Macelignan inhibits bee pathogenic fungi Ascophaera apis growth through HOG1 pathway

Ascosphaera apis is a bee pathogen that causes bee larvae infection disease, to which treatment is not yet well investigated. The aim of this study was to investigate antifungal susceptibility in vitro against A. apis and to identify a new antifungal agent for this pathogen through minimal inhibitory concentration (MIC) assay and western blot analysis. Macelignan had 1.56 and 3.125 μg/mL MIC against A. apis after 24 and 48 h, respectively, exhibiting the strongest growth inhibition against A. apis among the tested compounds (corosolic acid, dehydrocostus lactone, loganic acid, tracheloside, fangchinoline and emodin-8-O-β-D-glucopyranoside). Furthermore, macelignan showed a narrow-ranged spectrum against various fungal strains without any mammalian cell cytotoxicity. In spite of miconazole having powerful broad-ranged anti-fungal activity including A. apis, it demonstrated strong cytotoxicity. Therefore, even if macelignan alone was effective as an antifungal agent to treat A. apis, combined treatment with miconazole was more useful to overcome toxicity, drug resistance occurrence and cost effectiveness. Finally, HOG1 was revealed as a target molecule of macelignan in the anti-A. apis activity by inhibiting phosphorylation using S. cerevisiae as a model system. Based on our results, macelignan, a food-grade antimicrobial compound, would be an effective antifungal agent against A. apis infection in bees.


Introduction
Pollination is a very important process for reproduction of many crops, fruit trees, and wild plants. Among pollinators, honeybees play a major role and are the most economically valuable pollination mediators for agricultural and horticultural plants worldwide. Due to their role in nature, honeybees have a great economic impact (1).
Honey bees are continually exposed to numerous threats: pests and parasites (such as the Varroa mite or Nosema), bacterial diseases (foulbrood), fungal diseases (chalkbrood), viral diseases (invertebrate iridescent virus), and pesticides. Even if the reduction of the bee population is not fully understood, microorganism infections are one of the main reasons. Chalkbrood is a bee larvae infection disease caused by the fungus Ascosphaera apis. Spores of this fungus grow within the digestive track of infected bees and then develop into the mycelial form. This causes the death of the larva, which will have the appearance of chalk. The chalky larva is highly infective to the other larvae within the nest (2,3). A. apis can also infect adult bumble bees, which increases the potential to infect native bees via pathogen spillover (4). Except for the natural essential oils (5), no other compound that inhibits A. apis growth is reported.
Natural plants and medicinal herbs are a valuable resource for novel anti-microbiological agents because many have abundant biological activities and are pharmacologically safe (6)(7)(8). Natural compounds for the control of chalkbrood fungus would be a welcome alternative to synthetic fungicides. A broad range of compounds have been tested in honey bee colonies and on A. apis culture in an attempt to find a control for chalkbrood (2). Some of these compounds are natural plant-derived antimicrobial products (2,5,9,10). Essential oils containing citral, geraniol and citronellal were reported to have the best inhibiting effect on fungal growth in vitro (5). However, most herb extracts showed their best inhibitory effects at relatively low concentrations. For example, 1.5-3.5% (w/v) of Cinnamomum cassia and Piper betel extracts inhibited fungal growth (10) and more than 0.025% (w/v) of essential oils were used for chalkbrood control (5).
Since single compounds showed better efficacy and controllable activity compared with crude extract, we are focusing on highly effective single compounds for chalkbrood disease treatment. In this paper, we tried to identify a putative antifungal compound against A. apis among the elements from natural plants. Our medicinal herb extracts pool was screened to find new functional antifungal agents.
A. apis was maintained with SDA medium (11)(12)(13) at 35°C containing ampicillin (100 mg/mL) under normal atmosphere. Growing fungi were regularly monitored for contamination. Other molds and yeasts were maintained in yeast peptone dextrose (YPD) plates.

Broth microdilution antifungal testing
For A. apis, the broth microdilution assay suggested by the European Committee on Antimicrobial Susceptibility Testing (EUCAST) was applied. Stock compound solutions were prepared in dimethyl sulfoxide (DMSO) with 10 mg/mL concentration. The compound was serially two-fold diluted with Sabouraud dextrose (SD; TaKaRa, Japan), Sabouraud dextrose+0.2% yeast extract (SDYE), YPD (TaKaRa) or RPMI medium (RPMI 164, Sigma, USA) without sodium bicarbonate and with L-glutamine buffered to pH 7.0 with 0.165 M morpholinopropanesulfonic acid and added with 18 g of glucose per liter to make a final concentration of 2% (11)(12)(13)(14)(15). One hundred microliter of the medium containing each compound ranging from 400-0.39 mg/mL were added to the 96-well flat-bottomed microtitration plates (SPL, Korea). The last well was used for sterility and growth controls. One hundred microliter of the mold inoculum suspension containing 1 Â10 5 spores/mL were added to each well of microdilution plates. Plates were incubated at 35°C for 24 and 48 h. Minimum inhibitory concentration (MIC) values were defined as the lowest concentration of drug that completely inhibited cell growth.
For aspergillus species, the same broth microdilution assay described for A. apis antifungal testing was used, but with RPMI 1640 medium (14,16).
For Candida and Saccharomyces species, the broth microdilution assay recommended in the NCCLS document M27-A was used as described previously (15)(16)(17)(18). Each compound (concentration ranged from 200-0.2 mg/mL) was serially two-fold diluted with 100 mL of RPMI 1640 medium. The diluted drug was dispensed into 96 well round bottom microdilution plates (SPL) and 100 mL of yeast inoculum was added with a concentration of 1 Â10 4 cells/ mL to each well of microdilution plates. Plates were incubated at 35°C for 24 and 48 h. Tests were performed at least three times.

Antifungal synergy testing
Antifungal synergy between miconazole and macelignan against A. apis was tested using the broth microdilution assay and the checkerboard method (19). After miconazole and macelignan serial two-fold dilution, macelignan was dispensed into a 96-well microtitre plate with total 100 mL of drug containing SD medium per well. A. apis was inoculated with 100 mL of the suspension containing 1 Â10 5 spores/mL.
Final drug concentrations of 12.5-0.1 mg/mL for miconazole and of 12.5-0.05 mg/mL for macelignan were obtained. MIC values of individual drugs were determined on the same plate. Tests were performed at least three times.
The fractional inhibitory concentration index (FICI) was used to determine synergistic effects of compounds using the equation: FICI=(Ac/Aa)+(Bc/Ba) where Ac and Bc are the MICs of drugs A and B in combination, and Aa and Ba are the MICs of drugs alone. FICI values p0.5 indicate synergy, and values 0.5-4 indicate no interaction (19).

Cytotoxicity assays
The cytotoxicity of the test compounds was done using HepG2 cells with slight modifications from the previously reported procedure using a cell-based MTT assay (HepG2 cells were kindly provided by Prof. T.H. Yi, Kyung Hee University) (20). Cells were seeded at 1 Â10 4 cells per well in 96-well plates. After 24 h of incubation, the cells were washed with PBS and cultured in Dulbecco's modified Eagle's medium (DMEM) containing various concentrations of macelignan (0-3 mM). After 48 h of incubation, MTT solution (3-(4,5-dimethyl-thiazol-2-yl)-2,5diphenyltetrazolium Bromide, Sigma) in PBS was added to a final concentration of 0.5 mg/mL, followed by incubation for 3 h at 37°C. After, the supernatant medium was removed, and cells were suspended in 100 mL of DMSO for 10 min. Absorbance was measured at 520 nm using a micro-plate reader (VersaMax, Molecular Devices, USA). The cell proliferation rates were calculated from the optical density (OD) readings and reported as percentages of the vehicle control.

Western blot analysis of MPK1 and HOG1 in S. cerevisiae
For HOG1 analysis, yeast cells (S. cerevisiae, BY4742; 10 8 cells/mL) were incubated with macelignan and 0.4 M NaCl for 1 h. After disruption of the cells, 100 mg total protein was extracted and separated with SDS-PAGE. p-HOG1 and total HOG1 were detected with p-p38 (Santa Cruz, USA; sc-17852) and Hog1 antibody (Santa Cruz; sc-6815), respectively. For MPK1 analysis, cells with 10 mg/mL of macelignan were heated to 39°C to give heat stress. After 2 h, cells were disrupted and 100 mg of total protein were loaded to SDS-PAGE. p-MPK1 was detected with p42/44 MAPK (ERK1/2) antibody (Sigma) and MPK1-FLAG was detected with anti-FLAG antibody (M2; Sigma).

Natural plant extracts
This study searched for a putative antifungal compound against A. apis among the elements from natural plants. First a pool of medicinal herb extracts was screened to find newly functional antifungal agents. The following nine plant extracts were identified: Crataegus pinnatifida (Bunge leaf), Rhododendron brachycarpum (fruit), the root of Polygonum multiflorum Thunb. (root), Cassia brewsteri (root), the root of Stephania tetrandra (root), Gentiana scabra Bunge (root), carthamus tinctorius Linne (seed), Saussurea lappa (root), and Schisandra chinensis (fruit). These extracts showed preventive effects for fungal cell growth but their effects were not strong enough (Data not shown).

Single compound analysis for anti-A. apis activity
Because of the negative results from the tested herb extracts, widely known active compounds of nine herbal extracts were purchased from Chemface (China) to find an effective antifungal agent. We selected macelignan, corosolic acid, dehydrocostus lactone, loganic acid, tracheloside, fangchinoline and emodin-8-O-b-D-glucopyranoside because of their unknown antifungal function. We applied SD, SDYE, YPD or RPMI1640 medium to test antifungal activity since the standard medium for broth microdilution assay of A. apis was not established yet. Luckily medium effects were not detected for MIC, thereby SD medium was used for anti-fungal assay (Data not shown). Finally, macelignan showed the strongest anti-A. apis activity among all extracts (Table 1). Macelignan had 1.56 and 3.125 mg/mL MIC against A. apis after 24
Corosolic acid showed a mild effect against A. apis (12.5 mg/mL MIC after 24 and 48 h, Table 3). Even if corosolic acid showed strong anti-C. neoformans activity (3.125 mg/mL MIC) after 24 h, MIC after 48 h was decreased to over 200 mg/mL. Unfortunately, dehydrocostus lactone did not show strong effects, but a weak antifungal activity against all of the tested fungus was observed (Table 4). MIC values for miconazole ranged from 1.56 to 25 mg/mL against all of tested fungus after 24-to 48-h incubation (Table 5).

Synergistic inhibition of A. apis by macelignan and miconazole
We applied the broth microdilution assay in a checkerboard method to determine if miconazole and macelignan synergistically inhibited the growth of A. apis (Table 6). Despite miconazole and macelignan not showing strong synergistic effects against A. apis (FICI=0.75 and 0.78 after 24-and 48-h incubation, respectively), miconazole MIC was reduced to 0.78 mg/mL combined with 0.39 mg/mL of macelignan. We also observed a weak growth of A. apis at 0.78 mg/mL miconazole with all macelignan concentrations tested and 0.78 mg/mL of macelignan with all miconazole concentrations tested.

Cytotoxicity of macelignan and miconazole
Cytotoxicity of macelignan and miconazole against HepG2 cells was tested to dertermine whether compounds also inhibit the growth of human cells ( Figure 1). Macelignan had low cytotoxicity against HepG2 cells (IC 50 =62.7 ± 5.3 mg/mL; Figure 1A). This result is in agreement with the report that showed that macelignan had weak cytotoxicity against the HepG2 and MDA cells (21). However, miconazole had a strong cytotoxicity (IC 50 =5.9 ± 5.5 mg/mL; Figure 1B).

Anti-A. apis target validation using S. cerevisiae
Although A. apis genome sequencing was completed at 2006, translational analysis was barely performed because of poor antibody availability against A. apis. Therefore, in this study S. cerevisiae was used to validate A. apis antifungal target. Initially, we used the A. apis genome sequencing project results. We found two candidate genes Scaffold65 and Scaffold592, both of which have short sequences. Compared with S. cerevisiae HOG1 that has 435 amino acids, predicted amino acid numbers for Scaffold65 is 228 and for the Scaf-fold592 is 94. But when we examined their sequences, we found that both are actually the same, and differences were because of the mRNA quality; moreover, there were 2 un-sequenced regions. Therefore, we prepared the genomic DNA of the A. apis to inspect its sequence, and generated the predicted amino acid sequence for A. apis HOG1 homologue. We performed an alignment with 3 different programs that showed 79.5% identity with 89.7% similarity at the conserved kinase region and Data are reported as minimum inhibitory concentration (MIC) in mg/mL. The experiment was repeated three times with essentially the same results. 63.7% identity if the c-terminal variable region was included. Therefore, we conclude that putative A. apis HOG1 was very similar to S. cerevisiae HOG1. Total protein and phosphorylated forms of MPK and HOG1 were observed after macelignan treatment (Figure 2). After macelignan treatment, phosphorylation of HOG1 showed a dose-dependent inhibition compared with total HOG1 protein. However, MPK1 phosphorylation was not changed even with the highest concentration of macelignan treatment.

Discussion
In the present study, macelignan exhibited the strongest growth inhibition against A. apis among tested compounds, followed by corosolic acid and dehydocostus lactone. Miconazole also had anti-A. apis activity. In addition, miconazole MIC value was reduced when combined with macelignan. Although macelignan was effective as a sole antifungal agent to treat A. apis, combined treatment with miconazole showed much better results.
Miconazole is an imidazole antifungal agent, which is commonly applied topically on the skin or to mucous membranes for treating fungal infections in human and veterinary medicine. It works by inhibiting the synthesis of ergosterol, a critical component of fungal cell membranes.   Even though miconazole is a well-known antifungal drug and is relatively cheap, miconazole has never been reported for A. apis growth inhibition. In spite of miconazole having powerful anti-fungal activity, it has strong cytotoxicity (22)(23)(24). Therefore, combination of miconazole and macelignan would strengthen the cost-effectiveness of miconazole while overcoming the cytotoxicity by the usage of a natural origin compoundmacelignan.
Previously developed antimicrobial compounds are usually active against various microorganisms. Overuse of broad-spectrum anti-microbial compounds facilitates the development of antibiotic-resistant infections and multidrug resistance. Moreover, growth of probiotic organisms, which help make the microbial flora healthier, is also inhibited by anti-microbial drugs. In other words, when broad range anti-microbial agents are applied to control the pathogenic microorganism, these agents also suppress the growth of favorable organisms. Accordingly, one of the trends in the development of anti-microbial agents is to find drugs that are narrow-ranged (25). Therefore, the narrow spectrum of macelignan would be advantageous to treat a fungal disease in honey bees.
MAPK cascade transmits signals from outer cell surface to the nucleus and is involved in fungal survival mechanisms against environmental stress conditions (35). MAPK signaling molecules can be good targets for antifungal drugs to avoid fungal survival against conventional drugs. Heitman's lab published research reporting that MPK1 and HOG1, MAPK pathway molecules, have antifungal drug sensitivity in C. neoformans (36,37). Jung and Bahn also examined stress-activated signaling pathways, including HOG and MPK1 pathways, that would be drug target components for treatment of C. neoformans (38). In the present study, macelignan had antifungal activity only against C. neoformans, other than A. apis, which might infer the inhibition of MPK1 and/or HOG pathways. Recently, a systems biology study confirmed MPK1 among 11 MAPK molecules as the best target for antifungal drugs (39). Therefore, inhibition of A. apis and C. neoformans growth by macelignan treatment can be assumed through regulation of MPK1 and/or HOG1 pathways. To validate which pathway was related with the anti-A. apis effect, S. cerevisiae was used for protein analysis because A. apis antibodies are not commercially available. When macelignan was used in MPK1 or HOG1 protein activated yeast cells, only HOG1 phosphorylation was inhibited dosedependently. Therefore, HOG1, homolog of mammalian p38, should be a target of macelignan.
In conclusion, macelignan exhibited a strong growth inhibition against A. apis, with a narrow range spectrum. When macelignan was combined with miconazole, each compensated the other's weakness, showing a synergistic effect. Macelignan, as a food-grade antimicrobial compound, would minimally influence the eco-system (40) and could be a good candidate for the treatment of bee diseases caused by microorganisms. However, macelignan is not inexpensive enough to be solely used compared with miconazole. Therefore, combined usage of macelignan and miconazole would compensate the relatively high cost of macelignan, the occurrence of antifungal-agent resistance and the cytotoxicity of miconazole. Field tests to confirm the effectiveness of macelignan will be performed, as well as the development of macelignan derivatives to increase compound stability and anti-A. apis activity.