Distribution of moniliformin in industrial maize milling and �aking process

Moniliformin (MON) is a widespread emerging mycotoxin often occurring in maize at not negligible levels. Few published studies investigated MON redistribution in maize derived products for human consumption; to better understand this issue, 5 maize lots with different level of MON contamination were processed following an industrial milling process to evaluate the redistribution of the mycotoxin in �nal products (grits), by-products destined to feed (bran and �our) and cleaning waste. A relevant MON reduction was obtained after sieve cleaning, scourer process and optical sorting, achieving a decrement of the concentration level close to 70%. The following other milling procedures showed a limited reduction from cleaned maize to small and large grits; considering the entire industrial process, the reduction percentage of MON contamination in the �nal products was 80.9 ± 9.3% and 81.0 ± 6.7% for small and large grits, respectively. The �aking process showed a very limited reduction of MON, close to 10%. Considering the widespread of MON occurrence in maize, the study highlights the importance of cleaning steps to achieve a low risk of exposure for the consumer.


Introduction
Moniliformin (MON) is an emerging Fusarium mycotoxin occurring in cereals with high levels found in maize.MON is mainly produced by F. subglutinans, F. temperatum, F. verticilloides, and F. proliferatum (Logrieco et al, 2002;Scarpino et al., 2015; Scau aire et al, 2012); it is a highly polar and acidic molecule (Fig. 1) and occurs as a water-soluble sodium or potassium salt (Springer et al, 1974).The Panel on Contaminants in the Food Chain of the European Food Safety Authority (EFSA-CONTAM, 2018) indicated that cardiotoxicity and hepatotoxicity are its major adverse health effects, resulting in a shortage of energy, respiratory stress, and myocardial loss of functionality.MON can also increase oxidative stress, reducing the activity of glutathione peroxidase and glutathione reductase (Rossi, F.; et al, 2020); nally, its toxicity in combination with fumonisins was studied (Fremy et al, 2019).No regulatory limits have been xed by EU Commission for MON, however EFSA recommended more data on its occurrence in cereals and derived products.Widespread contaminations were found in cereals (mainly maize) cultivated in both Northern and Southern Europe, showing that MON can be produced in different climate conditions (Scarpino et

Maize milling process and sampling
Five different commercial lots of maize cultivated in 2023 in Northern Italy were processed in an industrial mill.Initially, maize was subjected to different cleaning steps; at rst, a sieve cleaner was used to discard broken grains having a particle size inferior to 500 µm.Then, maize was passed to densimetric tables to separate defective grains and successively to a destoner to remove impurities.After a new passage to densimetric tables, maize was subjected to an intensive scourer with aspirator.A nal cleaning step involved an optical sorting to remove grains undetectable by visual control.Cleaned maize was processed using a tempering degermination system consisting in adding warm water to achieve a moisture level of about 20%; in these conditions, germ becomes elastic and maize can be degerminated.After that, maize was sifted to obtain bran, our, small and large size grits.Flour, bran, and a fraction of cleaning waste were processed for the production of animal feed our.The sampled products represented a maize lot of about 40 t and were collected during the milling process according to European Commission Regulation EC 401/2006 (European Commission 2006).From the 5 maize lots considered, unprocessed maize, two cleaning waste (after sieve cleaner and after densimetric tables), maize before and after optical sorting, germ, bran, our, small and large size grits were collected.For each product, the nal sample (about 5 kg) was obtained by blending 30 incremental samples (150-200 g each) collected at regular intervals for 1 hour by means dynamic sampling procedure.The samples were kept at -20°C until the analysis.Moreover, two lots of cleaned maize were collected before and after an industrial aking process (95°C for 35 min).

LC-MS/MS analysis for MON determination
After milling of samples using a cyclone hammer mill (sieve 1 mm) and their homogenisation, MON was extracted from 25 g using 100 ml mixture CH 3 CN:H 2 O 1 + 1 (v/v) for 60 min.After ltration on folded lter paper and centrifugation (3000 g for 10 min), the extract was puri ed on a MycoSep® 240 Mon cleanup columns (Romer Labs®, Tulln, Austria); then, after ltration through 0.45 µm lter, the extract was injected (10 µl) into the LC-MS/MS system.The HPLC-MS/MS system consisted of a LC 1.4 Surveyor pump (Thermo Fisher Scienti c, San Jose, CA, USA), a PAL 1.3.1 sampling system (CTC Analytics AG, Zwingen, Switzerland) and a Quantum Discovery Max triple quadrupole mass spectrometer; the system was controlled by an Excalibur 1.4 software (Thermo Fisher Scienti c, San Jose, CA, USA).MON was separated on a HILIC zwitterionic column (BEH-Z-HILIC, 2.5 µm, 2.1 x 100 mm, Waters) using as mobile phase a mixture 25 mM ammonium formate (pH 3.2):CH 3 CN 15 + 85 (v/v).The ionization was carried out with an ESI interface (Thermo-Fisher) in negative mode as reported by Bertuzzi et al (2019): spray capillary voltage was 3.5 kV, sheath and auxiliary gas 40 and 15 psi, respectively; skimmer 9 V, temperature of the heated capillary 350°C.
The mass spectrometric analysis was performed in selected reaction monitoring (SRM).For fragmentation of the [M − H] − ion (97 m/z), the argon collision pressure was set to 1.2 mTorr and the collision energy to 21 V.
The detected and quanti ed fragment ion was 41 m/z.Five MON standard solutions among 0.005 and 0.4 mg l − 1 were injected.The limit of detection (LOD) and of quanti cation (LOQ) were 7 and 20 µg kg − 1 , respectively.Three replicates were carried out for each sample.
For extraction and puri cation steps, quality control data are reported in our previous work (Bertuzzi et al, 2020); matrix effect was limited, below 6%.

Con rmatory LC-MS/MS method for MON determination
Considering the low molecular mass of MON and the production of one fragment ion, a derivatization with 1,2-diamino-4,5-dichlorobenzene (DDB) was carried out to increase the sensitivity and the accuracy of the analysis, in according to the study of Zollner et al (2003).A total of 20 samples were analysed using this method.Then, standard MON solutions and sample puri ed extracts (2 for each matrix) were derivatized as follows: 2 ml extracts were derivatized with 0.5 ml DDB solution (1mg/ml in HCl 1 M) at 60°C for 120 min.
After evaporation under gentle ow of nitrogen, the extract was redissolved with 0.5 ml CH 3 CN: and injected into the LC-MS/MS system.The HPLC-MS/MS system consisted of a Vanquish pump and autosampler, and a Fortis triple quadrupole mass spectrometer (Thermo Fisher Scienti c, San Jose, CA, USA); the system was controlled by an Excalibur 1.4 software (Thermo Fisher Scienti c, San Jose, CA, USA).MON was separated on a Betasil RP-18 column (2.5 µm, 2.1 x 100 mm, Thermo Fisher) using a gradient elution H 2 O (A):CH 3 CN (B), both acidi ed with 0.2% formic acid, as follows: A from 70-15% in 5 min, isocratic for 4 min.The ionization was carried out in positive mode with an H-ESI interface: sheath and auxiliary gas 35 and 12 psi, respectively; sweep gas 2 psi, temperature of heated capillary and vaporization 270 and 200°C, respectively.The mass spectrometric analysis was performed in selected reaction monitoring (SRM).For fragmentation of the [M + H] + ion (229 m/z), the dwell rime was 500 ms, the argon collision pressure was set to 1.5 mTorr and the collision energy to 30 V. The fragment ions were 124, 152 and 187 m/z.The retention time of MON was 8.4 min (Fig. 1).The limit of detection (LOD) and of quanti cation (LOQ) were 0.5 and 2 µg kg − 1 , respectively.

Analysis of data
The quanti cation of MON distribution in the different sub-products of maize was calculated considering the following ratio: MON ratio = MON concentration in the product/MON concentration in unprocessed maize

Results and discussion
The maize lots showed different levels of MON contamination, from about 150 to 1100 µg kg − 1 , close to ndings of surveys on maize cultivated in Northern Italy (Bertuzzi et al, 2020;Scarpino et al, 2015).The Table 1 shows contamination levels of MON (µg kg − 1 ) found in the collected samples and the ratio of distribution respect the unprocessed maize calculated for each lot.The samples (n = 20) analysed using the con rmatory LC-MS/MS method showed levels of contamination similar (+/-7%) to those obtained with the conventional method; using this method, lower LOD and LOQ were obtained (Fig. 2).
The results con rmed the e cacy of cleaning steps, mainly the use of sieve cleaner that removed high levels of MON with an average concentration in the waste about 6 times higher than in unprocessed maize (6.21 ± 2.40); the concentration in maize after this cleaning step was one third lower than in raw maize (mean: 0.67 ± 0.18).The waste of densimetric tables resulted less contaminated, at a concentration close to that found in the initial maize (mean: 1.01 ± 1.34).After the scourer process and the optical sorting, MON concentration in cleaned maize was about one third of that in the unprocessed maize (average ratio: 0.29 ± 0.07), achieving a global reduction of 71.3 ± 7.1%, remarkably higher that the values between 36-59% reported by Scarpino et al (2020).The use of the optical sorter con rmed to achieve a high reduction of MON contamination.
The following steps of the milling process showed high levels of contamination in bran fraction, reporting concentrations always higher in comparison with cleaned maize (average ratio: 3.93 ± 1.97).Germ was the fraction less contaminated, the average concentration ratio between germ and cleaned maize was 0.36 ± 0.19; higher ratios, but always inferior to 1, were calculated for both small and large size grits, showing very similar values, 0.65 ± 0.27 and 0.65 ± 0.15, respectively.Considering the entire process, MON concentration in both grits was one fth than that found in unprocessed maize (0.19 ± 0.06 and 0.19 ± 0.09); then, the reduction percentage in the nal product respect to the unprocessed maize was 80.9 ± 9.3% and 81.±6.7% for small and large grits, respectively, resulting very similar to 80% and 64% reported by Scarpino et al (2020) for aking and medium hominy grits.However, in our study no signi cant difference in concentration of MON was found between large and small grits.
Considering all the ve lots of maize, the average yield of the cleaning fractions was 3% for both sieve cleaner and densimetric table waste; generally, cleaned maize after optical sorting was 92% of the unprocessed maize.For the degermination and milling process, the average yields respect to the cleaned maize were: 2% for bran, 5% for germ, 20% for our, 55 and 10% for large and small size grits, respectively.
Table 2 shows the amount of MON found in unprocessed maize and in the collected fractions.The percentage of MON content found in siever cleaner and densimetric tables waste respect to unprocessed maize were different among the lots; the average values, with high standard deviations, were 23.9 ± 13.5% and 3.0 ± 4.1%, respectively.After the entire cleaning process, the average percentage of MON remained in the cleaned maize was 26.4 ± 6.5%; these data con rmed that intensive scourer and optical sorting were very effective steps to remove the mycotoxin.Regarding degerminating and milling process, the sum of MON in bran, germ, our and grits achieved 86.9 ± 6.5% respect to the MON in cleaned maize; this minor content found in the fractions can be due to the complexity of the sampling of several matrices.If compared to MON content in cleaned maize, MON distribution was: 39.1 ± 8.9% in large size grits, 7.1 ± 2.9% in small size grits, 30.1 ± 9.1% in our, 8.5 ± 4.3% in bran, 2.0 ± 1.0% in germ.These results showed that, after the milling process, a relevant amount, about 50%, of MON remained in nal products (grits), highlighting that the cleaning steps were more e cacious to reduce the MON presence.
The contamination levels of MON before and after the aking process are reported in Table 3.It is possible to observe that MON reduction was very low, (mean of 2 process was 10.1%), showing that this mycotoxin is stable at temperature close to 100°C.It is well known that fumonisins (FBs) and MON often co-occur in maize; FBs redistribution was largely investigated in the milling fractions showing high reduction both in cleaning and in milling process; unlike MON was rarely studied.This study underlines that the cleaning steps are very e cient in reducing the content of MON in maize, as for FBs, and these operations are essential to reduce the risk of contamination of these Fusarium toxins.Moreover, the moderate reduction of MON during the following steps of milling might lead to a not negligible risk of exposure for the consumers, if a not accurate cleaning was carried out.
The co-occurrence of MON and FBs was widely studied in maize samples, however there are few data regarding maize derived products intended for human consumption.The exposure to these mycotoxins could be relevant for consumers in the countries where maize is a staple food, for the baby food supply chain and for the celiac population.The con rmatory method for MON determination developed in this study could con rm its occurrence at trace levels obtaining a more accurate evaluation of risk.

Declarations Figures
Structure of moniliformin.
al, 2013; Bertuzzi et al 2019; Van Asselt et al, 2012; Herrera et al, 2017; Jajic et al, 2019; Janic et al, 2020).Although several studies on MON occurrence in maize have recently been published, few works investigated the distribution of this mycotoxin during the milling process.Pineda-Valdes reported a possible reduction of MON by heat processing and during alkaline cooking of maize (Pineda-Valdes 2002 and 2003); Tittlemier et al (2014) investigated the fate of MON during durum wheat milling, processing and cooking of spaghetti.Lately, Scarpino et al (2020) determined a relevant reduction of MON during maize large-scale dry milling process.In this work, distribution of MON during industrial dry milling process of different lots of maize intended for human consumption was evaluated.Five maize lots having different MON contaminations were processedand cleaning waste, by-products and nal products were collected and analysed; nally, the in uence of industrial aking process for corn akes production on MON contamination was also determined.

Table 1
MON contamination (µg/kg) in maize fractions collected during milling process.

Table 2
MON presence (µg) in maize fractions collected during milling process.