Total synthesis of antifungal lipopeptide iturin A analogues and evaluation of their bioactivity against F. graminearum

The pursuit of novel antifungal agents is imperative to tackle the threat of antifungal resistance, which poses major risks to both human health and to food security. Iturin A is a cyclic lipopeptide, produced by Bacillus sp., with pronounced antifungal properties against several pathogens. Its challenging synthesis, mainly due to the laborious synthesis of the β‐amino fatty acid present in its structure, has hindered the study of its mode of action and the development of more potent analogues. In this work, a facile synthesis of bioactive iturin A analogues containing an alkylated cysteine residue is presented. Two analogues with opposite configurations of the alkylated cysteine residue were synthesized, to evaluate the role of the stereochemistry of the newly introduced amino acid on the bioactivity. Antifungal assays, conducted against F. graminearum, showed that the novel analogues are bioactive and can be used as a synthetic model for the design of new analogues and in structure–activity relationship studies. The assays also highlight the importance of the β‐amino acid in the natural structure and the role of the stereochemistry of the amino fatty acid, as the analogue with the D configuration showed stronger antifungal properties than the one with the L configuration.


| INTRODUCTION
2][3] Microbial biological control agents (MBCAs), such as Bacillus sp.5][6] However, their commercial use as living organisms has faced several challenges, given their dependence on environmental factors, their delayed action and their lower persistence in the environment, when compared with chemical pesticides. 7The use of the isolated bioactive compounds produced by the bacteria is a promising alternative approach to overcome these limitations.
The antifungal activity of Bacillus spp.has been linked to the production of specific cyclic lipopeptides (CLPs) as secondary metabolites, belonging to the iturin, fengycin and surfactin families. 8,91][12] Moreover, their cyclic structure confers a reduced conformational flexibility to the peptides, which aids the alignment of the required structural motif to its cellular target. 124][15][16] Its structure includes a chiral LDDLLDL sequence of seven α-amino acids and one β-amino acid with an alkyl side chain. 9,17The length of the β-amino acid can vary (C 14 -C 17 ) and a ramification can be present in the alkyl side chain (normal, iso and anteiso isoforms). 17,18The isoforms are referred to as iturin A1 through iturin A8 (Figure 1A). 18 has been proposed that the mechanism of action of iturin A involves the formation of hydrogen bonds between the hydroxy group of D-Tyr 3 and the sterols present in the fungal membrane, followed by penetration of the cell membrane by the alkyl side chain.
The subsequent formation of pores and the leakage of potassium ions lead to cell death. 19,20In fact, the O-methylation and O-acetylation of the D-Tyr 3 residue causes total activity loss. 21Further evidence suggests that iturin A leads to the accumulation of reactive oxygen species (ROS), thus triggering oxidative stress and cell apoptosis. 16,22reover, it has been shown that the biological activity of iturin A is enhanced with increasing carbon chain length of the fatty acid side chain. 23,24 addition to its pronounced antifungal activity, iturin A has numerous other potential applications. 25The lipopeptide displays a broad antibacterial effect against both Gram-positive and Gramnegative strains. 26,272][33] Finally, the lipopeptide exhibits surfactant characteristics and its use as an emulsifier has been suggested. 25,34though the bioactivity of iturin A is well studied, the challenges associated with its isolation and purification from the bacterial extracts have hindered its commercial use.To address this challenge, efforts have been made to enhance the yield of lipopeptide production in Bacillus spp.cultures. 256][37][38] Biochemical methods have also been used to introduce structural modifications in the native structure, as fluorinated analogues have been obtained via precursor-directed biosynthesis, albeit in low yields and with a challenging purification process. 39On the other hand, the development of a method for the total synthesis of the lipopeptide faces the obstacle of the laborious multi-step synthesis of the β-amino fatty acid. 23,40,41Past efforts to simplify the synthesis by replacing the β-amino acid, with an α-amino acid, via cysteine lipidation through native chemical ligation (NCL), yielded lipopeptides that were inactive when tested against C. albicans. 42It must be noted that these reported synthetic analogues introduce an ester moiety in the lipid side chain, which might alter the hydrophobic properties of the lipopeptide, inhibiting its ability to penetrate the cell membrane.Nevertheless, efforts to develop new synthetic methodologies for production of lipopeptides are necessary given the immediacy of the challenges in agriculture regarding antimicrobial resistance and sustainability.
The aim of the present work was to establish a synthetic method to obtain bioactive iturin A analogues, via solid-phase peptide synthesis (SPPS) techniques and via a one-step alkylation of a Boc-protected cysteine residue. 43,44Because the naturally occurring β-amino acid is replaced by an α-amino acid, it was deemed necessary to examine the effect of the configuration of the newly introduced amino acid on the bioactivity of the lipopeptide, as the removal of a methylene group might alter its alignment with the fungal cell.Hence, two lipopeptides were synthesized, one mimicking the configuration of the natural compound containing a D-alkyl cysteine and one with the opposite configuration, containing an L-alkyl cysteine (Figure 1B).These bioactive analogues can be used as a model to conduct structure-activity relationship (SAR) studies, which can help to shed further light into the mode of action of iturin A and can lead to the development of more potent analogues of this lipopeptide.
F I G U R E 1 Structures of (A) natural iturin A isoforms and (B) iturin A analogues synthesized in this work.

| General methods
All protected amino acids (98% purity or higher) were purchased from Iris Biotech GMBH (Marktredwitz, Germany).Solvents and reagents (reagent grade or better) were purchased from Merck KGaA (Darmstadt, Germany).Chemical reactions were monitored using analytical thin-layer chromatography, performed using aluminium-backed silica plates (60 F254) and the stated eluents.Visualization was accomplished using a potassium permanganate stain.Product purification by flash column chromatography was performed using silica gel (Davisil, 230-400 mesh, 40-63 μm).
1 H NMR spectra were recorded using a Varian VnmrS 400-MHz spectrometer.Samples were dissolved in d 6 -DMSO and referenced to TMS (0.00 ppm). 13C NMR spectra were recorded using a Varian VnmrS (101 MHz) spectrometer.Samples were dissolved in d 6 -DMSO and referenced to a residual solvent peak (39.52 ppm).Spectra were analysed using MestreNova 14.0.Chemical shifts are reported in parts per million (ppm) and coupling constants (J) are given in Hertz.Multiplicities are abbreviated as b (broad), s (singlet), d (doublet), t (triplet), q (quartet), m (multiplet) or combinations thereof.
Cyclic lipopeptides were purified using an Agilent 1260 reversedphase high-performance liquid chromatography system (RP-HPLC) with a C 18 column (Agilent Zorbax SB-C 18 , 9.4x250 mm) using a binary solvent system with a linear gradient of acetonitrile and water (both containing 0.1% TFA), changing over 40 min from 10%-100% acetonitrile, followed by a final isocratic hold for 5 min.The flow rate was set at 4 mL/min with UV detection at 220 nm.

LC-MS data were acquired in ESI+ mode on an Agilent 6546
QToF system coupled with an Agilent 1260 Infinity Prime II LC system.Chromatography was carried out with a C 18 column (Agilent Zorbax Eclipse Plus, 2.1 Â 50 mm) using a binary solvent system with a linear gradient of acetonitrile and water (both containing 0.1% formic acid) changing over 5 min from 10%-90% acetonitrile, followed by a final isocratic hold for 5 min.The flow rate was set at 0.6 mL/min.Data were processed with the Agilent Masshunter software.Target compounds were searched via compound matching using the Agilent FBF (Find-By Formula) algorithm, matching for singly and doubly charged monomeric ion species for common ions such as [M+H] + and [M+Na] + .
The mixture was suspended in 15 mL of H 2 O, and the pH was adjusted to 6 using a 1 M HCl solution.The acidified mixture was extracted with EtOAc, and the organic phase was washed with brine and dried with Na 2 SO 4 .The solvent was removed under reduced pressure, and the product was purified via flash column chromatography (DCM/MeOH 9:1) to afford 1 (535 mg, yield 63%) and 2 (550 mg, yield 65%) as white solids.The resulting solids contained a small amount of bromoundecane, as confirmed by NMR analysis but were used in subsequent steps with no further purification.Small aliquots of the solids were dissolved in MeOH (10 mg/mL) and purified via HPLC to aid in the NMR characterization of the product.The method consisted of an isocratic run of 35 min with a mobile phase of 30% H 2 O and 70% acetonitrile (both containing 0.1% TFA, retention time: 28 min).NMR spectra of alkylated cysteines (1) and ( 2) are reported in the Supporting Information.Coupling and deprotection were routinely monitored via the ninhydrin (Kaiser) test. 45he resulting resin (200 mg, 0.22 mmol, 1.1 mmol/g) was coupled with products 1 and 2 (248 mg, 0.66 mmol) in the presence of DIC (0.17 mL, 1.1 mmol), HOBt (0.15 g, 1.1 mmol) and DMAP (15 mg, 0.66 mmol).The mixtures were shaken overnight at room temperature, and after solvent removal through filtration, the resin was washed three times with DMF and DCM.Cleavage of the resin and simultaneous removal of the remaining protecting groups was performed with a solution of 95% trifluoroacetic acid (TFA) and 5% scavengers (triisopropylsilane, thioanisole, H 2 O).After repeating the cleavage procedure, the mixture was shaken for 2 h.The resulting solution was collected and concentrated under N 2 gas.Ether precipitation followed by centrifugation afforded the final crude peptide mixtures as white solids, containing peptides 3 and 4. The crude peptide mixtures were used for subsequent reactions without further purification.

| Antifungal assays
The antifungal assays against F. graminearum were performed as previously reported, with some modifications. 47,48The fungal strains were streaked into a Saboraud dextrose agar plate and incubated for 72 h at 26 C. The culture was blended with 200 mL of sterile water, and the resulting mixture was diluted with Saboraud dextrose broth to obtain an OD 6oo of $0.3.This final diluted mixture was used as the inoculum for the antifungal assays and was routinely streaked onto Saboraud dextrose agar plates to ensure cell viability and lack of contamination.Stock solutions of the cyclic lipopeptides were prepared by dissolving 4 mg of the compounds in 40 μL of DMSO, followed by further dilution with 960 μL of Saboraud dextrose broth.The final stock solution concentrations were 4 mg/mL of lipopeptide with 4% DMSO (v/v).The assays were performed in flat-bottom 96-well plates (Greiner) with a total plate volume of 100 μL, using the broth microdilution method.The range of lipopeptide concentrations tested was 0.25-2 mg/mL.C 14 -iturin A was used as a positive control and plates with no lipopeptides present served as the negative control.The effect of the solvent was also monitored by inoculating well plates containing 2% DMSO (v/v), which was the maximum concentration of solvent present in the well plates containing the lipopeptides.Cell growth inhibition was monitored by measuring the OD 6oo at T0 and after 36 h of incubation, using an Epoch Microplate Spectrophotometer (BioTek).The OD 600 at T0 was $0.2 for every well plate.The assays were performed in triplicate and stated as mean ± standard deviation.Paired sample t-test was employed to determine significant differences between means.The null hypothesis was set as the mean of the

| Design and synthesis of cyclic lipopeptides
This study presents the total synthesis of two cyclic lipopeptides that mimic the structure of the natural cyclic lipopeptide iturin A. Because the challenge in the synthesis of the natural structure lies in the laborious synthesis of the β-amino acid with the lipid moiety, it was decided to replace it with an α-amino acid.This would facilitate the synthesis of iturin A mimetics and unlock new opportunities in the development of novel analogues and in conducting structureactivity relationship studies.Previously reported iturin A analogues containing an α-amino acid also contained an ester moiety in the structure, which diminishes the lipophilic properties of the alkyl side chain. 42The aim of this study was to synthesize analogues lacking an ester moiety and with varying stereochemistry.Cysteine was chosen as the replacement amino acid, as its nucleophilic properties and wellreported alkylation protocols made it an attractive choice. 44,49e cysteine alkylation reaction was performed in solution, on a Boc-protected cysteine, in the presence of bromoundecane and DIPEA in DMF (Figure 2). 43The use of fewer equivalents of the bromoalkane resulted in an incomplete reaction and the use of Boccysteine methyl ester as the starting material did not improve the yield over two steps (alkylation and saponification).No racemization and no oxidation of the resulting thioether was observed during NMR F I G U R E 3 Solid-phase peptide synthesis, resin cleavage and cyclization reaction scheme for cyclic lipopeptides 5 and 6.R = (CH 2 ) 10 CH 3 .analysis of the crude product.It is worth noting that the crude mixture of this alkylation reaction was used in the subsequent solid-phase peptide synthesis step without any further purification.The unreacted reagents (bromoundecane and DIPEA) did not influence the coupling of the alkylated cysteine on the resin and were just filtered out after the coupling reaction was completed.
The solid-phase synthesis of the linear octapeptide was performed using standard Fmoc-SPPS techniques and the cyclization was performed under high dilution to promote the intermolecular coupling and avoid dimerization (Figure 3). 50Neither the linear peptide nor the dimerization product was detected following LC-MS analysis of the crude mixture after the cyclization reaction.HPLC purification afforded the final cyclic lipopeptides with a final yield of 8.4% for product 5 and 10.3% for product 6.Their structures were confirmed by ESI LC-MS analysis (Figure 4).

| Antifungal assays against F. graminearum
F. graminearum is ranked among the top five most destructive fungal pathogens that affect agroecosystems. 51It causes fusarium head blight, a concerning disease that affects crops such as wheat and barley and leads to loss of yield and quality of the cereals. 52,53Iturin A has been shown to strongly inhibit the growth of this pathogenic fungus and its possible use as a sustainable fungicide for agricultural use has been suggested as one of the most promising applications for this lipopeptide. 13,54,55Hence, it was decided to test the antifungal activity of the newly synthesized iturin A analogues against F. graminearum to assess their suitability as bioactive mimetics of the natural lipopeptide (Figure 5).The bioactivity was assessed quantitatively by measuring spectrophotometrically the OD 600 of the cell culture in the presence and in the absence of the lipopeptides.The synthetic lipopeptide 5, containing the L alkylated cysteine, inhibited the growth of F. graminearum by 81% when present in 2 mg/mL (1.7 mM) concentration, although its inhibitory effect diminished drastically at the lower concentrations tested.The synthetic lipopeptide 6, containing the D alkylated cysteine, inhibited the growth of F. graminearum by 96% at a concentration of 2 mg/mL (1.7 mM) and by 57% at a concentration of 1 mg/mL (0.84 mM), but it did not show a significant inhibitory effect at lower concentrations tested.
Natural C 14 -iturin A showed an inhibitory effect of more than 92% at all the concentrations tested (0.25-2 mg/mL).The maximum amount In light of these results, it can be reasoned that although the β-amino acid present in the naturally occurring lipopeptide appears to have an important role in the antifungal activity, its replacement with an α-amino acid still yields a bioactive lipopeptide, albeit with a diminished antifungal effect.The introduction of a sulphur atom in the side chain of the newly introduced amino acid might also play a role in its bioactivity, as it might diminish the lipophilicity of the alkyl side chain, due to its electronegativity.Moreover, from the two diastereomeric lipopeptides that were synthesized, the one mimicking the native conformation of iturin A (containing the D alkylated cysteine) had a more pronounced bioactivity, compared with its diastereomer (containing the L alkylated cysteine).

| CONCLUSIONS
Two new iturin A analogues were synthesized and evaluated as antifungal agents against F. graminearum.An alternative synthetic strategy to that reported in the literature was employed, consisting of a one-step alkylation of cysteine, followed by the SPPS of the linear octapeptide and a head-to-tail cyclization, yielding the desired cyclic lipopeptides.A key difference between the analogues generated by this route compared with those reported recently by Yim et al. is that there is no ester moiety in the alkyl side chain, so they bear a closer resemblance to natural iturin A, whereas the overall yield of the two methods is comparable. 42oassay experiments confirmed the importance of the naturally occurring β-amino acid in the bioactivity of iturin A, as the synthetic analogues are less bioactive than the native lipopeptide.Additionally, the effect of the configuration of the newly introduced α-amino acid showed that the bioactivity of the lipopeptide containing an alkylated cysteine with a D configuration was more pronounced.A to the fungal cell membranes, as this approach has been recently used to identify potential protein targets of iturin A. 56,57

Fusarium
graminearum (ATCC 36016) was obtained from DSMZ, German Collection of Microorganisms and Cell Culture GmBH and maintained on Saboraud dextrose agar.Bacillus sp.CS93 (NRRL β-21974) was acquired from the Microbial Genomics and Bioprocessing Research Unit, National Centre for Agricultural Utilization Research, Peoria, IL, USA, and maintained on tryptic soy agar.Saboraud dextrose agar (4% glucose), tryptic soy agar and Saboraud dextrose broth were prepared as per manufacturer's instructions.The medium optimized for lipopeptide production (MOLP) was prepared as previously reported.

2. 2 . 1 |
Bacillus sp.CS93 cultivation and isolation of natural iturin A Bacillus sp.CS93 was grown in 50 mL of MOLP medium in a 250-mL Erlenmeyer flask.The cultures were left shaking at 300 rpm for 72 h at 30 C. The cultures were then centrifuged at 9500 rpm for 30 min at 4 C.The supernatant was recovered, and the lipopeptides were precipitated in acid conditions by dropwise addition of a 5 M HCl solution until pH 2 was reached.The precipitates were incubated on an ice-water bath for 6 h and pelleted at 9500 rpm for 30 min at 4 C.The supernatant was discarded and the precipitate was stored at À20 C. Prior to HPLC purification, the pellets were suspended in 10 mL MeOH for 2 h at room temperature.The extracts were filtered and the filtrate was dried under nitrogen gas.The lipopeptide crude mixture was then dissolved in MeOH at a final concentration of 10 mg/mL.HPLC purification of the mixture yielded pure C 14 -iturin A as a TFA salt.C 14 -iturin A: retention time 16.4 min.HRMS calcd for C 48 H 75 N 12 O 14 [M+H] + : 1043.5526found 1043.5518.HPLC trace and MS spectrum of purified natural C 14 -iturin A are reported in the Supporting Information.
OD 600 of the fungal culture in the absence of lipopeptides (control) remains unchanged in the presence of the lipopeptides.P-value of <0.05 was considered a significant difference.Standard deviation and Student t-test calculations were performed in Microsoft Excel for each sample set.

F I G U R E 2
Alkylation of Boc-L-Cys-OH and Boc-D-Cys-OH with bromoundecane and DIPEA in DMF.

F
I G U R E 4 HPLC traces (top) and LC/(ESI+) MS spectra (bottom) of the purified linear and cyclic lipopeptides.(A) Linear lipopeptide 3, (B) linear lipopeptide 4, (C) cyclic lipopeptide 5 and (D) cyclic lipopeptide 6. R = (CH 2 ) 10 CH 3 .Calculated isotopic abundances and permitted ppm deviation ranges from predicted formula masses are shown as red bounding boxes around MS spectral peaks. of solvent used in the well plates (2% DMSO) was determined to not have any inhibitory effect.
The new analogues are improved model compounds for structure-activity relationship studies.Future work should focus on the further elucidation of the mode of action of iturin A and on the development of a robust and facile protocol for the synthesis of the natural compound.Incorporating electron-withdrawing atoms on the tyrosine ring, such as fluorine, has the potential to yield more powerful analogues characterized by enhanced hydrogen bond donation capacities.Computational techniques and molecular modelling have the potential to provide further insights on the binding of iturin