The Inactivation by Curcumin-Mediated Photosensitization of Botrytis cinerea Spores Isolated from Strawberry Fruits

Photosensitization is a novel environmentally friendly technology with promising applications in the food industry to extend food shelf life. In this study, the natural food dye curcumin, when combined with visible light (430 nm), was shown to be an effective photosensitizer against the common phytopathogenic fungi Botrytis cinerea (the cause of grey mould). Production of the associated phytotoxic metabolites botrydial and dihydrobotrydial was measured by our newly developed and validated HRAM UPLC-MS/MS method, and was also shown to be reduced by this treatment. With a light dose of 120 J/cm2, the reduction in spore viability was directly proportional to curcumin concentrations, and the overall concentration of both botrydial and dihydrobotrydial also decreased with increasing curcumin concentration above 200 µM. With curcumin concentrations above 600 µM, the percentage reduction in fungal spores was close to 100%. When the dye concentration was increased to 800 µM, the spores were completely inactive and neither botrydial nor dihydrobotrydial could be detected. These results suggest that curcumin-mediated photosensitization is a potentially effective method to control B. cinerea spoilage, and also to reduce the formation of these phytotoxic botryane secondary metabolites.


Introduction
Strawberries are one of the most popular fruits consumed worldwide, with global production in excess of 9M tons per annum and an estimated annual production value of USD 18B worldwide [1]. These fruits are rich sources of antioxidants and vitamin C [2], but their commercial value is restricted by limited shelf-life due largely to susceptibility to microbial infections, particularly Botrytis grey mould. Botrytis cinerea is the causal agent of this disease, which effects not only strawberries, and is considered one of the most important phytopathogens of horticultural crops. This fungus is found in more than 200 ornamental and agriculturally important plants in the world, and ranks second in the list of 'top 10' fungal plant pathogens [3].
Antimicrobial photosensitization is an effective and promising approach to elicit cell death and kill microorganisms such as Gram-negative and Gram-positive bacteria and yeasts, as well as parasites and viruses [4][5][6]. The basic principle of photosensitization is a photochemical process in which a photosensitizer dye such as crystal violet, methylene blue and safranin O are energized to an unstable singlet excited-state after absorbing light photons, and then loses its excess energy to produce free radicals including superoxide and hydroxyl radicals (Type I pathway) and singlet oxygen (Type II pathway). These reactive oxygen species (ROS) cause oxidative damage to biomolecules including amino acids, lipids, nucleic acids, with ensuing cell death [7,8]. The application of natural plant additives such as curcumin as photosensitizer is considered a clean and green technology with minimal reported adverse impacts [9]. By comparison fungicides, and other methods, such as irradiation, biocontrol agents, etc., that are currently used to extend food shelf life, show unfavourable side effects that can impact on consumer health. Photosensitization is thus being investigated as a novel technology to accomplish microbiological decontamination and extend the shelf life of food in an environmentally friendly way.
B. cinerea has many unique characteristics in its phytopathological behaviour, including the capacity to kill host cells through both the production of toxins and also oxidative burst (reactive oxygen species) generated by both the host and the pathogen [10]. This fungus is both a pathogenic and saprophytic organism, initially infecting the weak or dead parts of the plant, and then extending to the rest of healthy plant tissue [11]. B. cinerea produces characteristic fungal metabolites based on the botryane skeleton, principally botrydial (1) and its metabolite dihydrobotrydial (2) (Figure 1), which have been isolated from fungal culture medium [12]. Botrydial is a particularly powerful phytotoxin, produced during plant infection, which can induce chlorosis and cell collapse. It has been demonstrated to have phytotoxic and cytotoxic activity at ID 50 as equal to or lower than 5 µg/mL and antibiotic activity at 100 ppm [13]. A further B. cinerea phytotoxin botcinic acid has also been studied and shown to have a redundant role in virulence with botrydial [14]. Even though photosensitization has been demonstrated to have significant antimicrobial efficiency against an array of microorganisms, there is limited literature available as to whether photosensitization will inhibit the growth of B. cinerea, and moreover the impacts of photosensitization on the production of the phytotoxic metabolites botrydial and dihydrobotrydial. Previous studies have investigated the antifungal effect of 405 nm light at 60 mW cm −2 for 24 h on this phytopathogen [15]. These authors suggested that the excitation of endogenous porphyrins and subsequent accumulation of singlet oxygen contribute to the light-mediated photoinactivation of grey mould, and suggested such treatment as a means of controlling plant disease. Such lengthy irradiation would, however, not be appropriate with a food commodity such as strawberries, and moreover these authors did not examine the impacts of this treatment on the phytotoxic metabolites.
In this study, B. cinerea was isolated from Queensland (Australia) grown strawberries and used to evaluate the effects of photosensitization on the growth of this fungus under a range of experimental conditions including different concentrations of photosensitizer (curcumin), incubation time, and light dose with irradiation periods of 10-30 min. At the same time, a new high-resolution accurate mass (HRAM) UPLC-MS/MS method was developed and validated to study the relative formation of the metabolites botrydial and dihydrobotrydial under these photosensitization conditions.

Extraction of Botrydial and Dihydrobotrydial from B. cinerea Cultures
Solvent extraction of B. cinerea cultures obtained from Queensland strawberries provided a mixed sample of the fungal metabolites botrydial and dihydrobotrydial after repeated silica chromatography. Unfortunately, attempts to further purify these compounds resulted in further degradation of these metabolites. Botrydial and dihydrobotrydial are known compounds for which both 1 H and 13 C NMR have been reported [16,17], and literature data were used to both confirm the identity of the two components in our extract and also enabled quantification of the amount of each component within the mixed sample. Integration of 1 H NMR resonances with a measured aliquot of dioxane as internal standard were used to determine the amount of botrydial/dihydobotrydial present in the sample. The isolated mixed standard was determined to contain botrydial and dihydrobotrydial in a 1:9 ratio, and used without further purification as a mixed external standard in the development of and validation of a new HRAM UPLC-MS/MS analysis method for both metabolites.

HRAM UPLC-MS/MS Analysis of Botrydial and Dihydrobotrydial
HRAM UPLC-MS/MS analysis of the mixed botrydial/dihydobotrydial standard provided two peaks with molecular ions consistent with the molecular formula of the two components botrydial (1)  The effect of increasing curcumin dye (D) concentrations with a photosensitization light dose (L) of 120 J/cm 2 on B. cinera growth on agar plates incubated at 26 • C for 8 days was investigated. The B. cinerea control D − L − treatment (D − = no curcumin; L − = no light) and D − L + treatment (no curcumin, light dose 120 J/cm 2 ) produced characteristic prolific greyish-brown spores, and visible differences in fungal growth was readily apparent with increasing curcumin concentrations after photosensitization treatment and incubation for 8 days ( Figure S2). With increasing curcumin concentration, the growth of B. cinerea decreased. When treated with a curcumin concentration above 800 µM and a light does of 120 J/cm 2 , the fungus was not able to survive ( Figure S2). As shown in Figure 3, a significant percentage (p < 0.001) reduction in the spores was also observed when the spores were treated with curcumin-mediated photosensitization. Treatment with light or curcumin alone (D − L + or D + L − ) was not effective. The curcumin-mediated photosensitization (D + L + ) reduction in spore germination was directly proportionate to curcumin concentrations, with increasing concentrations of curcumin photosensitizer acting to drastically inhibit B. cinerea conidia germination. With curcumin concentrations above 600 µM and light dose of 120 J/cm 2 , spore germination was completely (100%) inhibited (p < 0.001) ( Figure 3).  . Effect of curcumin-mediated photosensitization with different curcumin concentrations (0-1000 µM) and light dose (120 J/cm 2 ) on the concentration of botrydial and dihydrobotrydial in spores after incubation at 26 • C for 11 days (n = 6). The significances of difference in concentration of (a) botrydial and (b) dihydrobotrydial and ratio of (c) dihydrobotrdial/botrydial between different curcumin concentrations and light-only treatment (D − L + , curcumin = 0 µM) p < 0.05 (*), and compared to the control (D − L − ) p < 0.05 ( # ), tested with t-test.

Effect of Curcumin-Mediated Photosensitization on the B. cinerea Spore Using Different Light Doses
As shown in Figure S3, when treated with a light dose of 120 J/cm 2 and curcumin concentration of 800 µM, no spore germination was observed, which indicates that the spores were completely inactivated by this photosensitization mediated curcumin treatment. By comparison, the spores grew well after treatment with 800 µM curcumin without light (D + L − ) and under control conditions (D − L − ). The percentage reduction in total spores reached 100% with the curcumin-mediated light doses of 120, 240, and 360 J/cm 2 ( Figure 5). Similarly, neither botrydial nor dihydrobotrydial could be detected when light and dye (D + L + ) are used at the same time with light doses of 120, 240 and 360 J/cm 2 ( Figure 6). The botrydial concentration was observed to increase slightly from 6.2 ± 2.2 to 8.5 ± 0.9 µg/g in the curcumin-only treatment (D + L − ), comparing to the control (D − L − ) (p < 0.05). However, the dihydrobotrydial concentration was not significantly different between the curcumin-only treatment (D + L − ) and the control (D − L − ) ( Figure 6).

The Growth of B. cinerea over Different Incubation Time After Photosensitization
In studies of effect of different treatments on B. cinerea with increasing incubation time ( Figure S4), it was observed that in the control treatment (D − L − ) mycelial growth was fast with white hyphae apparent within two days incubation. With prolonging incubation time, the three treatments (D − L − ; D − L + ; D + L − ) display the expected regular fungi morphology. In contrast, with the curcumin treatment (D + L + ), the spores did not germinate during the 16 days incubation. As shown in Figure 7, in general the light-only and dye-only treatments (D − L + ) and (D + L − ) demonstrated a general increase in production of both secondary metabolites reaching 8.31 ± 0.54 and 11.18 ± 1.53 µg/g of botrydial, and 179.69 ± 30.04 and 205.29 ± 53.71 µg/g of dihydrobotrydial, respectively, at 15 days. The control treatment (D − L − ) produced a more variable botrydial concentration ( Figure 7) with a maximum of 7.41 ± 3.56 µg/g. At all incubation times, all samples from treatment (D + L + ) were below limits of detection for both botrydial and dihydrobotrydial (Figure 7), which is most likely due to the destruction of all B. cinerea spore by this treatment. In order to eliminate interference due to variability between plates, the ratio of dihydrobotrydial/botrydial was also calculated. The average metabolite ratio in the three treatments (D − L − ; D − L + ; D + L − ) is 19.42 ± 3.62, 21.83 ± 3.52 and 15.34 ± 3.20, respectively.

LC-MS/MS Analysis of Metabolites Botrydial and Dihydrobotrydial
High-resolution accurate mass (HRAM) spectrometers, such as the time-of-flight (TOF) and the Orbitrap spectrometer, has become more accessible and affordable to many laboratories in recent years. One of the biggest strengths of a HRAM instrument is its ability to acquire accurate mass data, which allows the user to relate the observed accurate mass to an equivalent elemental composition/s and molecular formula/s. Molecular formulas alone, however, are not sufficient to identify specific compounds. Without applicable standards, erroneous identifications can result due to the many possible compounds with the same molecular formula, and for this reason we have chosen to isolate authentic botrydial/dihydrobotrydial to use as standards in this study. These metabolites are not commercially available but can be obtained from B. cinerea cultures as described here.
Literature reports of UPLC-MS/MS analysis of similar authentic botrydial standards are lacking, but there are two recent literature reports on the detection of botrydial through the untargeted HRAM approach. Niedzwiecki et al. [18]  Notably, no loss of AcOH was observed, which in itself is inconsistent with botrydial structure (as is the loss of 2 × CO without loss of H 2 O). The MS 2 fragmentations of the metabolite detected by El Fellah et al. [19] is seemingly different to that of our authenticated botrydial (Figures 2 and S1), and casts doubt on the water analysis results published in their report. Similarly the [M+Na] + and [M+NH 4 ] + molecular ions reported by Niedzwiecki et al. [18] are not unique to botrydial and may or may not be actually botrydial.
Coincidentally, both the study by El Fellah et al. [19] and the current investigation were conducted on the Orbitrap mass spectrometer platform with similar elution solvent systems. We, however, find the HRAM approach not ideal for the detection of botrydial (without standard) as accurate mass fragmentation data in the literature is lacking to support the direct identification of botrydial by accurate mass alone. During our own investigations we frequently noted a second background contaminant in our samples and/or in our LCMS system which unexpectedly gives an accurate mass which is in agreement with the [M+H] + molecular ion of botrydial under 5 ppm mass errors. This background compound did not co-elute with our authentic botrydial standard, nor did it exhibit the same MS fragmentation as our authenticated botrydial despite apparently having the same molecular formula accordingly to HRAM data. Consequently, without a botrydial standard we find it unreliable to identify botrydial based purely on the untargeted accurate mass data. Interestingly, dihydrobotrydial, which is a detoxification by-product of botrydial, was not reported from the two previous HRAM studies [18,19]. With the obvious gap in the untargeted HRAM detection of botrydial, we have isolated the two diagnostic metabolites, botrydial and dihydrobotrydial, from B. cinerea and have provided here detailed HRAM and MS 2 accurate mass data for future comparison (Figures 2 and S1). The new analysis method described here also demonstrated good reproducibility and reliability in recoveries of both botrydial and dihydrobotrydial from spiked agar plates (Table 1), enabling the use of this method to study effects of photosensitization on B. cinerea metabolite production.

Effect of Curcumin-Mediated Photosensitization on B. cinerea
This study demonstrated that the use of curcumin followed by irradiation with blue light (430 nm) promoted a significant reduction in the germination of B. cinerea. This result is in agreement with previous studies that have reported similar effects of curcumin and other photosensitizers with different food-borne pathogens. For example, Penha et al. [4] reported that curcumin (75 µM) and blue illumination (470 nm) had synergistic effect on inactivation of food-borne pathogens with relative order of efficacy Aeromonas hydrophila > Salmonella aureus > Escherichia coli > Salmonella typhimurium. Josewin et al. [20] similarly utilized blue light-emitting diodes (405-460 nm) with sodium chlorophyllin as photosensitizer to inactivate Listeria monocytogenes and Salmonella spp., minimizing the contamination risk of food-borne pathogens by consumption of cantaloupe. These authors achieved similar reductions with and without added photosensitizer on cantaloupe rind and speculated about the presence of an endogenous photosensitizer [20]. Our research group found that curcumin combined with visible light (420 nm) was an effective treatment to inactivate spores of Aspergillus flavus both in vitro and in vivo in maize kernels [5,21]. In subsequent studies, the percentage reduction in Aspergillus niger, A. flavus, Penicillium chrysogenum and Zygosacharomyces bailii remained above 50% with curcumin concentrations (100 to 400 µM) followed by illumination with a light dose of 96 J/cm 2 [22]. In the present study, at low curcumin concentration (50 to 400 µM) and under a similar light dosage, the percentage reduction in B. cinerea varied from 21.5 to 99.7%, which may reflect the differing susceptibility of the fungal species studied. Penicillium griseofulvum by comparison demonstrated less susceptibility in previous studies [22], than that recorded here for B. cinerea. In all cases, the phototoxic effect seems to depend on the curcumin concentration, which is in agreement with the previous reports [23]. Fungi are usually enveloped by a thick rigid cell wall, which is made of chitin, mannoproteins and α-β glucans, and photoinactivation is therefore dependent on photosensitizer uptake through this cell wall and distribution to subcellular targets [24]. The phototoxicity of curcumin is then directed through reactive oxygen species formed by the reaction between the light-activated dye molecules and the excited states of oxygen molecules [25]. Sanita et al. [26] reported that photodynamic therapy mediated by curcumin associated with LED light significantly reduced the metabolism of the biofilm organized cells of Candida dubliniensis. They also reported that exposure to curcumin alone in the highest concentrations (30 and 40 µM) significantly reduced the biofilm viability. Further, while curcumin was rapidly taken up by Candida dubliniensis cells, curcumin was reported to require a longer time interval to penetrate into biofilm cells [26]. Araujo et al. [23] reported that curcumin has a toxic effect on cariogenic pathogens only at appreciable concentrations (5.0 g/L) upon photoactivation. Providentially, curcumin is a food additive (E100) and a naturally polyphenol extracted from Curcuma longa with no significant reported toxicity, well-known for its beneficial biological activities, including antimicrobial anti-proliferative, anti-inflammatory and antioxidant activities [27,28]. From an economic viewpoint, curcumin can be produced in high quantities at a reasonable cost, so it is an ideal photosensitizer. However, curcumin is a very hydrophobic molecule and requires some kind of formulation vehicle to enable higher curcumin load and improve interaction between dissolved curcumin and cell membranes [29]. For consumer acceptance any change in flavour and colour of the treated commodity as a result of curcumin addition would also need to be considered.
The present study also evaluated the effect of different exposure time or light dose on curcumin-mediated photosensitization of B. cinerea and demonstrated that only a short 10 min period of illumination was required. The D + L + treatment achieved complete inactivation of B. cinerea when utilizing 800 µM curcumin and illumination time of 10 min, equivalent to a light dose of 120 J/cm 2 ( Figure 5). This result is analogous to previous studies with other microbial species. For example, Penh et al. [4] reported that with increasing the illumination time (10 to 30 min) or light does (139 to 417 J/cm 2 ), curcuminmediated photosensitization induced a significant reduction in the counts of Gram-positive and Gram-negative bacteria, such as S. aureus, Pseudomonas aeruginosa and Aeromonas hydrophila. Similarly, Araujo et al. [23] reported that a curcumin illumination time of 2 or 5 min was equally effective in reducing Streptococcus mutans and Lactobacillus acidophilus. It is, however, worth noting that some researchers have observed photobleaching with consequent decreases in effectiveness with increased photosensitization illumination times. For example, in evaluation of photodynamic therapy using a curcumin solution on root canals contaminated with Enterococcus faecalis, curcumin as sensitizer was effective with 5 min irradiation but not with 10 min irradiation, which the authors related to curcumin photobleaching [30].
This current study aimed at the identification of the most effective curcumin-mediated photosensitization treatment against B. cinerea. Therefore, the B. cinerea morphology was observed for a period of 16 days after treatment with a light does of 120 J/cm 2 and curcumin concentration of 800 µM. In comparison to curcumin-only or light-only treatment (D + L − and D − L + ), this study demonstrated that curcumin-mediated photosensitization (D + L + ) not only influences the mycelial growth of B. cinerea, but also causes a delay or inhibition in the germination of the fungal spore population. Curcumin has been reported to exhibit antifungal activity against Cryptococcus neoformans, Phytophthora infestans, Rhizoctonia solani, Candida albicans, and Erysiphe graminis [31,32]. In addition, phototherapy has been suggested as a potential therapeutic alternative to antifungal treatment for the treatment of C. albicans biofilm infections. Many studies have demonstrated that blue light (a wavelength of 400-500 nm) alone exhibits significantly antimicrobial effects against methicillin-sensitive S. aureus [33], Acinetobacter [34] and C. albicans [35]. However, in the present study, neither curcumin nor light alone was effective in eliminating B. cinerea, which is a similar conclusion to the previous reports with A. flavus [21].
The exact mechanism of the photosensitizer inducer photodynamic inactivation has not been elucidated. However, many researchers believe that the photosensitization method is based on combined action of photosensitizer, visible light, and oxygen, producing the observed cytotoxic effect [36], in which light-activated photosensitizers transfer energy to oxygen to form reactive oxygen species and other free radicals which inactivate microorganisms by damaging proteins, cell membranes, and organelles [37].
In the present study, the effects of photosensitization on B. cinerea phytotoxin production were also considered. Phytotoxins, including alkaloids, terpenes, polyketides, non-ribosomal peptides, or metabolites of mixed biosynthetic origin, are secreted by necrotrophic pathogens to induce cell necrosis and leakage of nutrients [38]. B. cinerea is one of most broad-host range fungal necrotrophs, and infects a large number of vegetables and fruit crops with subsequent economic loss both pre-and post-harvest. B. cinerea produces phytotoxins, ROS, as well as cell wall-degrading enzymes to induce necrosis of plant tissues. B. cinerea is also reported to trigger hypersensitive responses and ROS production in the host, which cause a series of programmed cell death to promote the infection process [39]. Botrydial is the primary phytotoxic sesquiterpene metabolite secreted by B. cinerea, and triggers ROS production, hypersensitive response, and expression of defence genes. Therefore, it is considered that botrydial may act not only as a phytotoxin, but also as an elicitor of plant defence responses [40,41]. For this reason, botrydial, as well as other B. cinerea-originated sesquiterpene metabolites, is regarded as an economically important fungal toxin for agriculturally and ornamental crops [11]. Although dihydrobotydial has only moderate phytotoxic activity at high concentrations [40], it is one of the major B. cinerea metabolites and co-occurs with botrydial during isolation [13]. The inhibition of the growth of B. cinerea has also been reported to be directly proportional to the botrydial concentration, with B. cinerea transforming botrydial to the less active phytotoxins such as dihydrobotrydial, botryenedial, and secobotrytrienediol [42]. Therefore, both botrydial and dihydrobotrydial were chosen in the present study as detection index to investigate the function of our photodynamic treatment.
Measured metabolite concentrations in the present study did not exactly mirror the effects of photodynamic degradation of B. cinerea spores. With increasing concentration of curcumin between 0 and 200 µM, the botrydial concentration remained relatively unchanged despite the inhibition of fungal growth (Figure 4). However, when the concentration of curcumin was raised above 200 µM, the botrydial concentration decreased and the fungi growth was also inhibited. Interestingly the production of dihydrobotrydial was more impacted at the lower curcumin concentrations. One explanation is that fungal growth may cease or be inhibited as a result of relatively higher botrydial with lesser colonies and with the increase in the concentration of curcumin. Duran-Patron et al. [42] similarly commented on the complexity of botrydial degradation pathways, with botrydial production during initial fungal growth stages regulating growth at higher concentrations, and metabolism to less inhibitory compounds, such as dihydrobotrydial, allowing growth to resume. It is noted that the botrydial concentrations measured in our agar plates (Figure 7) are less than one tenth of the maximal botrydial levels reported in the liquid culture studies of Duran-Patron et al. [42], and have not reached the levels reported by Duran-Patron et al. to regulate fungal growth. On the other hand, light-only (D − L + ) or curcumin-only (D + L − ) treatments may also effect the B. cinerea metabolite production, as seen in the varied ratio of dihydrobotrydial and botrydial produced (Figure 7). Lineiro et al. [43] found that there is a connection between gene expression of virulence factors and B. cinerea culture and environmental conditions, whereby the botrydial biosynthesis can be inhibited when the culture uses a sole carbon, such as cellulose and tomato cell walls. Similarly, when the spore suspension was treated with either light or curcumin in this study, this change in B. cinerea culture conditions could lead to a changed relationship between the metabolites botrydial and dihydrobotrydial (and consequently the ratio of these metabolites).
In the food industry, the most widely employed postharvest fruit protection methods to reduce microbial spoilage include chemicals and physical methods such as controlled atmosphere, low temperature or modified atmosphere packaging. Photosensitization as an environmentally friendly method, represents a potential novel technology with application in food and beverage industries. De Oliveira et al. [44] reported that combination of UV-A light and curcumin can significantly reduce bacterial cross-contamination of fresh produce. Tao et al. [45] reported curcumin-based photosensitization inactivate E. coli on the surface of apple slices and preserve the quality of apple slices. Our in vitro study has demonstrated the effectiveness of curcumin-mediated photosensitization against B. cinerea, a common fruit and vegetable pathogen. Further studies are required to investigate the effectiveness of this treatment against B. cinerea in fresh fruit such as strawberries. Promisingly, recent studies in Australian-grown strawberries demonstrated that photosensitization extended the shelf-life and did not affect the physicochemical quality of the strawberry and retained key quality attributes [9].

Conclusions
The observed reduction in B. cinerea spore germination was directly proportionate to curcumin concentrations, and both botrydial and dihydrobotrydial decreased with increasing curcumin concentration. In the treatment of D + L + with an illumination time of 10 min at curcumin concentration of 800 µM, the spores were completely inhibited and these botryane secondary metabolites could not be detected even after 16 days incubation. Under the other three treatments (D − L − , D − L + and D + L − ), the spores grew normally, with both botrydial and dihydrobotrydial being produced throughout the 16 days incubation period. Curcumin-mediated photosensitization therefore represents a potentially effective method to control B. cinerea infection, and warrants further investigation in vivo in susceptible fruit and vegetables.

Fungal Materials
Five filamentous fungi (Botrytis cinerea, Pestalotiopsis theae, Cladosporium sp., Penicillium raistrickii and Mucor rudolphii) were isolated and identified from four commercial varieties (Fortuna, Festival, Ruby Gem and Red Rhapsody) of strawberries grown on a Queensland farm, Australia. A randomly selected sample of ten fruits was washed thrice with potable water, and drained on clean paper towels. Then, each fruit was cut into four equal pieces using a sterilized scalper, and placed on potato dextrose agar (PDA) (Thermo Fisher Scientific, Victoria, Australia) plates. The plates were incubated at 25 • C. The samples were observed daily in a laminar flow cabinet for any fungal growth, and fungi were isolated as they appear on the cut pieces using sterile needles. B. cinerea isolated from these strawberries was identified by colony morphology. Pure cultures were observed under a light microscope (Leica, Germany) after staining with cotton blue lactophenol [46]. Identification was confirmed by performing 18S rDNA analysis, as previously described [47]. Specifically, the fungus was cultured on Czapek Dox agar (Thermo Fisher Scientific, Victoria, Australia) at 26 • C for 15 days, and the spores were harvested by flooding the plate with 10-15 mL 0.1% Tween 80 solution. The spore solution was then filtered through sterilized double folded cheesecloth to remove the hyphal fragments. The spore solution was diluted with sterile distilled water, and concentration determined by plating 0.1 mL aliquots of ten-fold dilutions on Dichloran Rose-Bengal Chloramphenicol Agar (DRBC) (Thermo Fisher Scientific, Victoria, Australia). The spores were stored in 15% glycerine at −20 • C. For the photosensitization experiments, spore suspensions of 10 4 CFU per mL were used, and cultured on Czapek Dox agar.

Isolation of Botrydial and Dihydrobotrydial Standards
Isolation of fungal metabolites was based on an adaption of the method of Lineiro et al. [43]. B. cinerea was grown on PDA culture for 3 days and then transferred to individual 500 mL reagent bottles (20 bottles, 3 × 1 cm plugs of agar, ca. 0.75 g per bottle), each containing about 250 mL modified Czapek-Dox medium, which were continuously shaken at 250 rpm under constant room temperature (22 • C) with 12 h light/dark cycles for 13 days. All the liquid broths were combined (ca. 4.8 L) and then extracted with ethyl acetate (3 × 1 L) (Merck, New South Wales, Australia). The combined organic extract was dried over anhydrous Na 2 SO 4 (Merck, New South Wales, Australia) and concentrated under reduced pressure to obtain ca. 0.3 g oily residue. This residue was then subjected to silica flash column chromatography, eluting with an increasing polarity of hexane (Merck, New South Wales, Australia) and ethyl acetate mixtures from 100% hexane to 100% EtOAc. Flash column fractions (37 × 10 mL each) were collected and monitored by silica TLC under UV (254 nm) with vanillin and 2,4-dinitrophenyl-hydrazine (DNP) staining reagent applied to the air-dried TLC plates. Several fractions which were eluted with~70% hexane were subsequently combined and evaporated to provide a solid residue (10 mg), which was confirmed by HRAM UPLC-MS/MS and 1 H and 13 C NMR (as described below) to contain a mixture of botrydial (1) and dihydrobotrydial (2).

NMR Analysis of Botrydial and Dihydrobotrydial Standards
The botrydial/dihydrobotrydial mixture (10 mg) was analysed by 1 H and 13 C NMR in deuterated chloroform (Cambridge Isotope Laboratories, Inc.) on a 500 MHz Bruker Avance instrument (5 mm Selective Excitation Inverse (SEI) probe) with addition of a measured aliquot of dioxane as internal standard for NMR quantification. Integration of aldehyde and dioxane 1 H NMR resonances was used to determine the amount of botrydial present in the sample. The sample containing botrydial (0.8 mg) and dihydrobotrydial (7.2 mg) was recovered from the NMR solvent (CDCl 3 ), and used without further purification as a mixed external standard in HRAM UPLC-MS/MS analysis as described below.

Photosensitizer and Light Source
A stock solution (2000 µM) of natural curcumin was prepared by dissolving 73.8 mg curcumin (Sigma Aldrich, St. Louis, MO, USA) in 50 mL propylene glycol (99.5%, Sigma Aldrich, St Louis, MO, USA), which was then diluted 1:1 in sterile water (50 mL). Our preliminary study showed that this solution had no inhibition effect on the spores (data not shown). The curcumin stock solution was stored in a dark cool place, and diluted as appropriate with sterile water to provide required curcumin concentrations for photosensitizer experiments. The illumination source was a 500-Watt xenon arc lamp (Polilight, PL 500, Rofin Australia Pty Ltd., Victoria, Australia) equipped with an optical fibre light with a range of 370-680 nm filters. This source was configured according to the previously described method [22] and a wavelength of 430 nm was selected. The light dose (J/cm 2 ) was calculated as irradiation time (s) multiplied by the light power (w) divided by the area of irradiation (cm 2 ).

Determination of Spore Survival under Differing Photosensitization Conditions
A wide concentration range of curcumin (50, 100, 200, 400, 600, 800, and 1000 µM) together with different levels of light doses (0, 120, 240, and 360 J/cm 2 , corresponding to an illumination time of 0, 10, 20 and 30 min, respectively) was utilized to determine an efficient photosensitization mediated curcumin treatment regime. Aliquots (1 mL) of both the fungal suspension and curcumin solution (1:1 v/v) were mixed in a Petri dish (30 mm diameter), and the light source was placed approximately 10 cm above the surface of the Petri dish. A magnetic stirrer bar was placed in the mixture (2 mL) to ensure constant stirring during illumination to expose all spores uniformly to the light energy. After illumination, aliquots (100 µL) of the treated mixture were transferred onto Czapek Dox agar (3 replicates per treatment) and incubated with 12 h light/dark cycles at 26 • C for 8 days to determine the colony forming units (CFU), with plates incubated under the same condition for a further 3 days for botrydial and dihydrobotrydial analysis. The percentage reduction in fungal spores after treatment was calculated by the number of the colony (CFU) as given below, with the results of 3 replicates averaged for each treatment.
The selected optimized curcumin concentration and light dose were then utilized to investigate the effect of curcumin-mediated photosensitization on B. cinerea spores using treatment regimes of D + L + : curcumin (800 µM) and light (120 J/cm 2 ); D − L − (control): no curcumin and no light; D − L + : no curcumin and light (120 J/cm 2 ); D + L − : curcumin (800 µM) and no light followed by incubation at 26 • C for 2-16 days.

Determination of Toxin Production under Differing Photosensitization Conditions
B. cinerea-inoculated agar plates after photosensitization treatment and incubation (as described above) were also analysed by LC-MS/MS to determine botrydial and dihydrobotrydial production. A 1 cm diameter plug of agar, with an average weight of 0.25 g, was selectively removed from the respective B. cinerea-inoculated agar plates and extracted with acetonitrile (2.5 mL) by sonication for 10 min. The sonicated sample was then filtered (0.2 µm polypropylene membrane syringe filter) and analysed by HRAM UPLC-MS/MS with results reported as µg botrydial or dihydrobotrydial per g of agar extracted. From each agar plate, three plugs were analysed as replicates.
Spiked recovery samples for metabolite analysis were prepared by addition of aliquots of the standard botrydial/dihydrobotrydial solution to 1 cm diameter plugs of agar from uninoculated agar plates at three different concentrations (n = 4), which were then extracted as above.

HRAM UPLC-MS/MS Conditions
UPLC separations were conducted out on an UltiMate 3000 RS UPLC system (Thermo Fisher Scientific, Bremen, Germany) equipped with an RS pump, a temperature control column compartment and an autosampler. An Agilent ZORBAX Eclipse Plus C 18 (2.1 mm × 100 mm; 1.8 µm) column was used for the LC separations. The UPLC parameters consisted of column oven at 20 • C, an injection volume of 2.0 µL, a flowrate of 0.2 mL/min and gradient elution with mobile phase (MP) A: 0.1% formic acid in H 2 O and MP B: 0.1% formic acid in MeOH (MP B). The mobile phase gradient was 5% MP B (t = 0) increasing to 40% MP B at 0.8 min, 80% MP B at 1.5 min, 95% MP B at 2.2 min before reaching 100% MP B at 4 min and held there until 7 min. MP B decreased to 5% at 7.3 min and was maintained at 5% until 10 min.
The HRAM MS/MS was acquired on a Q-Exactive mass spectrometer (ThermoFisher Scientific, Bremen, Germany) with a heated electrospray ionization (HESI) probe. The MS data were collected in the parallel reaction monitoring (PRM) positive mode at 35,000 FWHM mass resolution, AGC target at 2.0e5, maximum IT at 100 ms, mass isolation window

Statistical Analyses
All experiments were performed in triplicate. The value were expressed as mean ± SD. The analysis of statistical difference was performed using the TTEST function in Microsoft Excel 2010. Two-tails and 'type' for the test were choosen by the correspondence and variance of the data with FTEST. p < 0.05 (significant), or p < 0.001 (highly significant) were applied to determine any significant differences between specific means.

Conflicts of Interest:
The authors declare that there are no conflicts of interest.