Identification of fungi in Tunisian olive orchards: characterization and biological control potential

Olive production is the main agricultural activity in Tunisia. The diversity of fungi was explored in two different olive groves located in two distant geographical zones in Sfax (Tunisia) with different management practices. Fungal isolation was made from soil and the major olive tree pests, namely the Olive fly, Bactrocera oleae Gmelin (Diptera: Tephritidae), and the Olive psyllid, Euphyllura olivina Costa (Homoptera: Psyllidae). A total of 34 fungal isolates were identified according to their phenotypic, genotypic, biochemical and biological activities. Twenty fungal species were identified belonging to six different genera (Alternaria, Aspergillus, Cladosporium, Fusarium, Lecanicillium and Penicillium) by the analysis of their ITS1–5.8S–ITS2 ribosomal DNA region. Different bioassays performed in this work revealed that 25/34 (73.5%) of the identified fungal isolates showed an entomopathogenic and/or antagonistic activity, 9/34 (26.5%) of them displayed phytopathogenic features. Fungal species that showed entomopathogenic and/or antagonistic potentialities and that are non-phytopathogenic, (17/34; 50%) of our fungal isolates, could be explored for olive protection against fungal diseases and pests, and might have a future application as biocontrol agents.

economic impact and displays social, environmental and landscape significance.
Unfortunately, these olive orchards are continuously damaged by major olive insect pests, namely the olive fly, Bactrocera oleae Gmelin (Diptera: Tephritidae), the olive moth, Prays oleae Bernard (Lepidoptera: Yponomeutidae), and the olive psyllid, Euphyllura olivina Costa (Homoptera: Psyllidae), which may cause high economic losses [10][11][12][13]; and plant pathogens such as Fusicladium oleaginum (Castagne) Ritschel & U. Braun (Dothideomycetes), Pseudocercospora cladosporioides (Sacc.) U. Braun (syn. Cercospora cladosporioïdes (Sacc.) (Dothideomycetes), or Verticillium dahliae Kleb (Sordariomycetes) [14,15]. The control of these pests and pathogens has been based on the application of chemical pesticides -e.g. those based on synthetic pyrethroidscharacterized by their damaging effects to the ecosystems [16] as, e.g., pest-resistance problems [17], sideeffects on auxiliary fauna [18,19] and human health [20]. However, these practices are being replaced by biological control methods. Among the possible biological control methods, the use of entomopathogenic and the antagonistic fungi is considered as a promising strategy. The entomopathogenic activity of fungi as well as their antifungal activity depend on several factors including hydrolytic enzymes as amylases, proteases, lipases, chitinases, esterases, glucanases, and catalases [21][22][23]. Insects and fungi are embroiled in complex interactions between them. Enzymes such as amylases, proteases and lipases are implicated in the infective process of the insect by entomopathogenic fungi. In fact, lipases are involved in the adhesion of fungal spores on the insect cuticle, which is a mandatory pre-step that initiates the degradation of fatty acids and alkenes in the insect epicuticle [21]. Proteases intervene in the degradation of the proteinaceous material of insect and are considered as the most important enzymes for the infective process [22]. The amylase activity degrading glycogen of the insect tissues allows its use for the growth and invasion of entomopathogenic fungi [24].
In the present study, the mycobiota associated with olive trees and pests were investigated as part of a research effort targeting ecosystem-friendly approaches for olive tree pest and pathogen management. Fungal isolates from olive orchards were characterized by molecular methods, and investigated for their enzymatic, antifungal, entomopathogenic and phytopathogenic activities.

Systematic characterization of the fungal isolates
Fungi isolation from the insect pest cadavers and soil samples in Taous and Torba allowed the identification of 34 fungal strains (Table 1, Fig. S1). Twenty fungal species were identified belonging to 6 different genera (Alternaria, Aspergillus, Cladosporium, Fusarium, Lecanicillium, Penicillium) and 5 families (Trichocomaceae, Pleosporaceae, Nectriaceae, Davidiellaceae, Cordycipitiaceae) by the comparison of their ITS1-5.8S-ITS2 sequences. All the 34 nucleotide sequences obtained from this region represented high degrees of identity with fungal ITS sequences available in the GenBank Database (scores > 98%).
The most strongly represented fungal taxonomic family was the Trichocomaceae. This family accounted for 22/34 (i.e. 64.7%) of the identified fungal strains that were assigned to 14 species. Aspergillus (N = 8 species), Penicillium (N = 6 species) and Cladosporium (N = 3 species) were the most prevalent genera, whereas Cladosporium cladosporioides (N = 4 isolates) and Aspergillus ustus (N = 3 isolates) were most prevalent at the species level.
The experimental plot of Taous that was managed conventionally and under artificial irrigation, displayed greater fungal richness than the Torba plot that was managed biologically and under rain-fed conditions. In fact, in the Taous plot, among a total of 17 fungal isolates obtained; 12 varied species were identified, belonging to 6 genera and 5 families. The family Trichocomaceae representing 47.0% (8/17) and the genus Aspergillus representing 29.4% (5/17) of the identified fungi were prevalent in this plot. Three genera were typically found only in the Taous plot, namely Alternaria, Fusarium and Lecanicillium.
The majority of fungal species (90%) tended to appear in very low numbers (singletons and doubletons) while only two taxa (10%) were represented by three or more isolates. The fungal diversity was evaluated through Simpson's Index and values of this index showed that the two sites (Torba and Taous) have the same fungal diversity (0.95). The fungal biodiversity analysis with respect to the comparison between species of insects and soil, demonstrated that insects (0.96) had greater fungal diversity than soil (0.86).

Screening for specific enzymes production
The 34 collected isolates were evaluated for their enzymatic activities, namely the amylase, protease and lipase enzymes (Fig. 1). Among these, amylase activities were most widespread being present in 58.8% (20/34) of the isolates. Most of amylase producing fungi belonged to the genera Aspergillus and Penicillium. Strains isolated from soil produced significantly more lipase than other isolates as, e.g., derived from olive fly and olive psyllid cadavers (p = 0.034). Proteolytic activity was present in 23.5% (8/34) of the isolated fungi which showed their ability to produce extracellular protease by the degradation halos of casein on solid medium. Aspergillus was the genus able to degrade casein in all samples, with (statistically not significant) differences in the protease activity between isolates.

Entomopathogenic activity
The 16 tested fungal isolates displayed significant differences in their entomopathogenic activity as determined 7 days after treatment by the number of dead fourth instar larvae fed on a diet containing a fungal suspension (10 7 Conidia/ml), with p < 0.05 (Fig. 2). One-way ANOVA was done to test significant differences between the means of mortality percentage. Results studying the effect of the treatment are significant (p < 0.05) as compared to the negative control. The mortality percentage of L4 larvae fed semolina contaminated with the fungal suspension varied from 13% up to 100%. From our fungal isolates, Aspergillus pseudodeflectus F13 and Lecanicillium aphanocladi F28' were the most pathogenic with mortality rates of 100 and 93.3%, respectively, followed by Cladosporium sphaerospermum F17 (86.6%). Treated cadavers incubated in adequate conditions resulted in confirming the fungal infection causing the death of the insect.

Richness of fungal taxa
In this work, the diversity of fungi isolated from major olive pests and soil samples was evaluated. The molecular identification using the ITS marker presents 20 fungal species, belonging to 6 genera and 5 families. To support the identification of Fusarium species, more genetic markers should be included. In fact, ITS did not present a good molecular marker to identify species of Fusarium since this region is rather well-conserved in the genus [25].
The genera Aspergillus and Penicillium were most abundant in the collected samples comprising 64.7% (22/34) at the strain and 70% (14/20) at the species level independently on production guidelines. These observations are in line with studies emphasizing the largest widespread of Penicillium and Aspergillus identified as the dominant genera in most of the ecosystems and were the most ubiquitous fungal species in nature (soils, plants and agricultural communities) [13,[26][27][28].

Biotechnological properties of fungal isolates Enzymatic activities
Not all isolates of the same fungal species necessarily produce the same enzyme and have the same intensity of the enzymatic activities. These properties are strain dependent [33]. In fact, the enzymatic production differs among fungi and is related to their natural habitat or ecological niche [34].
Our results showed that amylolytic activity was the most common among all our collected strains. These results are consistent with previous studies showing that amylase was the predominant secreted enzyme of endophytic fungi isolated from soybean [35]. Moreover, Penicillium species have been reported to be prominent producers of amylolytic activities [36], as is the case for Aspergillus fungi isolated from different seeds, soils or from Morinda citrifolia [37][38][39][40]. Consistently, among the fungal isolates investigated in this study, those belonging to the genera Aspergillus and Penicillium were the most prominent amylase producers.
In addition to the amylolytic activity, Aspergillus was significantly the most abundant genus producing protease and lipase activities as compared to the other genera identified here. Consistently, previous studies with 110 fungal isolates of Aspergillus, Fusarium, Cladosporium and other genera had revealed that 73% of them produced protease, lipase and urease [41]. Predominant among the lipase and protease producing isolates were Aspergillus sp., Cladosporium sp., Myrothecium sp. and Fusarium sp..
In addition to their well-established nutritional function [36], amylase, protease and lipase activities could play an important role in fungal ecology and during microbial infection exhibiting antagonistic, entomopathogenic, and phytopathogenic features. Secreted enzymes such as amylases, proteases and lipases are known to be implicated in the interaction between insect hosts and entomopathogenic fungi. Lipases are involved in the adhesion of fungal spores on the insect cuticle, amylase allows degradation of glycogen from insect tissues for growth and invasion, and proteolytic degradation has been identified as a key factor in fungal pathogenesis, including several species of Aspergillus [36] and the infective degradation of proteinaceous material of the host cuticle by infection hyphae of entomopathogenic fungi [21,22].

Antagonistic activity
Increasing antifungal activity is a potentially useful and effective strategy for fungal biocontrol. A fungal biocontrol agent may act against fungal pathogens by using antibiosis which is defined as the capacity of producing different antibiotics probably involved in the suppression of competitors. This mode of action is exercised by a broad number of fungal bio-controllers [42]. Moreover, other studies reported that the antagonistic activity of fungi were associated to competition for nutrients and space [43][44][45].
In this study, fungal isolates have clearly shown their antagonistic properties against the six tested fungal reference strains. Antagonistic activity of Penicillium species against these fungi was characterized by the production of toxic metabolites. According to previous reports, filamentous fungi, including some of the strains tested in our study as Penicillium chrysogenum F49', Aspergillus ochraceus F57 and Fusarium solani-like F59 may restrict and stop fungal growth. These filamentous fungi such as Aspergillus and Penicillium are known as antimicrobial due to their ability to produce toxic secondary metabolites including hydrolytic enzymes such as protease, amylase and lipase [46]. The susceptibility of Alternaria consortialis to 14 of the fungal isolates as, e.g., Penicillium chrysogenum, Penicillium crustosum, Penicillium pinophilum, and Penicillium polonicum tested in this study is consistent with previous studies reporting the antagonistic activity of antifungal compounds produced by Penicillium sp., (Macropherin A, Atrpinin A, Botryodiplodin and Brefeldin A) which allowed the inhibition of Alternaria sp. [3,47].

Entomopathogenic activity
From the fungal taxa identified in this study, Aspergillus pseudodeflectus F13 and Lecanicillium aphanocladii F28' were found highly pathogenic to Ephestia kuehniella larvae. Previous studies have associated Aspergillus pseudodeflectus with entomopathogenic activity and the stimulation of fungal aflatoxin production by the insect host. It has been shown that the addition of aflatoxin to the food of insect larvae kills them and promotes the growth of the fungus [48,49]. Consequently, experiments to evaluate its infecting ability towards vertebrates have to be performed.
According to our knowledge, Lecanicillium aphanocladii fungi have not previously been recorded in association with olive pests. Isolate L. aphanocladii F28' was characterized as a potent entomopathogen causing a high percentage (75%) of larvalmortality. L. aphanocladii has previously been found associated with diseased and dead lepidopteran insects [50], and there appears to be clear link between the isolation from insect cadavers and entomopathogenic activity. The insect surface sterilization before fungal isolation as well as insect susceptibility bioassays confirmed this activity. The efficiency of Lecanicillium aphanocladii to parasitize dipteran insects as, e.g., mosquito larvae has also been reported [51,52]. Furthermore, Lecanicillium spp. was used for 15 commercial preparations.

Phytopathogenic activity
In addition to these potential biocontrol agents which have a role in limiting pests, other fungi have antagonistic features for limiting the occurrence of phytopathogenic fungi in olive groves. Among the phytopathogenic fungi we have identified Alternaria consortialis F50 which have showed vascular tissue browning. This is consistent with the previous identification of Alternaria species as the cause of olive spoilage and a disease on olive shoots grown under greenhouse conditions [13]. Moreover, isolate Alternaria consortialis F50 has been shown to display entomopathogenic activity against E. kuehniella (larval mortality of 66.6%). These results are corroborated by the findings of previous studies reporting the efficiency of Alternaria spp. as entomopathogenic fungi where the insect may ingest them through infected leaves [53].
Alternaria strains displaying entomopathogenic, but lacking phytopathogenic activity could be used in pest management as whole organisms [54]. However, with both entomopathogenic and phytopathogenic activities being recorded as is the case for strain A. consortialis F50, investigation and identification of the insecticidal principle or mode of action could lead to the development of derived pest control agents not requiring application of the living fungus [55].

Conclusions
The present study describes for the first time the fungal diversity obtained from dead insect pests collected from Tunisian olive groves. The screening of secreted fungal enzymes showed highly variable activity profiles which might have impact for improving technological processes. With respect to the possible development of biological control agents, 50% of the fungal species from our collection showed entomopathogenic or antagonistic activity and were not phytopathogenic as determined in terms of browning length on cortex and vascular tissues. From the fungal taxa identified in this study, Aspergillus pseudodeflectus F13 and Lecanicillium aphanocladii F28' appear the most noticeable as they combine high entomopathogenic potential with lack of phytopathogenic activity against olive trees andin the case of A. pseudodeflectus -antagonistic activity against phytopathogenic fungi. As endemic isolates well adapted to the regional climatic conditions, these strains could be promising in view of the development of protection of olive orchards after their insecticidal bioassays on the olive major pests.

Field sites and fungal isolation
Samples were randomly collected from 2 different olive groves located in two distinct geographical zones (Taous and Torba) of Sfax governorate, Tunisia (GPS coordinates: (34°55′17″N, 10°36′47″W; 34°50′56.9″N10°27′ 33.6″E). These samples were collected once per month in the period ranging from December 2016 to February 2017. At each site, around twenty olive fruits and twenty branches infested respectively by the olive fly, B. oleae, and the olive psyllid, E. olivine, as well as 100 g of soil (from 10 cm of the upper soil layer) were sampled from two sampling locations within the olive orchards. Each sample was packaged in a sterile plastic bag (31 × 20 cm), transported to the laboratory under aseptic conditions and stored at 4°C until the isolation procedure of fungi was carried out. The olive orchards were chosen according to the pest and pathogen management strategy practiced: Whereas the experimental plot at Taous was maintained under conventional control measures and in irrigated, the Torba plot was conducted according to the organic farming system under rain-fed conditions. Sampled insect cadavers were surface sterilized in a 70% ethanol solution for 1 min in laminar chamber. The surface-sterilized insects were rinsed 3 times with sterile water for 1 min and blotted dried on sterile paper towels. Each sample was ground using a micro-pestle in sterilized distilled water. The soil samples were treated as follows: 10 g of soil was dissolved in 90 ml of sterile distilled water and mixed with a stomacher. Supernatants were diluted 10-fold and 100 μl of each dilution was spread on the potato dextrose agar media (PDA: 4 g l − 1 Potato extract, 20 g l − 1 Dextrose, 16 g l − 1 Agar). A total of 34 pure fungal cultures were isolated after continuous sub-cultivation. The pure cultures were recorded and maintained on PDA in active form for further investigation. For long-term storage, the fungal colonies were stored in paraffin oil at − 20°C.

Identification of fungal isolates
The isolates were initially identified based on morphological structures (color, texture of the mycelia and spore formation) performed under microscopic observation of fresh cultures [56]. For fungal DNA extraction, the mycelial mass obtained from the culture plate was placed in a 2 ml tube, submerged in liquid nitrogen and then ground using a micro-pestle in lysis buffer. The extraction protocol is continued using the DNeasy Plant Mini kit (Qiagen® GmbH, Hilden, Germany) following manufacturer's instructions. The quality of the genomic DNA was evaluated by a NanoDrop NT-100 UV spectrophotometer (NanoDrop technologies®, Montchanins, USA).

Phylogenetic analysis
The amplified ITS products were sequenced using ITS4 and ITS5 primers. Sequencing was performed by Star-seq® GmbH (Mainz, Germany). The ITS sequence information was used to match the most closely related fungal isolates with the NCBI BLAST algorithm from the GenBank database (http://www.ncbi.nlm.nih.gov). The analysis of the ITS sequences were performed using the MEGA 6 program [59] (http://www.megasoftware.net).

Screening for extracellular enzyme production
Qualitative tests were used to evaluate the production of fungal extracellular enzymes of interest in industrial biotechnology. The screened enzymes were lipase, protease and amylase.
Lipase production was tested on PDA (1 l) supplemented with olive oil (16.0 ml), and Rhodamin solution (10 mg.ml − 1 ) (Sigma Aldrich®, Germany). After development of the isolates on agar at optimal conditions, lipase activity was detected by irradiating with Ultra Violet light at 350 nm. Positive reaction is accompanied by the presence of precipitates around the fungal colony [60].
Protease production was revealed on skim milk agar. To determine the proteolytic potential, fungi were grown at 25°C for 72 h. The enzyme production was assessed by the appearance of a clear zone corresponding to casein hydrolysis surrounding the fungal colony [61].
To check for amylase activity, fungi were incubated in PDA medium supplemented with starch (1%). After incubation and fungal growth an iodine solution (0.3 g iodine and 0.6 g.l. − 1 ) was added to the plate. Amylase production is shown by the presence of a clear halo due to the degradation of starch and failure to form the colored starch-iodine intercalation complex [62].

Detection of antagonistic activity in vitro
The in vitro antagonistic activity of the collected fungal isolates was tested against the following pathogenic fungi provided by the Olive Tree Institute: Aspergillus calidoustus, Aspergillus pseudodeflectus, Aspergillus tamarii, Alternaria consortialis and Penicillium chrysogenum. These latter represent known pathogens of the olive tree and are, therefore, main target organisms for biocontrol in olive orchards.
The antagonism assays were conducted using dualculture tests. Petri dishes containing PDA medium were first inoculated with the conidia of an actively growing fresh fungal culture. The fungal suspension was standardized to approximately 10 6 conidia/ml using Malassez lame under microscope (Leitz Dialux 9 40 EB). They were obtained by scraping conidia from 15-day-old cultures of fungal isolates on PDA into an aqueous solution of 2% Tween 80 then filtered through cheesecloth to remove mycelium. 100 μl of each fungal suspension was spread onto the surface of the Petri dishes. Then, mycelial disc (10 mm in diameter) cut from the antagonistic fungal culture was positioned in the center of the plate and incubated at 25°C in the dark for 7 days. The assay was concluded when the pathogen's radial growth reached the antagonist. The only type of interaction monitored was the antibiosis determined by the presence of inhibition zones, which was measured and considered as indication for antifungal activity.

Screening for Entomopathogenic activity
The pathogenicity of fungal isolates was tested on the factitious host Ephestia kuehniella Zeller (Lepidoptera: Pyralidae). This moth was reared on whole wheat semolina in the Laboratory of Improvement and Protection of Olive Tree Genetic Resources, Olive Tree Institute (Sfax, Tunisia). Fourth instar larvae were used for fungal susceptibility tests. For each assay, 10 larvae were fed with 1 g of semolina mixed with 500 μl of a culture suspension (1 × 10 7 conidia ml − 1 ). The conidial suspension of each fungal strain was determined under the microscope using a Neubauer hemocytometer and adjusted to the cited concentration; a diet with 1 g sterile semolina mixed with 500 μl of sterile water was used as the control. Three replicates were performed for each fungal isolate, including the control [46]. The cumulative mortality was recorded by counting the number of dead larvae after 7 days of incubation under controlled conditions (temperature 22-25°C, 70 ± 10% RH, natural light). To satisfy Koch's postulates, the dead larvae were treated as follow: firstly, larval cadaver surfaces were sterilized with 5% sodium hypochlorite, rinsed twice and dried on filter paper, then incubated at 25°C under saturating conditions to allow the development of the tested fungi on the surface of the insect cadaver. The conformity of our results was justified after the tested fungi grew, were transferred to fresh PDA medium and incubated at 25°C. The obtained fungi were determined by microscopy (Leitz Dialux 40 EB).
Data were analyzed using ANOVA and values (mean ± standard error) were compared using Tukey test and differences were statistically significant at p < 0.05. Statistical analyses were performed using SPSS software for version 19.0 (SPSS Inc., Chicago).

Evaluation of Phytopathogenic activity
Phytopathogenic activity of fungal isolates against olive trees was evaluated using the excised shoot assays [63]. Shoots of approximately 250 mm length and 15 mm diameter were collected from olive trees. Shoot surfaces were sterilized for 10 min with 10% sodium hypochlorite solution rinsed with sterile water and dried using sterile paper towels. A plug of the same size (5 × 5 mm) of 7 days-old colony was incubated in the middle of the shoot after removing the cortex flap (5 × 8 mm) by a sterile scalpel. Five shoots were inoculated with each isolate. Inoculated shoots were then kept at 28-32°C, 70 ± 10% RH, natural light for 28 days. The assays were performed in 5 replicates. To reveal the trunk browning, the bark of the inoculated shoots was scraped with a sharp knife. Inoculated shoots were longitudinally cut at whole length into two halves in order to check wood browning in the internal vascular tissues. To comply with Koch's postulates, sections of brown tissues were cultured on PDA and incubated at 25°C for 7 days. Fungal cultures that grew up were identified microscopically (Leitz Dialux 40 EB).
Additional file 1: Supplementary Figure 1. Taous is an experimental orchard of the Olive Tree Institute, with an area of 126 ha. It contains olive, almond and pistachio fields. Olive cultivations include several varieties. In this study, insects were sampled from the variety Chemlali grown under artificial irrigation. The plantation density is 69 plants per hectare, with a spacing of 12x12m. Regular monitoring of the major olive pests is performed and organophosphate contact insecticides are applied if required. Torba is a private olive orchard with an area of 50 ha. Insects and soil were sampled from the olive cv. Chemlali cultivated under rainfed conditions. The density of plantation is 17 plants per hectare, with spacing of 24x24m. Regular monitoring of the major olive pests is performed and mass trapping is conducted if required. The Images depicted in Figure