Characterization of mycotoxigenic Alternaria species isolated from the Tunisian halophyte Cakile maritima

Copyright: © 2020 A. Chalbi, B. Sghaier-Hammami, G. Meca, J.M. Quiles, C. Abdelly, C. Marangi, A.F. Logrieco, A. Moretti, M. Masiello. This is an open access, peer-reviewed article published by Firenze University Press (http://www.fupress.com/pm) and distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.


INTRODUCTION
Cakile maritima (sea rocket) is an extremophile C3 halophyte (Brassicaceae) which is widely distributed on sandy coasts (Clausing et al., 2000;Kadereit et al., 2005). This plant, together with Xanthium italicum and Eryngium maritimum, is typical coastal vegetation of several regions including the Black Sea coasts, the Mediterranean basin, the Atlantic coasts of North Africa and Europe, and the North Sea and Baltic Sea coasts (Clausing et al., 2000). These plants grow a few meters from shorelines, and are useful barriers defending coastal ecosystems. Sea rocket is an annual succulent plant, growing in geographical areas characterized by high salinity and low soil fertility (Barbour et al., 1970). It is tolerant to abiotic stress conditions such as high salinity, water stress and high temperature, which are all the characteristics of halophyte plants (Debez et al., 2012). Cakile maritima is a good candidate model plant for understanding botanical biochemical and physiological mechanisms.
In Tunisia, C. maritima, in addition to its role in ecosystem coastal preservation, is also considered an edible plant, used traditionally as a green vegetable for human and animal consumption. Dried root powder is also mixed with cereal flour to make bread (Zarrouk et al., 2003;Debez et al., 2004). The whole plant, harvested at flowering for its high content of iron, ascorbic acid and iodine, is traditionally used in antiscorbutic, diuretic and purgative folk remedies (Kubiak-Martens et al., 1999).
Alternaria is ubiquitous and abundant in the atmosphere and in soil, seeds, and agricultural commodities. This genus includes plant pathogenic and saprophytic species that may affect crops in the field or can cause harvest and postharvest decay of plant products (Logrieco et al., 2009;Patriarca, 2016;Somma et al., 2019). Alternaria spores have been related to human infections, such as rhinosinusitis, asthma, cutaneous and subcutaneous infections, and oculomycosis (Pastor and Gaurro, 2008;Canova et al., 2013;Hattab et al., 2019). Mycotoxins are major food contaminants affecting global food security, especially in low and middleincome countries . Several studies have confirmed the toxic effects of Alternaria metabolites for animals and humans, but these toxins are not regulated by legislation, and acceptable levels in food are not determined. Alternaria mycotoxins are arousing interest worldwide, and the European Food Safety Authority (EFSA) has provided scientific opinion on the risks for animal and public health related to the presence of Alternaria toxins in food and feed (Arcella et al., 2016).
Among the secondary metabolites produced by Alternaria species, there are both phytotoxins, that can play an important role in the plant pathogenesis processes, and mycotoxins, that can be harmful to humans and animals (Logrieco et al., 2009). Therefore, since C. maritima is used in traditional food, consumption of Alternaria toxin-contaminated plants can represent a serious toxicological risk for consumers (Wang et al., 1996;Ostry, 2008;Logrieco et al., 2009;Lou et al., 2013).
The dibenzopyrone derivatives alternariol (AOH), alternariolmonomethyl ether (AME) and altenuene (ALT), and the tetramic acid derivative tenuazoic acid (TA), are among the most important mycotoxins produced by Alternaria species. Alternariol and AME are usually found in combination, and have teratogenic and fetotoxic effects (Fehr et al., 2009). Genotoxic activity on human colon cancer cell lines has also been observed (Wang et al., 1996;Ostry, 2008;Fehr et al., 2009;Bensassi et al., 2012). A recent study also reported that AOH was an androgen agonist in in vitro assay (Stypula-Trebas et al., 2017).
Tenuazonic acid is a well-known mycotoxin and phytotoxin, produced primarily by A. alternata (Fr.) Keissl. (1912) and by other phytopathogenic Alternaria species, including A. japonica Yoshii (1941), detected on wheat in Tunisia (Bensassi et al., 2009) Wilt-shire (1933). In central and southern Africa, TA has been associated with human hematologic disorder known as "onalay", a thrombocytopenia (Steyn and Rabie, 1976). The presence of these mycotoxins in wheat has also been related to elevated levels of human esophageal cancer in China (Liu et al., 1992).
Morphological identification of Alternaria species is often difficult due to interspecific similarities and intraspecific polymorphisms (Simmons, 1990;Simmons and Roberts, 1993;Roberts et al., 2000;Pryor and Michailides, 2002;Serdani et al., 2002;Belisario et al., 2004;Hong et al., 2006). The high sensitivity of Alternaria to environmental factors may lead to ambiguities if taxonomy is based only on phenotype (Andrew et al., 2009). "Polyphasic" approaches using morphological, molecular and chemical techniques could therefore be useful for characterization of Alternaria species (Andersen et al., 2008;Brun et al., 2013).
Although several studies have focused on the occurrence of Alternaria species and Alternaria metabolites in food commodities, little knowledge is available on the Alternaria species living on extremophile plants. Santiago et al. (2018) suggested that Alternaria species that colonize extremophile plant environments show great adaptation to survival under extreme conditions, such as high temperature, low water and nutrient availability, osmotic stress, desiccation, and exposure to high levels of UV radiation, so these fungi are interesting examples of fungal diversity.
Based on this lack of knowledge, the aims of the present study were: i) to isolate and estimate the distribution of Alternaria species on C. maritima growing in different bioclimatic conditions, in Tunisia; ii) to identify Alternaria at species level using a polyphasic approach; iii) to analyze the profile of the main mycotoxins produced by the species isolated from C. maritima; and iv) to test their pathogenicity on C. maritima leaves.

Origin of the samples
Leaves, stems and seeds of spontaneous C. maritima plants, showing typical symptoms of Alternaria infections, were collected arbitrarily in Tunisia. Approximately 20 plants were collected per province, and five portions of each plant were used for fungal isolation. A total of 100 C. maritima samples were collected per province. Four provinces, from the north to the south of Tunisia and with two pedo-climatic conditions, were considered: Rades and Raouad, characterized by temperate winters and annual rainfall of 200-700 mm (semi-arid regions), and Sfax and Djerba, characterized by temperate winters and annual rainfall of 100-200 mm (arid regions).

Fungus isolation and growth conditions
After surface-disinfestation with 2% sodium hypochlorite solution for 2 min and two washings with sterilized distilled water for 1 min, symptomatic plant tissues were cut into small pieces (5 mm diam.) with a sterilized scalpel and dried on a sterile filter paper in a laminar flow cabinet. Single tissue pieces were then transferred into Petri dishes containing Potato Dextrose Agar (PDA) amended with 0.10 g L -1 streptomycin sulphate salt and 0.05 g L -1 neomycin. The dishes were incubated at 25 ± 1°C for 7 d under an alternating light/darkness cycle of 12 h photoperiod. From the 100 samples selected per province, 50 isolates were kept as monoconidium and homogeneous cultures. Two hundred Alternaria strains were collected from the four sample sites.

Morphological characterization
Following procedures of Simmons (2007), Alternaria strains were grown on PDA and on Potato Carrot Agar (PCA) to characterize their morphological traits. After 7 d of incubation at 25 ± 1°C, under an alternating light/ darkness cycle of 12 h photoperiod, conidium dimensions, colour, septation, conidiophore branching, and dimensions of catenulate conidia were determined using a light microscope (×40 magnification).

Molecular characterization and phylogenetic analysis of Alternaria strains
Based on morphological traits, 79 representative strains were selected for phylogenetic and chemical analyses. Thirty-eight Alternaria strains were from the semiarid region (Rades: 21 and Raouad: 17) and 41 were from the arid region (Djerba: 19 and Sfax: 22).
Mycelium of 3-d-old colonies, grown on sterilized cellophane disks overlaid on PDA plates, was collected by scraping and then frozen. Genomic DNA was extracted and purified from powdered lyophilized mycelia (each sample, 10-15 mg) using the Wizard Magnetic DNA Purification System for Food kit (Promega Corporation), according to the manufacture's protocol. Quantity and integrity of DNA were checked at Thermo-Scientific Nanodrop (LabX, Midland, Ontario, Canada), and by comparison with a standard 1 kb DNA Ladder (Fermentas GmbH) on 0.8% agarose gel, after electrophoretic separation.
Each PCR product was purified with the enzymatic mixture Exo/FastAP (Exonuclease I, FastAP thermosensitive alkaline phosphatase, Thermo Scientific) and then sequenced using the Big Dye Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosystems), according to the manufacture's recommendations. Both strands were purified by filtration through Sephadex G-50 (5%) (Sigma-Aldrich) and sequenced in an "ABI PRISM 3730 Genetic Analyzer" (Applied Biosystems). The FASTA sequences were obtained with BioNumerics software (Applied Maths). Phylogenetic trees of single and combined genes, following a multi-locus sequence approach, were generated using the Maximum Likelihood statistical method and bootstrap analyses (1000 replicates, removing gaps) with MEGA5 (Tamura et al., 2011).
Phylogeny analysis was carried out considering the gene sequences of the following reference strains: A. alternata E.G.S.34.016, A. tenuissima E.G.S.34.015, A. arborescens E.G.S.39.128, A. infectoria ex type E.G.S.27.193, The gene sequences of these reference strains were downloaded from the National Center for Biotechnology Information (NCBI) and "Alternaria Genomes Database" (AGD). Alternaria infectoria ex type E.G.S.27.193 was used as the outgroup in phylogenetic analyses.

Mycotoxin extraction and chemical profile analyses
Chemical analyses of the 79 Alternaria strains were carried out to evaluate their capabilities to produce AOH, AME, ALT and TA. Following the methods of Li, et al. (2001), all the strains were inoculated onto sterilized rice grain and grown for 21 d at 25 ± 1°C in darkness. The cultures were then dried and powdered before mycotoxin extraction.
Mycotoxin analyses were based on the methods of Rubert et al. (2012), with modifications. The samples were each finely ground with an Oster Classic grinder (220-240V, 50/60 Hz, 600W). Five grams of each homogenized sample were weighed in a 50 mL capacity plastic tube, and 25 mL of methanol was added. The extraction was carried out using an Ultra Ika T18 basic Ultra-turrax for 3 min. The extract was centrifuged at 4000 rpm for 5 min at 5°C. The supernatant (1 mL) was filtered through a 13 mm × 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.
Alternariol, AME, ALT, and TA were determined using a Merk HPLC with a diode array detector (LC-DAD) L-7455 (Merk), at 256 nm, and Hitachi Software Model D-7000 version 4.0 was used for data analysis. A Gemini C18 column (Phenomenex) 4.6 × 150 mm and 3 μm particle size was used as the stationary phase. The mobile phase consisted of two eluents, eluent A (water with 50 μL L -1 trifluoroacetic acid), and eluent B (acetonitrile with 50 μL L -1 trifluoroacetic acid). A gradient programme with a constant flow rate of 1 mL min -1 was used, starting with 90% A and 10% B, reaching 50% B after 15 min and 100% B after 20 min which was maintained for 1 min. The gradient was then returned to 10% B for 1 min and allowed to equilibrate for 3 min before the next analysis (Myresiotis et al., 2015). The limit of detection (LOD) for this method was 0.01 mg L -1 and the limit of quantification (LOQ) was 0.1 mg L -1 . The data of mycotoxin production were statistically processed using the Prism 5 software (La Jolla; www.graphpad.com).

Pathogenicity tests
A set of 41 representative strains, belonging to all phylogenetic groups previously characterized ( Figure  2), were tested for pathogenicity on C. maritima leaves. Twenty-four strains of the A. alternata group, eight of the A. arborescens group, five of the A. mali group, and four strains of unknown identity were tested. Seeds were sown into plastic pots (diam, 29 cm; height, 23 cm), filled with inert sand, which were then watered each day with distilled water until germination. The pots were maintained in a greenhouse under controlled conditions of 16 h light/8 h dark cycles, 25 ± 1ºC and 60% relative humidity. Resulting seedlings were irrigated each day for 2 weeks with the nutrient solution (Hewitt, 1996(Hewitt, ). 1966. At 6 weeks after sowing, leaves of the plants were cut and placed on glass slides in Petri dishes (one slide per dish) containing a moist filter paper. The leaves were then inoculated with Alternaria conidium suspension (10 8 conidia mL -1 ). Control plants were grown in similar condition without fungal inoculation. Separate leaves (six replicates per inoculation treatment) were inoculated. A 10 μL droplet of conidium suspension (10 8 conidia mL -1 ; 0.05% Tween 20) was placed in the centre of each leaf. Controls were inoculated by a droplet of sterile distilled water 0.05% Tween 20 solution. The Petri dishes containing leaves were then incubated under laboratory conditions (25 ± 3°C; 12h light/12 h dark photoperiod). Disease progress was evaluated at 7 d post-inoculation (dpi) (Taheri, 2019).
After the incubation period, lesion appearance and external and internal lesion diameters were assessed. The degree of pathogenicity of each inoculated strain was assessed by calculating the diameter of the lesion. Statistical analyses of data were performed using the statistical package Statistix8. A one-way ANOVA was used to test the hypothesis of Alternaria species differentiation with respect to the lesion diameter, and means were compared at P < 0.05.

Host symptoms and morphological characterization of Alternaria strains infecting Cakile maritima
Symptomatic plants of C. maritima were sampled from two different pedo-climatic conditions (arid and semi-arid). All collected samples (stems, leaves and seeds) showed blight symptoms typical of Alternaria infections (Figure 1). Infected plant tissues had necrotic flecks surrounded by chlorotic halos which coalesced into nearly circular or irregularly shaped lesions.
Conidium lengths of Alternaria strains ranged from 6 to 29 μm (greatest average = 17.45 μm). Conidium widths ranged from 3 to 10 μm (greatest average = 5.52 μm) (Supplementary Table S1). The numbers of transverse septa varied from 4 to 11 and longitudinal septa varied from 1 to 3. Different margin types, (smooth and rough surfaces) were observed and the colour of margins varied from brown to dark brown or blackish brown. Different forms of conidia were also observed, including ellipsoidal to long ellipsoidal, subspherical, base conical, rounded or obtuse, and apex rounded.
Based on morphological traits, all the strains were identified has A. alternata, A. arborescens E.G. Simmons or A. tenuissima. However, variations in culture characteristics such as margin types and colony colours were observed.

Molecular and phylogenetic analyses of Alternaria strains
Seventy-nine Alternaria strains were identified by sequencing informative fragments of the alt-a1, gpd and tef genes. All PCR reactions gave amplicons of the expected sizes of approx. 620 nucleotides for alt-a1, 510 for gpd and 590 for tef.
For sequence alignment of the gpd gene, a fragment of about 470 nucleotides was analyzed; however, in the A. brassicicola (ATCC96836) sequence, beginning at position 104 of the fragment considered, a short deletion of 27 nucleotides was observed. For tef and alt-a1 sequence alignments, common fragments of about 490 nt (tef ) and 460 nt (alt-a1) were considered. The tef gene showed low variability with few polymorphic sites, whereas the alt-a1 gene showed greater variability among Alternaria species.
To further determine the identity of the strains, phylogenetic analysis of the concatenated sequences of the three fragments was carried out. The phylogenetic tree, obtained with Mega5 software using the Maximum Likelihood method, allowed definition of four wellseparated clades, corresponding to Section Alternaria, Section Porri, Section Brassicicola, and Section Infectoriae, as determined using reference strains (Figure 2). The resolution of all the clades was supported by high bootstrap values. Alternaria brassicicola ATCC 96836 and A. infectoria E.G.S.27.193 did not cluster with any Alternaria strains isolated from C. maritima. In the Section Porri, only reference strains were clustered together ( Figure 2). All Alternaria strains clustered in Section Alternaria. Forty-seven strains showed homology close to 100% with A. alternata E.G.S. 34.016, A. tenuissima E.G.S.34.015, A. limoniasperae BMP2335, A. turkisafria BMP3436 (Figure 2).
Based on these data, four phylogenetic Alternaria groups can be discriminated. The A. alternata group was the most frequent (60% of strains), followed by A. arborescens (16%), and A. mali (15%). Seven strains (9%), phylogenetically closely-related to A. mali, were not identified at species level ( Figure 2).

Mycotoxin production profiles of Alternaria strains
Data of mycotoxin production are summarized in Table 1 and in Supplementary Table S2. Altenuene was produced by two A. alternata strains (39.2 mg kg -1 by ITEM17789 and 43.1 mg kg -1 by ITEM17855) and one A. mali strain (77.3 mg kg -1 by ITEM17788). The amounts of AME, AOH and TA were very variable. These were: for AOH, from 4.4 mg kg -1 (A. alternata ITEM17782) to 1,856.8 mg kg -1 (A. alternata ITEM17791); for AME, from 53.4 mg kg -1 (A. alternata ITEM17848 to 9,149.4 mg kg -1 (A. mali ITEM17835); and for TA, from 344.4 mg kg -1 (A. mali ITEM17849) to 19,837.3 mg kg -1 (A. alternata ITEM17808). Between the four clades of Alternaria species, qualitative and quantitative differences in mycotoxin production were examined. However, close correlation between phylogenetic clade and mycotoxin production was not detected.

Pathogenicity tests
Symptoms of tissue browning followed by necrosis were observed at 3 and 7 d after inoculation. Necrosis with yellowing occurred at 3 dpi which became more intense at 7 dpi. The control plants did not show disease symptoms (Figure 3).

DISCUSSION
This study has characterized Alternaria species occurring on the halophyte C. maritima from different bioclimatic areas in Tunisia. Alternaria species are major contaminants and pathogens of a wide range of crops, causing economic losses for producers, and health risks for consumers due to the accumulation of toxic metabolites (Da Cruz Cabrala et al., 2016).
Morphological characterization of the representative strains Alternaria examined in this study has shown high variability in conidium size, shape, dimensions and septation, confirming that the species identification based on morphological characters is unreliable. Furthermore, there was no correlation between specific morphological traits of the different strains and their bioclimatic region origins. There was also no correlation with the different plant tissues from which the strains were isolated. Similar results were observed by Chethana et al. (2018), who found different morphological traits of Alternaria strains isolated from onion did not depend from their geographical or plant tissue origins. To perform a more discriminating identification at species level, a representative set of the Alternaria strains was investigated by using a polyphasic approach, based also on molecular and chemical characterization.
All the representative strains were phylogenetically identified in Section Alternaria (Lawrence et al., 2013),  et al., 2015). The three genes (gpd, tef, alt-a1) used for multi-locus sequence analyses were selected for their informative responses in Alternaria, as suggested by Somma et al. (2019). All the strains grouped into four phylogenetic clades: 13 strains were identified as A. arborescens, although they showed a great variability; the largest clade included 47 strains, with a high level of homology, and was identified as A. Alternaria; 12 strains were identified as A. mali; and seven strains were grouped in a clade closely related to, but different from, A. mali. These seven strains were not assigned to any species. The different identified Alternaria species could not be related to particular plant tissues and/or bioclimatic conditions. Alternaria alternata and A. mali strains colonize all types of C. maritima tissues, and were isolated from the two climatic regions. The exception was all the strains of the unidentified clade, (Alternaria spp.; Figure 2), which were isolated only from the seeds of C. maritima plants collected in the arid region of Sfax. Also, the A. arborescens strains were not detected in the semi-arid region of Raouad. Similar results were reported by Bensassi et al. (2009;, who identified A. alternata as the most frequent Alternaria species occurring on wheat growing in all regions of Tunisia sampled, which were characterized by different environmental conditions. Schiro et al. (2018) also showed that conidium production of several Alternaria spp. on wheat plants was not correlated with specific microclimatic conditions and was uniform in the field in all regions sampled. Several authors (Andersen et al., 2008;Brun et al., 2013;Somma et al., 2019), have shown that accurate identification of Alternaria spp. requires the polyphasic approach, using morphology, phylogeny, and mycotoxin profiling. Cakile maritima is also often used for human and animal consumption. The parts of the plants used for nutrition could become potential natural sources of Alternaria mycotoxins, if contaminated by Alternaria species. However, since there is still little knowledge of the natural Alternaria mycotoxin contamination in these plants, more investigations are required before conclusions can be drawn on the health risks associated to C. maritima consumption. Capability to produce mycotoxins by representative strains of the Alternaria population isolated in this study has showed that most (90%) of the strains, produced at least one of four assessed mycotoxins, sometimes at very high levels (Supplementary Table S1). This indicated the potential mycotoxicological risk of Alternaria strains isolated from these plants. Only three strains produced ALT, while production AME and AOH by the strains analyzed was proven strictly correlated (Supplementary  Table S2). This confirms previous reports (Bensassi et al., 2011;Ramires et al., 2018). Alternaria mycotoxins have been frequently reported as natural contaminants  of agri-food crops (Puntscher et al., 2018), and their production under natural conditions has been associated to several factors such as plant species and host cultivar, environmental conditions, cultivation techniques, plant growth stage, time of harvesting, and crop rotation (Lauren and Fadwa, 2008). The dominance of mycotoxigenic Alternaria species on C. maritima, reported here for the first time, suggests that this plant can potentially be a risk for human and animal health, if contaminated parts of plants are consumed. In addition, C. maritima could be a source of inoculum of Alternaria species, if the plants grow in the proximity of other important food plants such as wheat.
The pathogenicity tests showed that the levels of aggressiveness of different Alternaria strains from C. maritima leaves were different among species, with the pathogenicity of A. mali being less than for the other species.
The phylogenetic study indicated the occurrence of a set of strains closely-related to A. mali, that could represent a new phylogenetic species. This shows that C. maritima could be a source of Alternaria genetic diversity to be further investigated. This study also confirms that it is still important to provide genetically well-defined reference isolates for taxonomic investigations of Alternaria. These should be accurately validated by a standardized common approach in research dealing with Alternaria taxonomy (Somma et al., 2019). CONCLUSION These results reported here show that the Tunisian halophyte C. maritima can be commonly infected by mycotoxigenic Alternaria species. Morphological, molecular and mycotoxin analyses confirmed variability within Alternaria species of Tunisian origin from different climates. However, no clear links between species with host or geographic origins could be established. This research has shown that C. maritima can be a source of Alternaria genetic diversity to be further investigated, to provide new information on the taxonomy of the genus. ACKNOWLEDGMENTS This work was supported by a grant from the Tunisian Ministry of Higher Education, Scientific Research and Technology (LR15CBBC02) and by H2020-E.U.3.2-678781-MycoKey-Integrated and innovative key actions for mycotoxin management in the food end feed chain. Dr. Besma Sghaier-Hammami was supported by a pro-ject "Programme d'encouragement des jeunes chercheurs (19PEJC07-17)".