Mycotoxin risk management in maize gluten meal

Abstract

Maize gluten meal (MGM) is a by-product of maize starch and ethanol, produced by the wet milling process. Its high protein content makes it a preferred ingredient in feed. Given the high prevalence of mycotoxins in maize globally, they pose a significant challenge to use of MGM for feed: wet milling could concentrate certain mycotoxins in gluten components, and mycotoxin consumption affects animal health and can contaminate animal-source foods. To help confront this issue, this paper summarizes mycotoxin occurrence in maize, distribution during MGM production and mycotoxin risk management strategies for MGM through a comprehensive literature review. Available data emphasize the importance of mycotoxin control in MGM and the necessity of a systematic control approach, which includes: good agriculture practices (GAP) in the context of climate change, degradation of mycotoxin during MGM processing with SO₂ and lactic acid bacteria (LAB) and the prospect of removing or detoxifying mycotoxins using emerging technologies. In the absence of mycotoxin contamination, MGM represents a safe and economically critical component of global animal feed. With a holistic risk assessment-based, seed-to-MGM-feed systematic approach to reducing and decontaminating mycotoxins in maize, costs and negative health impacts associated with MGM use in feed can be effectively reduced.

Keywords:

• Maize
• maize gluten meal
• mycotoxin
• wet milling
• mycotoxin management

Introduction

Worldwide maize production and use

Maize, Zea mays L. (also called corn), is considered the third most important global crop after wheat and rice, based on the area harvested (Ramirez-Cabral, Kumar, and Shabani 2017); more recent publications rank it second, after wheat (Erenstein et al. 2022). It can be readily produced areas across a wide range of agroecologies, with a relatively limited number of rainfall events during the production season, and is grown in every continent except Antarctica. From 1961 to 2020, the global area under maize production has increased from 106 million to 201 million hectares, while production increased from 205 million to 1.16 billion tons (Food and Agriculture Organization). Average maize yield has increased from 1.94 ton/ha to 5.75 ton/ha. As of 2020, the U.S.A, China and Brazil are the top three maize-producing countries in the world, contributing 31.0%, 22.4%, and 8.9% respectively of total production (FAOSTAT 2021) (Figure 1).

Figure 1. Maize production a) and yields b) of the world, U.S., China and Brazil from 1961 to 2020 (FAOSTAT 2021).

Global use of maize as both food and feed has increased accordingly, with maize having high worldwide value, nutritionally and commercially. With an energy density of 365 kcal/100 g, maize contains approximately 72% starch, 10% protein, and 4% fat, (Ranum, Peña‐Rosas, and Garcia‐Casal 2014; Dula 2019). Commonly used as a staple food, maize accounts for about 15-56% of daily caloric intake in 25 developing countries (Dubey et al. 2020). In addition to direct human consumption, maize is extensively processed in wet milling, dry milling, distilling and feed compounding for a broad range of industries. Food and industrial products produced from maize include staple foods, snacks, starch, MGM, sweeteners, cooking oils, alcoholic beverages, glue, and industrial alcohol (Ranum, Peña‐Rosas, and Garcia‐Casal 2014). Therefore, any food safety issue that threatens maize production poses a significant global challenge across a broad range of food and feed applications.

MGM production and use

Given that maize is such an important global crop, innovations for use in food, feed and industrial applications represent a major focus for research, industry and global food security. Maize gluten meal (MGM) is a major by-product of wet milling for starch and sometimes ethanol production, illustrated in Figure 2. MGM is prepared by re-centrifugation, filtering and drying of the gluten slurry obtained from the maize starch gluten separation process. Starch and oil-rich germ are the main products from wet milling, comprising 75% on a dry weight basis. Following removal of these main products, fiber, and soluble components, MGM accounts for about 5% of the dry weight.

Figure 2. Maize wet milling process and MGM.

a) wet milling process, b) dry MGM, c) wet MGM. a) adapted from Blasi et al. (2001); b) and c) from Ribeiro et al. (2019) under CC BY 4.0.

MGM is a protein-rich feed containing over 60% crude protein (dry basis) and low crude fat. It is used extensively as a source of protein, xanthophyll, energy and pigments for livestock, poultry, pets and fish feed, due to its high protein digestibility and gross energy content (RFA. 2008). Indeed, due to its commercial value and benefit to livestock, MGM has wide acceptance in the feed and pet industry, with an annual usage of about a million tons (Erickson, Klopfenstein, and Watson 2012).

MGM has a wide range of important applications beyond being used as a feed additive. Historically, MGM has been used as both fertilizer and a pre-emergent weed killer, making it useful in organic agriculture (Pasupuleti and Demain 2010). MGM is the major source for extracting zein, which has diverse uses, from edible food coating to industrial applications (Zheng et al. 2006; Anderson and Lamsal 2011). In the food industry, enzymatic hydrolysis or microbial fermentation of MGM to obtain bioactive peptides (Zhou et al. 2015) is increasing the interest in MGM for new functional dietary products (Li et al. 2019).

Demand and high-value applications for MGM are increasing. The global volume of MGM was estimated to be 10.22 million tons in 2019, based on the amount of wet milling carried out (Rausch et al. 2019). The U.S.A. and China account for more than 50% of global maize production and lead in the production of MGM. Enabling full utilization of MGM, through reducing mycotoxin contamination, will help to meet the global demand for these protein resources and minimize health risks in animals, and to humans who may be exposed through contaminated animal-source food.

Major mycotoxins and worldwide occurrence in maize

Major mycotoxins in maize

Given the susceptibility of maize and its kernels to fungal colonization, maize is vulnerable to fungal toxin (mycotoxin) contamination during both pre- and post-harvest phases. Mycotoxins pose major health risks for both humans and animals. Exposure can result in acute or chronic consequences including carcinogenic, teratogenic, immunosuppressive, or estrogenic effects (Carbas et al. 2021). Mycotoxin contamination can greatly increase losses and management costs, comprising both direct economic loss from reduced yield and crop value, and indirect loss due to reduced animal productivity, human health costs, and the costs of risk management and research (Munkvold et al. 2019). As an illustration of the economic importance of this issue, it has been estimated that in a year of conducive climatic conditions, aflatoxin – only one category of mycotoxins – could cause up to $1.68 billion of losses to US maize production alone (Mitchell et al. 2016).

Maize is susceptible to different fungal pathogens from growing in the field to product storage (Giorni, Bertuzzi, and Battilani 2019). Environmental conditions have a strong influence on mycotoxin development during each phase. The diversity of fungi that colonize maize and its kernels, the toxins they produce, and the stages of production and storage affected are summarized below.

Fusarium mycotoxins

Maize is susceptible to attack by Fusarium spp. in the field, with resultant production of Fusarium mycotoxins. Fumonisins (FBs), Trichothecenes (type A, T-2/HT-2 toxins and type B, deoxynivalenol (DON)) and zearalenone (ZEN) are common mycotoxins in maize and are mainly produced by several Fusarium species.

FBs, mainly produced by F. verticillioides, are the most common mycotoxins with the highest risk, because of the worldwide distribution of F. verticillioides (Munkvold et al. 2019), with DON being one of the most frequently detected. In regions with cooler climates, the occurrence of DON in maize is considered as a typical agricultural issue, as weather conditions are favorable for F. graminearum and F. culmorum growth and DON production (Kos et al. 2017).

T-2/HT-2 toxins are mainly produced by F. sporotrichioides, F. langsethiae and F. poae, and belong to type A trichothecene mycotoxins. These Fusarium species are widespread in both South and North America, Asia, Africa, and Europe (Richard et al. 2007) and produce T-2/HT-2 under moist cool conditions prior to cereals harvest (Porricelli et al. 2016).

ZEN is a nonsteroidal estrogenic mycotoxin primarily produced by is F. graminearum. Contamination of cereal grains by ZEN has been reported worldwide, primarily in temperate climates.

Some Fusarium species and their mycotoxins can also develop in stored maize when storage conditions are less than ideal. For example, FBs increased from 20 to 40% during a six-month storage due to growth of F. verticillioides (Carbas et al. 2021). Similarly, ZEN was also increased when the moisture content of the storage environment was greater than 30-40% (Gupta 2007).

Aflatoxin

The most recognized aflatoxigenic fungi are A. flavus and A. parasiticus. Maize is one of the major sources of human exposure to dietary aflatoxin (AF) globally, with an approximate contribution of more than 10% (JECFA. 2017; Schaarschmidt and Fauhl-Hassek 2021). Pre- and post- harvest AF contamination of maize production is highly dependent on biological (biotic) and environmental (abiotic) factors. Drought stress, excessive rainfall during planting and excessive water and temperature during storage will promote the growth of mold and the production of aflatoxin (Bhatnagar-Mathur et al. 2015).

Ochratoxins

Ochratoxins (OTA) may be significant mycotoxins in maize in some parts of the world. OTA is produced mainly by Penicillium. verrucosum in temperate climates and Aspergillus ochraceus and the rare Aspergillus carbonarius in warm and tropical climates. These fungi can contaminate agricultural products before harvest, but more commonly during storage (Ozbey and Kabak 2012).

Mycotoxin prevalence of maize

A global dataset of mycotoxin prevalence in maize sampled in different continents (including America, Europe, Asia, Middle East and Africa), conducted by the company BIOMIN, is summarized below (Table 1). While samples sizes for some regions (e.g., Middle East and Northern Africa) are too small to allow conclusions, this survey does provide a sense of mycotoxin contamination levels for many regions. Munkvold et al. (2019) introduced the global compilation of several mycotoxins in maize, surveyed from 2014 to 2017 (including 11,237 maize samples collected in 75 countries) and Gruber-Dorninger, Jenkins, and Schatzmayr (2019) performed a large scale survey of mycotoxin contamination in maize samples collected from 100 countries between 2008 and 2017. Both authors reached the similar conclusion that FBs-contaminated samples were the most abundant positive samples (>80%), followed by DON > ZEN > AFs > T-2 > OTA. BIOMIN’s survey also found that Fusarium mycotoxins, FBs, DON and ZEN, were much more prevalent compared with AFs, OTA and T-2 toxin between 2019 and 2020. This survey also noted that T-2 toxin occurred frequently in the European Union and high positive rates were also found in the Middle East and North Africa. Other literature supported the finding that T-2 toxin is widespread in the European Union (Vulic, Pleadin, and Persi 2011; Pleadin et al. 2012).

Table 1. Mycotoxin prevalence in maize in different regions.

Continent	Mycotoxin	Year	Region	Number of samples	Positive samples (%)	Average of positives (ppb)	Median (ppb)	Maximum (ppb)	References
America	FBs	2020	North America	465	70	2434	1089	45548	BIOMIN (2020)
			South & Central America	3577	84	2235	1390	56000
		2019	North America	534	78	2839	892	72950	BIOMIN (2019)
			South & Central America	3304	90	3177	1700	170300
		2014-2017	North America	945	61	3169	900	154000	Munkvold et al. (2019)
		2009-2011	North America	466	39	1357	490.0	22900	Rodrigues and Naehrer (2012)
			South America	807	92	3226	2008.0	53700
		Range	America	–	39-92	1357-3226	490-2008	22900-170300	–
	DON	2020	North America	468	72	808	557	6003	BIOMIN (2020)
			South & Central America	2568	49	473	350	6330
		2019	North America	448	85	1011	573	9250	BIOMIN (2019)
			South & Central America	2808	56	499	340	5600
		2014-2017	North America	893	66	814	460	24792	Munkvold et al. (2019)
		2009-2011	North America	390	79	1085	565.0	24900	Rodrigues and Naehrer (2012)
			South America	322	17	214	172.0	939
		Range	America	–	17-85	214-1085	172-573	939-24900	–
	ZEN	2020	North America	471	40	352	139	13351	BIOMIN (2020)
			South & Central America	3124	28	178	60	4948
		2019	North America	534	55	285	145	4154	BIOMIN (2019)
			South & Central America	3472	27	124	53	2487
		2014-2017	North America	958	27	246	100	13700	Munkvold et al. (2019)
		2009-2011	North America	395	29	251	86.4	4787	Rodrigues and Naehrer (2012)
			South America	321	43	176	87.1	1800
		Range	America	–	27-55	124-352	53-145	1800-13700	–
	AFs	2020	North America	471	6	48	6	482	BIOMIN (2020)
			South & Central America	3839	12	8	4	179
		2019	North America	524	4	132	5	1327	BIOMIN (2019)
			South & Central America	4091	21	10	4	1264
		2014-2017	North America	952	7	39	9	980	Munkvold et al. (2019)
		2009-2011	North America	375	26	67	10.1	920	Rodrigues and Naehrer (2012)
			South America	809	25	7	1.8	273
		Range	America	–	4-26	7-132	1.8-10	179-1327	–
	OTA	2020	North America	471	1	67	3	196	BIOMIN (2020)
			South & Central America	518	8	5	4	26
		2019	North America	534	1	5	5	9	BIOMIN (2019)
			South & Central America	462	2	10	4	30
		2014-2017	North America	912	2	25	3	200	Munkvold et al. (2019)
		2009-2011	North America	126	10	5	2.3	18	Rodrigues and Naehrer (2012)
			South America	147	12	133	71.3	355
		Range	America	–	1-12	5-133	2.3-71	9-355	–
	T-2	2020	North America	447	2	85	70	220	BIOMIN (2020)
			South & Central America	1112	13	40	28	321
		2019	North America	426	4	135	88	751	BIOMIN (2019)
			South & Central America	1101	10	52	30	707
		2014-2017	North America	855	3	48	19	200	Munkvold et al. (2019)
		Range	America	–	2-13	40-135	19-88	200-751	–
Europe	FBs	2020	Europe	733	71	1153	555	13902	BIOMIN (2020)
		2019	Europe	568	73	1475	397	297300	BIOMIN (2019)
		2009-2011	Central Europe	30	60	2180	684.0	7680	Rodrigues and Naehrer (2012)
			Southern Europe	48	90	2271	1407.0	11050
		Range	Europe	–	60-90	1153-2271	397-1407	7680-297300	–
	DON	2020	Europe	943	70	808	469	8300	BIOMIN (2020)
		2019	Europe	705	83	726	355	8600	BIOMIN (2019)
		2009-2011	Central Europe	535	72	1421	716.0	26121	Rodrigues and Naehrer (2012)
			Southern Europe	59	47	985	523.0	3851
		Range	Europe	–	47-83	726-1421	355-716	3851-26121	–
	ZEN	2020	Europe	805	55	171	50	9455	BIOMIN (2020)
		2019	Europe	652	58	94	45	1834	BIOMIN (2019)
		2009-2011	Central Europe	379	39	123	78.5	849	Rodrigues and Naehrer (2012)
			Southern Europe	52	21	290	166.0	1546
		Range	Europe	–	21-58	94-290	45-166	849-9455	–
	AFs	2020	Europe	449	10	12	4	89	BIOMIN (2020)
		2019	Europe	427	9	8	4	54	BIOMIN (2019)
		2009-2011	Central Europe	16	31	2	1.5	3	Rodrigues and Naehrer (2012)
			Southern Europe	42	36	9	4.0	44	Rodrigues and Naehrer (2012)
		Range	Europe	–	9-36	2-12	1.5-4	3-89	–
	OTA	2020	Europe	438	11	7	3	52	BIOMIN (2020)
		2019	Europe	416	16	20	3	412	BIOMIN (2019)
		2009-2011	Central Europe	21	10	2	2.4	3	Rodrigues and Naehrer (2012)
			Southern Europe	31	29	15	9.3	46
		Range	Europe	–	11-29	2-20	2.4-9.3	3-412	–
	T-2	2020	Europe	494	39	57	20	1387	BIOMIN (2020)
		2019	Europe	439	40	67	16	4134	BIOMIN (2019)
		Range	Europe	–	39-40	57-67	16-20	1387-4134	–
Asia	FBs	2020	Asia	983	96	2320	1078	30872	BIOMIN (2020)
		2019	Asia	719	100	2083	1010	40090	BIOMIN (2019)
		2009-2011	North Asia	443	75	2816	1518.5	23499	Rodrigues and Naehrer (2012)
			South-East Asia	326	83	1568	1033.0	19289
			South Asia	108	74	845	541.0	6196
		Range	Asia	–	74-100	845-2816	541-1518	6196-40090	–
	DON	2020	Asia	985	80	669	520	5978	BIOMIN (2020)
		2019	Asia	719	88	499	359	8419	BIOMIN (2019)
		2009-2011	North Asia	477	92	1154	640.0	15073	Rodrigues and Naehrer (2012)
			South-East Asia	218	45	307	182.0	4805
			South Asia	106	22	278	190.0	1150
		Range	Asia	–	22-92	278-1154	182-640	1150-15073	–
	ZEN	2020	Asia	983	68	266	43	11786	BIOMIN (2020)
		2019	Asia	718	75	78	32	2233	BIOMIN (2019)
		2009-2011	North Asia	470	67	437	176.0	7446	Rodrigues and Naehrer (2012)
			South-East Asia	319	20	288	97.0	2601
			South Asia	108	9	269	78.5	1099
		Range	Asia	–	9-75	78-437	32-437	1099-11786	–
	AFs	2020	Asia	984	22	81	17	2495	BIOMIN (2020)
		2019	Asia	717	31	43	10	773	BIOMIN (2019)
		2009-2011	North Asia	447	12	114	7.0	4687	Rodrigues and Naehrer (2012)
			South-East Asia	330	71	146	38.0	6105
			South Asia	108	82	240	96.0	2230
		Range	Asia	–	12-82	43-240	7-96	773-6105	–
	OTA	2020	Asia	640	47	6	2	571	BIOMIN (2020)
		2019	Asia	448	2	7	3	28	BIOMIN (2019)
		2009-2011	North Asia	420	10	4	1.4	19	Rodrigues and Naehrer (2012)
			South-East Asia	218	12	9	3.0	80
			South Asia	107	27	31	7.4	400
		Range	Asia	–	2-47	4-31	1.4-7.4	19-571	–
	T-2	2020	Asia	639	9	25	18	167	BIOMIN (2020)
		2019	Asia	448	2	19	17	49	BIOMIN (2019)
		Range	Asia	–	2-9	19-25	17-18	49-167	–
Middle East & Africa	FBs	2020	Middle East & North Africa	20	95	1418	508	8586	BIOMIN (2020)
			Africa (without North Africa)	338	65	596	221	9481
		2019	Middle East & North Africa	30	47	1440	848	6559	BIOMIN (2019)
			Africa (without North Africa)	376	78	741	244	14346
		2006-2017	Africa (without South Africa)	79	92.4	–	1371	10018	Gruber-Dorninger, Jenkins, and Schatzmayr (2018)
			South Africa	281	80.1	–	177	16932
		Range	Middle East & Africa	–	47-95	596-1440	177-1371	6559-16932	–
	DON	2020	Middle East & North Africa	20	75	260	79	1633	BIOMIN (2020)
			Africa (without North Africa)	338	93	730	412	5557
		2019	Middle East & North Africa	30	87	330	328	670	BIOMIN (2019)
			Africa (without North Africa)	376	83	664	305	10607
		2006-2017	Africa (without South Africa)	79	41.8	–	0	4974	Gruber-Dorninger, Jenkins, and Schatzmayr (2018)
			South Africa	314	80.6	–	290	9176
		Range	Middle East & Africa	–	42-93	260-730	79-412	670-10607	–
	ZEN	2020	Middle East & North Africa	20	45	80	32	418	BIOMIN (2020)
			Africa (without North Africa)	338	54	62	25	2198
		2019	Middle East & North Africa	30	67	39	36	137	BIOMIN (2019)
			Africa (without North Africa)	376	56	67	24	1344
		2006-2017	Africa (without South Africa)	79	41.8	–	0	858	Gruber-Dorninger, Jenkins, and Schatzmayr (2018)
			South Africa	308	47.1	–	0	6276
		Range	Middle East & Africa	–	42-67	39-80	24-36	137-6276	–
	AFs	2020	Middle East & North Africa	18	17	2	2	2	BIOMIN (2020)
			Africa (without North Africa)	338	2	191	7	1032
		2019	Middle East & North Africa	30	37	2	1	5	BIOMIN (2019)
			Africa (without North Africa)	376	7	16	6	64
		2006-2017	Africa (without South Africa)	79	49.4	–	0	379	Gruber-Dorninger, Jenkins, and Schatzmayr (2018)
			South Africa	282	9.6	–	0	14
		Range	Middle East & Africa	–	2-49	2-191	1-7	2-1032	–
	OTA	2020	Middle East & North Africa	18	22	2	2	3	BIOMIN (2020)
			Africa (without North Africa)	338	1	3	2	5
		2019	Middle East & North Africa	30	33	4	3	12	BIOMIN (2019)
			Africa (without North Africa)	375	1	15	15	24
		2006-2017	Africa (without South Africa)	49	22.5	–	0	694	Gruber-Dorninger, Jenkins, and Schatzmayr (2018)
			South Africa	269	7.4	–	0	95
		Range	Middle East & Africa	–	1-33	2-15	2-15	3-694	–
	T-2	2020	Middle East & North Africa	17	35	12	9	30	BIOMIN (2020)
			Africa (without North Africa)	338	0	–	–	0
		2019	Middle East & North Africa	29	38	22	14	60	BIOMIN (2019)
			Africa (without North Africa)	375	2	10	10	10
		2006-2017	Africa (without South Africa)	79	5.1	–	0	61	Gruber-Dorninger, Jenkins, and Schatzmayr (2018)
			South Africa	273	0.7	–	0	80
		Range	Middle East & Africa	–	0-38	10-22	9-14	0-80	–

-: not available

Globally, at least one mycotoxin proved to be present in more than 90% of the samples tested. Contamination with multiple mycotoxins was also common, with 68% of samples contaminated with at least two mycotoxins (Leite et al. 2021). The co-occurrence percentage of aflatoxin B₁(AFB₁) and OTA in maize flours in the period between 2014 and 2016 was 7.1% − 34.4% (Torovic 2018). In 705 maize samples from 26 US states across 2015–2016, contamination with three mycotoxins was found in 21% of the samples (Hendel et al. 2017). During 2008–2017, FBs and DON, FBs and ZEN, and DON and ZEN co-occurred in 49%, 37% and 39% respectively of global maize samples tested (Gruber-Dorninger, Jenkins, and Schatzmayr 2019). In a study of mycotoxin occurrence in maize kernels in Spain from 2015 to 2019, co-occurrence of several mycotoxins (BIOMIN 2019; Munkvold 2003; Garcia-Diaz et al. 2020) reached quantifiable levels in 33.5% of samples (Tarazona et al. 2020). The synergistic effect of co-occurring mycotoxins could cause severe harm at relatively low levels. Therefore, multiple strategies are needed to mitigate risks, which increases the difficulty of mycotoxin risk control.

Mycotoxin incidence in maize varies between regions, due to differences in the geographic locations of maize growth including meteorological conditions (Perrone et al. 2020) and toxigenicity of fungal populations (Okoth et al. 2012). Each region has its own unique mycotoxin pattern. A key driver of this high correlation between mycotoxin prevalence and region is climatic characteristics (humidity, temperature and rain fall), which provide the conditions for toxigenic fungi infection and mycotoxin accumulation (Paterson, Russell, and Lima 2010). When climate conditions are favorable for fungal growth, contamination with mycotoxins is more severe (Mahdjoubi et al. 2020). For instance, Aspergillus fungi (AFs and OTA producers), especially A. flavus, are most common at latitudes 26-35 degrees North or South, mainly due to drought stress and high temperature (Munkvold et al. 2019).

The impacts of climate change on mycotoxin contamination of maize

It has been suggested that climate change is a worldwide factor driving emerging risks for food and feed safety, as well as for plant and animal health and nutritional characteristics. Reflecting this concern, the UN Environment Programme considered climate change-related increase of mycotoxin risk during their 2016 annual assembly (Harvey et al. 2016). Rainfall, humidity and temperature are major drivers of both maize stress (which increases susceptibility to fungal pathogens and mycotoxins) and fungal growth and mycotoxin production. The influences of climatic conditions on mycotoxigenic fungal infection and mycotoxin contamination of maize have been widely documented, and exploited for risk mapping and agricultural production simulation modeling (Moretti, Pascale, and Logrieco 2019; Chauhan et al. 2015). Multiple studies have shown that the fluctuation of weather conditions may cause interannual differences in observed mycotoxin incidences and concentrations and the presence of Fusarium spp. DNA (Gruber-Dorninger, Jenkins, and Schatzmayr 2018; Vandicke et al. 2019), translating to increased AFs risk in maize in Northern Italy and Eastern Europe over the last fifteen years (Perrone et al. 2020; Battilani et al. 2016). Vandicke et al. (2019) also showed that the variation of weather conditions could influence fungal populations and significantly impact the corresponding mycotoxin contamination of maize. FBs were found in only 2% of the samples in 2016, which was wet and cold, but in 61% of samples in the hot and dry year 2018 in Flanders, Belgium.

The potential effects that are exerted by climate change on toxigenic fungi and mycotoxin risks warrant further study. Multidisciplinary cooperation is required to address these new fields and develop appropriate intervention strategies, such as creating a predictive model for farmers to diminish the impact of mycotoxins and thus secure food and feed safety.

Mycotoxins that contaminate maize carry over to MGM during processing. Furthermore, MGM production can actually concentrate certain mycotoxins, increasing contamination levels and making safe use of MGM potentially more challenging than the original raw material. With the widespread use and economic importance of MGM, mycotoxin contamination has become a critical challenge for its use in food and feed. Given the complexities of mycotoxigenic contamination of maize both pre- and post-harvest, that climate change is aggravating this challenge, and that MGM is used for food and feed, a systematic strategy to control mycotoxin risk in MGM is crucial to ensure its safe and economically viable use.

In order to identify critical points at which mycotoxins can be reduced or eliminated as they relate to MGM, the following sections discuss control strategies for mycotoxin risk reduction in maize production and storage, follow the fate of mycotoxins through the MGM production process, and review the potential application of sulfur dioxide (SO₂) and lactic acid bacteria (LAB) for decontamination during processing. Finally, the potential application prospects, challenges, and opportunities for emerging technologies to control MGM mycotoxins are explored.

Mycotoxin distribution during MGM production

Mycotoxins are redistributed into different maize fractions during wet milling (Table 2). Different mycotoxins exhibit different migration characteristics, although the data suggest that most of the mycotoxin types found in maize are concentrated in the gluten fraction through the wet milling process–meaning that they are prone to enrichment during MGM production. Park et al. (2018) reported that 31.3% of FB₂ from maize was distributed into maize gluten, followed by OTA (30.6%), ZEN (27.7%), T-2 (24.2%), HT-2 (16.1%), AFB₁ (15.1%), FB₁ (12.7%) and DON (9%). Other studies reported that higher levels of ZEN (45-56%) were distributed into maize gluten (Romer 1984; Bennett and Anderson 1978). In addition, 11 − 25.3% of AFs has been observed reported that redistributed into gluten fraction (Yahl et al. 1971; Romer 1984; Schaarschmidt and Fauhl-Hassek 2021). Considering that the gluten content only accounts for 5% of the initial maize, the distribution rate of these mycotoxins demonstrates that the mycotoxins in MGM have not decreased, and some mycotoxins have accumulated significantly through wet milling (Schaarschmidt and Fauhl-Hassek 2021).

Table 2. Fate of mycotoxins during wet milling.

Mycotoxin	Mycotoxin concentration (µg/kg)						Mycotoxin distribution rate (%)					Scale	Analysis	Reference
	Maize	^aSoluble	Fiber	Germ	Gluten	Starch	Soluble	Fiber	Germ	Gluten	Starch
AFs	487*	263	^b56.4		124	42.3	54.4	11.6		25.3	8.70	Lab	HPLC	Aly (2002)
	–	–	–	–	–	–	42.0	30.0	10.0	17.0	1.00	–	–	Yahl et al. (1971)
	–	–	–	–	–	–	50.0	28.0	11.0	11.0	0.20	–	–	Romer (1984)
AFB1	120	610	340	140	140	2.20	39.5	38.0	6.00	13.0	1.00	Lab	TLC	Bennett and Anderson (1978)
	638*	2450	1600	798	768	9.00	41.5	30.0	10.0	17.0	1.00
	0.07	0.70 ± 1.00	0.20 ± 0.30	0.20 ± 0.30	0.20 ± 0.30	ND	^c65.0		19.9	15.1	Nd	Factory	LC-MS	Park et al. (2018)
FB1	648	4595 ± 3263	820 ± 633	1043 ± 817	1416 ± 818	6.80 ± 5.70	75.1		11.7	12.7	0.50	Factory	LC-MS	Park et al. (2018)
FB2	153	300 ± 220	277 ± 245	512 ± 498	825 ± 425	3.00 ± 4.60	43.5		24.3	31.3	0.90	Factory	LC-MS	Park et al. (2018)
ZEN	900	3900	2700	1700	6800	ND	20.4	19.0	9.10	51.5	ND	Lab	TLC	Bennett and Anderson (1978)
	4100	9100	3600	3600	13400	ND	26.3	14.7	10.2	48.8	ND
	9400	10600	6800	7500	20400	ND	17.4	15.4	10.9	56.3	ND
	69.0	23.8 ± 22.6	143 ± 141	299 ± 245	330 ± 238	7.90 ± 9.20	33.1		31.4	27.7	7.80	Factory	LC-MS	Park et al. (2018)
	–	–	–	–	–	–	26.0	18.0	11.0	45.0	ND	–	–	Romer (1984)
DON	327	7418 ± 4409	530 ± 313	359 ± 217	506 ± 211	2.90 ± 7.00	82.6		8.00	9.00	0.40	Factory	LC-MS	Park et al. (2018)
T-2	502*	276	385	975	440	33.0	69.8	9.00	10.6	7.50	3.10	Lab	GC-FID	Collins and Rosen (1981)
	2250*	660	1380	3850	1380	198	66.9	8.80	12.4	6.30	5.60
	8700*	1790	7630	14200	11300	652	64.0	10.5	8.30	13.3	3.90
	2.91	9.90 ± 8.10	3.20 ± 5.00	6.90 ± 7.60	12.2 ± 10.3	0.20 ± 0.30	55.2		17.3	24.2	3.30	Factory	LC-MS	Park et al. (2018)
HT-2	5.52	33.2 ± 18.3	3.90 ± 4.40	1.00 ± 2.50	15.4 ± 10.3	ND	82.5		1.40	16.1	ND	Factory	LC-MS	Park et al. (2018)
OTA	0.09	0.80 ± 0.40	0.20 ± 0.30	0.20 ± 0.20	0.50 ± 0.40	0.10 ± 0.10	25.0		11.7	30.6	32.7	Factory	LC-MS	Park et al. (2018)

^a Soluble consists of steeping water and processing water.

^b Combination data for fiber and germ fractions.

^c Combination data for soluble and fiber fractions.

* Contaminated maize with relative high levels of toxins obtained by artificial inoculation.

^–Not available.

ND: not detected.

Redistribution of mycotoxins during the wet milling process may be affected by a number of factors. ZEN is mainly redistributed in soluble, fiber, germ and gluten, which are relatively small proportion (<30%) in maize (Table 2). However, there is little or even no ZEN detected in starch, which accounts for the highest proportion in maize (>65%) (Blasi et al. 2001). Therefore, the level of ZEN in the maize raw material is lower than all downstream products except starch, and ZEN is therefore enriched in the non-starch products. This result may be related to the poor water solubility of ZEN and its interaction with maize protein. By contrast, DON, T-2 and HT-2, exhibit high distribution rates in soluble components: Lauren and Ringrose (1997) considered that the high concentration of DON in liquid fractions during wet milling was related to its high-water solubility. Collins and Rosen (1981) and Park et al. (2018) obtained similar results for T-2 redistribution through laboratory simulation and factory sampling.

There may be several reasons for the variation in distribution observed in different studies. Firstly, different process conditions were used for production of MGM. Scale was likely to have been an important factor as some of these studies were at factory scale whilst others were laboratory based. Secondly, the binding interactions between proteins and mycotoxins can make quantitative detection more variable because of the existence of hidden (or “masked”) mycotoxins. Recently, it was reported that ZEN in maize was underestimated (masked) because of the presence of zein-bound ZEN, and the amount of total detected ZEN significantly increased after hydrolysis (Tan et al. 2021). Thirdly, the effect of wet milling depends on characteristics of the maize batch, including how mycotoxins distribute between maize fractions during processing. Studies using different batches of “starting” maize with different levels of mycotoxin contamination have shown different degrees of mycotoxin distributed into maize gluten (Schaarschmidt and Fauhl-Hassek 2021). Finally, differences in sampling will also contribute to the different results.

Global mycotoxin occurrence data indicate that the prevalence of FBs, DON, ZEN and AFs are high in maize used for MGM production, especially FBs and ZEN (BIOMIN 2020, 2019). Due to mycotoxin enrichment during the wet milling process, MGM produced from a batch of compliant maize might be found to have excessive mycotoxins such as ZEN or FBs due to now elevated levels. Therefore, predicting the relationship between mycotoxin occurrence in “starting” maize and MGM, through specific and well-controlled milling processes, is critical for managing the risk of mycotoxin contamination of MGM. Furthermore, understanding the mechanism of mycotoxin re-distribution and accumulation will further provide directions to develop novel approaches for decontaminating mycotoxins in MGM.

Mycotoxin risk management strategies in maize and MGM

As goes the common adage, an ounce of prevention is worth a pound of cure. In other words, those who rely on clean raw materials for MGM production need to consider how that maize is produced in the first place and how that influences the risk of mycotoxin contamination. A systematic control strategy for maize pre-harvest, post-harvest and during processing is crucial to intercept, control and remove hazardous mycotoxins from the food chain. Multiple strategies have been established to manage either mycotoxin occurrence or fungal growth, or to eliminate and to degrade mycotoxins in foodstuffs (Xing et al. 2017). A sustainable and effective sourcing strategy may involve partnering with organizations and initiatives that specialize in promoting best maize production and postharvest practices.

Risk management strategies by GAP in maize

Good agricultural practices (GAP) refers to the best approaches, principles and standards applied to on-site farm production to assure the safety of food and non-food agricultural products. GAP has been demonstrated to be an effective strategy to control and manage contamination by mycotoxins from pre- to post-harvest. GAP is effective and practicable for mycotoxin control in maize in rural, commercial and sustainable applications. GAP has been recommended by the FAO with uniform guidance provided for all countries to consider (FAO 2003), with various agricultural development initiatives tailoring local, stakeholder-informed solutions for smallholder production systems. Given its documented effectiveness at reducing the risk of mycotoxin contamination in maize, as well as the win-win impacts of increasing yields and reducing post-harvest losses in general, GAP is an essential element that should be taken into consideration both within high risk regions for farmers and in industrial production (Thai Agriculture Standard 2010).

Pre-harvest strategies

Fungal infection starts at the pre-harvest stage in maize, so GAP needs to be implemented from the production stage to minimize mycotoxin contamination (Mahuku et al. 2019). Pre-harvest management includes mitigation of drought stress, insect infestation, and the primary inoculum. Humans have minimal control over environmental factors. However, good crop husbandry practices, such as crop rotation, proper tillage, irrigation, timely planting and harvesting, and pesticide usage where appropriate and safely practicable are preventative actions that reduce mycotoxin occurrence in crops (Erickson, Klopfenstein, and Watson 2012).

Pre-harvest can be divided into pre-planting, planting, and after-planting stages. During pre-planting, breaking the hardpan during tillage can reduce subsequent drought stress. During planting, healthy seeds of a variety adapted to the given agroecological conditions are recommended; these should be planted at an optimal planting time and plant density to avoid preferred conditions for mycotoxigenic fungi (Abbas et al. 2009).

Practices such as irrigation (where available), fertilizer use, insect control, weed control and fungicide use are critical for field management after planting, to reduce mycotoxigenic fungal contamination. It should be noted that many insects can spread or provide entry points for mycotoxigenic fungi in maize, for example, lepidopteran ear borer can spread A. flavus in harvested maize, and insect damage of maize can be a predictive indicator for Fusarium contamination. An appropriate pesticide or insect control can be beneficial in reducing mycotoxin contamination (Adegoke and Letum 2013). The capacity of actors in a given food system to appropriately and safely apply pesticides should be considered when assessing this intervention, given that 44% of farmers (385 million) are estimated to suffer from Unintentional Acute Pesticide Poisoning globally, many of whom are in developing countries (Boedeker et al. 2020).

Water stresses can be a driver of mycotoxin susceptibility. Excess moisture during flowering allows favorable conditions for the dissemination and infection by Fusarium; because of which, there should be appropriate restraint on irrigation during the flowering and crop maturation stages, where irrigation is used. For rainfed maize production, which represents much of production in developing countries, the use of adapted, drought-tolerant varieties can significantly reduce the risk of mycotoxin contamination before harvest.

The use of biocontrol agents have been developed to control mycotoxin production (Neaj, Habschied, and Mastanjevj 2021; Agriopoulou, Stamatelopoulou, and Varzakas 2020). Biocontrol of AF by competitive exclusion is a strategy in which non-toxigenic spores of A. flavus are introduced in large amounts into the soil of the fields where peanuts or maize are grown, thus outcompeting toxigenic spores and reducing AF accumulation. Some research reported that biocontrol could be fruitful in reducing aflatoxin-contamination of maize in Thailand, although results were variable (Pitt et al. 2015; Tran-Dinh, Pitt, and Markwell 2014, 2018). Pitt also discussed the advantages and related problems of the practical use of biocontrol to treat maize and reduce aflatoxin contamination in Africa (Pitt 2019). Other studies also reported high efficacy of biocontrol in reducing aflatoxin levels in maize, from field trials in Nigeria (Bandyopadhyay et al. 2019).

Post-harvest strategies

AF and OTA contamination can occur before and after harvest. Fusarium toxin contamination, such as with FB, ZEN and DON, is mainly a pre-harvest problem, but if the external environment used during storage, transportation, and processing of harvested crops meets the growth conditions for fungi, fungi will continue to grow and produce mycotoxins. Good post-harvest practices are, therefore, also important to avoid further mycotoxin contamination.

Timely harvest after crop maturity and adequate drying are recommended to reduce the risk of mycotoxin contamination in maize in high-risk regions, where fungal pathogens have a relatively shorter time to grow or potentially produce mycotoxins (Monda, Masanga, and Alakonya 2020). AFs will increase in maize when the plants proceed to maturity and harvest is delayed. One study found that AFs increased 4-7 fold after a 3-4 week delay in harvest (Hell et al. 2008). GAP also emphasizes proper drying after harvesting, i.e., avoid harvesting directly onto the soil, improved drying methods and good storage conditions can greatly reduce the risk of mycotoxin contamination, including in peanuts and maize (Turner et al. 2005). The common developing country practice of heaping maize in the field after harvest was found to significantly increase AFs contamination in farmers’ fields in Ghana (Manu et al. 2019). Contamination with AFs can increase 10 fold in three days when harvested maize is stored under high moisture content (Hell et al. 2008). When collecting and transporting the harvested crop from the field to the drying process, and/or to storage, the containers utilized should be clean and dry, meanwhile the vehicles should be checked to avoid insects, soil, or visible fungal growth (Demissie 2018).

De-husking and sorting, aiming to remove poor-quality ears, are also effective ways to reduce the spread of spores and subsequent infection. Xu et al. (2021) suggested that the leafy outer covering of the maize could be removed manually before drying, because the husk could still protect the cobs during transportation. It has been suggested that maize ears should be removed if ear damage is greater than 10% (Setamou et al. 1998), however given that AF levels can be high even in visibly clean grains this practice may still fail to avert high levels of contamination.

Mycotoxin contamination during storage is mainly related to fungal load, moisture content, temperature, relative humidity, insect activities and packaging methods (Wang et al. 2021). Good field control will help to reduce fungal load. Good practice to avoid mycotoxin production in storage involves controlling moisture, which should be kept below 14% at 30 °C (corresponding to water activity below 0.70), temperature (<20 °C) and relative humidity (<80%) and maintaining proper ventilation. Under these conditions, the growth of fungi, the production of mycotoxins and the number of insects can be reduced (Channaiah and Maier 2014). In addition, pest and mycotoxin control can also be improved by closed storage, atmospheric control and the use of pesticides. Storage of maize and other crops in hermetic bags or silos is a highly effective method to arrest aflatoxin contamination in properly dried grains (Suleiman et al. 2018) and has been demonstrated to reduce the “lean season” where smallholder farming communities lack sufficient food between harvests (Brander, Bernauer, and Huss 2021). Hermetic storage technologies such as appropriate packaging (for example the Purdue Improved Crop Storage bag, and the Super Grain Bag) are also effective at eliminating insect pests (Singano, Mvumi, and Stathers 2019; Williams, Baributsa, and Woloshuk 2014). The use of phosphine fumigation is also beneficial in maize drying during storage, which helps to reduce mycotoxin contamination (FAO 1989). However, the safety of chemical residues should be fully guaranteed (Chulze 2010) and capacity of food system actors to safely apply them considered.

An appropriate training strategy needs to be developed for farmers and maize value chain workers to deploy these methods and GAP guidelines, so that mycotoxin contamination of maize is minimized before processing. A recent study in Kenya (Joutsjoki and Korhonen 2021) suggested that trained farmers deployed good practice, which resulted in a significant decrease of mycotoxin levels in maize. Xu et al. (2021) provided practicable GAP for smallholder maize farmers to reduce aflatoxin contamination in high-risk regions. Massomo (2020) also recommended an integrated action plan to address AF contamination in maize in Tanzania. GAP is an important tool to minimize mycotoxin occurrence and continuous monitoring of crops should be encouraged. Further research on the application of good agricultural storage practices and strategies to mitigate the mycotoxin risk in maize is critical (Carbas et al. 2021).

Risk management strategies for processing MGM

GAP can reduce mycotoxin contamination in maize, but not eliminate it. Mycotoxins may also be reduced through the physical separation, chemical inactivation and biological degradation that occurs during MGM processing.

Pre-processing

Regulations

To protect consumers from the harmful impact of mycotoxins, many countries have established regulations for mycotoxins in food and animal feed since the late 1960s. At least 99 countries specified mycotoxin limits for food and/or feed by late 2003 (FAO 2004). Maximum levels (MLs) in maize and its products have been agreed for FBs (sum of FB₁ and FB₂) and DON by the Codex Alimentarius Commission; application to commodities moving in international trade was recommended (FAO 1995). Corresponding MLs have been applied in EU regulations. AFB₁, total AFs, AFM₁, DON, total FBs, OTA, T-2 and HT-2 toxins and ZEN are regulated in unprocessed maize and different maize derivatives consumed by humans, as well as products used as animal feed (European Commission (EC)) 2006, 2007; Schaarschmidt and Fauhl-Hassek 2021). MLs ensure that food producers and processors consider measures to minimize the content of mycotoxins in maize and maize products.

Measurement

Measurement of mycotoxins is one of the key factors in mycotoxin management for surveillance, and to understand the occurrence, fate, and mitigation of mycotoxins in raw maize before processing and producing MGM. Several studies have reported that sampling is often the largest source of variability (among several sources of variability) in mycotoxin quantification, due to the highly heterogeneous distribution in contaminated raw and processed foods, for example, sampling attributed nearly 90% of the error when measuring AFs in some studies (Munkvold et al. 2019; Whitaker 2003; Romer 2006; Baker et al. 2014). Representative sampling becomes increasingly difficult as production scale increases (Schaarschmidt and Fauhl-Hassek 2021). A common sampling procedure has been established by the EU to reduce the variability caused by sampling. Thus, the establishment of an adequate sampling procedure is critical to guarantee reliable and accurate results (Pereira, Fernandes, and Cunha 2014).

Detection methods can also exert influence because different methods involve different extraction and clean-up procedures (Munkvold et al. 2019), and specific parameters (e.g. limit of detection) differ among studies (Leite et al. 2021). Simultaneous contamination with a variety of mycotoxins increases the difficulty of detection. Furthermore, the existence of hidden (masked) mycotoxins also makes quantitative measurement more variable. Thus, the accurate measurement of the types and contents of mycotoxins is required for effective monitoring and implementation of corresponding detoxification measures of mycotoxins during processing.

Cleaning

The cleaning process used in maize wet milling can remove large impurities, dust and sand. The ability of the ordinary cleaning process to remove grains contaminated with mycotoxin is limited. To reduce mycotoxin contamination from raw materials, the development of special mycotoxin sorters has attracted the attention of scientists and engineers. Improved gravity and optical sorters can effectively reduce the contents of FBs and AFs in maize (Morales et al. 2018; Lu et al. 2020; Lee, Herrman, and Yun 2014), but so far these sorters have not been widely used in production lines due to the sorting capacity.

Wet milling processing

SO₂ steeping process

Mycotoxin distribution during wet milling was discussed above. This section focuses on the effect of chemicals on mycotoxins during wet milling. Steeping is a crucial and typical process performed in industrial wet milling using SO₂ solutions at moderate temperatures (around 50 °C) for 24-48 h in a set of tanks with counter current flow. In such a series of tanks, as the steeping liquid (also called “steep water”) is pumped from one tank to the next, the concentration of SO₂ increases (up to around 0.3%) (Schaarschmidt and Fauhl-Hassek 2021). Chemical/biological modifications or changes in binding will take place during SO₂ steeping; these alterations will accelerate removal of the microorganisms and separation of the cellular components of the endosperm. At the same time steeping can significantly reduce the content of mycotoxins in maize, especially DON which migrates to water in large quantities because of its high solubility. SO₂ promotes starch-protein disconnection, controls microbial multiplication, and can reduce the content of AFs in maize. The amount of degradation is related to SO₂ concentration (17% of reduction rate under 0.15% SO₂ and 56% of reduction rate under 0.3% SO₂) (Wood 1982), temperature (Tabata et al. 1994) and time. A recent study found that 25-31 μg/kg of AFB₁ in maize could be detoxified by steeping with 0.2-0.3% of H₂SO₃. The transformation product was identified as C₁₇H₁₄O₉S, named AFB₁-HSO₃, and its toxicity was less than AFB₁ (Yang et al. 2022). Wood (1982) found that the OTA in kernels was reduced by 37% after steeping in a system that simulated commercial wet milling steeping with 0.15% SO₂. However, it was also noted that 17% of the total OTA was undetected, due to degradation, binding or modification. Pujol et al. (1999) showed that FB₁ was not degraded/modified in pure SO₂ solutions (0.2% and 0.4%, without maize kernels) over seven days at 60 °C. Nevertheless, he suggested that the combined effect of compounds in steep water released from maize kernels with SO₂ might promote the decomposition or transformation of FB₁. The mode of action and products from degradation of OTA and FB₁ by SO₂ are not well documented, therefore, further research in this area is necessary.

Other solutes are also used in steeping experiments to remove mycotoxins, such as sodium hypochlorite, sodium chlorite, hydrogen peroxide and sulfur-containing food additives (Schaarschmidt and Fauhl-Hassek 2021); degradation products resulting from use of these solutes are unknown or unstable and these are considered less advantageous than SO₂. The optimal conditions, degradation mechanism and the safety of degradation products of different mycotoxins requires further research, to allow adaptation to complex environments containing a variety of mycotoxins.

Lactic acid bacteria fermentation

There is a natural Lactic acid bacteria (LAB) fermentation process in the initial stages of maize steeping. LAB hydrolyzes maize protein, which is conducive to protein separation and inhibition of pathogenic and spoilage microorganisms. Various strategies for mycotoxin detoxification using microorganisms such as bacteria, yeast and fungi have been widely reviewed (Wan, Chen, and Rao 2020). Bio-decontamination based on probiotic LAB provides an emerging and promising solution that can be applied directly in maize food products (Wang et al. 2019). Degradation of mycotoxins by LAB during fermentation may, therefore, have a useful application in detoxification of MGM. Diverse species of LAB have been studied for their decontamination capabilities by biodegradation and binding (Table 3).

Table 3. Mycotoxin decontamination potential of LAB in maize and maize products and related mechanisms.

Mechanism	Source	Strain	Commodities	Temperature (°C)	Time (d)	Target Mycotoxin	Decontamination Rate (%)	Reference
Biodegradation	Clover	Lactobacillus delbrueckii, Streptococcus lactis	Maize Meal	25	3	ZEN	68	Mokoena, Chelule, and Gqaleni (2005)
					4	FB1	74
	Dairy and plant	L. brevis L. casei L. plantarum L. jugurti	Maize	25	1	ZEN	47-50	Niderkorn et al. (2007)
						FB2	34
	Ogi	L. acidophilus L. brevis L. casei L. delbruekii L. plantarum	Maize grain	37	5	AFB1	31-46	Oluwafemi et al. (2010)
		Mix LAB species	Maize-based product	35–45	4	Citrinin and ZEN	˂99	Okeke et al. (2015)
	USDA	L. casei L. fermentum L. plantarum P. pentosaceus	Maize porridges	30	1	FB1	17-20	Nyamete, Mourice, et al. (2016)
	NA	L. plantarum CECT 749 CFS	Maize kernels	37	7	FB1	90.6	Nazareth et al. (2020)
					25-40	AFB1	97.5-100
		L. casei L. fermentum L. plantarum P. pentosaceus	Tanzania maize-based gruel	30	1	AFB1	45-55	Nyamete, Bennink, et al. (2016)
Binding	Ogi	L. acidophilus L. brevis L. casei	Maize and maize products	37	3	AFS	48-54	Oluwafemi and Da-Silva (2009)
	sour grain mash	L. delbruekii L. plantarum	Ogi and mahewu	30	6	FB1	75-100	Dawlal et al. (2017)
						FB2	63-71
	Yoba	L.rhamnosus yoba2012	ground maize	1	30	AFB1	83	Wacoo et al. (2019)
Binding and Biodegradation	NA	L. buchneri L. reuteri L. plantarum L. fermentum	Maize grain	20 ± 2	48	OTA	80	Zielińska and Fabiszewska (2017)
NA	NA	LAB starter culture S. cerevesiae	Maize meal and mahewu	22-25	2–4	AFB1, FB1 and ZEN	100	Chelule et al. (2010)

NA = Information not available.

The benefits of using LAB strains for decontamination of mycotoxins in maize and MGM include direct application of LAB and the low cost during maize processing; however, there are some concerns about the safety of degradation products and impact on the quality and safety of treated maize and MGM. Further research with specific LAB suitable for the processing conditions may be needed to improve the efficiency and safety of mycotoxin degradation characteristics of the chosen LAB, these may include high degradation efficiency (within maize steeping time of 24-48 h), and high-temperature resistance (steeping temperature around 50°C). It will also be necessary to analyze the safety of mycotoxin degradation products and their impact on the process. Finally, selecting LAB according to the type of mycotoxin and using compounded LAB with different decontamination functions may improve degradation efficiency.

Post-processing

Mycotoxin measurement in final products is the last line of defense to ensure food safety. The EU has set the MLs of mycotoxins in maize and maize-based products consumed by humans. The ML of AFB₁ in maize and all maize products is 2 μg/kg. For processed maize-based foods that are consumed by infants and young children and dietary infant foods used for special medical purposes, it is 0.1 μg/kg. In addition, the limit of AFs (the sum of B₁, B₂, G₁ and G₂) in maize and all maize products is regulated as 4 μg/kg (European Commission (EC) 2006). MGM should be tested and follow the MLs to ensure safety before entering the market. Proper packaging, storage, and transportation of MGM are also necessary.

No single critical control point can achieve mycotoxin risk management in MGM. Comprehensive strategies to reduce the contamination caused by mycotoxins from the field through to the end-processed product are needed. GAP and improved processing can support the decontamination of mycotoxins in MGM by reducing the mycotoxin risk in “starting” maize. Future research will continue to address the challenges associated with the distribution of mycotoxins from maize to MGM during the wet milling process. We believe that research will enable the development of a risk assessment or predictive tool to help control mycotoxin contamination of MGM.

Emerging mycotoxin decontamination methods for maize and MGM

Mycotoxin removal and detoxification methods for cereals have been extensively researched (Mir et al. 2021) and mycotoxin control methods specialized for maize have been initiated and reviewed (Grenier et al. 2014; Munkvold et al. 2019). The approaches that are employed to mitigate mycotoxins in maize and maize products are commonly classified into chemical, physical, and biological methods, which act through removal or inactivation of mycotoxins. Finding safe and efficient methods to remove mycotoxins has always been the goal of researchers. Three emerging techniques for mycotoxin decontamination are discussed below which show great potential for mycotoxin control in MGM.

Emerging sorting techniques

Sorting is the most common and effective process for mycotoxin removal (Liu et al. 2020), although its effectiveness varies depending on context and application. Sorting enables differentiation between sub-standard or heterogeneous kernels and intact ones based on differences in density, size, shape and color. Various sorting techniques (such as hand-, color- and optical sorting) have been developed (Marshall et al. 2020). However, removal efficiency and accuracy of mycotoxin detection are challenges for traditional sorting machines. Advanced sorting methods embedded with emerging techniques are being studied and developed to exclude infected maize kernels.

Morales et al. (2018) demonstrated that the correlation between kernel bulk density (KBD) and FBs levels in maize kernels was negative. This knowledge of the association between KBD and FBs has been used in the development of a low-priced grain sorter prototype (“DropSort”). This machine has been shown to be effective in reducing FBs concentrations to below 2 ppm, but the effectiveness in reducing AF levels in maize grain to under 20 ppb is still controversial, especially in heavily AF-contaminated grain (Aoun et al. 2020).

Methods that have been developed for optical sorting mainly include using ultraviolet (UV), visible or infrared wavelengths. A combination of infrared, visible, and UV light imaging is currently being used to sort contaminated maize kernels (Stasiewicz et al. 2017; Chavez et al. 2022; Cheng, Vella, and Stasiewicz 2019). However, these technologies have not been applied in production for maize wet milling due to failing to meet the processing capability.

Currently, research in sorting has progressed beyond detecting and removing indicators of fungal infection (such as the size, shape and color of contaminated grains), to targeting the mycotoxins directly. Fluorescence spectroscopy showed promising outcomes for sorting AF contaminated maize. In chromatographic sorting of AF contaminated maize, fluorescence emitted by aflatoxin and bright greenish yellow fluorescence emitted by kojic acid (another secondary metabolite of A. flavus and A. parasiticus) oxidation under UV light made nondestructive detection and sorting of AF contaminated grain possible (Lillehoj, Jacks, and Calvert 1986; Tao et al. 2018). The commercialization of fluorescence sorting has been promoted by the development of new technologies including hyperspectral imaging (HSI). HSI adds a spatial dimension to conventional spectroscopic techniques, enabling mapping of chemical components (also known as chemical imaging) in the tested sample, which is particularly useful for detection of unevenly distributed components (Saha and Manickavasagan 2021; Tao et al. 2018).

A method and detection system for AF in maize using fluorescence spectra has been patented (Yao et al. 2013). Smeesters et al. (2015) used one- and two-photon induced fluorescence spectrometry to detect maize contaminated with AFs. Han et al. (2019) suggested a fluorescence multispectral system with two cameras to screen AF contaminated maize. An accuracy of above 96.0% was obtained based on the fluorescence images at 436 nm and 532 nm. At present, a commercially available sorting technique uses LED light and fluorescence HSI to detect and sort contaminated maize at 15 tons per hour, with a reduction in AF contamination averaging at 85-90% in tests (Bühler 2018; Marshall et al. 2020). Another kind of laser induced fluorescence-based sorter has also started to detect AFs in incoming maize (Tomra 2022).

Fluorescence-based sorting devices have a sound scientific basis and research has provided a solid foundation for commercial promotion and inline application. The devices will play an important role in removing AF contaminated maize before MGM processing. The main limiting factors are cost, fluorescence intensity and image acquisition at processing speeds which affect the sorting accuracy; as a result industrial application is relatively slow (Lu et al. 2020). Further studies are needed to establish new modalities with faster image processing and analysis to fit with industrial scale treatment during maize wet milling for the reduction of mycotoxins risk.

Sorting is a promising technology. However, dealing with maize rejected by sorting must be considered by every responsible manufacturer and scientist and it is particularly important that rejected maize does not enter food chain. To reduce waste and environmental pollution, physical, chemical and biological decontamination strategies for rejected maize need further research (Temba et al. 2016; Bosch et al. 2017; Meijer et al. 2019).

Cold plasma

Cold Plasma (CP) is a potential decontamination technology that can be applied before crops enter the production process. The ability of cold plasma to inactivate the mycotoxigenic fungi (including Aspergillus spp., Fusarium spp., Penicillium spp. and Alternaria spp.) and detoxify mycotoxins (including FBs, ZEN, DON and AFs) in various food matrixes, including cereals, have been confirmed (Sharma and Singh 2020; Gavahian and Cullen 2019; Ten Bosch et al. 2017; Schluter et al. 2013).

CP has advantages over other mycotoxin reduction methods currently used in the agriculture and food industries, due to its lower cost and energy needs (Misra et al. 2019). One study was carried out to determine the efficacy of a self-designed Low-pressure CP device for the inactivation of Aspergillus spp. and Penicillium spp. in maize kernels; the results demonstrated a significant decrease of three log orders of magnitude for both species within 15 min (Selcuk, Oksuz, and Basaran 2008). Dasan, Boyaci, and Mutlu (2016) performed research on the fungicidal effect of fluidized bed plasma treatment on A. flavus and A. parasiticus inoculated onto maize kernels. After a five-minute plasma treatment, a maximum reduction was observed with 5.48 and 5.20 lgCFU/g in A. flavus and A. parasiticus, respectively.

Ten Bosch et al. (2017) showed that pure mycotoxins (FB₁, DON, ZEN, and T2-toxin) exposed to Cold Atmospheric Pressure Plasma (CAPP) were degraded almost completely within 60 s. The degradation rate varied with mycotoxin structure and the presence of matrix slowed the process, but did not prevent degradation. Another study demonstrated high degradation of FBs (93%), ZEN (100%), AFs (93%) and T2/HT-2 (90%) after eight minutes of exposure to CAPP (Hojnik et al. 2019). AFs in maize could be reduced by 62% and 82% when treated with High Voltage Atmospheric Cold Plasma (HVACP) for one and 10 min, respectively. Humid air (40-80% RH) resulted in higher AFs degradation compared with dry air (5% RH), in maize treated by HVACP; there were no significant differences in AF degradation at RH between 40 and 80%. Lack of stirring of maize and difficulty in penetration of the reactive gas species into maize kernels also affected degradation (Shi et al. 2017). Results show that optimized CAPP (10 min of treatment) resulted in significant reduction of AFB₁ (65%) and FB₁ (64%) in maize. Whilst CAPP is a promising technology for mycotoxin detoxification, critical questions concerning potential changes to the nutritional and safety status of the food matrix require further investigation (Wielogorska et al. 2019).

CP is considered a potentially feasible technology to inactivate fungi and degrade mycotoxins in food and feed. CP can significantly reduce the toxigenic fungi and mycotoxins in maize and other food substrates and has the advantages that it has less impact on the nutrition, sensory attributes and texture of food than some other methods. It carries no risk of chemical pollution, and most of the intermediate and final degradation products are nontoxic (Adebo et al. 2020). However, this new technology is still in the early assessment/laboratory stage, costs are high and there is a risk of oxidation in lipid rich foods due to the existence of active gas (Guo et al. 2020). Due to the cost and poor permeability of dried solid material to ionized gas, CP is not readily used for detoxification of many dried maize raw materials. Wet milling provides better feasibility for using CP to degrade mycotoxins and inactivate fungi in MGM and other maize by-products, as the wet environment is more conducive to the degradation of mycotoxins. Treatment of MGM and other by-products with CP before drying may improve the degradation effect on mycotoxins. Further studies are necessary to optimize CP processing conditions for MGM and other maize by-products, including plasma type, energy, working gas, stirring, and time. Concurrent studies on the safety of mycotoxin degradation products, the interaction between CP and the food matrix (such as lipid peroxidation) will help to promote establishment of a standardized system and the commercial application of CP in MGM production.

Photocatalysis

Photocatalysis is a chemical reaction induced by the absorption of photons by a solid material (the photocatalyst). The process does not cause secondary pollutants (new pollutants by photocatalyst reacting with existing components), can be easy to operate and cost-effective, and some applications have been found in the energy and environmental fields. Decontamination of mycotoxins by photocatalysis is mature (Murugesan et al. 2021); for example, combining UV-visible irradiation with semiconducting photocatalysts can increase the efficiency of AF degradation in liquid matrix (Sun et al. 2019; Guo et al. 2020). During photocatalytic degradation, photo-generated valence band holes (h⁺), hydroxyl free radicals (OH⁻) and superoxide radical (O₂⁻) can directly oxidize AFB₁. The process also leads to the formation of different reactive oxygen species that help degrade AFs and convert them into less harmful derivatives (Sun et al. 2019; Mao, Zhang, Wang, et al. 2018). The most extensively researched photocatalyst is titanium dioxide (TiO₂) because of its high activity under UV irradiation, high efficiency and long-term photostability (Sun et al. 2019; Magzoub et al. 2019).

The disadvantages of using TiO₂ alone as a photocatalyst are that it needs to be removed from the substrate considering its potential risk to health, especially genotoxicity concerns (TiO₂ banned as a food additive in the EU in 2022) (USDA 2022), which is a difficult process. In addition, this process requires UV irradiation to be effective (Guo et al. 2021). Therefore, researchers began to develop immobilization technology (Xu et al. 2018; Magzoub et al. 2019) and composite photocatalysts capable of degrading mycotoxins under visible light.

Composite photocatalysts can improve quantum efficiency, visible light response, and substrate separation. Carbon-based materials are mainly used in composite photocatalysts. Carbon-based materials increase the adsorption performance and specific surface area of AF, thus improving the efficiency of photosensitive catalysts. The degradation efficiency of an AC/TiO₂ (AC stands for activated carbon) composite catalyst for AFB₁ was higher than that of a single catalyst (TiO₂) (Sun et al. 2019). Mao et al. (2019), Mao, Zhang, Wang, et al. (2018), Mao, Zhang, Li, et al. (2018) synthesized z-scheme based ternary composites consisting of tungsten trioxide (WO₃) nanowires and g-C₃N₄ nanosheets to enhance the photodegradation efficiency of AFB₁ under visible light irradiation. Sun et al. (2021) showed that a reduction of 96% of AFB₁ could be achieved in maize oil using a magnetic graphene oxide/TiO₂ nanocomposite after illumination for 120 min under UV-Vis light. Graphene/ZnO hybrid photocatalysts exhibited excellent activity for photodegradation of DON under UV irradiation (Bai et al. 2017). Other types of compound materials, such as Sc doping by SrTi_0.7Fe_0.3O₃ (Jamil et al. 2017), iodine-TiO₂ thin film (Xu et al. 2018), Clew-like WO₃ decorated with nanoparticles (Mao et al. 2019), MGO doping by TiO₂ (Popovic et al. 2018) had good effects on the degradation of AFB₁. Upconversion nanoparticles@TiO₂ composites (Wu et al. 2019) and NaYF₄: Yb, Tm@TiO₂ composites (Zhou et al. 2020) were effective for DON degradation.

Photocatalysis has not yet been applied to the manufacture of MGM. Nevertheless, the soaking and wet milling processes used for MGM production provide an appropriate liquid matrix for photocatalytic degradation of mycotoxins, suggesting that this may become a promising scientific and commercial field. There are still some problems to be solved for this application: i) development and utilization of new photosensitive composites. For example, the development of new nano carbon-based photosensitive materials with mycotoxin degradation under visible light has good prospects; ii) improving the efficiency of photocatalytic degradation; the most suitable type and dosage of photocatalyst should be selected according to the wet grinding soaking processing conditions, pH, toxin type and concentration, and the composition of the solution; iii) stability and safety of photosensitive catalysts and the influence of photosensitive catalysts on product quality and the safety of degradation products must be studied in detail.

It is worthwhile noting that the emerging techniques discussed above would require regulatory risk assessment and approval processes based on possible changes to the nutritional, quality and safety impacts of any by-products of the process itself.

Conclusions

Mycotoxin contamination of maize can occur pre-harvest, at harvest, during post-harvest operations and in storage. Moreover, the enrichment of mycotoxins in MGM processing increases the risk of contamination of MGM at unacceptable levels. A systematic and thorough mycotoxin risk management and decontamination strategy for MGM needs to be considered, extending from the maize field throughout processing and in research on mycotoxin decontamination in MGM (Figure 3).

Figure 3. Risk management strategy for mycotoxins in MGM from the field to the end-product.

GAP is effective in reducing mycotoxin contamination during maize planting and storage and it is worth promoting in maize planting and production regions. Sorting before processing is also an effective approach to reduce mycotoxins by removing contaminated kernels, especially with the development of new sorting equipment; a safe alternative use or effective decontamination of the waste stream is needed to avoid unintended concentration of mycotoxin consumption by some people or animals. Wet milling can reduce the total amount of mycotoxins due to transfer to water during steeping and degrading by SO₂, however, the enrichment of mycotoxins in MGM is still a disadvantageous factor. A full understanding of the migration path of mycotoxins during maize processing, and of the interaction between protein and mycotoxins, together with improvements to steeping (e.g., use of LAB), can be emphasized in future research to contribute to mycotoxin risk management in MGM.

It should be noted that the wet milling process is a mature manufacturing process that has been used for starch production for many decades. The quality and yield of starch cannot be ignored when considering any improvement for by-products such as MGM. Therefore, the method of removing mycotoxins from MGM should not increase the production burden from a commercial perspective and maintain the commercial specifications of MGM such as color, odor and nutritional composition. Technologies including the development of new sorters, the use of photocatalysts and cold plasma, show good potential within this context.

Acknowledgements

We extend our sincere thanks to Peter Markwell, Boris Bolschikov, Ahmad Tanvir, Stephen French and Cui Wang for comments and scoping that greatly improved the manuscript. We also would like to thank Cargill for supporting the technique integration, providing professional industry experience and building collaboration with the Mars Global Food Safety Center.

Disclosure statement

JH is a consultant for the Mars Global Food Safety Center.

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.