffinity gel to provide a simple, quick, and yet precision, and sensitivity. Mean recoveries for the toxins
sensitive liquid chromatography technique for detecting at 50-150 μg/kg were 91% for T-2 and 87% for HT-2 with
aflatoxins. In developing the method, pepper samples (black an RSD of 0.2-4% for repeatability and intra-laboratory
and white) were extracted using a 4:1 v/v methanol-water precision; LOQs were 2.2 and 1.2 μg/kg for T-2 and HT-2,
mixture and then diluted using a 7% Triton X solution. respectively. The authors also reported the need to store
The resulting solution was then treated with dSPE, solutions of the two mycotoxins away from light in order
immunoaffinity gel, and trifluoroacetic acid to yield a 0.15- to avoid photodegradation.
0.29 μg/kg solid phase fluorescence derivatization detection
limit in white and black pepper with 61.7-87.8% recovery Meanwhile, Wood et al. (2020) developed a sophisticated
and intermediate precision <15% (Saito et al., 2020). The method for determining AFM1 in liquid milk, cheese, and
method demonstrated good performance, however the other selected milk-based proteins. The technique used
trifluoroacetic acid derivatization is not stable and this can automated online immunoaffinity clean-up with HPLC-
cause complications while analysing multiple samples on FLD. In developing this high-throughput approach for
one analytical sequence. the quantitative analysis of AFM1, Wood and colleagues
extracted AFM1 in protein products using 1% ethanoic

acid diluted in acetonitrile with citrate salts. The AFM1 Lastly, Gonçalves et al. (2020) examined the use of
in the resulting extract was subsequently concentrated immunoaffinity chromatography (IAC) clean-up coupled
and purified by utilising RIDA® CREST/IMMUNOPREP® with HPLC-post-column derivatisation-FLD in determining
cartridges (R-Biopharm, Pfungstadt, Germany) (Wood et DON and its key conjugates in cereals. The study used IAC
al., 2020). This method was found accurate and precise, with that interacted with DON-3G, 3-ADON, and 15-ADON for
81-112% recoveries and 4-12% repeatability. The application extraction, and the Hantzch reaction to form fluorescent
of automated online cartridge immunoaffinity provided dihydropyridine derivatives for FLD analysis. Validation
added value of decreased human error during complicated variables such as capacity, cross-reactivity, and analyte
extraction and purification protocols. competition were evaluated. The study then measured
DON and its conjugates in wheat, barley, and maize fortified
Belajová (2020) optimised various aspects of existing at 10-1000, 10-300, and 10-100 µg/kg of DON, DON-3G,
HPLC methods for the detection and quantification of the and the two ADONs, respectively. Fifteen samples of the
polyketide PAT and OTA in fruit purées. The study focused cereals were analysed. From the findings, the recovery rates
on infant fruit-based and fruit-cereal purées and optimised ranged from 87 to 110%. The intermediate precision was
two independent detection methods: (1) diode-array and less than 14% RSD, except for DON-3G in wheat (21%).
(2) fluorescence detection (FLD). The detection limits In addition, the analysis revealed DON conjugate levels
were 5.8 and 0.06 μg/kg, for PAT and OTA, respectively. that represented up to 9-60% of the overall DON content.
The recoveries varied between 79 and 92% for PAT and 78 Since the method did not use any organic solvent when
and 105% for OTA. preparing the sample, it does not require additional safety
precautions during sample preparation.
Luci (2020) developed a method utilising enzymatic
digestion with molecularly imprinted SPE extraction for 8. Multiplex biosensors
quantification of OTA in pig muscle, kidney, and liver with
HPLC-FLD detection. The mean OTA recovery exceeded Research in the area of multiplexed immunosensors
89% with repeatability <5% RSD for all the matrices analysed continues to grow. In recent years the greatest number
under optimised conditions. The authors claimed an LOQ of reports in this area have focused on multiplexed lateral
of 0.003 μg/kg, but the lowest fortification concentration flow immunoassays (LFIA). Increasingly, as is discussed
evaluated with full % recovery and repeatability data elsewhere in this review, aptamers are also used in such
provided was 0.1 μg/kg. The authors reported it was devices. In addition to LFIA, a wide variety of microarrays of
possible to reuse the molecularly imprinted SPE column up various formats have also been reported, with no particular
to seven times without any significant peak area reduction. format establishing dominance.
The study thus concluded that molecularly imprinted SPE
could be used as an alternative sample preparation method Multiplexed LFIA
for detecting OTA in pig products.
Multiplexing based upon immunochromatographic
Pernica et al. (2021) assessed the effectiveness of using devices has reached a level of maturity. Multiplexed LFIA
UHPLC-photodiode array detection (PDA) in determining reported in the past year included assays for detecting
the presence of PAT and potentially hazardous products between two and five analytes. Most such multiplexed
formed during processing (5-hydroxymethylfurfural) in assays were duplex assays. In multiplex LFIA the analytes
beverages. The methodology utilised MIP extraction, are typically distinguished from one another based upon
with separation of PAT from 5-hydroxymethylfurfural spatial location, or with multiple labels having different
within two minutes using a Luna Omega C18 column while colours or fluorescent properties. Improvements to LFIA
the wavelength of the PDA detector was set at 276 nm. generally fall into five categories: improvements to the
Validation parameters, including accuracy, precision, and antibodies, improvements to the competitor molecules (i.e.
quantitation and detection limits, were also evaluated. PAT toxin-protein conjugates, mimotopes, etc.), improvements
produced a linear calibration curve from 50 to 1000 ng/ml, to the labels (brighter detection lines, etc.), improvements
with good linearity. LOD and LOQ for PAT, determined to detection hardware, and improvements to data analysis/
as the concentration of standard solution at which the reporting. With the widespread availability of antibodies
signal to noise ratio equalled 3 and 10 were 4.9 and 6.6 µg/l, for the commercially important mycotoxins, most efforts
respectively. The recovery rate for PAT ranged between 62 have focused on the remaining areas.
and 109%, with a precision of less than 8%. These findings
confirmed the efficacy of UPLC-PDA in analysing PAT Colloidal gold nanoparticles (AuNP) are the most common
in the beverages sold in retail shops. Patulin method label used in LFIA. A 5-plex LFIA using AuNP was reported
development is of continued interest due to stability issues for application to wheat (Xing et al., 2020). An interesting
of the toxin during sample preparation and the lack of feature was the use of QR codes to upload calibration curves
patulin immunoaffinity columns. into a portable reader that was used for quantification.

Antibody-AuNP conjugates were dried onto microwells. luminogens have been used as labels for a duplex LFIA
Adding sample extract to the wells dissolved the conjugates for AFB1 and cyclopiazonic acid (CPA) (Hu et al., 2021).
for the competition reaction. The antibody-AuNP were Polystyrene nanospheres containing an aggregation-
captured by corresponding toxin-protein conjugates on induced emission-active tetraphenylethylene derivative
the membrane. With the reader the limits of detection was synthesised and then conjugated to AFB1 or CPA
for AFB1, FB1, T-2, DON, and ZEN in wheat were 4, 20, monoclonal antibodies (mAbs). The mAb conjugates were
10, 200, and 40 µg/kg, respectively. For visual estimation then incorporated into a traditional duplex competitive
the corresponding cut-off values were 20, 1000, 200, LFIA format. Peanut samples were extracted with 70:30
4000, and 400 µg/kg. Using the reader, the LOD for AFB1 methanol/water v/v with added NaCl, filtered, and diluted
would be acceptable for monitoring in some, but not all, 3-fold for testing by LFIA and LC-MS/MS. For AFB1 the
countries. However, for FB1, T-2, and DON the reader LOD was 0.003 ng/ml, with a linear range between 0.05 and
provided sensitivity below that required by most regulatory 1.2 ng/ml. For CPA the LOD was 0.01 ng/ml, with a linear
agencies. The limitation of the method appears to be in range between 0.8 and 50 ng/ml. Based on the extraction
the sample pre-treatment. Samples were extracted using and dilution ratios these are equivalent to a linear range
acetonitrile/water, but to avoid the effects of solvents between 0.4 to 9.6 µg/kg for AFB1 and 6.4 to 400 µg/kg
upon the antibodies, the extracts were centrifuged and for CPA. Peanut samples were spiked with AFB1 at three
dried under nitrogen. The dried extracts were dissolved in levels (0.05, 5, and 20 µg/kg) and CPA at three levels (20,
aqueous buffer for detection. The need to dry-down the 100, and 700 µg/kg). The recoveries ranged from 93 to
extracts might limit the field-portability of the test, which 110% for AFB1 and 89 to 98% for CPA, with RSDs of 3.2 to
is unfortunate given the portability of the strips and the 5.8%. Strips were stable for 180 days when stored at 4 °C.
reader. Despite this limitation, the incorporation of 5 toxins The assay was applied to 100 samples of peanuts, 12 of
into one LFIA is helping to move true multiplexed LFIA which were contaminated with AFB1, seven of which were
towards field applications. contaminated with CPA. Five samples were contaminated
with both toxins. Good agreement was obtained relative to
Several groups reported the means to improve sensitivity an LC-MS/MS method, suggesting the aggregation-induced
through the use of alternative labels. Quantum dots (QDs) emission luminogens-based LFIA can be used for screening
are often substituted for colloidal gold as labels in LFIA, peanuts for these two toxins.
and recently novel fluorescent QDs were developed and
tested in a duplex LFIA for DON and ZEN (Goryacheva et Multiplexed sensor arrays
al., 2020). The investigators prepared CdSe/CdS and CdSe/
CdS/ZnS core-shell QDs in various sizes, coated them with Recently a ‘hybrid’ format incorporating LFIA and
SiO2 and then functionalised them with either epoxy or microarrays was explored for the assay of five mycotoxins
carboxyl groups. DON antibodies were conjugated with red- (Charlermroj et al., 2021). The format involved spotting a
emitting QD-SiO2 while ZEN antibodies were conjugated 7×4 array, rather than the traditional lines, onto the signal
with orange-emitting QD-SiO2. The labelled antibodies pad of the LFIA device. Four spots were assigned to each
were then incorporated into a LFIA of the antigen- of the five toxin-protein conjugates, four were assigned
immobilised format. Responses were observed visually to a positive control, and four were assigned to a negative
under UV excitation. Cut-off levels for ZEN and DON control (bovine serum albumin, BSA). The presence of
were adjusted to 40 and 400 µg/kg, respectively. Samples of toxin was determined by the resulting pattern of spots
maize (18) and wheat (16) were tested qualitatively by the on the strip, while the concentration was estimated
duplex LFIA and quantitatively by LC-MS/MS, with good from the (inverse of ) the intensity of fluorescence. To
agreement between the two methods at the indicated cut- facilitate detection a large Stokes-shift organic reporter
off values. At the cut-offs, sensitivities, determined by the molecule based on a pentahelicene dye was synthesised
authors as the percentage of samples classified as positive and conjugated to a secondary anti-mouse antibody.
relative to the number of true positive samples, were found Detection was based on the formation of a toxin-BSA/
to be 96 and 96% for DON in maize and wheat samples, primary antibody/secondary antibody-fluorophore complex
respectively. Specificities, determined as the percentage excited at 312 nm. The five mycotoxins that were measured
of samples classified as negative samples relative to the were: AFB1, DON, FB1, T-2, and ZEN. The LODs were: 1.3,
number of true negative samples, were found to be 97% for 0.5, 0.4, 0.4, and 0.9 ‘ppb’, respectively. Unfortunately, it is
DON in maize and 95% for DON in wheat. Sensitivities and unclear what is meant by ‘ppb’, i.e. whether this refers to the
specificities were each found to be 100% for ZEN in both concentration of calibration solutions (µg/l), or whether it
maize and wheat. These satisfy the requirements of the refers to concentration in actual solid samples (µg/kg). The
European Commission (EC, 2014) for a screening method. difference, of course, is that the latter would incorporate
the extraction step and any dilutions of the extract, while
Certain fluorophores will give increased fluorescence the former would not. While recovery data were reported
when they aggregate. Such aggregation-induced emission and were good (ranging from 70.7 to 124.8%), the details

of the spiking were not provided, to the extent that even ranges needed for application to real samples. Application
the matrix used for the spiking was not identified. For this to naturally-contaminated samples or certified reference
reason, it is unclear if the recoveries represent recoveries materials (CRMs) was not reported.
from spiked extracts or recoveries from spiked samples.
Because the article did not report results from specific food Another approach to improving the sensitivity of microarrays
or feed samples, it seems likely that ‘ppb’ referred to µg/l through modification of the sensor surface was reported
of standard solutions. Unfortunately, this situation occurs by Li et al. (2020). To increase the surface area available
frequently in the literature, and authors are reminded to for biomolecular interactions, the substrate chosen was a
clearly specify how recovery studies were conducted. In TiO2-modified porous silicon. The TiO2-modified porous
the current article, the impact was likely minor for DON, silicon wafers were treated with (3-glycidyloxypropyl)
FB1, T-2, and ZEN, as the LODs were well below regulatory trimethoxysilane to provide epoxide groups to react and
levels. However, there may be an impact for AFB1, with a immobilise toxin-protein conjugates. Wafers modified
reported LOD of 1.3 ppb. If this refers to 1.3 µg/l of standard with TiO2 resulted in much higher signals than unmodified
(or spiked extract), the LOD may not be sufficient to meet wafers. Toxins and their associated antibodies were added,
European regulations (EC, 2006) when applied to food washed, and a secondary antibody labelled with the
samples. fluorophore Cy3 was used for visualisation. Visualisation
was performed by scanning the green fluorescence of the
Improving the signal obtained from microarrays can be Cy3. The linear detection ranges were reported to be 0.01
accomplished by a variety of optical and electrochemical to 1 ng/ml for OTA and FB1, and 0.001 to 1 ng/ml for AFB1.
means. This includes techniques to increase the number LODs were calculated as 0.433, 0.093, and 0.243 ng/ml for
of binding events, and techniques to amplify the reporting OTA, FB1, and AFB1, respectively. The latter values are
of binding events. The former includes modifications confounding, as they place the LODs for OTA and AFB1 near
to the surfaces of sensors to increase the number of the upper end of their reported linear ranges. Recoveries
interactions, while the latter includes novel reagents to were obtained from three commodities, maize, wheat,
enhance the detection of such interactions. Recently an and rice each spiked at three levels (0.01, 0.1, and 1 ng/g).
approach was used to perform both such functions. A Recoveries of AFB1 ranged from 64 to 120%. Recoveries
4-plex microarray was developed wherein the microarray of OTA ranged from 70 to 129%, and recoveries of FB1
substrate was modified through the deposition of silver ranged from 65 to 124%. The immunoassay included a
onto poly(methyl methacrylate) coated slides that had 1 h incubation with the antibody/toxin mixture and a 1 h
been treated with oxygen plasma to create 3D micro- incubation with the secondary antibody, which indicates
nanostructures (Koukouvinos et al., 2021). The micro- that the assay times were likely rather long. However, the
nanostructures helped to increase surface area, allowing report demonstrated how increasing the surface area, by
greater antigen binding, while the silver also amplified modifying it with TiO2, was able to improve signal, which
the fluorescence from the label that was used (AlexaFluor is often key for improving sensitivity.
647; Life technologies, Carlsbad, CA, USA). The array
was 4×4 with toxin-protein conjugates immobilised on Recently a new type of biosensor was developed based upon
the surface. Toxin specific antibodies and sample extracts a form of film bulk acoustic resonance using solid mounted
were incubated for 60 min, followed by addition of a resonators. The technology is similar conceptually to a
secondary antibody-biotin conjugate (45 min). Visualisation quartz crystal microbalance, except the resonance occurs
followed incubation with an AlexaFluor-streptavidin in piezoelectric films rather than crystals. Two devices
reagent (15 min). The process reportedly took about based upon the mass sensitive microarray architecture are
2 h, excluding the sample preparation and washing steps, manufactured by a Finnish company (Biomensio, Espoo,
which is long relative many other detection systems. Maize Finland). A 3-plex immunoassay for T-2, FB1, and ZEN
samples were extracted with acetonitrile/water (80:20, v/v). was evaluated using the two devices (Nolan et al., 2021).
The extracts were purified using a QuEChERS mixture, The two devices differed in the number of microfluidic
then diluted 1:5 v/v for analysis. Results indicated that channels, either one (portable), or four (non-portable).
the nano-microstructuring of the surface and deposition The two devices used the same type of microarray chips,
of silver caused a 2.8-fold increase in signal intensity for which consisted of a 4×16 array of sensors. Toxin-protein
AFB1 and OTA, a 5.6-fold increase for DON, and a 16- conjugates were immobilised unto the chip surface, and the
fold increase for FB1, without a corresponding increases binding of toxin-specific antibodies was measured. Simplex
in non-specific signals. Therefore, the goal of increasing and multiplex assays were conducted with toxin standards.
signal intensity through surface modifications was achieved. For simplex assays using the single channel (portable) device
The linear working ranges for FB1, AFB1, OTA, and DON the IC50s were 1.3, 2.0, and 6.8 ng/ml for T-2, FB1, and ZEN,
were reportedly 10-400, 0.5-20, 0.5-400, and 10-400 µg/kg, respectively. For multiplex assays using the single channel
respectively, with intra-assay RSDs below 7.2% and inter- device, the corresponding IC50s were 6.1, 3.6, and 2.4 ng/
assay RSDs below 9.1%. As such the assay would cover the ml, respectively. Sensor chips could be cleaned for re-use

up to three times. The assays with the single channel device New emerging technologies have been explored as well
were relatively rapid, circa 10 min, exclusive of extraction as strategies for improving the traditional approaches.
and pre-incubation steps. The sensitivities were sufficient However, when a lot of effort is put in immunoreagent
to suggest that further testing with foods is warranted, synthesis or adaptation of a given technology to mycotoxin
particularly if the technique can be combined with a rapid detection, less space is left to method testing in real samples.
sample extraction/clean-up procedure. It also seems likely For most of the reviewed cases, it would be interesting to see
that speed could be improved by finding ways to reduce next year, as a follow up, the application of those innovative
the pre-incubation which, at 45 min, is much longer than approaches in routine analysis in a real-world testing
such incubations used in LFIA or in many other biosensors. environment. If compared to the detection approaches,
less innovation can be found in sample preparation. In
In previous years several microbead-based immunoassays spite of the strong trend observed in commercially available
have been reported, wherein the identity of the toxin was products implementing solvent-free extraction procedures,
encoded based upon the colour or fluorescent properties of literature methods still stick to the ‘traditional’ extraction
the beads, with the properties measured by luminometers. protocols based on methanol/water (or acetonitrile)
Recently a duplex assay based upon microbeads was mixtures.
developed wherein the toxins (OTA or ZEN) were encoded
by using different sized beads (Zhang et al., 2021b). The Having available antibodies paves the way for new assay
level of toxin was associated with the brightness of the development. While a wide range of antibodies for
microbeads following the immunoassay. The microbeads regulated mycotoxins is available, research made some
were spotted onto a slide and then a cell phone was used steps forward also this year in the development of antibodies
to capture an image of the beads, process the image, and for new, unregulated mycotoxins. As an example, the
use the particle size and brightness values to estimate the development of mAbs directed against roquefortine C
concentrations of the two toxins. The immune reaction (ROC) and their application in a competitive ELISA has
was competition between toxin-urease conjugates and been reported by Maragos (2020). ROC is commonly found
toxin (sample) for binding to antibody-coated microbeads. in foods infested either intentionally or unintentionally
The competition incubation was at 37 °C for 20 min. After with P. roqueforti, and has been suggested as candidate
washing, urea was added and incubated, then bromocresol biomarker of intoxication by tremorgens, such as penitrem
purple was added, incubated for 1 min, and a picture was A. Evaluation of cross reactivity towards several compounds
taken with a mobile phone. The image was converted to with related substructures, showed a very high selectivity
grey scale, binarized, and used to distinguish the size of the of the developed antibody for ROC. The resulting ELISA-
particles and calculate their brightness, which was inversely based method, requiring no sample pre-treatment other
associated with toxin concentration. The LODs of standard than a 1:5 dilution in buffer, was sensitive for ROC, with a
solutions were 0.771 and 1.039 ng/l for OTA and ZEN, LOD of 0.026 ng/ml, as well as tolerant to significant levels
respectively. This was a very promising result. However, of methanol. Recoveries from 4 types of nut milks spiked
close examination reveals high non-specific binding of over the range of 0.25 to 2 ng/ml were in the range of 84 to
the ZEN-urease to the OTA antibody and of the OTA- 116%. Applicability of the assay was evaluated by conducting
urease to the ZEN antibody. The result is that when the a small survey of commercial nut ‘milks’ and ‘creamers’.
two assays were combined the calibration curves had very Data indicated 4 out of 35 samples contained ROC at levels
low slope and large variations in the brightness values. This so low that they were unlikely to be significant to human
made it much more difficult to clearly differentiate high health (<0.6 ng/ml). The assay was also applied to canine
and low concentrations of the two toxins. It seems likely serum. Recoveries from serum spiked over the range of 0.2
that reducing the non-specific binding of the toxin-urease to 5 ng/ml ranged from 98 to 123%, indicating its potential
conjugates to their opposing antibodies (i.e. ZEN-urease applicability to the screening of serum from dogs suspected
to OTA mAb), or using other mAbs with less non-specific to be suffering from mycotoxin-induced tremors.
binding, would improve this. The total analysis time was
30 min and the portability of the technique suggests that Among unregulated mycotoxins, Alternaria toxins have
it warrants further investigation. been included in European Commission monitoring
programmes since 2012. This official request for generation
9. Single mycotoxin, or single mycotoxin family, of occurrence data is driving the development of fit-for-
biosensors purposes analytical methods. The available data show
tenuazonic acid (TeA) to be the most frequently occurring
About 80 papers focusing on single mycotoxin/family Alternaria toxin and at the highest contamination levels.
immunoassays have been published over the period mid In the work by Liang et al. (2021) a sandwich ELISA was
2020 – mid 2021, if compared to the about 45 published realised using a pair of matched heavy chain variable
last year, it can be indeed stated that this is a quite dynamic region and light chain variable region from hybridoma cells
research area. producing an anti-TeA monoclonal antibody. A sandwich

ELISA detects the antigen (TeA in this case) between two in LFIAs. In complex matrices of a dark colour, such as
layers of antibodies (capture and detection antibody). The grape juice, traditional LFIA using AuNPs as signal probes
established ELISA was intended for TeA determination in usually suffers from poor detection performance due to
juice, wheat flour, and tomato ketchup. The assay exhibited a high background colour. Therefore, immunomagnetic
an LOD of 0.08 ng/ml, and was reported to be 13 times separation integrated into the LFIA analysis has been
more sensitive than a classical indirect competitive ELISA proposed as strategy to minimise the matrix interference
previously developed by the same group. The improved via the magnetic-mediated separation and concentration
sensitivity of the assay allowed extract dilution (5, 20 and of analytes. Specifically, MGNHs, new magneto-optical
50-fold dilutions) to manage matrix effects. Recoveries bifunctional nanomaterials that simultaneously possess
from spiked juice, wheat flour and tomato ketchup magnetic and colorimetric properties, have been used in
samples (spiking range from 8 to 50 µg/kg) ranged from this study. Anti-OTA mAbs were coupled with MGNHs
88 to 111%, with RSDr lower than 15%. A small set of to prepare the MGNH-mAb conjugates to be used as
real samples including beer (n=3), wheat flour (n=4), and competitors in the LFIA. Briefly, grape juice was diluted
tomato ketchup (n=3) was also analysed to compare the 1:1 with a buffer, then combined with MGNH-mAb. The
developed ELISA with a reference UPLC-MS/MS method. formed MGNH-mAb-OTA immunocomplex and free
For beer and wheat flour samples, consistent results were MGNH-mAbs were collected using an external magnetic
obtained, however, for tomato ketchup, inconsistency field and re-suspended in buffer for LFIA. Under optimum
between the two methods was observed. In spite of the conditions, the MGNH-LFIA assay showed a LOD of 0.094
sufficient method performances obtained for most of ng/ml and average recoveries ranging from 92 to 109% with
the targeted commodities, it is worth mentioning that a RSDr of below 12%. A comparison between the MGNH-
TeA extraction from tomato ketchup or wheat flour was LFIA method and a reference LC-MS method measuring
achieved by using chloroform, which would significantly 36 OTA-contaminated grape juice samples was carried out
limit method application on site or in laboratories with obtaining satisfactory correlation (r2=0.9579) and no false
limited analytical equipment or organic solvents waste positive/false negative results.
management procedures.
The use of functionalised magnetic beads in combination
Fluorescence polarisation immunoassay (FPIA) is a with semiconductor QD labelling has been proposed
widely used immunoassay, performed in one phase by Xuan et al. (2020) who developed a quantitative
without requiring the immobilisation of reagents. electrochemical immunoassay for AFB 1 detection,
Moreover, it has been recently proven to be suitable for integrating an electrochemical QD-based immunosensor
multiplexing. Therefore, FPIA applications to mycotoxin with an automated platform for enrichment, washing,
detection continue to deserve attention. Keeping the competing, and releasing steps. Samples (maize, wheat, husk
sample preparation procedure as simple as possible, the rice and peanut oil) were extracted with methanol/water
achievement of improved sensitivity and selectivity is all (70:30, v/v). An automated pre-treatment system was set up
about choosing the best tracer/antibody combination. In adopting the immunoaffinity magnetic beads as the capture
the study by Huang et al. (2020) different combinations of probe of AFB1 and QD-labelled AFB1 as antigen competitor
antibody/tracer have been tested to set up a FPIA for the (AFB1-BSA-QDs) as the signal probe. After a washing step,
detection of OTA in rice. OTA-amino-methyl fluorescein QDs captured on the probe were digested with an acid
resulted to be the best performing tracer. The assay required solution to release corresponding metal ions (Cd2+), whose
OTA extraction with 40% aqueous methanol, followed by signal was detected by a fast electrochemical detector based
dilution 1:4 with water. Analytical performances of the assay on anodic stripping voltammetry. Interesting features of
(recovery, intra-assay, and inter-assay reproducibility) were the developed platform are either the automatization of
evaluated in rice samples spiked with 0.5, 5, and 50 µg/ the sample pre-treatment steps and the detection on an
kg OTA. The average recovery range was 70-110% with a unmodified disposable screen-printed electrode that makes
RSDr equal to or less than 10%, thus fulfilling the mycotoxin the method potentially suitable for replication by other
screening and testing method validation and performance users as well as routine testing. Recovery rates evaluated
criteria set by the EU (EC, 2014). The trueness of the in spiked samples were between 84 and 118%. The intra-
developed FPIA-based method was assessed by comparison day and inter-day RSDr values were in the ranges of 5-18%
with a reference HPLC-FLD method on a set of 10 rice and 3-10%, respectively. The lowest spiking levels were
samples contaminated with OTA at levels ranging from 1 comparable with EU maximum limits.
to 15 µg/kg, obtaining a satisfactory correlation (r2=0.9966).
Semiconductor QDs are also used as luminescent labels
OTA has been targeted in grape juice by Hao et al. (2020), in biosensors based on Förster resonance energy transfer.
aiming at demonstrating the suitability of a novel bifunctional This technique is based on the energy transfer from a
magneto-gold nanohybrid (MGNH) as an alternative to donor in the excited state to a non-excited acceptor due
conventional gold nanoparticles for colorimetric detection to dipole-dipole interactions. The Förster resonance energy

transfer mechanism leads to quenching of donor emission Besides botanical medicines, animal-derived medicines is
and is strongly dependent on the distance between donor another important part of traditional Chinese medicine,
and acceptor. An antigen-antibody interaction enables that was indicated as a target for AFB1 determination by
donor and acceptor to come closer and to realise Förster Wang et al. (2021). The contamination level of AFs in
resonance energy transfer. This approach was explored animal medicines is reported to be significantly higher than
by Goryacheva et al. (2021) in a Förster resonance energy the average level of other traditional Chinese medicines,
transfer-based immunoassay for the detection of DON so the authors proposed the application of an ELISA
in wheat using DON-ovoalbumin labelled with AuNPs method for risk assessment purposes. Animal-derived
as energy acceptor and anti-DON antibody labelled with medicines considered in this study were ground beetle,
green-fluorescent QDs as donor. Synthesis of AuNP and cockroach, silkworm, and earthworm. The sample powder
QD labelled immunoreagents and characterisation of the was extracted with acetonitrile, then dried down and re-
QD-AuNP interaction is reported, as well as the process of suspended in phosphate buffer to make it compatible with
setting up the optimal assay condition in buffer solution. the ELISA test. A traditional indirect competitive ELISA
When applying the assay on wheat samples (extracted with was set up, showing sufficient analytical performances to
methanol/water, 70/30, v/v), extract dilution (1:10) was be applied to the analysis of a total of 138 animal material
needed due either to the strong influence of the matrix samples, collected from medicinal markets and stores,
components on the signal or the impact on antibody across four different provinces. Based on the detected AFB1
binding of the high concentration of organic solvent. contamination, quantitative risk assessment was performed
The calculated LOD for DON in wheat samples was by calculating the population cancer risk caused by animal-
28 μg/kg. No extensive method testing was performed derived medicines intake using the approach established
with wheat samples, therefore no data on method analytical by JECFA. This approach enables the estimation of the
performances are reported to enable its comparison with number of liver cancer cases per 100,000 individuals per
other immunoassays. year. Data showed that the population cancer risk related
to the consumption of ground beetle (risk of liver cancer
Another example of application of QDs for improved of 0.036 cancers per 100,000 individuals) exceeded the
detection in LFIA has been reported by Wu et al. tolerable risk reference values of carcinogens calculated
(2021). The aim was to improve the detection of AFB1 in according to the European Chemicals Agency (ECA, 2012).
botanical materials by colloidal gold based LFIA, which is
hampered by strong matrix effects. For this purpose the A further example of rapid immunoassay-based test
authors synthesised a QD-labelled anti-AFB1 antibody to application for human exposure assessment can be found
be embedded in the sample pad of the strip membrane. in the literature this year. Wang et al. (2020) developed
The antibody was selected in order to manufacture a strip a surface-enhanced Raman spectroscopy (SERS)-based
test capable to detect the sum of AFB1, AFB2, aflatoxin lateral flow immunosensor for rapid monitoring of AFM1,
G1 (AFG1), aflatoxin G2 (AFG2), AFM1, and aflatoxin M2 the AFB1 biomarker of exposure, in urine. To this purpose,
(AFM2). For method development a mix of 10 botanical the Raman molecule 5,5-dithiobis-2-nitrobenzoic acid
medicines was prepared and spiked with the target and an anti-AFM1 monoclonal antibody were conjugated
aflatoxins in the range 2-200 µg/kg. Sample were extracted with AuNPs to serve as the SERS nanoprobe. A micro-
with methanol/water (70:30, v/v) then diluted 1:7 to be well lateral flow format was employed for the competitive
applied to the strip test. Matrix-matched calibration curves assay. The AFM1-BSA-AuNP conjugate was immobilised
were used for quantitative analysis. Estimated LODs for on the test line of nitrocellulose membrane as the capture
individual aflatoxins ranged 1 to 2 μg/kg. To demonstrate reagent, while the anti-AFM1 SERS nanoprobe was dried in
potential applicability and usefulness in the analysis of a micro-plate well. Quantification of AFM1 was performed
real samples, a set of 50 botanical medicine samples were by the readout of Raman signal from the SERS nanoprobes
analysed by the developed QD-LFIA strip and the Chinese captured on the test line. Sample pre-treatment required
national standard HPLC method as a reference. Aflatoxins only urine centrifugation followed by dilution with buffer.
were detected in 22/50 samples at levels up to 16 µg/kg. The assay detection limit was about 2 pg/ml of AFM1
Interestingly, the paper also reports on the development of in urine, whereas spiking experiments yielded 94-111%
a strip test reader, with relevant software. The QD reader recovery with coefficients of variation below 17%. Overall,
mainly consisted of the optical system to generate excitation the method performances indicated its suitability for rapid
light, the photoelectric conversion system to convert the aflatoxin biomonitoring in human urine.
optical signal to a voltage signal, and the high-performance
microprocessor for signal processing and data calculation. Among selected research works it is finally worth to
The software system operating under Windows XP or include an example of smartphone-based biosensor.
higher versions of Windows, provided main functions of Indeed, smartphones represent an ideal device for onsite/
processing QD-LFIA strip images, analysis, data storage, online detection, as they combine pervasive distribution
and results printing. with rapidly developing technologies for connectivity,

customisable applications, image acquisition and processing. inexpensive screening technology for on-site testing
Integration with paper-based assay technology is already applications that require extensive sampling. Several reports
a reality and some platforms are available on the market. over the past year have focused on exploiting some of these
A fluidic integrated system was proposed by Zangheri et advantages, but unfortunately many others, while inventive
al. (2021). It consisted of a smartphone integrated self- and attractive from an academic standpoint, are unlikely
standing device comprising a low-cost and disposable to be practical for field use.
analytical cartridge containing all the reagents required
for the execution of the analysis. OTA detection on the A common strategy in alternative mycotoxin detection
strip test membrane was based on chemiluminescence. methods is to simply replace the antibody in an established
The smartphone CMOS (complementary metal-oxide assay with a MIP or aptamer affinity agent. For example,
semiconductor) camera was used as a light detector. ELISA is a staple detection method for mycotoxin analysis.
Chemiluminescence images were then analysed by the Equivalent enzyme-linked apta-sorbent assays (ELASA)
freely available image analysis software ImageJ. Target have been developed in recent years in the hopes of
matrices were red and white wine, as well as instant coffee lowering costs and improving shelf-life by replacing the
(dissolved in water), that were mixed with aqueous solutions capture and/or detection antibody with an equivalent
of NaHCO3/PEG 10000 to remove substances that could aptamer. Recently, Mukherjee and co-workers (2021)
interfere with the chemiluminescence detection (e.g. developed colorimetric and chemiluminescent ELASAs
polyphenols). Matrix-matched calibration curves were for OTA monitoring in food. In each assay, one of the
used for quantitative analysis. Recoveries of the optimised two antibodies in the sandwich format was replaced
assay ranged from 81 to 123% with RSDr lower than 15%, with an aptamer and compared to a traditional sandwich
whereas LODs ranged from 0.3 and 0.1 μg/l, respectively, ELISA; both ELASAs were comparable in sensitivity with
which are below the European regulations. The shelf life a broader detection range than the ELISA format when
of the analytical cartridges at 4 °C was estimated to be challenged under optimised buffer conditions. Limits of
about one year, however no data regarding the maximum detection for each format in spiked solutions were in the
number of samples that could be analysed with the same 1 pg/ml range; determination of the sensitivity in sample
cartridge were reported. Indeed, this is a critical factor in matrix was not described. Recovery studies from coffee
fluidic systems, that may significantly affect analysis costs beans and groundnuts spiked at the 1 μg/kg level showed
in routine use. acceptable recoveries, with the best performance from the
chemiluminescent ELASA platform (91-108% recovery).
10. Assays with antibody analogues
Electrochemical mycotoxin biosensors have potential for
Although immunoaffinity-based separations and assays real world application thanks to their advantages such as
remain the touchstone for mycotoxin analysis, certain low cost and potential for miniaturisation and portability.
practical limitations with antibody technology, such as MIP technology is especially attractive for use with
batch to batch variation, limited shelf life, and cold chain electrochemical detection because the MIP can be prepared
requirements, have continued to motivate the search directly on the electrode surface by electropolymerization
for alternative approaches. Competing approaches methods. Hatamluyi et al. (2020) prepared a MIP-based
based on synthetic receptors such as aptamers and patulin biosensor by electropolymerization atop a glassy
molecularly imprinted polymers (MIPs) remain of great carbon electrode that had first been decorated with
interest. Aptamers are biopolymers (typically single- graphene QDs and a gold nanoparticle-functionalised
stranded oligonucleotides) that are discovered through metal-organic framework (Au@Cu-MOF). Changes in
a combinatorial selection process for their affinity and the peak current of the Au@Cu-MOF could be observed
specificity for a target of interest. MIPs, also known after only 180 s of PAT incubation with the MIP sensor. The
as ‘plastic antibodies’, are synthetic polymers that are sensor presented a linear range of 0.001-70.0 ng/ml, a low
synthesised in the presence of a molecular template that detection limit of 0.0007 ng/ml, and good selectivity for PAT
leaves the materials with a pre-determined selectivity for over other targets such as 5-hydroxymethyl-furfural, AOH,
a particular analyte. OTA, and AFB1. Recoveries from spiked apple juice samples
(from 0.005 ng/ml to 50 ng/ml) ranged from 98 to 99% with
Given the familiarity and acceptance of conventional RSD values from 1.2 to 4.6%. Finally, sensor performance
antibody-based detection systems, aptamer and MIP-based on apple juice concentrate samples was compared to HPLC
strategies must find their niche in mycotoxin analysis by using AOAC Official Method 995.10 and there was no
profiting from their practical advantages, in particular statistically significant difference in the findings. Overall,
their relatively low cost of production and their extended the approach seems promising but could face barriers to
shelf-lives. In particular, alternative synthetic receptors translation because of the unconventional components in
could have a marked advantage in the development of the electrode material.

In a similar example, Munawar et al. (2020) developed a Akgönüllü et al. (2021) prepared a MIP SPR sensor for
MIP-based sensor for FB1 by first preparing a nanoMIPs OTA detection by photo-polymerising the OTA-imprinted
template with FB1 and then depositing those nanoparticles polymer film on the sensor chip surface. Using aqueous
onto a coated platinum electrode. The LOD from simple OTA solutions, a LOD of 0.028 ng/ml was calculated.
buffer solutions was determined to be 0.03 fM using Recoveries from spiked dried fig samples were tested in
electrochemical impedence spectroscopy and 0.7  fM using the range of 1.6 to 32 μg/kg both by this method and by
differential pulse voltammetry. FB1 recovery from maize HPLC. Recoveries from 98 to 100% with RSD values from
(samples from 8.3 to 62.5 μg/kg) showed recoveries of 96 1.8 to 2.5% were observed with SPR; similar values were
to 101%. This approach is more straightforward than the found when the samples were analysed by HPLC (90 to
one described by Hatamluyi et al. (2020) and the possibility 100% recovery with RSD values of 1.7 to 3.5%). A second
to drop the cost of the assay by using disposable screen study by Akgönüllü et al. (2020) used a more complex MIP
printed platinum electrodes and a handheld potentiostat approach to detect AFB1 by SPR. The MIP coating was once
system would be worth exploring. again prepared by photo-polymerisation on the chip surface,
however, gold nanoparticles were also embedded within
Aptamers are also well-suited for electrochemical the film to increase surface area and improve sensitivity.
biosensing based on the ease with which they can assemble In this case, the calculated LOD from buffer solutions was
into monolayers on the electrode surface by chemisorption 1 pg/ml. Recoveries from spiked maize and peanut sample
or covalent attachment. For example, Rahimi et al. (2021) extract solutions (1-10 ng/ml) ranged between 97 and 106%,
developed an electrochemical aptasensor for detection of which was comparable to what was found when the same
AFB1 by the chemisorption of an amine-modified AFB1 samples were analysed by HPLC (87-106% recovery).
aptamer onto a glassy carbon electrode modified with a
carbon dot/Cu2O nanocomposite. The ‘signal-off’ biosensor Aptamer-based sensing approaches can take advantage of
showed a loss of redox signal as AFB 1 concentration their unique nature as oligonucleotides to conceive of new
increased with a linear range of 3 ag/ml to 1.9 µg/ml and assays designs that would not be possible with antibody
a calculated LOD (3× standard deviation of the blank/ affinity agents. For example, aptamer assays can capitalise
slope of the calibration curve) of 0.9±0.04 ag/ml. Recovery on the ability of DNA to assemble in a predictable and
experiments from wheat samples fortified at the micromolar programmable way based on complementary base-pair
level (0.78 to 2.3 mg/kg) showed recoveries from 97 to 104% hybridisation. With this in mind, Xiong et al. (2021)
with RSD values from 2.4-3%. In this case, while the carbon developed a one-step fluorescence assay for detection of
electrode and nanocomposite are all relatively low-cost AFB1 and OTA called a DNA ‘tweezers’ nanomachine. A
components, the complexity of the design may restrict its DNA assembly comprised of aptamers and fluorophore-
implementation. labelled complementary sequences were assembled into
a nanoscale structure that keeps two fluorophores in
Gökçe et al. (2020) developed an aptamer-based close proximity to their respective quencher molecules
electrochemical sensor with disposable pencil graphite (tweezers closed). In this locked conformation, there is
electrodes by first electrografting 4-carboxyphenyl to the no fluorescence detectable. In the presence of one or
electrode surface and then covalently immobilising amine- both of the toxins, however, the aptamers release their
modified OTA aptamers via amide coupling chemistry. complementary sequences in order to bind their cognate
This simple biosensor assay used impedimetric detection target. This structural change (tweezers open) allows the
for OTA measurements in buffer with a LOD of 0.1 ng/ml. fluorophores to be spatially separated from their quencher,
Spiked beer samples (0.4-1.6 ng/ml) were also tested using leading to an increase in fluorescence signal. The target can
this biosensor with recoveries from 92 to 94% and RSD be identified by what fluorescence signal is increased and
values of 3.5 to 5%. Comparison of these results to those of the intensity is proportional to its concentration. Linear
similar electrochemical assays using the same aptamer but relationships were observed between the fluorescent
on screen printed carbon electrodes suggests comparable intensities and AFB1 and OTA concentrations from 0.08 to
analytical performances at greatly reduced fabrication costs. 10 μg/kg for AFB1 and 0.5 to 50 μg/kg for OTA. The LODs
in simple aqueous solution were found to be 0.035 μg/kg for
Another sensor platform that is very compatible with AFB1 and 0.1 μg/kg for OTA, respectively. This solution-
antibody analogue technology is based on surface plasmon based assay was used to analyse spiked food samples, such
resonance (SPR). SPR sensors have gained popularity in as maize, peanut, coffee, olive oil, and peanut oil. Recoveries
recent years for their ability to perform label-free, real-time between 95 and 106% with RSD values of 4.1-8.1% were
measurements and the recent development of benchtop or reported. Comparable values were obtained when the
portable SPR systems could allow for more widespread use same samples were analysed by HPLC. Assays capable of
in mycotoxin testing. Just as with electrochemical sensing, simultaneous, low-cost screening of multiple mycotoxins
MIP technology is well-suited for SPR assays as the MIP can could also be a niche for synthetic antibody analogues.
be polymerised directly on the SPR chip surface. Recently,

Another unique advantage of aptamer technology in to real world mycotoxin screening is the lack of analysis
particular, is that methods for the direct exponential of reference materials, comparison with reference
amplification of oligonucleotides can be exploited to methods, and/or interlaboratory studies to verify method
dramatically improve the sensitivity of the designed assay reliability. Often, these studies with antibody analogues
in ways that would not be possible with an antibody limit their testing to dilution series of aqueous solutions,
recognition element. For example, Santovito et al. (2020) or occasionally recovery experiments from a spiked
recently designed a magnetic bead-based sandwich-assay sample. Another major impediment to the widespread
for OTA with two phases of signal amplification. Firstly, adoption of aptamer and MIP technology to mycotoxin
binding of OTA to a capture aptamer triggers an isothermal analysis is the trend towards more complex and costly
rolling circle amplification (RCA) reaction. After 4 h of biosensor and assay platforms without consideration
incubation, the DNA amplification is so extensive that of the impracticality of these approaches in real-world
a DNA precipitate is visible in the sample. Secondly, the scenarios. While these studies often boast low detection
RCA product is a DNAzyme (catalytic DNA sequence) with limits, appreciable sensitivity, and innovative designs,
peroxidase activity, capable of producing a blue-colour by the need for exotic components, multi-step preparation,
oxidising ABTS [2,2’-azino-bis(3-ethylbenzothiazoline-6- and complex instrumentation negate the advantages that
sulfonic acid)] substrate that is visible by the naked eye and aptamer and MIP technology can offer. The field would
quantifiable by UV-vis spectroscopy. The authors posited be well-served by a renewed focus on portable, stable,
that this test could be used to detect OTA in biological low-cost approaches that are compatible with point-of-
fluids for biomonitoring and exposure assessments, thus use testing in order to maximise the benefit of antibody
they focused on urine as the matrix of interest. As a result analogues for mycotoxin testing.
of the considerable signal amplification, the authors claim
that the LOD in simple buffer solution was 1×10-12 ng/ml Acknowledgements
(<2 OTA molecules/ml). Recovery experiments from spiked
rat urine using more substantial OTA quantities (0.001, Mention of trade names or commercial products in this
0.01, 0.1, 1, and 10 ng/ml) yielded satisfactory overall publication is solely for the purpose of providing specific
average recoveries (>95%), with RSDs in the range 3.6-15%. information and does not imply recommendation or
Furthermore, naturally contaminated urine from rats fed endorsement by the U.S. Department of Agriculture.
OTA contaminated feed was analysed both by this assay This work was in part supported by the U.S. Department
and by HPLC-FLD with excellent agreement. of Agriculture, Agricultural Research Service. The U.S.
Department of Agriculture is an equal opportunity provider
Niazi et al. (2020) also used RCA to prepare a fluorescence- and employer. Milena Stranska was financially supported
based aptasensor for AFM1 detection in milk. In this by project GACR 20-14649S.
design, the presence of the target prevented the aptamer
from forming a complex with a rolling circle template, Conflict of interest
thus inhibiting the production of the RCA product and
leading to a lower fluorescence signal generated by a The authors declare no conflict of interest.
Eu3+ fluorescent nanoparticle. The multi-step process
required 3 h of incubation time but was able to achieve References
an LOD of 0.0194 pg/ml in simple buffer solution, with
good specificity against AFB1, AFB2, T-2, FB1 and ZEN. Akgönüllü, S., Armutcu, C. and Denizli, A., 2021. Molecularly
The recoveries of AFM1 in milk (0.001, 0.05, 0.1 and 1 ng/ imprinted polymer film based plasmonic sensors for detection
ml) were between 92 and 100% with RSD values less than of ochratoxin A in dried fig. Polymer Bulletin 2021. https://doi.
5%. Comparison of the assay to a standard ELISA showed org/10.1007/s00289-021-03699-6
high correlation. While the extended incubation time and Akgönüllü, S., Yavuz, H. and Denizli, A., 2020. SPR nanosensor based
the multiple steps are limitations of these RCA methods, on molecularly imprinted polymer film with gold nanoparticles for
this approach could be adapted for point-of-use mycotoxin sensitive detection of aflatoxin B1. Talanta 219: 121219. https://doi.
screening in food and feed given that RCA is isothermal org/10.1016/j.talanta.2020.121219
and obviates the need for thermal cycling equipment. Belajová, E., 2020. Optimization and applicability of high performance
Both the colorimetric and fluorometric approach, liquid chromatographic methods for quantification of patulin and
however, would require portable spectrophotometers or ochratoxin A in fruit purées. Journal of Food and Nutrition Research
fluorimeters, although these are becoming more affordable 59: 290-300.
and commonplace. Braun, D., Eiser, M., Puntscher, H., Marko, D., and Warth, B., 2021.
Natural contaminants in infant food: the case of regulated and
Studies highlighted here were chosen for having some emerging mycotoxins. Food Control 123: 107676. https://doi.
degree of potential applicability to in-field testing. A org/10.1016/j.foodcont.2020.107676
major barrier to the translation of these new approaches

Çetinkaya, M.A. and Çetinkaya, G., 2021. The mycotoxin proficiency European Committee for Standardization, 2019. EN17280:2019
testing schemes organized by Turkish National Food Reference Foodstuffs – Determination of zearalenone and trichothecenes
Laboratory between the years 2013 and 2020. Accreditation and including deoxynivalenol and its acetylated derivatives (3-acetyl-
Quality Assurance 26: 61-68. https://doi.org/10.1007/s00769-021- deoxynivalenol and 15-acetyl-deoxynivalenol), nivalenol, T-2 toxin
01463-x and HT-2 toxin in cereals and cereal products by LC-MS/MS.
Charlermroj, R., Phuengwas, S., Makornwattana, M., Sooksimuang, European Committee for Standardization: Brussels, Belgium.
T., Sahasithiwat, S., Panchan, W., Sukbangnop, W., Elliott, C.T. and Gbashi, S., Njobeh, P.B., Madala, N.E., De Boevre, M., Kagot, V. and
Karoonuthaisiri, N., 2021. Development of a microarray lateral flow De Saeger, S., 2020. Parallel validation of a green-solvent extraction
strip test using a luminescent organic compound for multiplex method and quantitative estimation of multi-mycotoxins in staple
detection of five mycotoxins. Talanta 233: 122540. https://doi. cereals using LC-MS/MS. Scientific Reports 10: 10334. https://doi.
org/10.1016/j.talanta.2021.122540 org/10.1038/s41598-020-66787-z
Chavez, R.A., Cheng, X., Herrman, T.J. and Stasiewicz, M.J., 2022. Goessens, T., Baere, S., Troyer, N., Deknock, A., Goethals, P., Lens,
Single kernel aflatoxin and fumonisin contamination distribution L., Pasmans, F. and Croubels, S., 2021. Multi-residue analysis of 20
and spectral classification in commercial corn. Food Control 131: mycotoxins including major metabolites and emerging mycotoxins
108393. https://doi.org/10.1016/j.foodcont.2021.108393 in freshwater using UHPLC-MS/MS and application to freshwater
Cheng, X., Chavez, R.A. and Stasiewicz, M.J., 2020. When to use ponds in flanders, Belgium. Environmental Research 196: 110366.
one-dimensional, two-dimensional, and Shifted Transversal Design https://doi.org/10.1016/j.envres.2020.110366
pooling in mycotoxin screening. PLoS ONE 15: e0236668. https:// Gökçe, G., Aissa, S. B., Nemčeková, K., Catanante, G., Raouafi, N. and
doi.org/10.1371/journal.pone.0236668 Marty, J.-L., 2020. Aptamer-modified pencil graphite electrodes for
De Girolamo, A., Ciasca, B., Pascale, M., and Lattanzio, V. M. T., the impedimetric determination of ochratoxin A. Food Control 115:
2020. Determination of zearalenone and trichothecenes, including 107271. https://doi.org/10.1016/j.foodcont.2020.107271
deoxynivalenol and its acetylated derivatives, nivalenol, T-2 Gonçalves, C., Mischke, C. and Stroka, J., 2020. Determination of
and HT-2 toxins, in wheat and wheat products by LC-MS/MS: deoxynivalenol and its major conjugates in cereals using an organic
a collaborative study. Toxins 12: 786. https://doi.org/10.3390/ solvent-free extraction and IAC cleanup coupled in-line with HPLC-
toxins12120786 PCD-FLD. Food Additives and Contaminants Part A 37: 1765-1776.
European Chemicals Agency (ECA), 2012. Guidance on information https://doi.org/10.1080/19440049.2020.1800829
requirements and chemical safety assessment. Chapter R.8: González-Jartín, J.M., Rodríguez-Cañás, I., Alfonso, A., Sainz, M.J.,
Characterisation of dose [concentration]-response for human Vieytes, M.R., Gomes, A., Ramos, I. and Botana, L.M., 2021. Multi-
health. Version 2.1 November 2012. Available at: https://tinyurl. analyte method for the determination of regulated, emerging and
com/2p9efkvt. modified mycotoxins in milk: QuEChERS extraction followed by
European Commission (EC), 2006. Commission Regulation (EU) No UHPLC-MS/MS analysis. Food Chemistry 356: 129647. https://
4001/2006 of 23 February 2006 laying down the methods of sampling doi.org/10.1016/j.foodchem.2021.129647
and analysis for the official control of the levels of mycotoxins in Goryacheva, O.A., Beloglazova, N.V., Goryacheva, I.Y. and De Saeger,
foodstuffs. Official Journal of the European Union L 70: 12-34. S., 2021. Homogenous FRET-based fluorescent immunoassay for
European Commission (EC), 2013. Mandate for standardisation deoxynivalenol detection by controlling the distance of donor-
addressed to CEN for methods of analysis for mycotoxins in food. acceptor couple. Talanta 225: 121973. https://doi.org/10.1016/j.
Available at: https://law.resource.org/pub/eu/mandates/m520.pdf talanta.2020.121973
European Commission (EC), 2014. Commission Regulation (EU) Goryacheva, O.A., Guhrenz, C., Schneider, K., Beloglazova, N.V.,
No 519/2014, of 16 May 2014, amending Regulation (EC) No Goryacheva, I.Y., De Saeger, S. and Gaponik, N., 2020. Silanized
401/2006 as regards methods of sampling of large lots, spices and luminescent quantum dots for the simultaneous multicolor lateral
food supplements, performance criteria for T-2, HT-2 toxin and flow immunoassay of two mycotoxins. ACS Applied Materials and
citrinin and screening methods of analysis. Official Journal of the Interfaces 12: 24575-24584. https://doi.org/10.1021/acsami.0c05099
European Union L 147: 29-43. Guo, W., Yang, J., Niu, X., Tangni, E.K., Zhao, Z. and Han, Z., 2021.
European Commission (EC), 2020. Chemicals strategy for sustainability A reliable and accurate UHPLC-MS/MS method for screening
towards a toxic-free environment, Progress report on the assessment of: Aspergillus, Penicillium and Alternaria mycotoxins in orange,
and management of combined exposures to multiple chemicals grape and apple juices. Analytical Methods 13: 192-201 https://doi.
(chemical mixtures) and associated risks. Available at: https:// org/10.1039/D0AY01787F
ec.europa.eu/environment/pdf/chemicals/2020/10/SWD_mixtures. Hao, L., Chen, J., Chen, X., Tongtong, M., Cai, X., Duan, H., Leng, Y.,
pdf Huang, X. and Xiong, Y., 2020. A novel magneto-gold nanohybrid-
European Commission (EC), 2021. Commission Regulation (EU) enhanced lateral flow immunoassay for ultrasensitive and rapid
No 2021/1399 of 24 August 2021 amending Regulation (EC) No detection of ochratoxin A in grape juice. Food Chemistry 336:
1881/2006 as regards maximum levels of ergot sclerotia and ergot 127710. https://doi.org/10.1016/j.foodchem.2020.127710
alkaloids in certain foodstuffs. Official Journal of the European
Union L 301: 1-5.

Hatamluyi, B., Rezayi, M., Beheshti, H.R. and Boroushaki, M.T., 2020. Limay-Rios, V. and Schaafsma, A.W., 2021. Relationship between
Ultra-sensitive molecularly imprinted electrochemical sensor for mycotoxin content in winter wheat grain and aspirated dust
patulin detection based on a novel assembling strategy using Au@ collected during harvest and after storage. ACS Omega 6: 1857-
Cu-MOF/N-GQDs. Sensors and Actuators B: Chemical 318: 128219. 1871. https://doi.org/10.1021/acsomega.0c04256
https://doi.org/10.1016/j.snb.2020.128219 Luci, G., 2020. A rapid HPLC-FLD method for ochratoxin detection
Hu, X., Zhang, P., Wang, D., Jiang, J., Chen, X., Liu, Y., Zhang, Z., Tang, in pig muscle, kidney, liver by using enzymatic digestion with
B. and Li, P., 2021. AIEgens enabled ultrasensitive point-of-care test MISPE extraction. MethodsX 7: 100873. https://doi.org/10.1016/j.
for multiple targets of food safety: aflatoxin B1 and cyclopiazonic mex.2020.100873
acid as an example. Biosensors and Bioelectronics 182: 113188. Maragos, C.M., 2020. Development and characterisation of a
https://doi.org/10.1016/j.bios.2021.113188 monoclonal antibody to detect the mycotoxin roquefortine C.
Huang, X., Tang, X., Jallow, A., Qi, X., Zhang, W., Jiang, J., Li, H., Food Additives and Contaminants Part A 37: 1777-1790. https://
Zhang, Q. and Li. P., 2020. Development of an ultrasensitive and doi.org/10.1080/19440049.2020.1781937
rapid fluorescence polarization immunoassay for ochratoxin A in Mukherjee, M., Nandhini, C. and Bhatt, P., 2021. Colorimetric and
rice. Toxins 12: 682. https://doi.org/10.3390/toxins12110682 chemiluminescence based enzyme linked apta-sorbent assay
International Standards Organization (ISO), 2015. ISO 13528 – (ELASA) for ochratoxin A detection. Spectrochimica Acta Part
Statistical methods for use in proficiency testing by interlaboratory A: Molecular and Biomolecular Spectroscopy 244: 118875. https://
comparison, 2nd ed. International Standards Organization, Geneva, doi.org/10.1016/j.saa.2020.118875
Switzerland. Munawar, H., Garcia-Cruz, A., Majewska, M., Karim, K., Kutner,
Kerry, R., Ingram, B.R., Garcia-Cela, E. and Magan, N., 2021. W. and Piletsky, S. A., 2020. Electrochemical determination of
Investigation of the potential to reduce waste through sampling fumonisin B 1 using a chemosensor with a recognition unit
and spatial analysis of grain bulks. Biosystems Engineering 207: comprising molecularly imprinted polymer nanoparticles. Sensors
92-105. https://doi.org/10.1016/j.biosystemseng.2021.05.001 and Actuators B: Chemical 321: 128552. https://doi.org/10.1016/j.
Koukouvinos, G., Karachaliou, C.-E., Kanioura, A., Tsougeni, K., snb.2020.128552
Livaniou, E., Kakabakos, S.E. and Petrou, P.S., 2021. Fluorescence Narváez, A., Izzo, L., Rodríguez-Carrasco, Y. and Ritieni, A., 2021.
enhancement on silver-plated plasma micro-nanostructured 3D Citrinin dietary exposure assessment approach through human
polymeric microarray substrates for multiplex mycotoxin detection. biomonitoring high-resolution mass spectrometry-based data.
Processes 9: 392-392. https://doi.org/10.3390/pr9020392 Journal of Agricultural and Food Chemistry 69: 6330-6338. https://
Kumar, A., Dhanshetty, M. and Banerjee, K., 2020. Development and doi.org/10.1021/acs.jafc.1c01776
validation of a method for direct analysis of aflatoxins in animal feeds Ndaw, S., Jargot, D., Antoine, G., Denis, F., Melin, S. and Robert, A.,
by ultra-high-performance liquid chromatography with fluorescence 2021. Investigating multi-mycotoxin exposure in occupational
detection. Journal of AOAC International 103: 940-945. https://doi. settings: A biomonitoring and airborne measurement approach.
org/10.1093/jaoacint/qsz037 Toxins 13: 54. https://doi.org/10.3390/toxins13010054
Kumphanda, J., Matumba, L., Monjerezi, M., Whitaker, T.B., De Niazi, S., Khan, I. M., Yu, Y., Pasha, I., Lv, Y., Mohsin, A., Mushtaq, B.
Saeger, S. and Makun, H.A., 2021. Are sample size and sample S. and Wang, Z. 2020. A novel fluorescent aptasensor for aflatoxin
preparation for mycotoxin quantitation in grain products getting M1 detection using rolling circle amplification and g-C3N4 as
trivialized? Food Control 130: 108400. https://doi.org/10.1016/j. fluorescence quencher. Sensors and Actuators B: Chemical 315:
foodcont.2021.108400 128049. https://doi.org/10.1016/j.snb.2020.128049
Laposha, O., Senin, S., Midyk, S., Iakubchak, O., Taran, T., Zabarna, I., Nolan, P., Auer, S., Spehar, A., Oplatowska-Stachowiak, M. and
Ishchenko, L., Ishchenko, V., and Ushkalov, V., 2020. Determination Campbell, K., 2021. Evaluation of mass sensitive micro-array
of T-2 and HT-2 toxin in wheat grain by HPLC with fluorescence biosensors for their feasibility in multiplex detection of low
detection. Methods and Objects of Chemical Analysis 15: 137-143. molecular weight toxins using mycotoxins as model compounds.
https://doi.org/10.17721/moca.2020.137-143 Talanta 222: 121521. https://doi.org/10.1016/j.talanta.2020.121521
Li, J., Cai, T., Li, W., Li, W., Song, L., Li, Q., Lv, G., Sun, J., Jiao, S., Pernica, M., Martiník, J., Boško, R., Zušťáková, V., Benešová,
Wang, S., Jin, Y. and Zheng, T., 2020. Highly sensitive simultaneous K. and Běláková, S., 2021. Determination of patulin and
detection of multiple mycotoxins using a protein microarray on hydroxymethylfurfural in beverages by UPLC-PDA. World
a TiO2-modified porous silicon surface. Journal of Agricultural Mycotoxin Journal 14: 41-48. https://doi.org/10.3920/wmj2020.2587
and Food Chemistry 69: 528-536. https://doi.org/10.1021/acs. Peters, J., Ash, E., Gerssen, A., Van Dam, R., Franssen, M.C.R. and
jafc.0c06859 Nielen, M.W.F., 2021. Controlled production of zearalenone-
Liang, Y., Wang, Y., Wang, F., Li. J., Wang, C., Dong, J., Ueda, H., Xiao, glucopyranoside standards with Cunninghamella strains using
Z., Shen, Y., Xu, Z. and Wang, H., 2021. An enhanced open sandwich sulphate-depleted media. Toxins 13: 366. https://doi.org/10.3390/
immunoassay by molecular evolution for noncompetitive detection toxins13060366
of Alternaria mycotoxin tenuazonic acid. Food Chemistry 361: Rahimi, F., Roshanfekr, H. and Peyman, H., 2021. Ultra-sensitive
130103. https://doi.org/10.1016/j.foodchem.2021.130103 electrochemical aptasensor for label-free detection of aflatoxin
B1 in wheat flour sample using factorial design experiments.
Food Chemistry 343: 128436. https://doi.org/10.1016/j.
foodchem.2020.128436

Rämö, S., Haapalainen, M. and Latvala, S., 2021. Development and Sulyok, M., Stadler, D., Steiner, D. and Krska, R., 2020. Validation of an
validation of a UHPLC-MS/MS method for the analysis of Fusarium LC-MS/MS-based dilute-and-shoot approach for the quantification
mycotoxins in onion. Food Analytical Methods 14: 1524-1536. of >500 mycotoxins and other secondary metabolites in food crops:
https://doi.org/10.1007/s12161-021-01992-8 challenges and solutions. Analytical and Bioanalytical Chemistry
Rausch, A., Brockmeyer, R. and Schwerdtle, T., 2021. Development and 412: 2607-2620. https://doi.org/10.1007/s00216-020-02489-9
validation of a liquid chromatography tandem mass spectrometry Tangni, E.K., Van Hove, F., Huybrechts, B., Masquelier, J.,
multi-method for the determination of 41 free and modified Vandermeiren, K. and Van Hoeck, E., 2021. Citrinin determination
mycotoxins in beer. Food Chemistry 338: 127801. https://doi. in food and food supplements by LC-MS/MS: development and use
org/10.1016/j.foodchem.2020.127801 of reference materials in an international collaborative study. Toxins
Righetti, L., Dreolin, N., Celma, A., McCullagh, M., Barknowitz, G., 13: 245. https://doi.org/10.3390/toxins13040245
Sancho, J.V. and Dall’Asta, C., 2020. Travelling wave ion mobility- Thompson, M., 2000. Recent trends in inter-laboratory precision at
derived collision cross section for mycotoxins: Investigating ppb and sub-ppb concentrations in relation to fitness for purpose
interlaboratory and interplatform reproducibility. Journal of criteria in proficiency testing. Analyst 125: 385-386. https://doi.
Agricultural and Food Chemistry 68: 10937-10943. https://doi. org/10.1039/b000282h
org/10.1021/acs.jafc.0c04498 Tittlemier, S. A., Brunkhorst, J., Cramer, B., DeRosa, M. C., Lattanzio,
Rivas Casado, M., Parsons, D.J., Weightman, R.M., Magan, N. V. M. T., Malone, R., Maragos, C., Stranska, M. and Sumarah, M.
and Origgi, S., 2009. Geostatistical analysis of the spatial W., 2021. Developments in mycotoxin analysis: an update for 2019-
distribution of mycotoxin concentration in bulk cereals. Food 2020. World Mycotoxin Journal 14: 3-26 https://doi.org/10.3920/
Additives and Contaminants Part A 26: 867-873. http://dx.doi. WMJ2020.2664
org/10.1080/02652030902788912 Tittlemier, S. A., Cramer, B., Dall’Asta, C., Iha, M.H., Lattanzio, V. M.
Rocchetti, G., Ghilardelli, F., Masoero, F. and Gallo, A., 2021. Screening T., Maragos, C., Solfrizzo, M., Stranska, M., Stroka, J. and Sumarah,
of regulated and emerging mycotoxins in bulk milk samples by M., 2020. Developments in mycotoxin analysis: an update for 2018-
high-resolution mass spectrometry. Foods 10: 2025. https://doi. 2019. World Mycotoxin Journal 13: 3-24 https://doi.org/10.3920/
org/10.3390/foods10092025 WMJ2019.2535
Saito, K., Ishii, J., Naniwa, M., Ishii, R., Kato, M., Kondo, T., Sakurai, H., Uhlig, S., Rangel-Huerta, O.D., Divon, H.H., Rolén, E., Pauchon, K.,
Taniguchi, M., Hashiguchi, S., Hayashi, T. and Ito, R., 2020. Residual Sumarah, M.W., Vrålstad, T. and Renaud, J.B., 2021. Unraveling
analysis of aflatoxins in spice by HPLC coupled with solid-phase the ergot alkaloid and indole diterpenoid metabolome in the
dispersive extraction and solid-phase fluorescence derivatization Claviceps purpurea species complex using LC-HRMS/MS diagnostic
method. Journal of AOAC International 103: 1521-1527. https:// fragmentation filtering. Journal of Agriculture and Food Chemistry
doi.org/10.1093/jaoacint/qsaa077 69: 7137−7148. https://doi.org/10.1021/acs.jafc.1c01973
Santovito, E., Greco, D., D’Ascanio, V., Sanzani, S. M. and Avantaggiato, Vidal, A., Belova, L., Stove, C., De Boevre, M. and De Saeger, S., 2021.
G., 2020. Development of a DNA-based biosensor for the fast and Volumetric absorptive microsampling as an alternative tool for
sensitive detection of ochratoxin A in urine. Analytica Chimica Acta biomonitoring of multi-mycotoxin exposure in resource-limited
1133: 20-29. https://doi.org/10.1016/j.aca.2020.07.078 areas. Toxins 13: 345. https://doi.org/10.3390/toxins13050345
Scheibenzuber, S., Dick, F., Asam, S. and Rychlik, M., 2021. Analysis Wang, C., Zhang, L., Luo, J., Qin, J., Jiang, J., Qin, L., Zhao, Z., Yang, S.
of 13 Alternaria mycotoxins including modified forms in beer. and Yang, M., 2021. Development of a sensitive indirect competitive
Mycotoxin Research 37: 149-159. https://doi.org/10.1007/s12550- enzyme-linked immunosorbent assay for high-throughput detection
021-00424-0 and risk assessment of aflatoxin B1 in animal-derived medicines.
Schmidt, J., Cramer, B., Turner, P., Stoltzfus, R.J., Humphrey, J.H., Toxicon 197: 99-105. https://doi.org/10.1016/j.toxicon.2021.04.009
Smith, L.E. and Humpf, H.-U., 2021. Determination of urinary Wang, J., Chen, Q., Jin, Y., Zhang, X., He, L., Zhang, W. and Chen,
mycotoxin biomarkers using a sensitive online solid phase Y. 2020. Surface enhanced Raman scattering-based lateral flow
extraction-UHPLC-MS/MS method. Toxins 13: 418. https://doi. immunosensor for sensitive detection of aflatoxin M1 in urine.
org/10.3390/toxins13060418 Analytica Chimica Acta 1128: 184-192. https://doi.org/10.1016/j.
Steiner, D., Malachová, A., Sulyok, M. and Krska, R., 2021. Challenges aca.2020.06.076
and future directions in LC-MS-based multiclass method Wood, J.E., Gill, B.D., Indyk, H.E., Rhemrev, R., Pazdanska, M., Mackay,
development for the quantification of food contaminants. Analytical N. and Marley, E., 2020. Determination of aflatoxin M1 in liquid
and Bioanalytical Chemistry 413: 25-34. https://doi.org/10.1007/ milk, cheese, and selected milk proteins by automated online
s00216-020-03015-7 immunoaffinity cleanup with liquid chromatography‒fluorescence
Steiner, D., Sulyok, M., Malachová, A., Mueller, A. and Krska, R., 2020. detection. Journal of AOAC International 104: 719-724. https://doi.
Realizing the simultaneous liquid chromatography-tandem mass org/10.1093/jaoacint/qsaa164
spectrometry based quantification of >1200 biotoxins, pesticides Wu, Y., Yu, B., Cui, P., Yu, T., Shi, G. and Shen, Z. 2021. Development of
and veterinary drugs in complex feed. Journal of Chromatography a quantum dot-based lateral flow immunoassay with high reaction
A 1629: 461502. https://doi.org/10.1016/j.chroma.2020.461502 consistency to total aflatoxins in botanical materials. Analytical
Bioanalytical Chemistry 413: 1629-1637. https://doi.org/10.1007/
s00216-020-03123-4.

Xing, C., Dong, X., Xu, T., Yuan, J., Yan, W., Sui, X. and Zhao, X., 2020. Zachariasova, M., Cuhra, P. and Hajslova, J., 2014. Cross-reactivity of
Analysis of multiple mycotoxins-contaminated wheat by a smart rapid immunochemical methods for mycotoxins detection towards
analysis platform. Analytical Biochemistry 610: 113928. https:// metabolites and masked mycotoxins: the current state of knowledge.
doi.org/10.1016/j.ab.2020.113928 World Mycotoxin Journal 7: 449-464. https://doi.org/10.3920/
Xiong, Z., Wang, Q., Xie, Y., Li, N., Yun, W. and Yang, L., 2021. WMJ2014.1701
Simultaneous detection of aflatoxin B1 and ochratoxin A in food Zangheri, M., Di Nardo, F., Calabria, D., Marchegiani, E., Anfossi,
samples by dual DNA tweezers nanomachine. Food Chemistry 338: L., Guardigli, M., Mirasoli, M., Baggiani, C. and Roda. C., 2021.
128122. https://doi.org/10.1016/j.foodchem.2020.128122 Smartphone biosensor for point-of-need chemiluminescence
Xu, J., Liu, T., Chi, J., Zhang, W., Lin, C., Lin, X. and Xie, Z., 2020. detection of ochratoxin A in wine and coffee. Analytica Chimica
Online high-efficient specific detection of zearalenone in rice by Acta 1163: 338515. https://doi.org/10.1016/j.aca.2021.338515
using high-loading aptamer affinity hydrophilic monolithic column Zhang, L., Zhang, Z., Tian, Y., Cui, M., Huang, B., Luo, T., Zhang, S. and
coupled with HPLC. Talanta 219: 121309. https://doi.org/10.1016/j. Wang, H., 2021b. Rapid, simultaneous detection of mycotoxins with
talanta.2020.121309 smartphone recognition-based immune microspheres. Analytical
Xuan, Z., Liu, H., Ye, J., Li, L., Tian, W. and Wang, S., 2020. Reliable and Bioanalytical Chemistry 413: 3683-3693. https://doi.org/10.1007/
disposable quantum dot-based electrochemical immunosensor for s00216-021-03316-5
aflatoxin B1 simplified analysis with automated magneto-controlled Zhang, X., Qu, J., Dai, Z., Lin, Y., Lu, G., Yang, S., You, Y., Liu, H., Wu,
pretreatment system. Analytical Bioanalytical Chemistry 412: 7615- Y. and Jiang, G., 2021a. Data-dependent acquisition based high-
7625. https://doi.org/10.1007/s00216-020-02897-x resolution mass spectrum for trace Alternaria mycotoxin analysis
Yi, Y., Fan, K., Wang, J., Fu, Q., Zhou, X., Zhang, Y. and Zhang, H., and sulfated metabolites identification. Food Chemistry 364: 130450.
2021. Primary research on sampling scheme for analyzing mycotoxin https://doi.org/10.1016/j.foodchem.2021.130450
distribution in wheat and rice fields. Journal of the Science of Food
and Agriculture 101: 4980-4986. https://doi.org/10.1002/jsfa.11141

Appendix. List of abbreviations
3-ADON 3-acetyl-deoxynivalenol LFIA lateral flow immunoassay

15-ADON 15-acetyl-deoxynivalenol LOD limit of detection
ag attogram LOQ limit of quantitation
AIF all ion fragmentation mAbs monoclonal antibodies
AFB1 aflatoxin B1 MIPs molecularly imprinted polymers
AFB2 aflatoxin B2 MGNH magneto-gold nanohybrid
AFG1 aflatoxin G1 MS mass spectrometry
AFG2 aflatoxin G2 MON moniliformin
AFM1 aflatoxin M1 MS/MS tandem mass spectrometry
AFM2 aflatoxin M2 nanoMIPs molecularly imprinted polymer nanoparticles
AME alternariol monomethyl ether NIR near-infrared
AOH alternariol NIV nivalenol
AuNP colloidal gold nanoparticles nm nanometres
BEA beauvericin NTA non-targeted analysis
BSA bovine serum albumin OTα ochratoxin α
CCS collisional cross section OTA ochratoxin A
CIT citrinin PAT patulin
CTV citreoviridin PDA photodiode array detection
CRM certified reference material PT proficiency test
CPA cyclopiazonic acid Q-ToF quadrupole-time of flight
DDA data-dependent acquisition QD quantum dot
DIA data-independent acquisition QuEChERS quick, easy, cheap, effective, rugged, and safe
DON deoxynivalenol RCA rolling circle amplification
DON-3G deoxynivalenol-3-glucoside ROC roquefortine C
dSPE dispersive solid phase extraction RSD relative standard deviation
ELISA enzyme-linked immunosorbent assay SERS surface-enhanced Raman spectroscopy
ELASA enzyme-linked apta-sorbent assay SPE solid-phase extraction
EU European Union SPR surface plasmon resonance
FB1 fumonisin B1 T-2 T-2 toxin
FB2 fumonisin B2 TeA tenuazonic acid
FB3 fumonisin B3 UHPLC ultra-high pressure liquid chromatography
FLD fluorescence detection UV ultraviolet
FPIA fluorescence polarization immunoassay VAMS volumetric absorptive microsampling
HPLC high-performance liquid chromatography v/v volume/volume
HRMS high resolution mass spectrometry ZAN zearalanone
HT-2 HT-2 toxin ZEN zearalenone
IAC immunoaffinity chromatography α-ZEL α-zearalenol
IC50 half maximal inhibitory concentration β-ZEL β-zearalenol

Development (2020-21)

Uploaded by

Copyright:

Available Formats

Development (2020-21)

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Development (2020-21)

Uploaded by

Copyright:

Available Formats

Wageningen Academic

World Mycotoxin Journal, 2022; 15 (1): 3-25 P u b l i s h e r s

Developments in mycotoxin analysis: an update for 2020-2021

Received: 8 November 2021 / Accepted: 4 February 2022

OPEN ACCESS REVIEW ARTICLE

1. Introduction We have continued with organising the review to cover

ISSN 1875-0710 print, ISSN 1875-0796 online, DOI 10.3920/WMJ2021.27523

4 World Mycotoxin Journal 15 (1)

World Mycotoxin Journal 15 (1) 5

6 World Mycotoxin Journal 15 (1)

World Mycotoxin Journal 15 (1) 7

8 World Mycotoxin Journal 15 (1)

World Mycotoxin Journal 15 (1) 9

10 World Mycotoxin Journal 15 (1)

World Mycotoxin Journal 15 (1) 11

12 World Mycotoxin Journal 15 (1)

World Mycotoxin Journal 15 (1) 13

14 World Mycotoxin Journal 15 (1)

World Mycotoxin Journal 15 (1) 15

16 World Mycotoxin Journal 15 (1)

World Mycotoxin Journal 15 (1) 17

18 World Mycotoxin Journal 15 (1)

World Mycotoxin Journal 15 (1) 19

20 World Mycotoxin Journal 15 (1)

World Mycotoxin Journal 15 (1) 21

22 World Mycotoxin Journal 15 (1)

World Mycotoxin Journal 15 (1) 23

24 World Mycotoxin Journal 15 (1)

Appendix. List of abbreviations

3-ADON 3-acetyl-deoxynivalenol LFIA lateral flow immunoassay

World Mycotoxin Journal 15 (1) 25

You might also like

World Mycotoxin Journal, 2022; 15 (1): 3-25 P u b l i s h e r s

ISSN 1875-0710 print, ISSN 1875-0796 online, DOI 10.3920/WMJ2021.27523

4 World Mycotoxin Journal 15 (1)

World Mycotoxin Journal 15 (1) 5

6 World Mycotoxin Journal 15 (1)

World Mycotoxin Journal 15 (1) 7

8 World Mycotoxin Journal 15 (1)

World Mycotoxin Journal 15 (1) 9

10 World Mycotoxin Journal 15 (1)

World Mycotoxin Journal 15 (1) 11

12 World Mycotoxin Journal 15 (1)

World Mycotoxin Journal 15 (1) 13

14 World Mycotoxin Journal 15 (1)

World Mycotoxin Journal 15 (1) 15

16 World Mycotoxin Journal 15 (1)

World Mycotoxin Journal 15 (1) 17

18 World Mycotoxin Journal 15 (1)

World Mycotoxin Journal 15 (1) 19

20 World Mycotoxin Journal 15 (1)

World Mycotoxin Journal 15 (1) 21

22 World Mycotoxin Journal 15 (1)

World Mycotoxin Journal 15 (1) 23

24 World Mycotoxin Journal 15 (1)

World Mycotoxin Journal 15 (1) 25