The Major Fusarium Species Causing Maize Ear and Kernel Rot and Their Toxigenicity in Chongqing, China

Fusarium verticillioides, F. proliferatum, and F. meridionale were identified as the predominant fungi among 116 Fusarium isolates causing maize ear and kernel rot, a destructive disease in Chongqing areas, China. The toxigenic capability and genotype were determined by molecular amplification and toxin assay. The results showed that the key toxigenic gene FUM1 was detected in 47 F. verticillioides and 19 F. proliferatum isolates. Among these, F. verticillioides and F. proliferatum isolates mainly produced fumonisin B1, ranging from 3.17 to 1566.44, and 97.74 to 11,100.99 µg/g for each gram of dry hyphal weight, with the averages of 263.94 and 3632.88 µg/g, respectively, indicating the F. proliferatum isolates on average produced about an order of magnitude more fumonisins than F. verticillioides did in these areas, in vitro. Only NIV genotype was detected among 16 F. meridionale and three F. asiaticum isolates. Among these, 11 F. meridionale isolates produced NIV, varying from 17.40 to 2597.34 µg/g. ZEA and DON toxins were detected in 11 and 4 F. meridionale isolates, with the toxin production range of 8.35–78.57 and 3.38–33.41 µg/g, respectively. Three F. asiaticum isolates produced almost no mycotoxins, except that one isolate produced a small amount of DON. The findings provide us with insight into the risk of the main pathogenic Fusarium species and a guide for resistance breeding in these areas.


Introduction
Fusarium species are the main pathogenic fungi causing maize ear and kernel rot worldwide, including F. verticillioides, F. graminearum species complex (FGSC), F. oxysporum, F. equiseti, F. subglutinans [1][2][3][4]. These pathogens not only cause grain rot, but also produce a variety of mycotoxins that are a direct threat to human and animal health [5,6]. Studies have shown that F. verticillioides and F. proliferatum mainly produce fumonisin B (FB) that contaminate grains and grain products, whereas members of the FGSC mainly produce trichothecene toxins that contaminate grains. These mycotoxins act as phytotoxins and virulence factors, interact with their hosts [7].
The high incidence of maize ear and kernel rot in Chongqing and surrounding areas is mainly due to its special geographical and climatic conditions, as well as cropping systems and resistance level of the major maize cultivars. The incidence of the maize ear and kernel rot is 20-40%, even reaching as high as 75%, which significantly decreases thousand-kernel weight. More seriously, mycotoxin contamination of the affected maize kernels is severe. Up to now, no systematic studies on pathogenic Fusarium toxins causing maize ear rot in these areas have been conducted. This study aimed to clarify the composition and distribution of Fusarium spp. causing maize kernel rot in Chongqing and surrounding areas, as well as the toxigenic chemotypes and their potentiality and capability. The results will provide effective information on the toxigenic genotype and toxin production capacity of major pathogenic Fusarium spp. causing maize kernel rot in the Chongqing and surrounding areas, as well as provide an early warning mechanism for regional maize production.
There were a few differences in the frequency of Fusarium isolates in different regions of Chongqing (Table 2). In Southeast Chongqing, the frequency of F. verticillioides, F. proliferatum, FGSC, and F. oxysporum was 42.86%, 10.07%, 14.29%, and 25.00%, respectively. Therefore, F. verticillioides and F. oxysporum were the predominant Fusarium species in Southeast Chongqing. However, F. oxysporum were not be found in West Chongqing. The conclusion should not be drawn for the Central Chongqing and the other regions, due to the smaller sample size.

Detection of Toxigenic Genes and Chemotypes
Using the specific primers, the FUM1 gene was detected in 47 F. verticillioides and 19 F. proliferatum isolates (Figures 3 and 4). The results showed that these isolates theoretically possessed the capacity to synthesize FBs.

Detection of Toxigenic Genes and Chemotypes
Using the specific primers, the FUM1 gene was detected in 47 F. verticillioides and 19 F. proliferatum isolates (Figures 3 and 4). The results showed that these isolates theoretically possessed the capacity to synthesize FBs.
The Tri13 gene-specific primer Tri13P1/Tri13P2 was used to conduct the PCR amplification of 19 members of the FGSC in Figure 5. A single 859 bp fragment was stably amplified in 16 F. meridionale isolates and three F. asiaticum isolates, indicating that all 19 members of the FGSC were of the NIV chemotype.

Detection of Toxigenic Genes and Chemotypes
Using the specific primers, the FUM1 gene was detected in 47 F. verticillioides and 19 F. proliferatum isolates (Figures 3 and 4). The results showed that these isolates theoretically possessed the capacity to synthesize FBs.    The Tri13 gene-specific primer Tri13P1/Tri13P2 was used to conduct the PCR amplification of 19 members of the FGSC in Figure 5. A single 859 bp fragment was stably amplified in 16 F. meridionale isolates and three F. asiaticum isolates, indicating that all 19 members of the FGSC were of the NIV chemotype.

Analysis of FBs
Mycotoxin assays showed that all F. verticillioides and F. proliferatum isolates could produce toxins FB1, FB2, and FB3. Except for F. verticillioides isolates D61-1 and D63, the other isolates exhibited a significantly higher FB1 yield than that of FB2 and FB3 (Tables 3 and 4). In the F. verticillioides isolates, FB production (all toxin production expressed in micrograms per gram of mycelial dry weight in this paper) was 5.76-2015.19 µg/g, with an average of 344.81 µg/g. Among these, the production of toxin FB1 ranged from 3.17 to 1566.44 µg/g, with an average of 263.94 µg/g; the production of FB2 toxin was between 1.07 and 156.52 µg/g, with an average of 24.70 µg/g; and the production of FB3 toxin varied from 1.52 to 356.15 µg/g, with an average of 56.17 µg/g (Table 3).
Among the F. proliferatum isolates, the production of FB1, FB2, and FB3 was within the ranges of 97.74-11,100.99 µg/g, 16.01-1554.83 µg/g, and 9.11-381.4 µg/g, with the corresponding averages of 3632.88, 402.31, and 177.78 µg/g, respectively (Table 4). Tables 3 and 4 showed the significant differences in FB production among various isolates. For the F. verticillioides isolates, 42.6% of the isolates had <100.00 µg/g toxin production, whereas 68.4% of the F. proliferatum isolates exhibited >1000.00 µg/g toxin production, indicating that the toxigenicity of F. proliferatum in these areas was higher than that of F. verticillioides (Table 5).   The Tri13 gene-specific primer Tri13P1/Tri13P2 was used to conduct the PCR amplification of 19 members of the FGSC in Figure 5. A single 859 bp fragment was stably amplified in 16 F. meridionale isolates and three F. asiaticum isolates, indicating that all 19 members of the FGSC were of the NIV chemotype.

Analysis of FBs
Mycotoxin assays showed that all F. verticillioides and F. proliferatum isolates could produce toxins FB1, FB2, and FB3. Except for F. verticillioides isolates D61-1 and D63, the other isolates exhibited a significantly higher FB1 yield than that of FB2 and FB3 (Tables 3 and 4). In the F. verticillioides isolates, FB production (all toxin production expressed in micrograms per gram of mycelial dry weight in this paper) was 5.76-2015.19 µg/g, with an average of 344.81 µg/g. Among these, the production of toxin FB1 ranged from 3.17 to 1566.44 µg/g, with an average of 263.94 µg/g; the production of FB2 toxin was between 1.07 and 156.52 µg/g, with an average of 24.70 µg/g; and the production of FB3 toxin varied from 1.52 to 356.15 µg/g, with an average of 56.17 µg/g (Table 3).

Analysis of FBs
Mycotoxin assays showed that all F. verticillioides and F. proliferatum isolates could produce toxins FB 1 , FB 2 , and FB 3 . Except for F. verticillioides isolates D61-1 and D63, the other isolates exhibited a significantly higher FB 1 yield than that of FB 2 and FB 3 (Tables 3 and 4). In the F. verticillioides isolates, FB production (all toxin production expressed in micrograms per gram of mycelial dry weight in this paper) was 5.76-2015.19 µg/g, with an average of 344.81 µg/g. Among these, the production of toxin FB 1 ranged from 3.17 to 1566.44 µg/g, with an average of 263.94 µg/g; the production of FB 2 toxin was between 1.07 and 156.52 µg/g, with an average of 24.70 µg/g; and the production of FB 3 toxin varied from 1.52 to 356.15 µg/g, with an average of 56.17 µg/g (Table 3).
Among the F. proliferatum isolates, the production of FB 1 , FB 2 , and FB 3 was within the ranges of 97.74-11,100.99 µg/g, 16.01-1554.83 µg/g, and 9.11-381.4 µg/g, with the corresponding averages of 3632.88, 402.31, and 177.78 µg/g, respectively (Table 4). Tables 3 and 4 showed the significant differences in FB production among various isolates. For the F. verticillioides isolates, 42.6% of the isolates had <100.00 µg/g toxin production, whereas 68.4% of the F. proliferatum isolates exhibited >1000.00 µg/g toxin production, indicating that the toxigenicity of F. proliferatum in these areas was higher than that of F. verticillioides (Table 5).

Determination of the Toxigenicity of FGSC
The DON, ZEN, and NIV assay based on UHPLC-MS/MS were in agreement with the molecular detection of the Tri13 gene in members of the FGSC. The toxin assay showed that none of the three F. asiaticum strains produced the toxins NIV and ZEN, but only the CP5 isolate produced DON with 4.50 µg/g of dry hyphal weight, suggesting that the F. asiaticum produces almost no mycotoxins in Chongqing (Table 6). NIV was detected in 11 out of the 16 F. meridionale isolates, which showed a mycotoxin-producing range of 17.40-2597.34 µg/g of dry hyphal weight. ZEN was detected in 11 isolates, and toxin production ranged from 8.35 to 78.57 µg/g. DON was detected in four isolates, i.e., CP1, CP4, D14, and D46, with the toxin production range of 3.38-33.41 µg/g. Isolates D38, D58-2, and D85-1 expressed 3-AcDON toxin, with the corresponding productions of 7.81, 5.83, and 3.10 µg/g, respectively. 15-AcDON was only detected in D38 isolate, with toxin production of 7.90 µg/g. The results show that F. meridionale mainly produces NIV, but weakly does ZEN, DON, 3-AcDON or 15-AcDON.

Discussion
Maize is an important food crop in China, and it is also a significant energy crop and industrial material. Numerous studies have shown that in most countries and regions, Fusarium spp. are the main causative pathogens for maize kernel rot. F. verticillioides and F. graminearum sensu stricto are the predominant species in Huang-Huai-Hai and northeast China [2,33]. However, our study indicated that besides F. verticillioides, F. proliferatum, and F. meridionale were also the predominant pathogens that caused maize ear rot in Chongqing areas, indicating the characteristic composition of pathogenic Fusarium species causing maize ear rot in these areas. These discrepancies may be caused by particular environmental conditions. Chongqing is located in the southwest maize growing area of China, and the area is hilly and mountainous, with the highest elevation reaching up to 2800 m. Different ecological zones are present, thereby forming the unique pathogen community. The warm, rainy, and humid weather conditions are suitable for infection, growth, and reproduction of Fusarium spp. in these areas. It is an important factor contributing to the serious maize ear rot. Besides, cropping system is probably also a major factor. In Chongqing areas, crop planting patterns usually incude corn monoculture, wheat and corn rotation, and rice and corn rotation, and so on. Undoubtedly, inoculum production increases with corn monoculture. In addition, F. verticillioides, F. proliferatum, FGSC, and F. oxysporum are also important pathogenic fungi in wheat and rice. Therefore, corn rotation with wheat or rice hardly reduces the prevalence of these fungi, even leading to the accumulation of the above Fusarium species in these areas.
FBs, DON, NIV, ZEA, and other mycotoxins are the major causes of toxin contamination by Fusarium species. However, both F. verticillioides and F. proliferatum produce FBs, the former can cause FB contamination mainly in maize, whereas the latter can cause toxin contamination in a variety of crops. Mycotoxin assays showed that all F. proliferatum and F. verticillioides isolates could produce toxins the FB 1 , FB 2 , and FB 3 , with FB 1 as the predominant mycotoxin. However, the average toxin production of F. proliferatum isolates was 12.22-fold higher than that of F. verticillioides, and hence, potential contamination with F. proliferatum should always be fully considered. In the present study, PDB liquid medium was used in culturing the Fusarium strains, and whether the toxin production of F. proliferatum was the highest in vivo will be investigated in our future study. In the field or in storage, mycotoxin contamination from maize ears and kernels is heavily influenced by multiple factors, such as pathogens, environmental conditions (temperature, humidity, pH, and lighting), host resistance, and so on. Therefore, mycotoxin production from these isolates in the laboratory primarily represents their toxigenic potential.
In this study, the results of the Tri13P1/Tri13P2 specific primer assay and toxin detection indicated that 16 F. meridionale and 3 F. asiaticum isolates were of the NIV chemotype, thereby representing geographical characteristics. Kuppler et al. reported that among 63 FGSC strains from Germany, only two belonged to NIV type, and the remaining were of the DON type [34]. In France, only 14.6% of the members of the FGSC were of the NIV type, and the remaining 85.4% belonged to the 15-ADON type [35]. In Brazil, among the 92 strains of the FGSC isolated from barley, 61 (66.3%), 4 (4.4%) and 27 (29.3%) belonged to 15-ADON, 3-ADON, and NIV chemotype, respectively [31].
In China, studies on the population structure and toxigenicity of the FGSC have mostly focused on wheat and rice, whereas studies on maize are very limited. Shen et al. found that among 530 FGSC strains isolated from the main winter wheat-producing areas of China, 182 F. graminearum sensu stricto strains were mainly distributed in North China, and 348 F. asiaticum strains were mainly distributed in South China. Among these, a high isolation frequency of the 15-ADON strains was observed in North China, and the NIV and 3-ADON strains were more common in South China [28]. Similar studies also have proven that the NIV and 3-ADON strains are mostly distributed in warmer regions [36,37]. Our findings that FGSC isolates causing maize ear rot in Chongqing areas were of the NIV genotype also support the above conclusions.
Studies have shown that various mycotoxins have different toxicological properties. Compared to the toxin DON, NIV poses a more serious threat to humans and animals health, which requires a more stringent limit of daily intake [38,39]. NIV was detected in 57.9% of the FGSC, and three isolates had a relatively high NIV-producing capacity (>1000 µg/g of dry hyphal weight) were F. meridionale. Compared to NIV, these isolates produced a small amount of ZEA and DON toxins. These findings indicate that NIV is likely to be the predominant trichothecene contaminant in Chongqing areas.
In the present study, we found that several F. meridionale isolates, such as D58-2, D66, D73, D85-1, and D91-2, and F. asiaticum strains D99, CP5, and D57-2 harbored the gene responsible for NIV-production but did not secrete NIV toxin. These phenotypes could be explained by a mutation in the NIV producing gene sequences, or by altered expression of the NIV producing genes. Also, the amount of NIV toxin produced by these isolates is probably beyond the detection limit of our assays. In addition, the lack of NIV in these isolates may also be due to the growth medium used.
Based on our study, various Fusarium species show distinct differences in their toxigenicity. Therefore, in Chongqing areas, the potential maize food and feed safety threat caused by F. proliferatum and F. meridionale is probably more serious than that by F. verticillioides, and F. asiaticum, respectively. However, maize germplasm and varieties are usually merely screened for resistance to ear rot caused by F. verticillioides and F. graminearum sensu stricto in China. Therefore, the risk of growing the selected "resistant" varieties remains. Although F. proliferatum is not the firstly major causal pathogen of ear and kernel rot, this species should also be included in germplasm screening for resistance and crop breeding for disease resistance, particularly in Chongqing areas. Also, maize ear and kernel rot caused by F. meridionale deserves attention. The maize germplasm resistant to F. meridionale should be selected for cultivation in these areas. In addition, F. proliferatum contamination may be utilized as an important indicator of the quality and safety of grains produced in these particular areas.

Sample Collection and Isolation and Identification of Pathogenic Fungi
A total of 103 maize ear or kernel samples (five symptomatic maize ears or 500 g of kernels for each sample) were collected from production fields at harvest in 103 towns of 34 counties in Chongqing and surrounding areas in 2014 and 2015 ( Figure 6 and Table S1). About 30 seeds collected from each sample were soaked in 20% sodium hypochlorite solution for 3 min, and rinsed with sterile water thrice. These seeds were dried with sterile filter paper and placed on a potato dextrose agar (PDA) (potato infusion 200 g, dextrose 20 g, agar 20 g, distilled water 1000 mL) plate for culture for 3 days at 25 • C. Hyphae from typical Fusarium colonies on PDA were transferred to a fresh poor-nutrient potato dextrose agar (half-PDA) (potato infusion 100 g, dextrose 20 g, agar 20 g, distilled water 1000 mL) plate and the culture was grown for 5 to 7 days. Upon emergence of conidia, a single spore was isolated on PDA by the plate dilution method. Finally, the single spore was transplanted onto a PDA plate to culture single-spore isolates. The other Fusarium species that could not be determined by species-specific PCR were analyzed using the translation elongation factor (TEF)-1α gene sequences. TEF-F/R: 5′-ATGGGTAAGGARGACAAGAC-3′/5′-GGARGTACCAGTSATCATGTT-3′ [45]. Each PCR reaction system (50.0 µL) consisted of a DNA template (5.0 µL), upstream and downstream primers (2.5 µL each), 2× Taq PCR Master Mix (25.0 µL), and ddH2O (15.0 µL). Reactions were performed using a

Identification of Pathogenic Fungi
The morphological identification of fungal cultures was conducted based on general characteristics and conidial morphology [40]. In order to confirm the morphological identification, genomic DNA was extracted from collected aerial mycelia using the Rapid Fungi Genomic DNA Isolation Kit (SK8230, Sangon Biotech, Shanghai, China) according to the manufacturer's instruction and were validated by species-specific polymerase chain reaction (PCR) for identification (Table 7). Reactions were performed using a GeneAmp PCR System 9700 thermal cycler (ABI, Norwalk, CT, USA) programmed for 94 • C for 5 min; followed by 35 cycles of 95 • C for 50 s, 58-60 • C for 50 s, and 72 • C for 60 s; and a final extension at 72 • C for 10 min. Electrophoretic analysis of the PCR-amplified products was performed on a 1% agarose gel.
The other Fusarium species that could not be determined by species-specific PCR were analyzed using the translation elongation factor (TEF)-1α gene sequences. TEF-F/R: 5 -ATGGGTAAGGARG ACAAGAC-3 /5 -GGARGTACCAGTSATCATGTT-3 [45]. Each PCR reaction system (50.0 µL) consisted of a DNA template (5.0 µL), upstream and downstream primers (2.5 µL each), 2× Taq PCR Master Mix (25.0 µL), and ddH 2 O (15.0 µL). Reactions were performed using a GeneAmp PCR System 9700 thermal cycler programmed for 94 • C for 5 min; followed by 35 cycles of 95 • C 50 s, 53 • C for 50 s, and 72 • C 60 s; and a final extension at 72 • C for 10 min. The amplified PCR products were bi-directionally sequenced by Sangon Biotech, and the sequences were compared with Fusarium sequences in the Fusarium Center's database at Penn State. Using MEGA 5.0 software (ASU, Phoenix, AZ, USA, 2011), a phylogenetic tree was constructed via Test Maximum Likelihood Tree clustering method based on the TEF-1α gene sequences, and the bootstrap analysis was performed with 1000 replicates for statistical support of branches.

Detection of Mycotoxin Production
Equivalent Fusarium spp. were cut from half-PDA and placed in a sterilized conical flask containing 150 mL of potato dextrose broth (PDB). Each fungal isolate was cultured in triplicate, and the sterile liquid medium with no inoculant was used as control. F. verticillioides and F. proliferatum were grown in a 15 day static culture in PDB with pH 8.0 at 25 • C, and FGSC was grown in a 15 day shaking culture (100g) in PDB with pH 3.0 at 25 • C [32]. The inoculated culture medium was filtered with a Whatman GF/A glass fiber filter paper, the filtrate was then stored at −80 • C or sterilized under high pressure, and the hyphae were collected, dried, and weighed.
For all Fusarium isolates, 20 mL of the filtrate was collected and used in the toxin assays. Immunoaffinity column purification and HPLC analysis of F. verticillioides and F. proliferatum were performed to measure FB production [48]. The eluent was dried with nitrogen and dissolved in 1.5 mL of 80% methanol solution. FBs were tested using a C 18 reverse-phase liquid chromatography/fluorescence detector after O-phthaldialdehyde (OPA) derivation and quantified via an external standard method. DON, ZEN, and NIV production of FGSC was determined using UHPLC-MS/MS [49]. Samples were extracted with an 80% acetonitrile water solution, purified via a multifunction decontamination column, isolated via a Waters ACQUITY UPLC BEH C18 chromatographic column, tested by multireaction ion monitoring of quadrupole mass spectrometry, and quantified by an external standard method. Statistical analysis was performed with SPSS 10.0 software (SPSS Inc., Chicago, IL, USA, 2007).

Conflicts of Interest:
The authors declare no conflict of interest.

Abbreviations
The following abbreviations are used in this manuscript: