A Plumieridine-Rich Fraction From Allamanda polyantha Inhibits Chitinolytic Activity and Exhibits Antifungal Properties Against Cryptococcus neoformans

Cryptococcosis is a fungal infection caused mainly by the pathogenic yeasts Cryptococcus neoformans and Cryptococcus gattii. The infection initiates with the inhalation of propagules that are then deposited in the lungs. If not properly treated, cryptococci cells can disseminate and reach the central nervous system. The current recommended treatment for cryptococcosis employs a three-stage regimen, with the administration of amphotericin B, flucytosine and fluconazole. Although effective, these drugs are often unavailable worldwide, can lead to resistance development, and may display toxic effects on the patients. Thus, new drugs for cryptococcosis treatment are needed. Recently, an iridoid named plumieridine was found in Allamanda polyantha seed extract; it exhibited antifungal activity against C. neoformans with a MIC of 250 μg/mL. To address the mode of action of plumieridine, several in silico and in vitro experiments were performed. Through a ligand-based a virtual screening approach, chitinases were identified as potential targets. Confirmatory in vitro assays showed that C. neoformans cell-free supernatant incubated with plumieridine displayed reduced chitinase activity, while chitinolytic activity was not inhibited in the insoluble cell fraction. Additionally, confocal microscopy revealed changes in the distribution of chitooligomers in the cryptococcal cell wall, from a polarized to a diffuse cell pattern state. Remarkably, further assays have shown that plumieridine can also inhibit the chitinolytic activity from the supernatant and cell-free extracts of bacteria, insect and mouse-derived macrophage cells (J774.A1). Together, our results suggest that plumieridine can be a broad-spectrum chitinase inhibitor.

Cryptococcosis is a fungal infection caused mainly by the pathogenic yeasts Cryptococcus neoformans and Cryptococcus gattii. The infection initiates with the inhalation of propagules that are then deposited in the lungs. If not properly treated, cryptococci cells can disseminate and reach the central nervous system. The current recommended treatment for cryptococcosis employs a three-stage regimen, with the administration of amphotericin B, flucytosine and fluconazole. Although effective, these drugs are often unavailable worldwide, can lead to resistance development, and may display toxic effects on the patients. Thus, new drugs for cryptococcosis treatment are needed. Recently, an iridoid named plumieridine was found in Allamanda polyantha seed extract; it exhibited antifungal activity against C. neoformans with a MIC of 250 µg/mL. To address the mode of action of plumieridine, several in silico and in vitro experiments were performed. Through a ligand-based a virtual screening approach, chitinases were identified as potential targets. Confirmatory in vitro assays showed that C. neoformans cell-free supernatant incubated with plumieridine displayed reduced chitinase activity, while chitinolytic activity was not inhibited in the insoluble cell fraction. Additionally, confocal microscopy revealed changes in the distribution of chitooligomers in the cryptococcal cell wall, from a polarized to a diffuse cell pattern state. Remarkably, further assays have shown that plumieridine can also inhibit the chitinolytic activity from the supernatant and cell-free extracts of bacteria, insect and mouse-derived macrophage cells (J774.A1). Together, our results suggest that plumieridine can be a broad-spectrum chitinase inhibitor.

INTRODUCTION
Cryptococcosis is a neglected fungal infection caused predominantly by Cryptococcus neoformans and Cryptococcus gattii (Kwon-Chung et al., 2017). C. neoformans is considered an opportunistic pathogen that infects mainly immunocompromised patients, while C. gattii is also capable of infecting immunocompetent individuals (Kwon-Chung et al., 2014). The infection occurs when dry cryptococci cells or spores are inhaled and reach the lungs, where they can either be controlled by the host immune system in case of an immunocompetent patient, or they may remain latent for a certain period (Giles et al., 2009;Velagapudi et al., 2009;Sabiiti and May, 2012;Ballou and Johnston, 2017). Under conditions of immunocompromise, the pathogen can disseminate from the lungs to the brain through different mechanisms and frequently cause fatal meningitis (Srikanta et al., 2014;Denham and Brown, 2018).
Cases of cryptococcosis are closely related to the pandemic events of AIDS, with less than 300 reports in the late 1950s and more than a million in 2006 (Park et al., 2009). The Joint United Nations Programme on HIV/AIDS (UNAIDS) estimates a mortality rate of 1.4 million AIDS patients annually, of which 15% are due to cryptococcosis alone (Rajasingham et al., 2017). Moreover, reports of cryptococcosis in immunocompetent patients are increasing (Chen et al., 2008;Suchitha et al., 2012;Poley et al., 2019).
For hyphal branching and growth, autolysis, and morphogenesis, fungi employ several enzymes, such as glucanases and chitinases (Adams, 2004;Duo-Chuan, 2006;Ene et al., 2015). Chitinases are responsible for the hydrolysis of the chitin β-1,4 linkage resulting in monomers and oligomers of GlcNAc (Howard et al., 2003). These enzymes have been reported in a variety of organisms beside fungi, including plants, bacteria and humans (Grover, 2012;Junges et al., 2014;Rathore and Gupta, 2015). In C. neoformans, chitinases have been reported to be essential for sexual development but not vegetative growth and asexual reproduction (Baker et al., 2009). Furthermore, it has been suggested that cryptococcal chitinases are indirectly enrolled in modifications of the melanin scaffold (i.e., through the regular cell-wall remodeling activity) that lead to the detachment of melanin granules detachment from the cell wall (Camacho et al., 2019).
Several chitinase inhibitors have been described in the literature. For instance, the natural peptides argifin and argadin can inhibit Aspergillus fumigatus, Serratia marcescens, and human GH18 chitinases (Rao et al., 2005b). Caffeine was shown to be a chitinase inhibitor for the fungus Clonostachys rosea chitinase CrChi1. Also, conservation in the binding site is crucial for the effectiveness of this inhibitor (Yang et al., 2010). Methylxanthines, which includes caffeine, harbor anti-inflammatory properties and are also reported as GH18 chitinase inhibitors (Rao et al., 2005a). A chemical class named acetazolamide has been reported as a chitinase inhibitor in the pathogenic fungus A. fumigatus (Schüttelkopf et al., 2010). Noteworthy, several chitinolytic enzymes are not intracellular, which makes it possible to explore inhibitors that do not need to cross the cell wall and plasmatic membrane (Hamid et al., 2013).
The current treatment of cryptococcal meningoencephalitis consists of an induction, consolidation, and maintenance regimen (Mourad and Perfect, 2018a). The Infectious Disease Society of America (IDSA) recommends a 2-week treatment with amphotericin B (AMB) and flucytosine (5-FC) followed by treatment with fluconazole with time frame and dose depending on the patient's response (Perfect et al., 2010). The combination of AMB and 5-FC shows more fungicidal activity than the sole treatment with AMB (Sloan and Parris, 2014). Although not always effective, this is still the best treatment available nowadays, which unfortunately, is not commercialized worldwide (Pappas, 2010). Additionally, the currently used drugs for cryptococcosis treatment have some disadvantages as C. neoformans strains are innately heteroresistant to fluconazole in vitro, even producing highly resistant subpopulation (Sionov et al., 2013). Thus, the administration of fluconazole as monotherapy (in case of AMB and 5-FC lack of availability) or at the end of the combination treatment may lead to occurrence of heteroresistant strains in vivo (Sionov et al., 2013). Furthermore, hepatoand nephrotoxicity are also reported during the cryptococcosis treatment, a side effect mostly caused by AMB administration (Krysan, 2015). Thus, there is a need for drugs with less or no toxicity but still effective against Cryptococcus spp. In this context, natural products are an interesting starting point for alternative treatment drugs and studies of natural compounds with anti-cryptococcal activity are increasing (da Silva et al., 2016;Teixeira et al., 2018).
The Allamanda genus (Apocynaceae: Gentianales) comprises 15 plant species distributed in South America (Sakane and Shepherd, 1986;The Plant List, 2020). Included in this genus, Allamanda polyantha is endemic of the Brazilian Atlantic Forest (Flora do Brasil, 2019). Plants of this genus are used in popular medicine to treat several illnesses, with potential antifungal, diuretic, antidiabetes and antiparasitic properties (Petricevich and Abarca-Vargas, 2019). Several iridoid compounds have been isolated from Allamanda spp., especially from A. cathartica. The iridoids found in A. cathartica include, but are not limited to, plumiericin, isoplumiericin, plumieride, and plumieride coumarate (Petricevich and Abarca-Vargas, 2019). However, the aglycone configuration of plumieride, plumieridine, has not been reported for the Allamanda spp. On the other hand, isolation of plumieridine was reported from Plumeria obtusa, another Apocynaceae (Saleem et al., 2011).
Recently, our research group identified anticryptococcal activity in the seed's extract of A. polyantha Müll. Arg. This antifungal activity was attributed to a fraction rich in plumieridine (Bresciani et al., 2020). However, the target and mode of action of this potential antifungal agent are still unknown. To address the potential targets of plumieridine, a ligand based virtual screening was performed and indicated that it targets C. neoformans chitinases. Thus, several in vitro and in silico assays were employed to evaluate the mechanism of action of this molecule in C. neoformans chitinases. Furthermore, the activity of plumieridine against insect, bacteria, and mousederived macrophage chitinases was also evaluated.

Plumieridine Isolation and Purification
Plumieridine isolation was performed as recently published (Bresciani et al., 2020). Briefly, A. polyantha seeds were crushed in a kitchen blender. Crushed seeds were placed in contact with ultrapure water (10 g/20 mL) for 4 h, under agitation. The liquid suspension was centrifuged (for 10 min at 7168 × g). The resulting supernatant was filtered in filter paper and polypropylene prefilter (AP 25, Millipore). The aqueous extract was completely lyophilized at −50 • C and 0.040 mbar (Christ Alpha 1-4 LD plus, Germany) and stored at -80 • C. The lyophilized crude extract was subjected to silica gel column chromatography (70-320 mesh, Merck), using a gradient elution of dichloromethane: methanol (95:5 to 80:20) as the mobile phase to obtain plumieridine. The fractions were chromatographed over preparative TLC (20 cm × 20 cm, 0.5 mm layer, SiO 2 F254 plates -Merck) using a mixture of dichloromethane:methanol (80:20) as eluent. Fractions were subjected to another chromatographic column as mentioned above to obtain the compound in a higher degree of purity. Fractions with antifungal activity were submitted to nuclear magnetic resonance (NMR) recorded in CD 3 OD, on a Varian spectrometer, operating at 400 MHz. Peaks of residual water were used as an internal standard in 1 H NMR spectra and the solvent peak was used as an internal standard in 13 C spectra (Gottlieb et al., 1997). Results were analyzed with MestraNova software (v. 6.0.2).

Antifungal Susceptibility Assay
Minimum inhibitory concentration (MIC) of the plumieridinerich fraction was determined against C. neoformans strain H99 according to the Clinical and Laboratory Standards Institute M27-A2 (NCCLS, 2002) and compared to amphotericin and fluconazole values. The compound was resuspended in Milli-Q water with 10% DMSO (stock concentration: 20 mg/mL; final concentration of DMSO in the experiments was usually less than 1%), and filtered before use (polyvinylidene difluoride filter, 0.22 µm pore size, Millipore). MIC assays were performed in 96well plates (Corning R , Corning, NY, United States). Plumieridine was serially diluted, starting with 1 or 1.25 mg/mL, in RPMI 1640 (pH 7; Gibco R Life Technologies, Waltham, MA, United States) buffered with MOPS (Acros Organics, Geel, Belgium). Plates were incubated at 37 • C for 72 h. To ensure reproducibility, MIC assays were performed with every new batch of the plumieridine-rich fraction and the obtained values were compared with previous results (Bresciani et al., 2020).

Virtual Screening
To predict potential plumieridine targets, an ad hoc ligand based virtual screening approach was performed. The pharmACOphore software, which allows the alignment of active compounds, was used to search for similar ligands bound to proteins in the Protein Data Bank (PDB) (Hessler et al., 2010).
Molecular docking of all four chitinases and plumieridine was simulated using AutoDock Vina (Trott and Olson, 2010), with the UCSF Chimera interface (Pettersen et al., 2004). The best ligand position was considered as the lowest energy pose, thus, plumieridine was manually docked in the same position in all C. neoformans chitinases. To infer the way plumieridine interacts with the active site of chitinases, two orientations were assayed arbitrarily named inward and outward.

Molecular Dynamics Simulations
Molecular dynamics simulations were performed on complexes obtained from molecular docking using GROMACS 5 software (Van Der Spoel et al., 2005) with the aid of the CHARMM force field (Vanommeslaeghe et al., 2010). Plumieridine-chitinase complexes were placed inside a cubic box large enough to allow for a minimum of 1.0 nm of space from the protein to the box. The solvent properties were mimetic using the TIP3P water model. The system had the charge neutralized with the addition of ions at the physiological concentration (0.15 µM). Volume (NVT) and pressure (NPT) equilibrium simulations were geometrically optimized in the solvated system. During the simulation, the temperature was kept constant at 300 • K coupling the system with a V-rescale thermostat with a 13coupling time of 0.1 ps. The pressure was also kept constant at 1 bar with the Parinello-Rahman coupling algorithm. Molecular dynamics simulations were performed during 2600 ps. ensuring the stabilization of root-mean-square deviation (RMSD).

Chitinase Activity and Inhibitory Assays
To evaluate the inhibitory activity of plumieridine on chitinases, a variety of models, such as C. neoformans, Bacillus subtilis, Tenebrio molitor, and mouse-derived macrophage cells (J774.A1) were employed. C. neoformans cells were grown on either YPD (1% yeast extract w/v, 2% dextrose w/v, and 2% Bacto peptone w/v) or YPGlcNAc (1% yeast extract w/v, 2% N-acetylglucosamine w/v, and 2% Bacto peptone w/v) on shaker for 24 h at 30 • C. Both media were used to compare whether there is a difference in chitinase activity due to a change in the carbon source (Baker et al., 2009). After incubation, the culture was centrifuged (9000 × g for 10 min) and the supernatant was collected and lyophilized (secreted fraction). Lysis buffer (50 mM Tris-HCl pH 8.0, 20 mM EDTA, 200 mM NaCl, 1% SDS, and 1% Triton; 5 min of incubation at room temperature followed by vortex for 2.5 min) was used to release the proteins attached to the cell membrane. After centrifugation (9000 × g for 10 min), the supernatant was collected (soluble cell fraction) as well as the resulting pellet (insoluble cell fraction). All samples were lyophilized and solubilized in phosphate-buffered saline (PBS; 20 mg/mL). B. subtilis strain ATCC6633 was grown in Luria-Bertani (LB) broth (Sigma-Aldrich Co., St. Louis, MO, United States) medium at 37 • C for 24 h. The supernatant was lyophilized and resuspended to the concentration of 20 mg/mL in PBS. This solution was used in the assay. For the insect model, eight and a half grams of whole T. molitor larvae were dried frozen and ground to a powder using liquid nitrogen. The powder was homogenized in PBS 1:2 (w/v) for 15 min under agitation at room temperature. T. molitor crude extract was centrifuged (9000 × g for 10 min) and filtered with qualitative filter paper (Unifil, Brazil). The resulting supernatant was lyophilized and resuspended in PBS (2 mg/mL). Lastly, J774.A1 cells, obtained from Banco de Células do Rio de Janeiro (BCRJ; accession number 0121) were cultured in DMEM (Dulbecco's modified Eagle's medium; Gibco R Life Technologies) supplemented with 10% heat-inactivated fetal bovine serum (FBS; Gibco R Life Technologies), 1 mM L-glutamine (Gibco R Life Technologies), 1 mM sodium pyruvate (Gibco Life Technologies), 1% nonessential amino acids (Gibco R Life Technologies) and incubated at 37 • C and 5% CO 2 for 3 days. After this step, the cell culture was centrifuged (9000 × g for 10 min) and the resulting supernatant was lyophilized and subsequently resuspended in PBS (20 mg/mL).
Chitinase activity assays were performed employing 4methylumbelliferyl β-D-N,N ,N -triacetylchitotrioside (Sigma-Aldrich Co.) as substrate (Boldo et al., 2009). A standard curve was created using 4-methylumbelliferyl (4MU) (Sigma-Aldrich Co.). The assays were performed in 96-well coated microplates (Greiner CELLSTAR R Sigma-Aldrich Co.) and consisted of 100 µL of McIlvaine buffer pH 6.0, 5 µL of the substrate (0.8 mM), and 10 µL of the sample. The reaction was incubated at 37 • C for 30 min. The fluorescence was read at 355 nm excitation and 460 nm emission on SpectraMax I3. Inhibitory assays employed an increasing plumieridine concentration diluted in McIlvaine buffer pH 6,0. Plumieridine was added in the following concentrations: 0, 33, 100, 160, and 260 µg/mL in a final volume of 200 µL. Quantification of samples was performed according to relative fluorescent units (RFU), using the standard curve previously generated.

Quantitative Real-Time PCR (qRT-PCR)
Fungi were grown in the same two media previously described for 4 h at 30 • C and 200 rpm in the presence and absence of plumieridine using the sublethal dose determined through MIC (156 µg/mL). Total RNA extraction was performed using glass beads and Trizol treatment (Invitrogen, Carlsbad, CA, United States), following the manufacturer's guidelines. cDNA was synthesized using 1 µg of DNase-treated RNA and ImProm-II TM Reverse Transcription System (Promega, Madison, WA, United States), following the manufacturer's guidelines. Quantitative PCR reactions were conducted at a final volume of 20 µL, containing 2 µL of cDNA (4 ng/µL), 2 µL SYBR Green (1:1000) (Invitrogen), 0.1 µL dNTP (10 mM), 2 µL PCR buffer 10x, 1.2 µL MgCl 2 , 0.05 U Platinum Taq DNA Polymerase (Invitrogen), and 1 µL of each primer (5 pmol/µL). The experiments were carried out on an Applied Biosystems 7500 Fast Real-Time PCR System R with thermal cycling conditions set to an initial step at 95 • C for 10 min, followed by 40 cycles at 95 • C for 15 s, then 60 • C for 15 s and, lastly, 72 • C for 60 s. A melting curve analysis was performed at the end of the reaction to confirm the presence of a single PCR product. All experiments were performed using three independent cultures, and each cDNA sample was analyzed in triplicate for each set of primers. Data were normalized to levels of ACT1, which was included in each set of PCR experiments. Relative expression was determined using the 2 − Ct method. The primers used can be found in Supplementary Table S1.

Confocal Microscopy
Overnight cultures grown in YPD were washed with PBS, diluted to 10 6 cells/mL in minimal media (15 mM glucose, 10 mM MgSO 4 .7H 2 O, 29.4 mM KH 2 PO 4 13 mM glycine and 3 µM thiamine), plated (1 ml/well) in triplicate in 24well plates with the following concentration of plumieridine: 0, 156, 312, and 625 µg/mL, and incubated at 37 • C and 5% CO 2 for 72 h. Cells were washed with PBS and fixed with paraformaldehyde, followed by incubation with 5 µg/mL of Wheat Germ Agglutinin (WGA) conjugated with Alexa 488 (green) for 30 min at 37 • C. Cells were washed again with PBS and incubated with 5 µg/mL Calcofluor white (blue) for 30 min at 37 • C. As treatment with plumieridine reduces cell count, it was not possible to observe the impact of treatment with different drug concentrations in the same number of cells. Cell count was performed and a percentage ratio between the total cell count and cells with apparent phenotypes was calculated for all treatments.

Time Kill Assay
Overnight cultures grown in YPD were washed with PBS and inoculated (10 5 cells/mL) in RPMI 1640 buffered with MOPS with and without plumieridine supplementation (312, 625, and 1250 µg/mL). Cultures were incubated at 37 • C and 200 rpm for 0, 2, 4, 8, and 24 h. After incubation, 100 µL of each sample was taken and diluted 1:10 with PBS. Thirty µL of the dilution was plated in YPD agar plates for colony forming unit (CFU) determination. The experiments were performed in biological and technical duplicates. Water and DMSO 1.25% were employed as control and control vehicle, respectively (Klepser et al., 2000).

In silico Prediction of Rat Oral Toxicity
An in silico prediction of oral toxicity in rats (Lethal Dose 50 -LD 50 ) was calculated by TEST (Toxicity Estimation Software Tool; version 4.2.1) based on plumieridine structure, using default parameters. TEST employs known experimental toxicity values and applies Quantitative Structure-Activity Relationship (QSAR), against internal and external datasets, to predict the toxicity values for the query structure (Martin et al., 2012).

Statistical Analysis
All experiments were performed in biological triplicates. Oneway analysis of variance (One-way ANOVA) was used to evaluate triplicates from the same experiment, while Two-way ANOVA was used to perform comparisons among experiments. All graphs were generated in Prism -GraphPad 8.0 (GraphPad Software, Inc., San Diego, CA, United States). Letters in the graphs indicate statistical significance between samples evaluated.

Plumieridine Putatively Targets Chitinases
The virtual screening approach resulted in 14,993 predicted targets and, among these, 38 hits belonged to A. fumigatus (

Plumieridine Interacts With Chitinase Catalytic Residues
Signal peptide prediction analysis indicates that the C. neoformans chitinases Chi4, Chi21, and Chi22 can be secreted, while Chi2 does not have a predicted signal peptide sequence (Supplementary Table S3). TMHMM analysis shows that only Chi2 has a transmembrane domain (Supplementary Table S3) that expands from residues 21-43, a feature previously observed by Baker et al. (2009). To create the models for docking, the chitinases sequences from C. neoformans were searched against PDB templates. Since the best-identity hits for these sequences against available templates were around 30%, molecular models for C. neoformans chitinases were created using different approaches (SwissModel, Phyre2, and Robetta). The best models (all created by SwissModel) were selected based on parameters presented by the SwissModel Evaluation tool and all models presented 80% or more residues in Ramachandran favored areas. Notably, folding conservation in the active site of all four chitinases can be observed (Supplementary Figure S1).
Molecular binding analysis indicates that plumieridine putatively interacts with amino acid residues in the chitinases active binding site (Figure 1). NCBI Conserved Domains analysis shows that C. neoformans Chi4 and Chi2 have a GH18_chitinaselike domain (cl10447), while Chi21 and Chi22 have a  Table S3). Both domains are part of the Glycoside Hydrolase 18 superfamily. This superfamily has the characteristic motif DxDxE observed through sequence alignment (Supplementary Figure S2) (Vaaje-Kolstad et al., 2004). In the binding simulations was observed that plumieridine interacts closely with the DxDxE motif. As these residues are directly involved in catalysis, the predicted interaction may be responsible for the loss in the chitinases catalytic activity.

Plumieridine Inhibits Chitinase Activity in the Secreted and Cell-Soluble Fractions of C. neoformans
Through NMR spectra, plumieridine was identified as the main compound present in the chromatographic fraction obtained from the seed extract of A. polyantha with anti-cryptococcal activity (Bresciani et al., 2020). To obtain enough plumieridine for all experiments, several batches of purification were needed. To ensure reproducibility, every new batch of the plumieridinerich fraction was assayed against C. neoformans, aiming for MIC concentrations close to the one obtained previously (Bresciani et al., 2020). The MIC values for each batch varies between 125 and 312 µg/mL, but all batches used in this work had a MIC of 312 µg/mL (Supplementary Table S4). NMR experiments were also repeated and included as Supplementary Figures S3, S4. Through sequence and structure comparison, chitinase 42 from Trichoderma harzianum was chosen as a positive control for chitinolytic activity assays (Supplementary Figure S5). A solution of 1 mg/mL of the Lysing Enzymes from T. harzianum (Sigma-Aldrich Co.) was used in each experiment as a control. For C. neoformans, chitinolytic activity was significantly higher in the insoluble cell fraction, for cells grown either in YPD or YPGlcNAc, when compared to the other fractions evaluated (Figures 2A,B). For fungal grown in YPD, the chitinolytic activity in the secreted fraction did not show statistical difference when compared to the soluble cell fraction. While fungal grown in YPGlcNAc presented higher chitinolytic activity in the soluble cell fraction than in the secreted fraction (Figures 2A,B).
Secreted and soluble cell fractions of C. neoformans, grown in both media, showed a significant reduction in chitinase activity in the presence of plumieridine (Figures 2C,D and  Supplementary Figures S6A,B). Noteworthy, the secreted fraction (C. neoformans grown in YPD) displayed a constant reduction in the chitinase activity in the presence of plumieridine between 33 and 160 µg/mL (Figure 2C). Similarly, a reduction in chitinolytic activity was also observed (between 33 and 100 µg/mL) for cells grown in YPGlcNAc (Figure 2D). In the soluble cell fraction, a reduction in chitinolytic activity was observed in the assays employing 260 µg/mL of plumieriedine for C. neoformans grown in YPD and YPGlcNAc (Supplementary Figures S6A,B).
Conversely, the inhibition of the chitinolytic activity in the insoluble cell fraction did not present the same dose-dependent pattern. For the insoluble cell fraction obtained from cells grown in YPD, the maximum inhibitory activity was observed with 100 µg/mL of plumieridine (Supplementary Figure S6C). While, for YPGlcNAc, a reduction in the chitinolytic activity was observed with 300 µg/mL of plumieridine (Supplementary Figure S6D).

Transcriptional Levels of CHI22 Are Reduced in the Presence of Plumieridine
The qRT-PCR results revealed that the most expressed chitinase gene in C. neoformans, independently of the culture media, was CHI4 (Figures 3A,B). Although CHI4 showed the highest relative expression levels, it was reported that mutants expressing only CHI4 or CHI21 did not present chitinolytic activity (Baker et al., 2009). CHI2, CHI21, and CHI4 transcriptional levels were not influenced by the carbon sources tested (glucose or GlcNAc) or treatment with plumieridine (Figures 3A,B and Supplementary Figure S7). Notably, CHI22 was the only chitinase in which transcriptional levels were influenced by the carbon source evaluated (higher when the yeast was grown in YPGlcNAc). Remarkably, plumieridine treatment negatively affected the expression levels of CHI22 when the fungus was grown in YPGlcNAc (Figures 3C,D). Bresciani et al. (2020) reported that aqueous extracts of A. polyantha seeds induced morphological alterations in Cryptococcus spp. The morphological alterations were putatively attributed to plumieridine and plumieride. Thus, we repeated the experiments with the plumieridine-rich fraction to confirm the previous suggestions. Confocal microscopy revealed that plumieridine reduces cell count in a dose-dependent manner (Figure 4). Cell counts of 31, 29, 23, and 7 cells per field were observed in assays employing 0, 156, 312, and 625 µL/mL of plumieridine, respectively. Cells treated with 312 e 625 µL/mL of plumieridine have incomplete motherdaughter separation, evidenced by a group of three cells lined up (Figure 4 and Supplementary Figure S8, white arrow). Changes in chitin from control and treated cells were not observed through calcofluor white staining (Figure 4). However, WGA appears in one or, more frequently, two FIGURE 3 | Relative expression levels of C. neoformans chitinases. Relative expression levels of CHI2, CHI21, CHI22, and CHI4 after cryptococcal growth in YPD (A) and YPGlcNAc (B). CHI22 expression levels after C. neoformans growth in the presence or absence of plumieridine in YPD (C) and YPGlcNAc (D). Letters above bars indicate statistical difference among different treatments. The expression data are relative to ACT1 level. dots per cell, which can be described as a polarized pattern. Nonetheless, 6.4% (2/31) of the cell count in the control presented a diffuse WGA staining in the cell wall (Figure 4). The diffuse staining pattern increases with higher plumieridine concentrations: 34% (10/29) in the treatment with 156 µg/mL, 43% (10/23) with 312 µg/mL and 57% (4/7) with 625 µg/mL of plumieridine (Figure 4).

Plumieridine Harbors Fungistatic and Fungicidal Activities
In the confocal microscopy assay a reduction in the fungal loads was observed, which pointed for a potential fungicidal activity of the plumieridine-rich fraction. To evaluate the potential fungistatic and fungicidal activities of plumieridine treatment in C. neoformans, a time kill assay was employed and the reduction in the fungal loads was followed through CFU counting (Figure 5). When C. neoformans was treated with the MIC concentration, a fungistatic activity was observed with minimal fluctuation in the fungal loads for up to 24 h (Figure 5). On the other hand, a fungicidal activity was observed in the treatments employing 625 and 1250 µg/mL (Figure 5).

Plumieridine Exerts Inhibitory Activity Against Several Chitinases From GH18 Superfamily
The local alignment showed that human chitinase (PDB 1HKI) (Rao et al., 2003) and mouse chitinase (PDB 1VF8) (Tsai et al., 2004) possess 48% of identity and 67% of similarity (data not shown). Structure comparison reveals that these chitinase structures present a superposition of 0.692 Å (Supplementary Figure S9). Based on this similarity, the chitinases from mouse-derived macrophage cells (J774.A1) were employed as a model in the assays of chitinolytic inhibitory activity, and these results can potentially be applied for humans. Chitinase inhibitory assays employing J774.A1 supernatant revealed a constant reduction in the chitinase activity after treatment with plumieridine (between 33 and 260 µg/mL), whereas higher plumieridine concentrations failed to reduce the chitinolytic activity even more ( Figure 6A). Similar results were found employing B. subtilis ATCC6633 supernatant (Figure 6B), where chitinolytic activity was inhibited up to 260 µg/mL, with no further enhancement with a higher amount of plumieridine. T. molitor supernatant showed significant inhibition of chitinase activity in treatment employing the plumieridine concentrations up to 160 µg/mL with no further enhancement with higher concentrations of the compound (Figure 6C). Given the promiscuous inhibitory activity presented by plumieridine on different organisms (all harboring GH18 chitinases) and that the simulations showed the interaction of the compound with catalytic residues of chitinases from several sources, our results suggest that plumieridine can be, potentially, a broad-spectrum GH18 superfamily inhibitor.

Plumieridine Interacts Differently With C. neoformans Chitinases
All plumieridine-chitinase complexes reached system equilibrium evidenced by RMSD analysis (Supplementary Figure S10). Molecular dynamics reveals that the inward plumieridine orientation is the most energetically favorable binding configuration (Supplementary Figure S11). For Chi2-, Chi21-, and Chi22-plumieridine complexes, in the initial outward orientation, we observed that plumieridine rotates and changes to the inward orientation at the end of the simulation (Figure 7 and Supplementary Figure S12). This suggests that plumieridine, only in the inward orientation, can inhibit the activity of C. neoformans chitinases.
Controversially, Chi4-plumieridine complexes in both orientations had the inhibitor expelled from the active site. This suggests that plumieridine is prone to inhibit GH18 superfamily chitinases, however, interactions potentially happen according to affinities for residues in the active site (Figure 7 and Supplementary Figure S12). It also suggests that plumieridine inhibits GH18 chitinases selectively and in different levels, thus, these results can also explain the residual chitinase activity observed in the assays employing the higher concentrations of the compound (Figures 2C,D, 5 and Supplementary Figure S6).

In silico Toxicity Assay
To explore the potential toxicity of the plumieiridine in rats, an in silico approach was employed. The prediction pointed to an LD 50 value of 79,04 mg/kg (Supplementary Table S5), which was based on similarities between plumieridine and known compounds in the databases (Supplementary Figure S13). According to the Hodge and Sterner Scale of toxicity classes, plumieridine can be classified as moderately toxic FIGURE 5 | Time to kill assay. C. neoformans viability was determined by CFU counting after growth in RPMI 1640 with and without plumieridine supplementation (312, 625, and 1250 µg/mL) at different time points. The experiments were performed in biological and technical duplicates. Water and DMSO 1.25% were employed as control e control vehicle, respectively. *p < 0.05 and **p < 0.01 for comparisons of control results to the other conditions by one-way ANOVA with post hoc Dunnet test for each time point. (C) Inhibition of chitinolytic activity in the filtered crude extract of the mealworm larvae T. molitor. Letters above bars indicate statistical difference among different treatments.
FIGURE 7 | The interaction of the complexes chitinases-plumieridine after dynamics simulation with initial inward/outward ligand's orientation. Panels with inward (A) and outward (B) plumieridine orientations regarding the active site of each C. neoformans chitinases. Plumieridine (red). (Hodge and Sterner, 1949;Erhirhie et al., 2018). The compounds are classified as moderately toxic when the oral LD 50 ranges between 50-500 mg/kg (single dose to rats) (Hodge and Sterner, 1949).

DISCUSSION
There is a constant need for new, cheaper, less toxic, and widely available drugs for cryptococcosis treatment (Coelho and Casadevall, 2016;Mourad and Perfect, 2018a,b). In this way, South America's biodiversity can be a rich source of new molecules (Valli et al., 2012). The antifungal activity displayed by plumieridine against C. neoformans, led us to investigate the potential drug-target interactions. The time from the discovery and trials of a potential drug is estimated to be around 14 years (Song et al., 2009) and costs US$ 800 million (Lavecchia and Di Giovanni, 2013). Additionally, detailed information about drug-target interactions can also consume several years. Virtual screening approaches (as those applied here to identify chitinases as potential plumieridine targets) can reduce substantially the research time, providing detailed information on drug-target interactions (Kitchen et al., 2004). To support the results found through the virtual screening approaches several chitinolytic assays were conducted. Remarkably, relative chitinolytic activity in C. neoformanssoluble fractions was reduced by the presence of plumieridine, proving the efficiency of our approach.
Although previous reports have pointed that chitinases are not required for asexual reproduction in C. neoformans KN99a and KN99α (Baker et al., 2009), our results suggest that partial impairment of chitinolytic activity can lead to reductions in the cell count due to asexual reproduction impairment, as well as to other potential morphological alterations. When chitinase activity is absent due to plumieridine interaction, cell aggregation and incomplete cytokinesis can be observed. Moreover, the MIC concentration (325 µg/mL) displayed fungistatic activity, while the treatment with 625 or 1250 µg/mL of the plumieridine-rich fraction exerted fungicidal activity. Notably, treatment of C. neoformans with other chitinase inhibitors (methylxanthines) resulted in phenotypes similar to the ones observed for the treatments with the aqueous extracts of A. polyantha or the plumieridine-rich fraction (Tsirilakis et al., 2012;Bresciani et al., 2020).
Even though the molecular docking experiments have predicted that plumieridine would be effective against all C. neoformans chitinases, chitinolytic activity in the insoluble fraction was not constant with increasing plumieridine concentrations. As predicted by TMHMM, Chi2 possesses a transmembrane helix and may be responsible for chitinolytic activity detected in the insoluble cell fraction. This suggests that Chi2 activity might not be inhibited by plumieridine. Additionally, molecular dynamics simulations indicated that Chi4 may not interact with plumieridine in the same way that other C. neoformans chitinases interact. The weaker or lack of affinity observed in the simulations suggests that plumieridine can have specificity for some residues in the active site of chitinases.
Allosamidin is a known chitinase inhibitor isolated from Streptomyces spp. and this compound can also regulate chitinase production in these bacteria. Besides regulating chitinase production, allosamidin does not inhibit the chitin-hydrolytic activity of Streptomyces chitinases (Suzuki et al., 2006). On the other hand, plumieridine reduced C. neoformans chitinolytic activity but affected only the CHI22 expression levels. These results suggest that plumieridine is not involved in cryptococcal chitinase transcription regulation.
The virtual screening approach employed here to identify potential targets of plumieridine pointed to the complex chitinase B1-tripeptide VR0 (PDB ID 3CHE) (i.e., VR0 is derived from the chitinase inhibitor argifin) (Andersen et al., 2008). Structural analyses revealed that the interactions in C. neoformans chitinaseplumieridine complexes occur with residues conserved in the chitinase catalytic motif (DxDxE) or in the vicinities of the conserved motif (Supplementary Figure S12). The initial position of plumieridine was the same as the VR0 in A. fumigatus chitinase, however, after dynamic simulations, the inhibitor's position varied in the active site. Although, interactions between the catalytic motif and plumieridine were still observed in Chi2, Chi21, and Chi22 (Supplementary Figure S12). Thus, plumieridine binding is similar to argifin (i.e., a general GH18 superfamily inhibitor) (Andersen et al., 2008) suggesting a conserved inhibitory activity behavior. To test this hypothesis, in vitro and in silico experiments were conducted with supernatant and cell-free extracts of bacteria, insect and mousederived macrophage cells. Inhibition of chitinolytic activity was observed in these assays, further supporting plumieridine as a broad-spectrum chitinase inhibitor and pointing for applications beyond cryptococcosis treatment. In this way, numerous chitinase inhibitors have been reported with several applications (Rao et al., 2005a;Schüttelkopf et al., 2010;Chen et al., 2014;Christy and Jayaprakash, 2017). The already described argifin has an insecticide potential demonstrated in Periplaneta americana (Omura et al., 2000). Furthermore, chitinase inhibitors can also be employed to relieve the symptoms of respiratory diseases, since chitinolytic activity and chitinase expression levels increases during pulmonary inflammations, aggravating it (Létuvé et al., 2010;Mazur et al., 2019). For instance, the already described allosamidin was observed to reduce the asthma inflammatory process by reducing lymphocyte and eosinophil recruitment to mouse lungs (Zhu et al., 2004).
Finally, to explore the potential toxicity of plumieridine, an in silico oral toxicity assay was employed. According to TEST prediction, plumieridine is moderately toxic. The in silico approach adds up with fibroblast viability assay conducted previously, employing aqueous extracts of A. polyantha (Bresciani et al., 2020). For the aqueous extracts, concentrations lower than 70 µg/mL did not affect cell viability, while 563 µg/mL (C. neoformans strain H99 crude extract MIC concentration) reduced cell viability around 82% (Bresciani et al., 2020). However, the toxicity of the pure plumieridine still needs further investigation with cytotoxic assays, to correctly set the potential applications of this compound.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/Supplementary Material.