Molecular Identification and Mycotoxin Production by Alternaria Species Occurring on Durum Wheat, Showing Black Point Symptoms.

Black point is a fungal disease of wheat, mainly associated with mycotoxigenic Alternaria species. Affected wheat kernels are characterized by dark brown discolouration of the embryo region and reduction of grain quality. Potential risk is the possible accumulation of Alternaria mycotoxins, alternariol (AOH), alternariol-monomethyl ether (AME), tenuazonic acid (TA), and altenuene (ALT), provided by haemato-toxic, genotoxic, and mutagenic activities. One hundred and twenty durum wheat samples belonging to 30 different genotypes grown in Bologna and Modena areas, in Italy, showing black point symptoms, were analyzed for Alternaria species and their mycotoxin contamination. Alternariol was selected as an indicator of the capability of the Alternaria species to produce mycotoxin in vivo in field conditions. The data showed that Alternaria species occurred in 118 out of 120 wheat kernels samples, with the incidence of infected kernels ranging between 1% and 26%. Moreover, AOH was detected by using a HPLC with a diode array detector (LC-DAD) in 98 out of 120 samples with values ranging between 24 and 262 µg Kg-1. Ninety-two Alternaria representative strains, previously identified morphologically, were identified at species/section level using gene sequencing, and therefore were analyzed for their mycotoxin profiles. Eighty-four strains, phylogenetically grouped in the Alternaria section, produced AOH, AME, and TA with values up to 8064, 14,341, and 3683 µg g-1, respectively, analyzed by using a LC-DAD. On the other hand, eight Alternaria strains, included in Infectoriae Section, showed a very low or no capability to produce mycotoxins.

evaluating not only the toxigenic risk associated with infected kernels, but also to understand if such metabolites are produced in the field during the infection process.
Although the genus Alternaria is considered an important wheat colonizer able to produce the abovementioned mycotoxins in vivo, the occurrence of Alternaria mycotoxins in cereals has been largely neglected for long time. This is probably due also to the lack of legislation on Alternaria mycotoxin occurrence in commodities. On the other hand, very poor studies have been devoted to evaluate the susceptibility of wheat to black point disease, have for many years considered only a problem for the quality and nutritive decline of wheat products, and not possibly related to safety. Therefore, the present work mainly aimed to: (a) Identify the main fungal species infecting kernel wheat samples showing black point symptoms, collected in Italy; (b) Detect the presence of AOH in wheat kernels in the field; (c) Study the genetic variability of the Alternaria species isolated, by sequencing the genes coding for allergen alt a1 (alt-a1), glyceraldeyde-3-phosphate dehydrogenase (gpd) and translation elongation factor 1-α (tef ), by following a multi-locus gene sequence approach; (d) Analyze the mycotoxin profile of selected strains of the Alternaria species identified.

Black Point Disease Symptoms on Wheat Samples
At harvest, black point disease symptoms were evaluated on wheat plants, in both Bologna and Modena fields, according to an empirical scale with ten classes of severity (0 = absence of infections, 9 = extensive symptoms of infection). Data on black point symptoms are reported in Tables 1 and 2.  In the Bologna field, with the exception of the G17 genotype, exposed to fungicide treatment, all theses showed black point symptoms. Black point symptoms were more severe on plants not exposed to fungicide treatment. Indeed, comparing the same genotypes, treated or untreated with fungicide, black point symptoms were more evident on untreated plants (Table 1). On untreated plants, symptoms ranged between 0.5 (G5 genotype) and 8 (G9 genotype), with a mean value of 3.5. Treated plants showed symptoms ranging between 0 (G17) and 4 (G1 and G16 genotypes), with a mean value of 1.9. Data on commercial varieties are summarized also in Table 1 and confirmed that fungicide treatment reduced drastically black point symptoms.
Data from the Modena field are summarized in Table 2 and confirmed that the highest black point symptoms were recorded in untreated plots, although they were lower than in the Bologna field.

Alternariol Contamination of Wheat Samples
Data on contamination of kernels by AOH are reported in Table 1. In the Bologna field, 88% of the samples were contaminated by AOH. Three samples exposed to fungicide spraying (G5, G9, G23) and four untreated samples (G4, G8, G11, Iride) were not contaminated from AOH. In the plots exposed to fungicide treatment the contamination values ranged from 24 µg kg −1 (G3) to 92 µg kg −1 (Saragolla), with a mean value of 39 µg kg −1 . In the untreated samples the contamination values ranged between 40 (G22) and 148 (G2) µg kg −1 , with a mean value of 84 µg kg −1 . With the exception of the samples G4, G8 and Iride, in which chemical treatment did not influence AOH contamination, the treated plots showed a lower contamination of AOH than untreated plots (Table 1).
In the Modena field, AOH contamination was higher than in the Bologna field. In treated plots the AOH contamination ranged between 64 (G9) and 262 (G5) µg kg −1 , with mean value of 109 µg kg −1 ; in an untreated plot the level of AOH contamination ranged between 68 (G20) and 230 (G6) µg kg −1 , with a mean value 66 µg kg −1 . In the Modena trial, a strict correlation between AOH contamination and chemical treatment was not observed. In 17 out of 30 untreated samples the AOH contamination was lower than the same genotypes or commercial cultivar spayed with chemicals ( Table 2).

Detection of Fungal Infection
Fungal infection was detected in all wheat samples, with values ranging between 21% and 83% in the Bologna field and between 71% and 98% in the Modena field (data not shown). Un-treated samples were slightly more infected than treated samples. Particular attention was focused on Alternaria and Fusarium genera. Among the fungal colonies morphologically identified as belonging to Alternaria and Fusarium genera, 382 monoconidial representative isolates of Alternaria spp. and 177 of Fusarium spp. were recovered from the infected kernels.
Infection by Alternaria species occurred in 118 out of 120 wheat samples, ranging from 1% to 26% in the Bologna field (Table 1), with mean value of 8% and 10% for treated and untreated theses, respectively; the incidence of infected kernels ranging between 2% and 20% were detected in Modena field (Table 2), with mean value of 8% and 12% in treated and untreated theses, respectively. The infection rate at the Alternaria species level was also evaluated.
In 88 out of 120 wheat samples, strains belonging to the Fusarium genus were also detected; however, Fusarium spp incidence of infected kernels was very low, up to 7%, with mean values in treated and untreated plots of 1.2% and 2.3%, respectively (data not shown).
Total fungal infection and therefore Alternaria infection rate was lower in treated than untreated theses, according with the black-point symptoms assessed in fields during the field inspections (Table 1;  Table 2). However, between the symptoms of black point disease and average of Alternaria species infection, a strict correlation was not observed.

Morphological Characterization
All the 382 mono-conidial strains, representative of the whole population found on the 120 samples, grew very quickly on potato dextrose agar (PDA) and potato-carrot agar (PCA), usually covering the whole surface of the Petri plates, and showing abundant sporulation. Conidia and conidiophore branch morphologies were used to identify the Alternaria species, according to Simmons [21]. Almost all 382 isolates were related to four morphospecies: A. alternata, A. tenuissima, A. arborescens, A. infectoria. In detail, 125 isolates were identified as A. alternata, 152 were identified as A. tenuissima, 72 were identified as A. arborescens and 33 were identified as A. infectoria.

PCR Amplification and Phylogenetic Analysis
All the 92 selected Alternaria strains gave PCR products of the expected size for the alt-a1 and tef genes tested. In all Alternaria strains, morphologically identified as A. infectoria, a detection of 55 nt in the first part of the gpd region sequence amplified, was observed.
Fragment of tef gene showed a lower variability with few polymorphic regions. The gene showing the higher degree of variability was alt-a1. In particular, the higher variability was observed in all the alt-a1 sequences where a 6 nt deletion in A. infectoria group compared to A. alternata, A. tenuissima and A. arborescens groups was observed. Moreover, the final part of the resulting fragment was highly polymorphic.
The analysis of the combined sequences of the three genes resulting in a Maximum Parsimony tree is shown in Figure 1. The phylogenetic tree, obtained with Mega5 allowed us to define two well-separated clades, corresponding to Alternaria section (clade A) and Infectoriae section (clade B). Clade B grouped all the Alternaria reference strains belonging to Infectoriae section and eight out of 92 strains tested (Figure 1). A high variability was observed; however, seven out of eight strains of this clade and A. ventricosa reference strain, formed a well-supported (99 bootstrap value) group, that differed from the A. triticina reference strain. Among Alternaria strains grouped in this clade, strain ITEM17335 was highly similar to A. rosae E.G.S. 41-130 reference strain and was well separated by Infectoriae section.
In sub-clade A2, strains shared a very high level of homology and therefore it was not possible to distinguish between A. alternata and A. tenuissima species, in terms of morphological definitions. On the contrary, a great variability was observed in sub-clade A1 which grouped A. arborescens reference stains and 30 of the our Alternaria strains (Figure 1).

Design of Infectoriae Section Specific PCR Primers
Based on the deletion of 55 nucleotides observed in the first part of the obtained gpd sequences of Alternaria strains belonging to Infectoriae section, specific primers pair was designed by using Primer3 software [34]. In detail, forward primer (Ainf29Fw-5 CGTCTTCCGCAATGCTATCG3 ) overlapped the deleted region ( Figure 2) and reverse primer (Ainf277Rev-5 ACCTTGATCTCGCCCTTGAA3 ) was designed in a region common to all Alternaria strains. PCR conditions were optimized to amplify the specific fragment of 249 nt, as follows: an initial stage at 95 • C for 2 minutes; 30 cycles each consisting of a step at 95 • C for 30 sec, 58 • C for 30 sec and 72 • C for 30 sec, and a final stage at 72 • C for 7 minutes. Based on the deletion of 55 nucleotides observed in the first part of the obtained gpd sequences of Alternaria strains belonging to Infectoriae section, specific primers pair was designed by using Primer3 software [34]. In detail, forward primer (Ainf29Fw -5′CGTCTTCCGCAATGCTATCG3′) overlapped the deleted region ( Figure 2) and reverse primer (Ainf277Rev -5′ACCTTGATCTCGCCCTTGAA3′) was designed in a region common to all Alternaria strains. PCR conditions were optimized to amplify the specific fragment of 249 nt, as follows: an initial stage at 95 °C for 2 minutes; 30 cycles each consisting of a step at 95 °C for 30 sec, 58 °C for 30 sec and 72 °C for 30 sec, and a final stage at 72 °C for 7 minutes.

Mycotoxins Production of Alternaria Strains
Mycotoxins, with the exception of the A. infectoria ITEM 17344 strain, were detected in all rice cultures, as reported in Table 3. In all the strains, AOH and AME production was very variable, ranging between 1 and 8064 mg kg -1 and between 1 and 14,341 mg kg -1 for AOH and AME, respectively. In general, as shown in Table 3, even considering the species identification obtained by phylogenetic analysis, mycotoxin production of both AOH and AME was very variable. Mean values of AOH and AME were also reported in Table 3 for each phylogenetic clade or sub-clade. Alternariol and AME production of strains belonging to sub-clade A2, showed mean values of 774 and 2099 mg kg -1 , respectively. Similar mean values were reported for sub-clade A1 (884 and 1449 mg kg -1 ) and sub-clade A3 (651 and 1773 mg kg -1 ).
Only the strains belonging to clade B, assigned to A. infectoria species group, were low or no producers of toxins. The quantity of mycotoxins, with the exception of the strain ITEM17356 (AOH 223; AME 151 mg kg -1 ). ranged between 0 and 68 mg kg -1 (mean value 16) for AOH and between 1 and 64 mg kg -1 (mean value 15) for AME.
With regard to ALT production, only 2 out of 73 strains tested, A. arborescens ITEM 17230 (2.5 mg kg -1 ) and A. alternata ITEM 17244 (25 mg kg -1 ), were able to synthetize low content of this mycotoxin.
Tenuazonic acid was detected in 50 out of 73 stains with values ranging between 23 mg kg -1 (A.infectoria ITEM 17372 strain) and 3684 mg kg -1 (A. mali ITEM17349 strain). In particular, the highest TA amount was detected among the strains grouped in the sub-clade A3. All the strains belonging to A3 sub-clade showed a great capability to produce the TA with values ranging between 736 mg kg -1 and 3684 mg kg -1 (mean value 1733 mg kg -1 ). Among the strains identified as A. alternaria and A tenuissima, 8 strains were not able to produce TA and 27 out of 35 strains were high TA producers, with values ranging between 183 mg kg -1 and 2388 mg kg -1 (mean value 723 mg kg -1 ). Alternaria arborescens strains produced TA with values ranging between 157 and 3043 mg kg -1 (mean value of 730 mg kg -1 ); however 9 out of 22 strains were not a TA producer. With regard to strains grouped with A. infectoria reference strain, except the strain ITEM 17372 (TA content of 23 mg kg -1 ), all strains were not producers.

Infectoriae Section
Alternaria Section

Mycotoxins Production of Alternaria Strains
Mycotoxins, with the exception of the A. infectoria ITEM 17344 strain, were detected in all rice cultures, as reported in Table 3. In all the strains, AOH and AME production was very variable, ranging between 1 and 8064 mg kg −1 and between 1 and 14,341 mg kg −1 for AOH and AME, respectively. In general, as shown in Table 3, even considering the species identification obtained by phylogenetic analysis, mycotoxin production of both AOH and AME was very variable. Mean values of AOH and AME were also reported in Table 3 for each phylogenetic clade or sub-clade. Alternariol and AME production of strains belonging to sub-clade A2, showed mean values of 774 and 2099 mg kg −1 , respectively. Similar mean values were reported for sub-clade A1 (884 and 1449 mg kg −1 ) and sub-clade A3 (651 and 1773 mg kg −1 ).
Only the strains belonging to clade B, assigned to A. infectoria species group, were low or no producers of toxins. The quantity of mycotoxins, with the exception of the strain ITEM17356 (AOH 223; AME 151 mg kg −1 ). ranged between 0 and 68 mg kg −1 (mean value 16) for AOH and between 1 and 64 mg kg −1 (mean value 15) for AME.
With regard to ALT production, only 2 out of 73 strains tested, A. arborescens ITEM 17230 (2.5 mg kg −1 ) and A. alternata ITEM 17244 (25 mg kg −1 ), were able to synthetize low content of this mycotoxin.
Tenuazonic acid was detected in 50 out of 73 stains with values ranging between 23 mg kg −1 (A. infectoria ITEM 17372 strain) and 3684 mg kg −1 (A. mali ITEM17349 strain). In particular, the highest TA amount was detected among the strains grouped in the sub-clade A3. All the strains belonging to A3 sub-clade showed a great capability to produce the TA with values ranging between 736 mg kg −1 and 3684 mg kg −1 (mean value 1733 mg kg −1 ). Among the strains identified as A. alternaria and A tenuissima, 8 strains were not able to produce TA and 27 out of 35 strains were high TA producers, with values ranging between 183 mg kg −1 and 2388 mg kg −1 (mean value 723 mg kg −1 ). Alternaria arborescens strains produced TA with values ranging between 157 and 3043 mg kg −1 (mean value of 730 mg kg −1 ); however 9 out of 22 strains were not a TA producer. With regard to strains grouped with A. infectoria reference strain, except the strain ITEM 17372 (TA content of 23 mg kg −1 ), all strains were not producers.

Discussion
In the present study, AOH contamination was detected on 82% of wheat samples with values ranging between 24 and 262 µg Kg −1 . These data show that the occurrence of AOH in kernels collected in the field was extended and maybe it was related to the colonization process of the Alternaria species, as shown by Wenderoth et al. [33] in apple, citrus and tomato for Alternaria alternata. However, if AOH is a virulence and colonization factor for Alternaria infection on wheat, this needs to be confirmed. Moreover, the toxigenic risk has to be taken in account, since the AOH contamination of samples was high.
Fungicide treatments showed that Alternaria was less influenced by the fungicide application compared to total fungal infection. The reduction of fungicide efficacy against Alternaria species was previously reported by Avenot et al. [35] as a consequence of resistant strains occurring in the field. Further studies could be useful to explain the different response of fungicides against the main fungal species recovered on wheat.
Alternaria incidence of infected kernels, was rather lower than expected, in both experimental fields, if compared to the symptoms of black point disease. Even if black point symptoms, detected in the field, ranged from low to moderate visible infection (mean value of severity scale = 2.5), a low correlation was observed between disease symptoms and Alternaria infection.
This divergent result could be explained by the implication of other fungal species in the black point disease, or by abiotic damages associated with the dark brown discolouration of the wheat kernels.
The lack of congruence in Alternaria infection, disease symptoms and mycotoxin accumulation in wheat samples could be also linked to a different capability of the species to colonize plant tissue or to produce mycotoxins in field conditions. In addition, plant defence mechanisms, host and different pedo-climatic conditions could interfere in fungal colonization. Indeed, mycotoxin production by Alternaria strains revealed that almost all Alternaria strains tested showed a great capability to produce AOH and AME when grown in vivo on autoclaved rice, in which no mechanical barriers occur and many of the grain defences, that are thermos-labile, have been suppressed [36]. Although small-spored Alternaria species are of high economic relevance, they are poorly investigated among pathogenic fungi, mainly due to the difficulties of differentiating between species [37]. The morphological identification of Alternaria population carried out in the present study, performed according to Simmons [21], lead to assign all the strains to four main Alternaria species: A. tenuissima, A. alternata, A. arborescens and A. infectoria. These species are widely reported as the most occurring Alternaria species on wheat [8,12,14]. However, for a more accurate identification, we carried out a poly-phasic approach, using also molecular and chemical techniques [38,39].
Through our phylogenetic analyses A. alternata and A. tenuissima morpho-species were not distinguishable. This result is in accordance with several studies aimed to identify Alternaria species by using different molecular tools, such as AFLP [45] and RAPD-PCR [46]. A wide investigation on the Alternaria genus by phylogenetic analysis was carried out by Woudenberg et al. [16]. Although eleven genes sequences were used, A. alternata and A. tenuissima were yet not resolved, suggesting validity of the hypothesis that considers these two morpho-species as the same phylogenetic species [16]. On the other hand, in the present study, it was possible to distinguish A. arborescens strains from the A. alternata species group, all belonging to Alternaria section as defined by the recent taxonomic rearrangement of Alternaria genus [16]. At the same section we also assigned the group, named in this work as sub-clade A3, including strains with high homology with reference strains of A. citriarbusti, A. radhina and A. mali.
Our results indicate that population structure of Alternaria species associated with black point disease, is prevalently represented by A. tenuissima and A. alternata (about 50% of total strains tested), as observed in previous studies [7,9]. Moreover, A. arborescens was also identified in wheat in accordance with other recent studies in Argentina [18,47], Tunisia [48], Germany, and Russia [49].
Finally, in this work, a small number of A. infectoria strains were also isolated from wheat, confirming a previous report from Italy [7], and other reports from Norway, Argentina, Germany and Russia [23,49,50].
Thus, Alternaria species composition on wheat is evolving, influenced by agricultural practices and environmental conditions. Moreover, great concern in the scientific community exists, because of the ability of Alternaria species to infect a huge number of field and processed wheat products and their ability to produce a range of mycotoxins, with AOH and AME being the most frequent and dangerous metabolites [12,51].
In our study, Alternaria strains examined showed a very variable amount of AOH and AME production, with mean values of about 800 and 1800 mg kg −1 for AOH and AME, respectively, except for the group (eight strains) assigned to the A. infectoria section. These strains showed very low toxin production abilities, confirming that A. infectoria strains are weak or no mycotoxin producers, as previously reported [8,45,49]. Thus, while A. alternata, A. tenuissima and A. arborescens, phylogenetically clustered in the same Alternaria section, and represent a serious risk for human heath since almost all strains produce AOH and AME [8], A. infectoria is not a particular threat for mycotoxin risk in wheat, due to the lack of mycotoxin production. For this reason, diagnostic tools based on the molecular detection of A. infectoria strains can allow a rapid evaluation of Alternaria mycotoxin risk assessment in wheat samples.
Consumption of wheat invaded by Alternaria does not necessarily imply mycotoxin presence in the grains but it indicates a potential risk for contamination. Mycotoxins of Alternaria have not yet received the same attention as mycotoxins produced by other fungal genera [49] although they are associated with many different diseases in humans, animals and plants [8,52]. However, nowadays consumers demand food and feed productions with a high qualitative standards to guarantee human and animal health. In this respect, and considering the continuously changing species composition, a future regulation of Alternaria mycotoxins could be expected, making more urgent, deeper and wider investigations on Alternaria species and related mycotoxins.

Origin of Wheat Samples
The 120 durum wheat samples analyzed in this work, were collected from two experimental fields, "San Lazzaro di Savena" (Bologna) and "Castelfranco Emilia" (Modena), located in the Emilia Romagna region, in Northern Italy. The susceptibility to black point disease of 24 durum wheat genotypes obtained with cross-breeding activities, was evaluated and compared with 6 commercial cultivars. In both fields, each genotype was grown in two replicates by using a randomized block design. In particular, one replicate was grown under high-input (fungicide application) and one replicate under low-input (no fungicide application) agronomic strategies. The treated plants were exposed to a first treatment with fungicide containing a mixture of Trifloxystrobin and Cyproconazole at the dose of 0.5 litre per hectare (L ha −1 ), at raising growth stage, and a second treatment with a formulate containing Prothioconazole, at the dose of 0.8 L ha −1 , at earing stage.

Evaluation of Black Point Symptoms
At kernels maturity stage, black point symptoms were evaluated according to an empirical scale with ten classes of severity (0 = absence of infections, 9 = extensive symptoms of infection).

Fungal Isolation and Growth Conditions
Representative kernels of the wheat samples (100 kernels), randomly collected at harvest time, were stored at 4 • C and then tested to detect the fungal incidence of infected kernels.
After superficial disinfection in 2% sodium hypochlorite solution for 2 min and two washing with distilled sterilized water for 1 min, 100 representative kernels for each sample, randomly collected, were plated (10 kernels/plate) on potato dextrose agar (PDA) added with 0.2 g L −1 of penthacloronitrobenzene (PCNB), 0.10 g L −1 of streptomycin sulphate salt and 0.05 g L −1 of neomycin. Incubation was performed at 25 ± 1 • C for 5 days under an alternating light/darkness cycle of 12 h photoperiod.
Total fungal incidence of infected kernels, and Alternaria, Fusarium, Aspergillus and Penicillium infection was calculated for each sample. Pure cultures of representative isolates belonging to Alternaria and Fusarium species, were isolated. To obtain mono-conidial isolates, conidia were spread at low density on water agar (WA: 20 g L −1 agar Oxoid n. 3) and singly collected after germination, using a dissection microscope, on PDA. Isolates were identified on the basis of colony and conidia morphology on PDA and potato-carrot agar (PCA: infusion from 20 g peeled and sliced white potatoes, 20 g carrot kept at 60 • C for 1 h; 15 g L −1 agar Oxoid n. 3), after 7 days of incubation at 25 ± 1 • C under an alternating light/darkness cycle of 12 h photoperiod, according to Simmons [35].

Molecular Characterization of Alternaria Strains
Ninety-two Alternaria strains, were selected for molecular characterization, among the set of strains isolated from wheat genotypes and commercial cultivars characterized.
The mycelium of three-days old fungal colonies, grown on cellophane disks overlaid on PDA plates, was collected by scraping and lyophilized.
Genomic DNA was extracted and purified from powdered lyophilized mycelia (10-15 mg) by using the "Wizard Magnetic DNA Purification System for Food" kit (Promega Corporation, Madison, WI), according to the manufacture's protocol. Quantity and integrity of DNA were checked at Thermo-Scientific Nanodrop (LabX, Midland, ON, Canada) and by comparison with a standard DNA (1 kb DNA Ladder, Fermentas GmbH, St. Leon-Rot, Germany) on 0.8% agarose gel after electrophoretic separation.
The PCR products were visualized with UV after electrophoretic separation in 1X TAE buffer, on 1.5% agarose gel.
Sequencing of the fragments, previously purified with the enzymatic mixture Exo/FastAP (Exonuclease I, FastAP thermosensitive alkaline phosphatase, Thermo Fisher Scientific-Waltham, MA, USA) was performed with Big Dye Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosystems, Foster City, CA, USA), according to the manufacturer's recommendations. Both strands were purified by filtration through Sephadex G-50 (5%) (Sigma-Aldrich, Aldrich, Saint Louis, MO, USA) and sequenced in "ABI PRISM 3730 Genetic Analyzer" (Applied Biosystems, Foster City, CA, USA). The FASTA sequences were obtained with BioNumerics software (Applied Maths, Kortrijk, Belgium). Phylogeny analysis was carried out including even the deposited gene sequences of the reference strains, as reported in Ramires et al. [7].
Phylogenetic trees of single and combined genes were generated by using Maximum Parsimony method and bootstrap analyses (1000 replicates, removing gaps) with MEGA5 [54].

Mycotoxin Extraction
All 120 wheat samples, collected from each plot and tested for fungal infection were analyzed for the natural occurrence of AOH contamination. Only AOH occurrence was evaluated in order to select an indicator of the capability of the Alternaria species isolated in this study of producing mycotoxin in vivo in the field. Chemical analyses on the 92 Alternaria strains, molecularly characterized, were also carried out to evaluate the capability of Alternaria strains to produce AOH, AME, ALT and TA. According to Li et al. [55], all the strains were inoculated on sterilized polished rice and grown for 21 days at 25 ± 1 • C in darkness. The cultures were then dried and powdered before mycotoxin extraction.
The method used for mycotoxins analysis is based on that described by Rubert et al. [56] with some modifications. The samples were finely ground with an Oster Classic grinder (220-240V, 50/60 Hz, 600W; Madrid, Spain). Five grams of each homogenized sample was weighed in a 50 mL plastic tube and 25 mL of methanol was added. The extraction was carried out using an Ultra Ika T18 basic Ultra-turrax (Staufen, Germany) for 3 min. The extract was centrifuged at 4000 rpm for 5 min at 5 • C. One mL of the surnatant was filtered through a 13mm/0.22 µm nylon filter and diluted before injection into high performance liquid chromatography associated with a diode array detector (LC-DAD). All the extractions were carried out in triplicate.

HPLC Analysis
Alternariol, AME, ALT and TA were determined using Merk HPLC with a diode array detector (LC-DAD) L-7455 (Merk, Darmstadt, Germany) at 256 nm and Hitachi Software Model D-7000 version 4.0 was used for data analysis. A Gemini C18 column (Phenomenex, Torrance, USA) 4.6 × 150 mm, 3 µm particle size was used as the stationary phase. The mobile phase consisted of two eluents, namely, eluent A (water with 50 µL/L trifluoroacetic acid) and eluent B (acetonitrile with 50 µL/L trifluoroacetic acid). A gradient program with a constant flow rate of 1 ml/min was used, starting with 90% A and 10% B, reaching 50% B after 15 min and 100% B after 20 min. 100% B was maintained for 1 min. Thereafter the gradient was returned to 10% B in 1 min and allowed to equilibrate for 3 min before the next analysis [57].