Honokiol, magnolol and its monoacetyl derivative show strong anti-fungal effect on Fusarium isolates of clinical relevance

The antifungal activity of magnolol and honokiol, two naturally occurring hydroxylated biphenyls, and of their synthetic derivatives was evaluated on a collection of representative isolates of Fusarium oxysporum, F. solani and F. verticillioides of clinical and ecological concern. The tested compounds were proposed as a ‘natural’ alternative to conventional fungicides, even though a larger range of concentrations (5–400 μg/ml) was applied. The activity of magnolol and honokiol was compared with that of terbinafine (0.1–10 μg/ml), and fluconazole (1–50 μg/ml), two fungicides widely used in treating fungal infections on humans. Magnolol showed similar fungicidal activity compared to fluconazole, whereas honokiol was more effective in inhibiting mycelium growth compared to this fungicide on all tested clinical Fusarium spp. isolates. Compared to terbinafine, honokiol showed similar antifungal activity when tested on clinical F. solani isolates, whereas magnolol was less effective at all selected concentrations (5–400 μg/ml). The different position of the phenol-OH group, as well as its protection, explain different in vitro activities between magnolol, honokiol, and their derivatives. Furthermore, magnolol showed mycelium dry weight reduction at a concentration of 0.5 mM when tested on a set of agricultural isolates of Fusaria, leading to complete inhibition of some of them. Magnolol and honokiol are proposed as efficient and safe candidates for treating clinically relevant Fusaria.


Introduction
Natural occurring polyphenols are a great source of biologically active compounds beneficial to human and animal health [1]. Magnolol (1, 5,5'-diallyl-2,2'-dihydroxybiphenyl) and honokiol (2, 5,5'-diallyl-2,4'-dihydroxybiphenyl) are two hydroxylated biphenyl-type neolignans PLOS  plant extracts are effective against human fusarioses [22], the scarce availability of natural sources from which large amounts of them can be obtained at stable concentrations hamper their application. In our previous studies devoted to the search of effective sustainable fungicides in agriculture, we observed antifungal activity of magnolol 1 against F. graminearum and F. culmorum, both in vitro [23] and in silico [24]. Magnolol 1 manifested the highest anti-fungal activity among a wide range of natural occurring phenols, with fungal growth inhibition at concentrations as low as 0.125 mM. Furthermore, magnolol 1 and honokiol 2 are generally recognised as harmless for humans and animals [25]. Cytotoxicity (IC 50 ) against normal human lymphocytes was set at 38.6 μM for magnolol 1 and 16.1 μM for honokiol 2 [26]. Both compounds are metabolized by extrahepatic and hepatic pathways, producing mainly glucuronic derivatives. Moreover, magnolol 1 is not phytotoxic at concentrations up to 1.5 mM [23]. Considering that many antifungal drugs exert toxic effects on human cells, these characteristics make magnolol 1 and honokiol 2 excellent candidates as sustainable and commercially available fungicides against human and animal pathogens. However, the poor aqueous solubility of magnolol 1 and honokiol 2, common in natural polyphenols, has hampered their broad clinical application. In order to overcome this issue, encapsulation of honokiol has been assessed [27] and synthetic derivatives were prepared [28]. Nevertheless, derivatives and improved formulations of these compounds are often limited by toxicity and drug interaction, hence preventing their use in medical practice. The pro-drug approach [29] represents a versatile strategy to improve the bioactivity of such molecules by transformation of the hydroxyl groups of magnolol 1 and honokiol 2 in an ester group with a mono-and diacetyl functional group or in an acetal group with a glucose unit, respectively. These ester or acetal groups would undergo in vivo biotransformation trough chemical or enzymatic cleavage, thus favouring the delivery of the active compound with a higher yield.
Aims of the present work were: 1) to evaluate the inhibitory activity of magnolol 1 and honokiol 2 over a collection of Fusarium spp. isolates of clinical and phytopathogenic interest, and 2) to develop derivatives of magnolol 1 and honokiol 2 with improved water solubility and equal inhibitory activity compared to the parental compounds. In a preliminary experiment, the spectrum of activity of magnolol 1 was tested against representative isolates of Fusarium species of relevant human concern or of phytopathogenic interest. Subsequently, magnolol 1, honokiol 2, and their derivatives (compounds 3-8, Fig 1) were tested for their anti-fungal activity and specificity towards human and nosocomial isolates. The activity of these compounds was compared with that of terbinafine and fluconazole, two common fungicides. To our knowledge, this is the first time that magnolol 1 and honokiol 2 are assayed over a collection of Fusarium spp. relevant to both plant and human health.

Strains and culture
In a preliminary screening, thirty-two Fusarium spp. isolates from the collection of the University of Sassari (PVS-Fu) sampled from soil (10 isolates), diseased plants (9 isolates), and from human specimen (13 isolates) and water tap (1 isolate) in different Italian hospitals were tested against magnolol 1 (Table 1).
Subsequently, a selection of F. oxysporum, F. solani and F. verticillioides isolates from the PVS-Fu collection, collected from human samples, water tap, and basin sink in different Italian hospitals and identified morphologically and molecularly by analysis of the elongation factor gene sequence (TEF-1α) was tested (Table 2).

Lipophilicity estimation
Lipophilicity was estimated using the logarithm of the partition coefficient for n-octanol/water (log P), implemented in the ChemBioDraw Ultra 13.0 software.

Evaluation of antifungal activity in Vogel's medium
In a preliminary assay, the antifungal activity of magnolol 1 at 0.5 mM was evaluated according to Pani et al. [23] on thirty-two isolates of Fusarium spp. (Table 1). Vogel's medium was amended with 0.5 mM of magnolol 1 and sonicated for one hour at room temperature. Three replicates of 10 4 conidia in 8 ml of Vogel's/magnolol 1 medium for each strain were cultured at 25˚C in the dark without shaking. After 14 days, mycelia were recovered, filtered and dried at 85˚C for 48 h. Inhibition was expressed as dry weight percentage of the untreated test. from different Italian hospitals ( Table 2). Concentration of compounds 1-4 and 8 ranged from 5 to 400 μg/ml, while concentrations between 5-100 μg/ml were used for compounds 5-7. The concentrations of terbinafine and fluconazole ranged from 0.1-10 μg/ml and 1-50 μg/ ml, respectively. Antifungal activity was measured after three days of growth on potato dextrose agar (PDA) medium at 25˚C in the dark in the presence of the different concentrations (μg/ml) of each compound and expressed as the colony diameter in percentage relative to control. Concentrations of conventional fungicides were selected according to clinical dosage and standard experimental procedures [30]. PDA was amended with each inhibitor suspension previously sonicated for an hour at room temperature. Conidial suspensions of each strain were prepared by growing the fungi on carboxymethyl cellulose (CMC) medium [31] for 7 days at 25˚C and 170 rpm in the dark. Cultures were filtered, the spores were collected by centrifugation and the concentration was adjusted to~10 6 colony-forming units (CFU)/ml in sterile water. Ten microliters of the conidial suspension of each strain were plated on the amended PDA. Three replicates were prepared for each inhibitor and concentration.

Measurement of Minimal inhibitory concentration (MIC)
The inhibitory activity of each compound was expressed as Minimal Inhibitory Concentration (MIC) and represents the lowest concentration of active ingredient (μg/ml) that is sufficient to inhibit fungal growth.

Measurement of Lethal Dose 50 (LD 50 )
The Lethal Dose 50 of each compound was calculated as the concentration of active ingredient (μg/ml) able to reduce by 50% the growth of the fungus in vitro.

Statistical analysis
Data on mycelium weight (mg of dry weight per Petri plate) and mycelium growth (mm per Petri plate), obtained from separate experiments, were expressed as percentage of the relative control treatment and pooled to perform statistical analyses. To compare the activities of different concentrations of each of the selected compounds, a one-way analysis of variance (ANOVA) was performed, followed by multiple comparisons by Tukey HSD test at P = 0.01 using Minitab for Windows, release 17. Prior to performing the ANOVA, normality of

Results
Antifungal activity of magnolol 1 on a collection of representative isolates of Fusarium spp.
The inhibitory effect of magnolol 1 on the dry weight of thirty-two Fusarium strains tested in Vogel's medium is reported on Table 3.  25 Fusarium isolates underwent a mycelium dry weight reduction greater than 50% when exposed to magnolol 1 at 0.5 mM (Table 3).
Complete or almost complete inhibition was observed in eight isolates, four of them belonging to clinical Fusarium collected from human specimen (PVS-Fu, 864; 95; 112; 113, Table 2 and Table 3). Magnolol 1 at 0.5 mM was very effective on F. guttiforme 864 cultured from a paronychial infection and against F. verticillioides 112 and F. verticillioides 113, collected from pathological human blood, leading to complete inhibition of mycelium growth. Complete inhibition was also observed for F. solani 97, whereas more than 50% reduction of mycelium dry weight was obtained for F. solani 105 and F. solani 93. Similarly, onycomycosiscausing F. oxysporum 89 and F. oxysporum 126 were inhibited by over 75%, whereas F. oxysporum 127 was inhibited only by 20%.

Selection and activity of two conventional fungicides targeted at Fusarium isolates of clinical relevance
Terbinafine and fluconazole, two conventional fungicides, were assayed over a range of 0.1-10 μg/ml and 1-50 μg/ml, respectively, on a set of F. verticillioides, F. solani and F. oxysporum isolates from human samples and clinical environment ( Table 2). With different levels of intensity, F. verticillioides isolates were the most sensitive, whereas F. solani isolates showed the lowest sensitivity towards both fungicides (Fig A and Fig B in Supporting Information). In particular, F. verticillioides isolates were highly sensitive to terbinafine, with MICs ranging from 1 to 5 μg/ml and LD 50 between 0.1 and 1 μg/ml (Table 4).
At 25 μg/ml, fluconazole was able to completely inhibit the vegetative growth of all F. verticillioides isolates, whereas almost 60% mycelium growth inhibition was observed for F. solani 502 and for all F. oxysporum isolates. The mycelium growth of F. solani 94 and F. solani 96 was not affected by fluconazole even at the highest concentration of 50 μg/ml. Even at lower concentrations (5-10 μg/ml), terbinafine showed more effective antifungal activity (growth Table 4. LD 50 and MIC values (in μg/ml) of terbinafine, fluconazole, magnolol 1 and honokiol 2 tested against Fusarium spp. inhibition being > 50%) than fluconazole in all isolates belonging to the three Fusarium species complexes (Fig A and Fig B in Supporting Information).

Antifungal activity of magnolol 1, honokiol 2 and their derivatives 3-8
Isolates belonging to F. oxysporum were more sensitive to magnolol 1 compared to the other species (Table 4). Magnolol 1 affected all isolates belonging to the three species complexes with a mycelium growth inhibition >50% at 100 μg/ml; the trend was not significantly different at 200 μg/ml and for F. verticillioides 604, F. solani 502, F. oxysporum 89, F. oxysporum 90 and F. oxysporum 91 even at 400 μg/ml (Fig C in Supporting Information). MIC of magnolol 1 was >400 μg/ml for all the tested Fusarium isolates (Table 4). Over a concentration range of 10-100 μg/ml, honokiol 2 completely inhibited mycelium growth in all 15 fungal strains investigated (Fig D in Supporting Information). More than 50% mycelium growth inhibition was observed in all Fusarium spp. when honokiol 2 was amended at 10 μg/ml. In F. verticillioides 84, a fungus responsible for chronic sinusitis, honokiol 2 showed strong antifungal activity with MIC between 5 and 10 μg/ml (Table 4). Honokiol 2 was more effective than magnolol 1 at inhibiting F. solani, with MIC ranging from 10 to 100 μg/ml, except for F. solani 93, whose MIC was comprised between 5 and 10 μg/ml. Besides F. oxysporum 91, antifungal activity of magnolol 1 was comparable to that of honokiol 2 in F. oxysporum isolates, with a LD 50 ranging from 5 to 10 μg/ml. However, magnolol 1 could not completely inhibit fungal growth at the tested concentrations, whereas honokiol 1 did over a range of 10-100 μg/ml. Honokiol mono acetate 5 and 6 proved less effective than the parent compound, with LD 50 and MIC values higher than those displayed by honokiol 2 (Table 5). At 100 μg/ml, except for F. solani 94, F. solani 96 and F. solani 93, honokiol 2-mono acetate 5 and honokiol 4'-mono acetate 6 inhibited by 50% the vegetative growth of all Fusarium isolates (Fig G and Fig H in Supporting Information). The colony growth of F. verticillioides 87, F. verticillioides 88 and F. oxysporum 90 was inhibited by 80% in the presence of honokiol 2-mono acetate 5 at 100 μg/ ml. Honokiol diacetate 7 was less effective against all the Fusarium species complexes (Table 5 and Fig I in Supporting Information). Complete protection of both phenol-OH groups in magnolol 1 and honokiol 2 determined a reduction of their efficacy (

Comparison of antifungal activity of compounds 1-8 with the conventional fungicides
When the activity of magnolol 1, honokiol 2 and their derivatives 3-8 was compared with that of terbinafine and fluconazole, a different trend of efficacy was observed (Table 4 and Table 5 and  Table A in Supporting Information reports a comparison of μg/ml and molarity for all investigated compounds with that of fluconazole and terbinafine. Furthermore, the growth inhibition data of terbinafine, fluconazole, magnolol 1 and honokiol 2 were collected in four graphics where concentrations in μg/ml were converted to molarity (Fig 2A,  2B, 2C and 2D). In Fig 2A the summary effect of terbinafine, fluconazole, magnolol 1 and honokiol 2 on F. oxysporum, F. solani and F. verticillioides is shown. Terbinafine and honokiol 2 were both able to inhibit mycelium growth effectively. Nevertheless, terbinafine, at the tested clinical dosages, could not completely inhibit fungal growth as honokiol 2 did on all Fusarium isolates over a range of 37.5-375 μM (10-100 μg/ml). The effect of magnolol 1 and fluconazole is superimposable over a range of concentrations, roughly between 19 and 160 μM, though magnolol 1 was more effective as vegetative growth inhibitor.

Discussion
The objective of this work was to propose 'natural' alternatives to the use of conventional fungicides, even if this approach required the use of higher concentrations of the identified compounds. The compounds we tested, namely magnolol 1 and honokiol 2, are commercially available in large amounts and have been used in modern clinical practice in Eastern Asia, while in the United States and Europe are considered cosmetic additives. In a preliminary screening, the selection of Fusarium spp. from agricultural and human sources provided a general information of the potential antifungal activity of magnolol 1 towards these fungi, including ubiquitous representatives of the F. solani and F. oxysporum species complexes. With different ranges of intensity, magnolol 1 was able to control all F. oxysporum and F. solani isolates sampled from human specimen. Magnolol 1 also proved a potential candidate to control F. subglutinans and, to a lesser extent, F. verticillioides 400, two species able to infect corn. Data reported in Table 3 highlight a different sensitivity of isolates within the same species complex, regardless of their origin.
Subsequently, this study was carried out on Fusarium spp. of clinical interest, aiming at testing the antifungal activity of magnolol 1, honokiol 2 and their derivatives 3-8 as sustainable alternatives to current clinical drug treatments, which suffer from frequent emergence of resistant pathogen populations. Two commercial fungicides with different mode of action, i.e. terbinafine and fluconazole, were selected in order to compare their antifungal activity with that of compounds 1-8 against all clinical Fusarium spp. isolates. Terbinafine, a tertiary allylamine and one of the most effective antimycotic drugs used in human therapy, inhibits the ergosterol biosynthesis by acting on the squalene epoxidation and is widely used for the treatment of nail infections [32]. Fluconazole, an inhibitor of the lanosterol 14α-demethylase, belongs to the azole SBI class I, whose activity is weaker than that of terbinafine. Although acquired resistance against fluconazole was observed in Fusarium spp. of agricultural interest, the fungicide is still being used in the clinical treatment of these fungi [33,34]. Concentrations of terbinafine and fluconazole were selected according to literature [30].
Magnolol 1, honokiol 2 and their derivatives 3-8 were tested at concentrations ranging from twenty-five-fold lower and, roughly three-fold higher than those used to amend Vogel's medium in the preliminary assay. This concentration range allowed us to compare the species diversity and the antifungal susceptibility profile of the clinical isolates with all tested. Over an acceptable range of variability, percentage of growth inhibition of F. verticillioides 88, F. solani 93, F. oxysporum 89 and F. oxysporum 516 observed in Vogel's assay with magnolol 1 was comparable to that observed on PDA, suggesting that growth inhibition is independent from the cultural medium. In our screening, magnolol 1 and its structural isomer, honokiol 2, showed effective antifungal activity. These compounds may prove useful in antimicrobial drug development, based on their recognised multitarget bioactivities, market availability and almost safe metabolic profile in human body [25,26]. Moreover, the capacity of magnolol 1 and honokiol 2 to act on multiple cell targets can be the key to control infection by resistant fungi which do not respond to conventional therapies.
When the concentrations applied were analysed in terms of molarity, a clearer comparison of the antifungal activities of terbinafine, fluconazole, magnolol 1 and honokiol 2 was obtained. Even though different ranges of molarity were taken into account, a very similar trend in the activities of terbinafine and honokiol 2 was observed. In the case of F. oxysporum 91, honokiol 2 was very effective, thereby appearing as a promising antifungal agent in the treatment of onychomycoses. Interestingly, the antifungal activity of honokiol 2 towards F. solani 93, isolated from foot dermatomycosis, and, to a lesser extent, against F. solani 502, isolated from hospital basin sink, was comparable to that exerted by terbinafine. Due to the acceptable metabolic profile of honokiol 2 taken in oral dosage, these results suggest a potential application of this compound in washing water contaminated by F. oxysporum and in the sanitation of sinks.
A pro-drug approach [29] was applied as a versatile strategy to improve the bioactivity of these compounds by transformation of the hydroxyl groups of magnolol 1 and honokiol 2 in a mono-and di-acetyl ester, respectively, and, in the case of magnolol 1, in a mono glucosyl acetal 8. The antifungal activity of honokiol 2 dropped when one of the phenol-OH was protected with an acetyl group (i.e. in compounds 5 and 6). The effect was less remarkable between honokiol 4'-mono acetate 6 and honokiol 2-mono acetate 5, further highlighting the importance of both phenol-OH groups in the activity of honokiol 2. On the contrary, magnolol mono acetate 3 showed higher inhibitory activity than the corresponding parental compound, whereas the bioactivity dropped when phenol-OH groups were protected with two acetyl groups.
Due to the presence of a glucosyl unit, magnolol mono glucopyranoside 8 was more hydrophilic compared to compounds 1-7. Although a free phenol-OH group is present in the structure of this compound, it is possible that its poor antifungal activity is a result of its failure to reach its target in the fungal cell through the lipophilic membrane.
Magnolol 1 and honokiol 2 are structural isomers, but the different position of one phenol-OH group confers distinct conformations and electronic effects and, as a result, a different reactivity (e.g., antioxidant activity) [7]. A C 2 -symmetry axis in magnolol 1 allows only one monoester isomer, whereas two distinct isomers can be produced after monoesterification of honokiol 2. Phenol-OH groups in magnolol 1 have different acidity due to the formation of an intramolecular H-bond between the two phenol-OH. On the contrary, in honokiol 2, the large dihedral angle and the distance of the two phenol-OH groups reduce the conjugation effect. These features reflect the different activity that monoesters and diesters of magnolol 1 and honokiol 2 have displayed during in vitro assays. According to the position of the two phenol-OH groups in the biphenyl structure, a different mechanism of action between magnolol 1 and honokiol 2 might take place when these compounds interact with fungi.
Several modes of action of magnolol 1 and honokiol 2 were described towards fungi [35][36][37]. Magnolol 1 interacts with ergosterol present in the cell membrane, inducing a partial disruption of the structure, but its mechanism of action is likely to differ from that of other inhibitors of sterol biosynthesis as it has proven effective also on fluconazole-resistant Candida spp. isolates [38]. Cell wall components such as β-1,3-glucans were proposed as potential target of magnolol, similarly to fungicides of the echinocandin family [39].
It is reasonable to hypothesize that other mechanisms underlay the antifungal activity of magnolol 1 and honokiol 2 that may justify a differential action towards the three investigated Fusarium species complexes. Recently, the role of honokiol 2 as activator of mitochondrial ROS by mitochondrial dysfunction and depolarisation of mitochondrial membrane potential in C. albicans was highlighted [40]. Moreover, honokiol 2 is able to hinder the high content of pro-oxidant iron ions in fungi by sequestering the ion [41]. Having different targets and multiple mechanisms of action, it is likely that honokiol 2 acts more effectively than magnolol 1 on selected fungal targets. In fact, candidates that are able to induce dysfunction of mitochondrial membrane and alteration of iron homeostasis represent thenew pharmacological leads currently under development [42]. The scarce toxicity of magnolol 1 upon dermal and oral administration, which is even lower than that of honokiol 2 [25,26], opens new straightforward clinical applications for this compound.
In conclusion, our findings suggest that magnolol 1, honokiol 2 and magnolol acetate 3 may represent promising alternatives for the treatment of fungal infections on both plants and humans: they share a different mode of action compared to conventional antifungal drugs, whose clinical application is jeopardized by the onset of resistance phenomena in fungal populations. The low toxicity of magnolol 1 and honokiol 2, their low cost and their efficacy against both yeast and filamentous fungi prompt further investigation on other fungal pathogens relevant in human or plant clinics and, to a large extent, on fungal pathogens of ecological concern.
Supporting information S1 Text. Supplementary material and methods, synthesis and references. Comparison of μg/ml and molarity (μM or mM) for compounds 3-8 with that of fluconazole and terbinafine (Table A). Raw datasets of the effect of compounds 1-8, terbinafine and fluconazole on mycelium growth of Fusarium oxysporum, Fusarium verticillioides and Fusarium solani (Table B). Antifungal activity of terbinafine (Fig A). Antifungal activity of fluconazole (Fig B).Antifungal activity of compound 1 (Fig C). Antifungal activity of compound 2 (Fig D). Antifungal activity of compound 3 (Fig E). Antifungal activity of compound 4 (Fig F). Antifungal activity of compound 5 (Fig G). Antifungal activity of compound 6 ( Fig H). Antifungal activity of compound 7 (Fig I). Antifungal activity of compound 8 (Fig K). Photos of mycelium growth of five Fusarium oxysporum isolates in the presence of magnolol 1 at 5 and 400 μg/ml in comparison with control ( Fig L).
Photos of mycelium growth of six Fusarium verticillioides isolates in the presence of magnolol 1 at 5 and 400 μg/ml in comparison with control (Fig M). Photos of mycelium growth of four Fusarium solani isolates in the presence of magnolol 1 at 5 and 400 μg/ml in comparison with control ( Fig N). (ZIP)