First report of Fusarium solani species complex infection on mandarin and maltese in Tunisia and semi-quantitative analysis of mycotoxins production in vitro

Abstract

Fusarium spp. are well known plant pathogens that cause wilt and rot in roots and seeds. This study is the first to report Fusarium infection in mandarins and maltese in Tunisia. Samples of these varieties with Fusarium infection symptoms were collected from 5 orchards in northern Tunisia, resulting in a list of 12 isolates. The pathogenicity test of Fusarium isolates on oranges proved three highly virulent strains (L1, L3, and K6), which were subsequently identified using the sequence of the internal transcribed spacer (ITS) region to belong to Fusarium solani species complex (FSSC). Qualitative analysis of the enzymatic activities of the pathogenic isolates showed the potential of strains L1. L3, and K6 to produce amylase and protease. The metabolic profiling enabled the putative identification (level 2) of several secondary metabolites. L1 predominantly produced depudecin, while L3 and K6 produced fonsecin and sambutoxin. Moreover, correlation analysis suggested no clear trend between pathogenicity and fonsecin and sambutoxin production, and a negative trend with depudecin. Therefore, further investigations are required to validate the production of these metabolites in planta and to assess their potential implications for citrus health and food safety.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12866-026-05041-4.

Introduction

Oranges are one of the major citrus fruit crops worldwide, with Tunisia ranking as a leading exporter of maltese to the European market, particularly France, as the principal destination [1].

Fusarium is one of the top 10 most important phytopathogenic fungi worldwide that causes important economic losses due to its ability to infect diverse crops [2, 3]. While Fusarium sp. is well studied in cereals and other crops, it is less explored in citrus fruits [4]. In citrus, Fusarium spp. is well established as a causal agent of root rot along with Phytophthora spp [5].

Although Fusarium species are mainly soil-borne fungi, their conidia can be spread by air and water enabling fruit infection [2, 6–8]. Moreover, Fusarium species use several infection strategies that vary according to environmental conditions and metabolic cues [2, 9]. Therefore, Fusarium can also infect fruits and cause fruit and stem end rot. Several species are associated with fruit infection such as F. oxysporum, F. solani, and F. proliferatum [10].

Fusarium species are a major concern due to their ability to produce diverse mycotoxins, including fumonisins, trichothecenes, and zearalenones [11–15].

Therefore, this study aimed to investigate the causal agents of Fusarium fruit rot on mandarins and Maltese in Tunisian orchards for the first time, to assess, qualitatively, their ability to produce hydrolytic enzymes and mycotoxins, and to explore the possible association between the pathogenicity and toxigenicity of the fungal isolates.

Materials and methods

Isolation of fungi from fruits

Sampling was conducted during the winter season of 2021 in 5 orchards located in the Capbon region of the Nabeul governorate. Maltese and mandarin fruits that presented rot symptoms of brownish sunken necrotic lesions close to the stem end and fruits with advanced infection that showed wet brown rot with white mycelia were collected (Fig. 1).

Fig. 1.

Naturally infected citrus fruits with Fusarium rot. A infected maltese fruit that present wet brownish rot almost half the fruits surface with darker brown sunken lesions close to the stem, (B) Mandarin infected fruit that present wet brownish rot with white mycelia

Fungi isolation

The fruit samples were washed with 20% sodium chloride. Afterwards, they were washed with sterile distilled water. After the fruits dried, fragments from the necrosis were cut and placed on a potato dextrose agar (PDA; Oxoid Ltd., Basingstoke, UK) plate amended with streptomycin at a concentration of 3 mg/l, to inhibit the growth of the bacteria, and incubated at 25 °C for 7 days.

Colonies with Fusarium characteristic features, including colony color and aspect, were selected [16] and purified using the single-conidium subculture technique on PDA plates under the same incubation conditions [17].

In vivo pathogenicity test

To fulfill Koch’s postulates, artificial inoculation of the isolated strains in maltese oranges was conducted. A representative strain from each geographical location was selected for virulence testing in order to confirm the pathogenic potential of the fungi recovered from infected fruits in each location. Hence, the strains K6, L1, L3, L9, and G3 were selected for the pathogenicity test.

The pathogenicity test was conducted as described by Garganese et al. [18], with slight modifications. The fruits were washed with 5% sodium chloride and then with sterile distilled water. A 2 mm deep wound was created with a sterile yellow tip on the dried oranges (2 holes per fruit). Subsequently, 10 µL of conidial suspension with a fixed concentration of 10⁵ conidia/ml was injected into the holes (4 replicates per fungal strain). The control was inoculated with sterile distilled water. The inoculated fruits were incubated under humid conditions in firmly closed plastic boxes at 25 °C for 10 days.

In order to complete Koch’s postulates, after symptoms development, fragments of the lesion were cut, surface sterilized and inoculated on PDA plates to re-isolate the causal pathogen. Afterwards, the resulting colonies were morphologically identified and compared to the original isolates to confirm their identity.

Molecular identification

The molecular identification of the virulent strains started by extracting the genomic DNA from fungal cultures of 7 days old with the GF-1 fungal DNA extraction kit (Vivantis, Malaysia).

Subsequently, amplification of the internal transcribed spacer (ITS) region was performed using the primers ITS1 (5’ TCCGTAGGTGAACCTGCGG 3′) and ITS 4 (5′ TCCTCCGCTTATTGATATGC 3’) [19]. The PCR reaction was conducted using GoTaq Hot Green Master Mix (Promega, USA), which included dNTPS, Buffer, MgCl2, and Taq Pol. The PCR was programmed at 94 °C for 3 min as the initialization step, followed by 35 cycles of 94 °C for 30 s and 72 °C for 30 s. The final extension was performed at 72 °C for 10 min.

The PCR products were subsequently sequenced and compared with ITS sequences available and published in the NCBI BLAST platform (https://blast.ncbi.nlm.nih.gov/Blast), Mycobank (http://www.mycobank.org), and CBS-KNAW databases (http://www.westerdijkinstitute.nl). The reference ITS sequences from GenBank used for phylogenetic analysis are listed in Table A1.

In vitro detection of hydrolytic enzymes

The ability of the pathogenic isolates to produce hydrolytic enzymes, including amylase, lipase, protease, pectinase, and cellulase, was, qualitatively, tested in vitro.

The production of hydrolytic enzymes was conducted in a minimal synthetic medium to which the enzyme substrate was added as the sole carbon source. Cellulase production was detected by adding carboxymethyl cellulose and submerging the culture in Congo red [9]. Pectinase production was confirmed by adding pectin as the sole carbon source in the medium and copper acetate after incubation [10, 11]. Protease, lipase, and amylase were detected by adding skimmed milk, Tween 80, and starch as the sole carbon source, respectively [10, 12].

Mycotoxins extraction

The extraction of the mycotoxins from the fungal cultures and the UPLC analysis were conducted using acetonitrile (99.9%), ethyl acetate (99.9%), formic acid (99%), acetic acid (AcOH), and methanol (99.9%) (MeOH) provided by VWR Chemicals (Radnor, United States).

The mycotoxins produced by Fusarium strains were extracted from axenic cultures on PDA plates, as reported by Riolo et al. [13]. The strains were transplanted into PDA and incubated at 25 °C for 10 days. Subsequently, MeOH was added to the cultures at a proportion of 1:5 (w/v) and agitated for 24 h at room temperature. Subsequently, the extracts were centrifuged at 4000 rpm for 15 min at 4 °C. The supernatant was collected and filtered through 0.22 μm filter. Finally, the filtered extracts were stored in glass vials until the day of analysis.

Identification of Fusarium mycotoxins through UPLC-ESI-MS-TOF analysis

The mycotoxin extracts were analyzed using a UPLC system (1290 Infinity LC, Agilent Technologies) linked to a quadrupole time-of-flight mass spectrometer (Agilent 6546 LC/Q-TOF) and an Agilent Dual Jet Stream Electrospray Ionization module that operates in positive ionization mode (Dual AJS ESI). An Agilent Zorbax RRHD SB-C18 column (2.1 × 50 mm, 1.8 μm) was used for separation with two mobile phases acidified with 0.1% formic acid; mobile phase A and mobile phase B were composed of Milli-Q water and acetonitrile, respectively.

Gradient elution was performed first at 2% B, which increased to 95%B at 22 min. Afterwards, it decreased to 5% at 25 min, followed by an equilibration period that lasted 3 min before each analysis. 5 µl of each sample were injected with a flow rate maintained at 0.4 mL/min.

The parameters of the Dual AJS ESI source were set at 325 °C: gas temperature, 10 L/min; gas flow rate, 40 psig; nebulizer pressure, 295 °C; sheath gas temperature, 12 L/min; sheath gas flow, 4000 V; capillary voltage, 500 V; nozzle voltage, 120 V; fragmentor voltage, 70 V and skimmer voltage. The product scan range was set 100 to 1500 Da. The MS and MS/MS scan rates were 5 and 3 spectra/s, respectively. Two maximum precursors per cycle and 10, 20, and 40 eV were set.

Calibration of the UPLC-ESI-MS-TOF analysis method

The detection and identification tool for mycotoxins produced by Fusarium strains was calibrated using two mycotoxins of two identified fungi, enniatin B of Fusarium spp. and alternariol monomethyl ether (AME) of Alternaria alternata, which were provided by Sigma-Aldrich (St. Louis, MO, USA). The accuracy and precision of UPLC-ESI-MS-TOF were evaluated based on inter- and intraday variations.

Intact orange peel and juice samples spiked with the selected standards were injected into the UPLC-ESI-MS-TOF three times to determine the mean recoveries. Moreover, injections were administered for one day and three consecutive days to evaluate inter-day variation.

Data analysis

The mean necrosis diameter caused after the artificial inoculation of the fungal strains was compared using analysis of variance (ANOVA) followed by Tukey’s post-hoc test at α = 0.05. The variance analysis was performed using IBM SPSS Statistics version 28.0 (2021) and JASP version 0.19 (2021).

Microscopic features of the fungal isolates and the length and width of the corresponding conidia were measured using ImageJ (NIH, Bethesda, MD, USA) [14].

The ITS sequences of the fungal isolates from this study were aligned with the chosen reference sequences using MUSCLE, and the phylogenetic tree was constructed using MEGA 12 software (version 12, Pennsylvania State University, USA). Phylogenetic analysis was conducted using the ITS sequences, and node support was assessed using bootstrap analysis with 1000 replicates.

Metabolic analysis of fungal cultures was conducted in triplicate. Data integration, processing, and metabolite identification were conducted using the MassHunter Qualitative Analysis B. 08.00 software (Agilent Technologies, Santa Clara, CA, USA) and library PCDL Manager B.08.00 (Agilent Technologies, Santa Clara, CA, USA).

The relative abundance of the detected metabolites is presented as a heatmap with a color gradient. Moreover, correlation coefficients were calculated between necrosis diameter and mycotoxin peak intensities across pathogenic strains. Due to the small sample size (n = 3 isolates), identical coefficients may arise from limited variability in the dataset, therefore, correlation coefficient results were not interpreted as indicative. If the coefficient values r were between − 0.8 and 0.8 the correlation was considered absent and data was interpreted as no clear trend. Meanwhile, if r values were higher than 0.8 the data was interpreted as positive trend. Finally, if the r values were lower than − 0.8 negative trend was concluded [19].

The calculation, principal component analysis (PCA), and heatmap construction were conducted using MetaboAnalyst 6.0 (Wishart Research Group, Edmonton, Canada).

Results

Isolation and morphological characterization of Fusarium pathogens

The isolation of Fusarium strains from oranges resulted in a list of 12 fungal isolates that originated from 5 orchards located in the municipalities Beni Khaled, Grombalia, and Bouargoub from the governorate of Nabeul in the northern region of Tunisia (Capbon). Dominance of isolates from mandarins was noted with a proportion of 66.66%, while 33.33% of isolates originated from maltese. The macroscopic features of the colonies when incubated on PDA showed growth rate of 7–10 days at 25 °C resulting in a colony with white aerial hyphae initially that developed pinkish and creamy color later for the strains K6 and L1, respectively (Table 1; Fig. 2). The reverse side of the colonies had a yellowish color. Under a 100x objective, abundance of microconidia produced from short conidiophores were observed. The microconidia had smooth walls and oval to ellipsoid format that showed 1–2 internal divisions due to a transverse septation in the conidium and formed in clusters at the tips of phialides. Measurement of the microconidia showed a length ranging from 8 μm to 9.98 μm and width ranging from 1.4 μm to 1.99 μm, which is in line with recently reported morphological description of Fusarium solani, a member of Fusarium solani species complex (FSSC) [5].

Table 1.

Identification of the fungal isolates, their geographical distribution, their varietal origin

Order	Isolate code	Fungal species	Accession number	Origin	Orchard number	Geographical location
1	L1	Fusarium solani species complex (FSSC)	PV876380	Mandarin	1	Beni khaled, gouvernorate Nabeul
2	L2	Fusarium sp.	-	Mandarin	1	Beni khaled, gouvernorate Nabeul
3	L3	Fusarium solani species complex (FSSC)	PV876381	Maltese	4	Grombalia, gouvernorate Nabeul
4	L5	Fusarium sp.	-	Mandarin	1	Beni khaled, gouvernorate Nabeul
5	L7	Fusarium sp.	-	Mandarin	1	Beni khaled, gouvernorate Nabeul
6	L9	Fusarium sp.	-	Maltese	3	Grombalia, gouvernorate Nabeul
7	W1	Fusarium sp.	-	Mandarin	2	Beni khaled, gouvernorate Nabeul
8	W2	Fusarium sp.	-	Mandarin	2	Beni khaled, gouvernorate Nabeul
9	G3	Fusarium sp.	-	Mandarin	2	Beni khaled, gouvernorate Nabeul
10	W5	Fusarium sp.	-	Maltese	5	Bouargoub, gouvernerate Nabeul
11	K2	Fusarium sp.	-	Maltese	5	Bouargoub, gouvernerate Nabeul
12	K6	Fusarium solani species complex (FSSC)	PV876382	Mandarin	5	Bouargoub, gouvernerate Nabeul

Fig. 2.

Macroscopic features of the identified pathogenic isolates

Pathogenicity assay of Fusarium isolates on orange fruits

The artificial inoculation of Fusarium strains in oranges demonstrated the virulence of the strains K6, L1, and L3 by causing brown lesions on oranges (Fig. 3B). Meanwhile, the strains G3 and L9, the representatives of orchards two and three, respectively, did not cause any lesions on orange.

Fig. 3.

Necrosis diameter and symptom development induced by Fusarium isolates on oranges. A Mean necrosis diameters caused by Fusarium isolates on oranges after artificial inoculation. Error bars represent standard deviation (SD) for n = 4. Diagrams followed by different letters have values with significant difference (p ˂0.001) according to LSD post hoc test. (B) Symptoms caused by Fusarium strains after 10 days incubation

Lesion diameters were measured 10 days after inoculation and expressed as mean ± standard deviation (SD) based on 4 replicates. The strain K6 caused largest lesion diameter (22.5 ± 0.80 mm), significantly greater than the other strains (p < 0.001), making it the most pathogenic strain. The isolate L1 was the second most pathogenic strain, with a diameter of 12.5 ± 1.04 mm and L3 caused mean necrosis diameter equal to 2.5 ± 0.38 (Fig. 3A).One-way ANOVA followed by LSD post hoc statistical analysis revealed significant differences among the isolates (p < 0.001). According to the letter-based grouping analysis, isolates were classified into four statistically distinct groups: K6 (d), L1 (c), L3 (b), and G3/L9 (a).

The pathogenic strains K6, L1 and L3 were successfully re-isolated from the lesions developed after the artificial inoculation of the original isolates in fruits. Thereby, Koch’s postulates were fulfilled. Control fruits did not develop any symptoms throughout the experiment.

Molecular identification of L1, L3, and K6 strains

The pathogenic strains were selected for molecular identification via ITS sequencing. The ITS sequences of L1 and L3 were approximately 483 bp, while the ITS sequence of K6 was 483 bp. The strains L1, L3, and K6 showed 100% similarity after BLAST and sequence alignment to several members of the Fusarium solani species complex such as Fusarium solani (KY128174) and Fusarium falciforme (KY436188). Based on these results, phylogenetic analysis was conducted using the maximum likelihood method on the strains’ ITS sequences, compared to 19 Fusarium species sequences, including reference CBS strains [17] from Mycobank (http://www.mycobank.org) and type/representative strains from GenBank (https://www.ncbi.nlm.nih.gov) databases.

The phylogenetic tree showed that FSSC NRRL 43,536, NRRL 54,966, and NRRL52832 and F. cf. solani (CBS 121450, CBS124627, and CBS 318.73) are the closest relatives to K6, L1, and L3 (Fig. 4). The ITS sequences of L1, L3, and K6 were assigned the accession numbers PV876380, PV876381, and PV876382, respectively, after submission to GenBank (Table 1).

Fig. 4.

Phylogenetic tree using the maximum likelihood method of Fusarium isolates and other Fusarium species based on the ITS sequences. Numbers above nodes represent bootstrap support values from 1000 replicates

The ability of L1, L3 and K6 strains to produce hydrolytic enzymes in vitro

The qualitative in vitro assays to detect the extracellular production of the hydrolytic enzymes showed that the isolates L1, L3, and K6 showed a similar ability to produce the enzymes amylase and protease resulting in a halo circling the fungal colony that signifies the absence of the substrate used to detect the enzyme. Moreover, the three strains were found not to produce the enzymes cellulase, pectinase, and lipase due to the absence of halos suggesting low or undetectable activity under the tested conditions (Table 2). These assays provide a preliminary screening of the potential hydrolytic capabilities of Fusarium isolates.

Table 2.

Results of the in vitro production of hydrolytic enzymes cellulase, amylase, protease, pectinase, and lipase

Strain	Cellulase	Amylase	Protease	Pectinase	Lipase
L1	-	+	+	-	-
L3	-	+	+	-	-
K6	-	+	+	-	-

Identification of mycotoxins produced by the isolates in vitro

The identification of the detected metabolites was based on high-resolution mass spectrometry data, including accurate mass measurements, retention time, and database matching. Therefore, the identification corresponds to level 2 (putatively annotated compounds) (Table 3).

Table 3.

Putative identification of the mycotoxins produced by pathogenic Fusarium isolates L1, L3, and K6 in PDA medium

Name of the component	Molecular formula	Molecular weight (g/mol)	Retention time (min)	m/z	level of identification	Diff (Tgt, ppm)	Score (Tgt)	Chemical class
Depudecin	C11 H16 O4	212.11	5.56	230.14	Lv. 2a Probable structure*	0.4	98	Polyketide lactone
Fonsecin	C15 H14 O6	290.08	5.75	291.09	Lv. 2a Probable structure*	0.52	100	Polyketide pigment
Sambutoxin	C28 H39 N O4	453.29	11.5	476.28	Lv. 2a Probable structure*	0.01	98.37	Polyketide

*Identification level follows Schymanski et al. [20] classification (Level 2: putatively annotated compounds based on accurate mass and database matching; no confirmation with reference standards)

The potential of the strains L1, L3, and K6 to produce mycotoxins was studied using the metabolic extracts from their cultures, which were analyzed with UPLC-ESI-MS-TOF. Three mycotoxins were tentatively detected in the cultures of the strains (Table 3). The toxigenic profiles, presented as a color gradient of each mycotoxin relative abundance in a heatmap showed discrepancy across strains. The strains L3 and K6 demonstrated higher apparent toxigenic potential compared to L1 producing 2 mycotoxins out of 3 with high relative abundance (colored in red in the heatmap). The predominant mycotoxins in FSSC L3 and K6 cultures were annotated as sambutoxin and fonsecin. Meanwhile, depudecin was the most abundant mycotoxin in the culture of L1 (Fig. 5).

Fig. 5.

Heatmap of the relative abundance of the mycotoxins produced by Fusarium strains on PDA plates after 10 days incubation. The peak area of each mycotoxin was log₁₀ transformed to construct the semi-quantitative heatmap that represents its relative abundance. The latter was presented on a red-green color scale that shows the highly abundant compounds and lowly abundant compounds colored red and green, respectively. Meanwhile, the absent compounds are colored black

It is important to note that mycotoxin production in this study was assessed under in vitro conditions using PDA medium, which may not fully reflect fungal behavior in planta. Environmental factors such as host tissue composition, water activity, and biotic interactions can significantly influence secondary metabolite biosynthesis [21]. Therefore, the detected metabolite profiles should be interpreted as indicative of the toxigenic potential of the isolates rather than direct evidence of mycotoxin occurrence in infected citrus fruits.

Correlation analysis across Fusarium mycotoxins and between chromatograms peak intensities and isolates pathogenicity

The preliminary exploratory correlation analysis using Pearson r-test between the detected mycotoxins in Fusarium cultures suggested a positive trend between fonsecin and sambutoxin. Whereas depudecin showed negative trend with all other mycotoxins (Fig. 6). These findings are in line with Fig. 5 in which the heatmap showed the low abundance of the other mycotoxins in L1 culture in which depudecin was predominant.

Fig. 6.

Pearson correlation of mycotoxin profiles. Correlations between the peak intensities of individual mycotoxins were calculated using Pearson correlation coefficients (n = 3 mycotoxins per strain). Bars represent the correlation coefficient (r) of each mycotoxin against the other detected mycotoxins. Correlations are interpreted as absent (-0.8 < r > 0.8), positive (r ≥ 0.8),and negative (r ≤ -0.8). Panels show correlations for (A) Depudecin, (B) Sambutoxin, and (C) Fonsecin

Moreover, no clear trend was found between the pathogenicity of Fusarium strains and the production of fonsecin and sambutoxin. The Pearson r test between the diameters of necrosis caused by FSSC L1, L3 and K6 and the peak intensities of fonsecin and sambutoxin showed that all the coefficient values r were between − 0.8 and 0.8 therefore the results were interpreted as no clear trend due to the limited number of isolates [19]. However, depudecin production was found to have a negative trend with necrosis diameter (r= -0.87, p-value = 0.02) (Table A2).

In conclusion, while the negative trend of depudecin and the diameter of necrosis may be associated with the moderate pathogenicity of L1 in the test conducted during this study, there was no sufficient statistical evidence in this study to claim any association between the pathogenicity and the toxigenicity of the fungal strains L3, and K6. Therefore, the reason for remarkable difference in the pathogenic potential of K6 and L3 resulting in the former causing a diameter of necrosis 10 times greater than L3, despite having similar toxigenic profile remains unsolved. This insufficient evidence may be attributed to the lack of favorable conditions for producing mycotoxins for more in-depth study of their effect on pathogenic potential [21].

Overall, considering the limited number of isolates (n = 3), these findings should be considered preliminary, and no definitive conclusions can be drawn regarding the relationship between pathogenicity and mycotoxin production.

Discussion

Fusarium species are typically related to rhizosphere and root infection. However, recently it has been increasingly acknowledged as a citrus fruit pathogen that affects the stem and the rind in which it causes necrotic lesions and soft rot, respectively. It also causes internal dry or watery decay which subsequently leads in some cases to mycelia growth [22, 23]. Fusarium species can cause these symptoms through cell wall degrading enzymes production such as pectinases and secondary metabolites, including mycotoxin-like compounds [24].

In this first report of Fusarium infection on citrus fruits in Tunisia, the pathogenicity assays proved the virulence of three strains out of five isolated from infected mandarin and maltese. The molecular identification of the strains L1, L3, and K6 was conducted using the ITS of ribosomal RNA (rRNA). The ITS region sequences of the pathogenic isolates from this study identified the strains L1, L3, and K6 as part of Fusarium solani species complex (FSSC)and grouped them with reference strains such as F. cf. solani CBS 121,450, CBS124627, and CBS 318.73. Within the FSSC, several phylogenetic species such as Fusarium falciforme have been reported as plant pathogens. However, the available ITS data do not allow precise discrimination among these taxa [25, 26]. These findings are in line with Zhu et al., (2025) that reported the pathogenicity of F. solani, a member of FSSC, on sour orange and sweet orange trees and causing root rot.

While, Fusarium species are widely known as a plant pathogens that causes plant fusariosis with symptoms such as roots and seeds wilt and rot [27]. It was also reported to infect fruits such as peach, apple, pear, date, banana, melon, and pineapple [27–30]. Several reports of the first detection of Fusarium infection on citrus fruits were recorded. Several species infected oranges such as F. solani that was reported in Egypt, F. tricinctum in China, F. solani and F. sambucinum in Korea, and F. concentricum in Bangladesh [28, 31–33]. While our findings align with the report of F. solani detection in oranges in Egypt, they converge in the host orange variety since it was recorded in Navel [32]. Moreover, in Tunisia, F. solani was associated with taxonomically unrelated host such as pistachio, watermelon, and apple trees [34–36]. Meanwhile, in citrus trees Fusarium wilt caused by F. oxysporum was reported for the first time in 2014 [37]. In Tunisia Fusarium is mainly associated with cereals infection. The major Fusarium cereals pathogens are F. culmorum, F. graminearum, F. avenaceum, F. acuminatum and members of the Fusarium incarnatum–equiseti and FIESC groups. They are causal agents of several symptoms such as foot and root rot, seedling blight and Fusarium head blight [38–40]. Therefore, to our knowledge, no reports of its infection on citrus fruits were recorded in Tunisia.

The in vitro detection of the hydrolytic enzymes in this study was based on qualitative plate assays in which the presence of a halo around colonies was used as indicator of the enzyme production without exact measurement of enzyme activity. It should be noted that medium composition, substrate type, pH, and incubation conditions influence the detection of the enzymes activity [41, 42]. Therefore, the qualitative assays revealed the potential of the isolates L1, L3 and K6 of the FSSC to produce amylase and protease. These findings are in accordance with previous studies reported amylase and protease production by F. solani, a member of FSSC [43–46]. Moreover, Hussein et al. [47] that studied the enzyme profile of the FSSC member, Fusarium solani isolated from potato tubers, reported its ability to produce amylase, cellulase, pectinase, lipase, and protease and stated that the pathogenicity levels of F. solani strains were positively correlated to their enzyme profiles. The discrepancy between this report and our findings can be due to genotype-dependent differences.

This study aimed to evaluate the toxigenic potential of Fusarium isolates by putatively identifying mycotoxins produced by the fungal isolates under in vitro conditions on PDA medium through UPLC-ESI-MS-TOF. The latter detected the presence of annotated depudecin, fonsecin and sambutoxin in the metabolic extracts of FSSC isolates. The identification of the detected metabolites corresponded to level 2 (putatively annotated compounds) based on database matching and accurate mass measurements.

Previous reports recorded the detection of fusaric acid, beauvericin, deoxynivalenol, enniatins, fumonisins, moniliformin, zearalenone, fonsecin and fonsecin B in the exo-metabolome of Fusarium. kuroshium and F. oxysporum f. sp. Psidii [48, 49]. In this context, these reports are in line with the presence of fonsecin in this study.

The in vitro putative identification of sambutoxin and depudecin in the metabolic extractes of the FSSC strains L1, L3 and K6 is noteworthy. Sambutoxin has historically been associated with other Fusarium species, yet its production by members of the FSSC remains insufficiently documented [50]. Therefore, despite this study suggestion of a broader metabolic capacity of the genus, additional targeted analytical approaches such as reference standards and/or MS/MS fragmentation, are needed to confirm their attribution to the isolates within FSSC.

In contrast, depudecin detection in this study is in line with previous studies that reported not only the presence of the gene cluster including the efflux pump (DEP3), and the transcription factor (DEP6) genes responsible for depudecin biosynthesis in the genome of Fusarium species but also its production in PDA medium [48, 51].

The relationship between toxigenicity and pathogenicity remains complex and context dependent. A study conducted on F. solani to investigate the correlation between its pathogenicity and toxigenicity proved that the disease severity was positively correlated to the levels of deoxynivalenol, and T2- toxin detected in the medium [21] These findings contradict the exploratory results of this study that suggested negative trend between depudecin and the pathogenicity of the FSSC strains, and no clear trend in the case of fonsecin and sambutoxin. This contradiction can be explained by substantial differences in the experiments conducted. In the previous report the researchers extracted the mycotoxins from rice based medium inoculated with the strains that were isolated from rice plants which means that the medium simulated the natural habitat of the fungal strains. Meanwhile, in this study mycotoxins were extracted from cultures on standard PDA plates. This difference can be further confirmed by the fundamental differences of mycotoxins detected in both cases, knowing that the environmental conditions affect mycotoxins production [21]. Therefore, the putative results of the present study are an indication of the potential of the isolates to produce these compounds and not evidence of their presence in planta.

In addition, the limited number of isolates analyzed in this study (n = 3) restricts the generalization of the observed trends and precludes robust statistical conclusions regarding the relationship between pathogenicity and toxigenicity.

In summary, this study is the first report of Fusarium infection on mandarin and maltese oranges in Tunisia. It explored the characteristics and potential ability to produce hydrolytic enzymes and mycotoxins by the 3 isolates of the FSSC (K6, L1, L3).

However, this study has its limitations including the molecular identification using ITS sequencing that allowed the placement of the isolates within the FSSC, but it did not provide sufficient resolution to distinguish closely related species within the genus Fusarium. Therefore, additional loci such as TEF1-α need to be added for precise species identification. Only one isolate per orchard was tested in pathogenicity assays because the objective of the present study was exploratory, aiming to confirm the pathogenic potential of Fusarium isolates recovered from symptomatic citrus fruits observed during field sampling. Future studies involving a larger number of isolates will be necessary to better evaluate intra-orchard variability and the prevalence of the pathogen in citrus production systems in Tunisia. Additionally, the absence of in vivo detection of mycotoxin in oranges and higher sample size are required to more accurately study fungal behavior in planta and the correlation between the pathogenicity and toxigenicity of Fusarium pathogens. Therefore, further research on mycotoxins production by Fusarium fungi on oranges are needed to directly study mycotoxin occurrence in infected citrus fruits.

Conclusion

This study is the first report of Fusarium infection on mandarin and maltese in Tunisia. The isolation of fungal strains from oranges showing Fusarium infection symptoms from five orchards in Tunisia resulted in a list of 12 isolates from which five were chosen for the pathogenicity test, one representative from each orchard. The pathogenicity test proved the pathogenecity of the strains L1, L3, and K6 that were subsequently identified by ITS region sequencing as part of the FSSC. Fusarium pathogens were found to potentially produce hydrolytic enzymes such as amylase, protease, and mycotoxins in vitro such as fonsecin, depudecin, and sambutoxin. While these compounds are known mycotoxins, their presence in the edible parts of oranges was not assessed in this study, and therefore no direct food safety risk to consumers can be inferred. Moreover, the correlation analysis of the necrosis diameter caused by the strains L1 L3, and K6 and the peak intensities of the mycotoxins suggested a negative trend between the pathogenicity of the strains and depudecin, and no clear trend in the case of fonsecin and sambutoxin.

Abbreviations

AcOH

Acetic acid

AME

Alternariol monomethyl ether

AJS ESI

Agilent Jet Stream Electrospray Ionization

BLAST

Basic Local Alignment Search Tool

CBS KNAW

Centraalbureau voor Schimmelcultures–Koninklijke Nederlandse Akademie van Wetenschappen (Westerdijk Fungal Biodiversity Institute)

dNTPs

Deoxynucleotide triphosphates

ESI

Electrospray ionization

FSSC

Fusarium solani species complex

ITS

Internal transcribed spacer

MeOH

Methanol

MgCl₂

Magnesium chloride

PCDL

Personal Compound Database and Library

PDA

Potato dextrose agar

Q TOF

Quadrupole time–of–flight

SPSS

Statistical Package for the Social Sciences

UPLC

Ultra–performance liquid chromatography

UPLC–ESI–MS–TOF

Ultra–performance liquid chromatography–electrospray ionization–mass spectrometry–time–of–flight

Authors’ contributions

D.S. (Conceptualization, methodology, Data curation, Formal analysis, visualization, writing original draft, writing—review and editing), S.A. (Conceptualization, Data curation, Formal analysis, visualization, writing—review and editing), V.D. (Conceptualization, writing—review and editing), M.H. (Conceptualization, writing—review and editing), G.M. (Conceptualization, Methodology, writing—review and editing, Supervision, funding acquisition, Project administration). N.S.-Z. (Conceptualization, Methodology, writing—review and editing, Supervision, funding acquisition, Project administration).

Funding

This research did not receive funding.

Data availability

All data generated or analyzed during this study are included in this published article and its supplementary information files. The ITS region sequences of this study (accessions L1:PV876380, L3:PV876381, K6:PV876382) have been submitted to the NCBI GenBank database ( [PV876380:PV876382[accn] - Nucleotide - NCBI] (https://www.ncbi.nlm.nih.gov/nuccore/?term=PV876380:PV876382%5baccn%5d) ).

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.