Phytotoxic Metabolites Isolated from Neufusicoccum batangarum, the Causal Agent of the Scabby Canker of Cactus Pear (Opuntia ficus-indica L.)

Six phytotoxins were obtained from the culture filtrates of the ascomycete Neofusicoccum batangarum, the causal agent of the scabby canker of cactus pear (Opuntia ficus-indica L.) in minor Sicily islands. The phytotoxins were identified as (−)-(R)-mellein (1); (±)-botryoisocoumarin A (2); (−)-(3R,4R)- and (−)-(3R,4S)-4-hydroxymellein (3 and 4); (−)-terpestacin (5); and (+)-3,4-dihydro-4,5,8-trihydroxy-3-methylisocoumarin, which we named (+)-neoisocoumarin (6). This identification was done by comparing their spectral and optical data with those already reported in literature. The absolute configuration (3R,4S) to (+)-neoisocoumarin (6) was determined using the advanced Mosher method. All six metabolites were shown to have phytotoxicity on the host (cactus pear) and non-host (tomato) plants, and the most active compounds were (±)-botryoisocoumarin A (2), (−)-terpestacin (5), and (+)-neoisocoumarin (6).


Introduction
Cactus pear (Opuntia ficus-indica (L.) Mill.), Cactaceae family, is believed to be native of Mexico [1], and after the discovery of America, it was introduced into the Mediterranean Basin where it is now naturalized [2]. In Sicily, it has become an economically important fruit crop and a characteristic feature of the landscape. Cactus pear is also cultivated, mostly as productive living fences, in minor islands around Sicily, and as a fruit crop in Sardinia, Apulia, Calabria, and Basilicata Italian regions. The causal agent of the scabby canker of cactus pear found in minor islands of Sicily was identified as Neofusicoccum batangarum Begoude, Jol. Roux & Slippers, a fungal species not reported previously in Europe, whose distribution includes Africa, Brazil, and USA [4][5][6]. In Brazil, N. batangarum was reported as an aggressive pathogen of cashew (Anacardium occidentale) and cochineal cactus (Nopalea cochenillifera (L.) Salm-Dyck, syn. Opuntia cochenillifera (L.) Mill.), a relative of cactus pear [6,7]. On cochineal cactus in Brazil, N. batangarum, alone or in association with other fungi, including different species of Bothryosphaeriaceae, causes a disease that was named cladode brown spot and constitutes a serious threat to the cultivation of cochineal cactus as a fodder for livestock [7][8][9][10]. In preliminary cross-pathogenicity tests, the fungus isolated from infected cactus pear plants in minor islands of Sicily was able to reproduce the disease symptoms on the host plant and to induce a disease reaction on other plant species, such as Aleppo pine (Pinus halepensis), almond (Prunus dulcis), sweet orange (Citrus sinensis), citrange (Citrus sinesis Poncirus trifoliata), and holm oak (Quercus ilex) trees. These results show that N. batangarum has a wide host range like many other Botryosphaeriaceae species, which are frequently reported as causal agents of different important crop diseases, including the grapevine Botryosphaeria dieback [11][12][13][14][15].
The purpose of this study was to characterize the agent of the scabby canker of cactus pear for its ability to accumulate in culture biologically active substances, namely phytotoxins, which may have a role in the production of disease symptoms. The chemical identification of phytotoxins, obtained from the culture filtrates of N. batangarum, and their toxic effects on host and non-host plants are reported.

Results and Discussion
The organic extract of the N. batangarum culture was purified using both CC (Column Chromatography) and TLC (Thin Layer Chromatography) as reported in detail in the Materials and Methods, to produce five pure metabolites. They were identified, by comparing their 1   The causal agent of the scabby canker of cactus pear found in minor islands of Sicily was identified as Neofusicoccum batangarum Begoude, Jol. Roux & Slippers, a fungal species not reported previously in Europe, whose distribution includes Africa, Brazil, and USA [4][5][6]. In Brazil, N. batangarum was reported as an aggressive pathogen of cashew (Anacardium occidentale) and cochineal cactus (Nopalea cochenillifera (L.) Salm-Dyck, syn. Opuntia cochenillifera (L.) Mill.), a relative of cactus pear [6,7]. On cochineal cactus in Brazil, N. batangarum, alone or in association with other fungi, including different species of Bothryosphaeriaceae, causes a disease that was named cladode brown spot and constitutes a serious threat to the cultivation of cochineal cactus as a fodder for livestock [7][8][9][10]. In preliminary cross-pathogenicity tests, the fungus isolated from infected cactus pear plants in minor islands of Sicily was able to reproduce the disease symptoms on the host plant and to induce a disease reaction on other plant species, such as Aleppo pine (Pinus halepensis), almond (Prunus dulcis), sweet orange (Citrus sinensis), citrange (Citrus sinesis Poncirus trifoliata), and holm oak (Quercus ilex) trees. These results show that N. batangarum has a wide host range like many other Botryosphaeriaceae species, which are frequently reported as causal agents of different important crop diseases, including the grapevine Botryosphaeria dieback [11][12][13][14][15].
The purpose of this study was to characterize the agent of the scabby canker of cactus pear for its ability to accumulate in culture biologically active substances, namely phytotoxins, which may have a role in the production of disease symptoms. The chemical identification of phytotoxins, obtained from the culture filtrates of N. batangarum, and their toxic effects on host and non-host plants are reported.

Results and Discussion
The organic extract of the N. batangarum culture was purified using both CC (Column Chromatography) and TLC (Thin Layer Chromatography) as reported in detail in the Materials and Methods, to produce five pure metabolites.
(−)-(R)-Mellein (1) and its derivatives, including 3 and 4, belong to the class of isocoumarins and are widely found in the fungal kingdom [13,15,[19][20][21][27][28][29][30][31]; they have also been produced by plants and insects [32][33][34][35]. They are reported as phytotoxic metabolites of the fungi pathogen of forest and ornamental plants [36,37], and have been shown to possess a plethora of biological activities, including antiviral and antiparasitic activities [27,[38][39][40]. Recently, in an advanced review of phytotoxins produced by grapevine pathogens [15], 1, 3, and 4 are reported as known phytotoxic metabolites synthesized by Botryosphaeriaceae species, inducing grapevine trunk disease, and 1 has also been found in infected grapevine wood [20]. These results suggested the potential use of these compounds as predictive biomarkers for early recognition of the disease [15]. However, the role of 1, 3, and 4 in the pathogenic process has not yet been clarified. Considering the potential application, some isocoumarins, such as chenisocoumarin, have been tested for the control of noxious weeds [41,42].
(±)-Botryoisocoumarin A (2) was isolated, together with 3,8-dihydroxy-3-methylisochroman-1-one and 4,8-dihydroxy-3-methylisochroman-1-one from Botryosphaeria sp. F00741 [16]. Then, it was also obtained together with five other metabolites from the mangrove Kandelia candel. This is an endophytic fungus Botryopspheria sp. KcF6 studied during a screening carried out to find new metabolites for drug development. Compound 2 showed COX-2 inhibitory activity (IC50 = 6.51 µ M) but no cytotoxic activity. However, the structure drawn did not correspond to that of 2 [43]. Successively, 2 was isolated from the marine mangrove-derived fungus Aspergillus ochraceus, together with three new metabolites and another eleven known ones. In this study, some efforts were made to assign its absolute configuration (AC) at C-3 using X-ray methods but the results demonstrated the racemic nature of 2, which was then named (±)-botryoisocoumarin A [44]. Compound 2, together with a new acetate derivative and four already known compounds, were also isolated from Aspergillus westerdijkiae SCSIO 05233, a deep see fungus, but 2 did not show antibiotic or cytotoxic activities [45].
(−)-(R)-Mellein (1) and its derivatives, including 3 and 4, belong to the class of isocoumarins and are widely found in the fungal kingdom [13,15,[19][20][21][27][28][29][30][31]; they have also been produced by plants and insects [32][33][34][35]. They are reported as phytotoxic metabolites of the fungi pathogen of forest and ornamental plants [36,37], and have been shown to possess a plethora of biological activities, including antiviral and antiparasitic activities [27,[38][39][40]. Recently, in an advanced review of phytotoxins produced by grapevine pathogens [15], 1, 3, and 4 are reported as known phytotoxic metabolites synthesized by Botryosphaeriaceae species, inducing grapevine trunk disease, and 1 has also been found in infected grapevine wood [20]. These results suggested the potential use of these compounds as predictive biomarkers for early recognition of the disease [15]. However, the role of 1, 3, and 4 in the pathogenic process has not yet been clarified. Considering the potential application, some isocoumarins, such as chenisocoumarin, have been tested for the control of noxious weeds [41,42].
(±)-Botryoisocoumarin A (2) was isolated, together with 3,8-dihydroxy-3-methylisochroman-1-one and 4,8-dihydroxy-3-methylisochroman-1-one from Botryosphaeria sp. F00741 [16]. Then, it was also obtained together with five other metabolites from the mangrove Kandelia candel. This is an endophytic fungus Botryopspheria sp. KcF6 studied during a screening carried out to find new metabolites for drug development. Compound 2 showed COX-2 inhibitory activity (IC 50 = 6.51 µM) but no cytotoxic activity. However, the structure drawn did not correspond to that of 2 [43]. Successively, 2 was isolated from the marine mangrove-derived fungus Aspergillus ochraceus, together with three new metabolites and another eleven known ones. In this study, some efforts were made to assign its absolute configuration (AC) at C-3 using X-ray methods but the results demonstrated the racemic nature of 2, which was then named (±)-botryoisocoumarin A [44]. Compound 2, together with a new acetate derivative and four already known compounds, were also isolated from Aspergillus westerdijkiae SCSIO 05233, a deep see fungus, but 2 did not show antibiotic or cytotoxic activities [45].
(+)-3,4-Dihydro-4,5,8-trihydroxy-3-methylisocoumarin (6) was previously obtained, together with two already known compounds, from Phomopsis sp. (No. ZH-111) during a screening aimed at finding new compounds from endophytes of the South China Sea [26]. When assayed on zebrafish, the tetrasubstituted-3,4-dihydroisocoumarin significantly accelerates the growth of vessels but showed only weak cytotoxicity on the two cancer cell lines tested [26]. However, Yang et al. [26] assigned only the 3R*,4S* relative configuration to this compound based on the correlation measured in the NOESY spectrum. After recording the NOESY spectrum of 6 in the same conditions, we observed the expected correlations between H-6 and H-7, H-3 and Me-11, and H-4 and Me-11, and thus 6 had the same relative configuration that was assigned by Jang et al. [26].
Thus, to assign the absolute configuration at C-3 and C-4, 6 was firstly converted in the corresponding 5,8-O,O -dimethyl ether derivative (7), whose spectroscopic data ( 1 H and 13 C NMR and ESI MS) were fully consistent with the structure of 6. The 1 H NMR spectrum of 7 differed from that of 6 due to the two methoxy groups at The reaction mechanism that could explain this result is reported in Figure 3. The driving force of the reaction was the hydroxyl anion generated from KOH where pyridine was preserved in non-dry conditions. Toxins 2020, 12, x FOR PEER REVIEW 4 of 13 (+)-3,4-Dihydro-4,5,8-trihydroxy-3-methylisocoumarin (6) was previously obtained, together with two already known compounds, from Phomopsis sp. (No. ZH-111) during a screening aimed at finding new compounds from endophytes of the South China Sea [26]. When assayed on zebrafish, the tetrasubstituted-3,4-dihydroisocoumarin significantly accelerates the growth of vessels but showed only weak cytotoxicity on the two cancer cell lines tested [26]. However, Yang et al. [26] assigned only the 3R*,4S* relative configuration to this compound based on the correlation measured in the NOESY spectrum. After recording the NOESY spectrum of 6 in the same conditions, we observed the expected correlations between H-6 and H-7, H-3 and Me-11, and H-4 and Me-11, and thus 6 had the same relative configuration that was assigned by Jang et al. [26].
Thus, to assign the absolute configuration at C-3 and C-4, 6 was firstly converted in the corresponding 5,8-O,O′-dimethyl ether derivative (7), whose spectroscopic data ( 1 H and 13 C NMR and ESI MS) were fully consistent with the structure of 6. The 1 H NMR spectrum of 7 differed from that of 6 due to the two methoxy groups at δ 3. with that of 7. The reaction mechanism that could explain this result is reported in Figure 3. The driving force of the reaction was the hydroxyl anion generated from KOH where pyridine was preserved in non-dry conditions.  (Table 1), the Δδ (8-9) values for all the protons were determined and reported in Figure 4.  (Table 1), the ∆δ (8-9) values for all the protons were determined and reported in Figure 4.   1 The chemical shifts are in values (ppm) from TMS (Tetramethylsilane). 2 These two signals could be exchanged.  Appling model A as reported in Cimmino et al. [48], the (R) configuration was assigned at C-3, and consequently, on the basis of NOESY data, the (S) one was at C-4. Then, 6 was formulated as (+)-(3R,4S)-3,4 dihydro-4,5,8-trihydroxy-3-methylisocoumarin and named (+)-neoisocumarin.
All the isolated compounds (1-6) were screened for phytotoxic activity as described in detail in Materials and Methods. All phytotoxins tested at the highest concentrations induced necrosis around inoculation points after 7 days, on cladodes of cactus pear ( Figure 5), as well as on tomato leaves ( Figure 6). However, 1, 3, and 4 were only phytotoxic at the highest concentrations on cactus pear and on non-host tomato plants (Tables 2 and 3). All three metabolites were found to be produced by N. parvum, another species in the genus, and other fungi in the Botryospaeriaceae family associated with grapevine trunk diseases, confirming that they may be virulence factors in plant diseases caused by these fungi, although their role in the pathogenesis is still controversial [19,20]. (±)-Botryoisocoumarin A (2), (+)-neoisocoumarin (6), and (−)-terpestacin (5) showed phytotoxic activity, both on tomato (non-host) and cactus pear (host) plants in biological assays (Tables 2 and 3 for 2 and 6). The metabolites 2, 5, and 6 thus proved to have almost the same spectrum of phytotoxic activity as they showed a comparable activity against host and non-host plants. Appling model A as reported in Cimmino et al. [48], the (R) configuration was assigned at C-3, and consequently, on the basis of NOESY data, the (S) one was at C-4. Then, 6 was formulated as (+)-(3R,4S)-3,4 dihydro-4,5,8-trihydroxy-3-methylisocoumarin and named (+)-neoisocumarin.
All the isolated compounds (1-6) were screened for phytotoxic activity as described in detail in Materials and Methods. All phytotoxins tested at the highest concentrations induced necrosis around inoculation points after 7 days, on cladodes of cactus pear ( Figure 5), as well as on tomato leaves ( Figure 6). However, 1, 3, and 4 were only phytotoxic at the highest concentrations on cactus pear and on non-host tomato plants (Tables 2 and 3). All three metabolites were found to be produced by N. parvum, another species in the genus, and other fungi in the Botryospaeriaceae family associated with grapevine trunk diseases, confirming that they may be virulence factors in plant diseases caused by these fungi, although their role in the pathogenesis is still controversial [19,20]. (±)-Botryoisocoumarin A (2), (+)-neoisocoumarin (6), and (−)-terpestacin (5) showed phytotoxic activity, both on tomato (non-host) and cactus pear (host) plants in biological assays (Tables 2 and 3  The results showed that (±)-botryoisocumarin A (2), (−)-terpestacin (5), and (+)-neoisocoumarin (6) were by far more phytotoxic than melleins (1, 3, and 4) when tested on cladodes of cactus pear and tomato leaves. Compounds 2 and 6 were the most active at concentrations in a range from 10 −3 M to 10 −4 M, inducing a necrosis area around the inoculation points in both host and non-host plants, followed by 5. The ability of 2, 5, and 6 to induce large necrotic lesions on cactus pear cladodes  The results showed that (±)-botryoisocumarin A (2), (−)-terpestacin (5), and (+)-neoisocoumarin (6) were by far more phytotoxic than melleins (1, 3, and 4) when tested on cladodes of cactus pear and tomato leaves. Compounds 2 and 6 were the most active at concentrations in a range from 10 −3 M to 10 −4 M, inducing a necrosis area around the inoculation points in both host and non-host plants, followed by 5. The ability of 2, 5, and 6 to induce large necrotic lesions on cactus pear cladodes The results showed that (±)-botryoisocumarin A (2), (−)-terpestacin (5), and (+)-neoisocoumarin (6) were by far more phytotoxic than melleins (1, 3, and 4) when tested on cladodes of cactus pear and tomato leaves. Compounds 2 and 6 were the most active at concentrations in a range from 10 −3 M to 10 −4 M, inducing a necrosis area around the inoculation points in both host and non-host plants, followed by 5. The ability of 2, 5, and 6 to induce large necrotic lesions on cactus pear cladodes suggests Toxins 2020, 12, 126 7 of 13 that these toxins may be involved in the scabby cankers disease syndrome. These results also indicate the ability of fungi in the Botryosphaeriaceae family to produce phytotoxins that are active against non-taxonomically related plants, such as cactus pear and tomato, and could explain the wide host spectrum of many species in the family. Moreover, it can be speculated that the allelopathic, antifungal, and antibacterial activity of terpestacin demonstrated in previous studies [25,44] might enhance the ecological fitness of N. batangarum and might explain the ability of this fungus to rapidly colonize the cactus pear cladode and sporulate on infected tissues [3] before they are invaded by other saprophytes or opportunistic weak pathogens of the plant biosphere. Due to its allelopathic activity, N. batangarum was indicated as a potential biocontrol agent [49,50]. In a recent review, Masi et al. [51] questioned the possibility of using terpestacin as a biopesticide because of its toxicity and stressed that, although the functions of this mycotoxin in nature have not been clearly established, its allelopathic activity would suggest a role in eliminating other microorganisms competing in the same environment. Table 2. Phytotoxicity on tomato of 1-6 tested at various concentrations using the leaf puncture assay and evaluated based on the size of the necrotic lesions.   The absolute configuration of (+)-neoisocoumarin was determined by using the advanced Mosher's method. Considering the phytotoxic activity on both host and non-host test plants, it can be deduced that all these metabolites were involved in the syndrome of scabby cankers disease of cactus pear, and like other Botryosphaeriaceae, N. batangarum has a wider host range than previously thought.

General Experimental Procedures
A Jasco P-1010 digital polarimeter (Jasco, Tokyo, Japan) was used to record the optical rotations in CHCl 3 , or as stated otherwise. Electrospray ionization (ESI) mass spectrometry and liquid chromatography/mass spectrometry (LC/MS) analyses were performed using an LC/

Fungal Strain
The strains of N. batangarum used in this study were collected from infected host plant tissues on the islands of Favignana, Lampedusa, Linosa, and Ustica from 2013 to 2018 and maintained in potato dextrose agar (PDA, Fluka, Sigma-Aldrich Chemic GmbH, Buchs, Switzerland) and stored at 4 • C in the strain collection of the Dipartimento di Agricoltura, Alimentazione e Ambiente, Università di Catania, Catania, Italy. N. batangarum was obtained from scabby cankers on cladodes of cactus pear (Opuntia ficus-indica L.) recovered from minor islands of Sicily (Lampedusa, Linosa, Favignana, and Ustica). Four representative isolates, one from each island, were deposited at CBS-KNAW Biodiversity Centre, strains code nos. CBS143023, CBS143024, CBS143025, and CBS143026 [3].

Production, Extraction, and Isolation of Secondary Metabolites
The isolate CBS143023 (Neofusicoccum batangarum) was grown in liquid culture to obtain the liquid filtrate. In detail, the mycelium of fungal cultures, grown on potato dextrose agar for 5-7 days at 25 • C, was homogenized with sterile distilled water. Three milliliters of this suspension was distributed individually into 1 L Roux flasks containing 170 mL of modified Difco™ Czapeck-Dox (Benton, Dickinson and Company, Sparks, MD, USA) medium with 0.5% yeast and 0.5% malt extract (pH 5.75) and incubated at 25 • C for 28 days in the dark [15]. At harvest, the liquid cultures were filtered, initially with a double layer of gauze to reduce the fungal biomass, and subsequently with suction filters using porous membrane filter complexes Stericup Millipore ® (pore diameter = 0.22 µm). A total of 20 L of culture filtrate was obtained and stored at −20 • C until use. It was concentrated under reduced pressure at 35 • C to 1 L and extracted with EtOAc (3 × 1 L). The combined organic extracts were dried (Na 2 SO 4 ) and evaporated under reduced pressure. The oily residue (1.84 g) was fractioned via column chromatography eluted with chloroform/iso-propanol (9:1), obtaining twelve groups of homogeneous fractions. The residues (24.4 mg and 23.1 mg) of the first and second fractions were combined and further purified using preparative TLC eluted with hexane/ethyl acetate (8.5:1.5), yielding two pure metabolites as amorphous solids that were identified as (−)-(R)-mellein (1, 8.

Biological Assays
The metabolites isolated (1-6) were tested on young cladodes of host plant cactus pear (Opuntia ficus-indica (L.) Mill.) and on non-host tomato plant (Solanum lycopersicum L.) leaves. Pure metabolites were first dissolved in methanol, and then diluted with distilled water (final concentration of methanol, 4%) up to the desired concentrations. For each metabolite, 50 µL of the solution were pipetted into cladodes at concentrations of 0.05, 0.1, 0.25, 0.5, or 1 mg/mL. The phytotoxicity of the pure metabolites of N. batangarum was also tested using a puncture assay on tomato leaves. A droplet (10 µL) of each metabolite, at concentrations of 0.05, 0.1, 0.25, 0.5, or 1 mg/mL, was placed on the leaf lamina previously punctured by a needle. One cladode of a single plant of cactus pear and four leaves of tomato were used as replicates, and the test was repeated twice. Methanol (4% v/v) and sterile distilled water were used as controls. Symptoms of phytotoxicity of inoculated cladodes and leaves, kept in a climatic chamber under controlled conditions, were observed each day for 7 days. The size (mm 2 ) of necrotic the area surrounding the punctures was measured.