Azaphilones Pigments from the Fungus Penicillium hirayamae

: The use of fungal pigments as dyes is attractive for various industries. Fungal pigments arise a strong interest because they are suitable for large-scale industrial production and have none of the drawbacks of synthetic pigments. Their advantages over synthetic or vegetal dyes mark them as a prime target. Azaphilones are fungal polyketides pigments bearing a highly oxygenated pyranoquinone bicyclic core produced by numerous species of ascomyceteous and basidiomyceteous fungi. In order to ﬁnd new azaphilones dyes, the fungal strain Penicillium hirayamae U. , a known producer of azaphilone but, chemically, barely studied so far, was investigated by molecular networking and led to the isolation of three new azaphilones, penazaphilone J-L, along with the known penazaphilone D, isochromophilone VI, and sclerketide E. Their structures were determined based on extensive NMR and the absolute conﬁgurations by ECD. All compounds were evaluated for their cytotoxic activity against human cell lines and human pathogenic-resistant strains.


Introduction
Fungal pigments are of great interest to food colorant industries because they are suitable for large-scale industrial production and do not have the disadvantages of synthetic pigments. Fungal pigments have the advantage of being produced using sources of carbon and nitrogen, which can even be obtained from food by-products or from agro-industrial waste. Moreover, public awareness of the adverse effects of synthetic dye increased the demand for natural pigments, even though only 25 natural colorants are approved by the European Union so far [1].
Azaphilones are fungal polyketides pigments bearing a highly oxygenated pyranoquinone bicyclic core produced by numerous species of ascomyceteous and basidiomyceteous fungi [2]. In Asia, Monascus-fermented rice is used to produce azaphilone pigments for centuries. Azaphilones are one of the most promising classes of fungal pigments in research as industrial pigments, especially because they stand out for their yellow, orange, and red colors. Thus, they received an increasingly great deal of research interest, especially for their applications in the dyeing, cosmetic, and printing industries [3][4][5].
In the search for new azaphilone pigments produced by fungi, we used molecular networking and the latest Sirius 5.5.5 version for fungal metabolites annotation in order to efficiently identify the presence of azaphilones in fungal extracts and optimize the search 2.7. Computational Details DFT calculations have been performed using Gaussian 16 W [9]. Prior to geometry optimization, all compounds were submitted to a conformational analysis using the GMMX package with the MMFF94 forcefield (energy threshold = 1 kcal/mol). After geometry optimization and frequency calculation at the b3lyp/6-31g(d) level. ECD spectra were plotted using GaussView 6.
2.8. Biological Activities 2.8.1. Cytotoxic Activities HCEC-1CT human colon epithelial cell line was purchased from Evercyte and maintained in DMEM (Gibco) in addition with cosmic calf serum (2%), EGF (20 ng/mL), Insulin (10 µg/mL), Apo-Transferrin (2 µg/mL), sodium-selenit (5 nM) and Hydrocortisone (1 µg/mL). Cell viability was assessed by CellTiter Glo 2.0 Luminescent Viability assay (Promega). At J0, 2500 cells were seeded in a white 96-well plate in a final volume of 100 µL. At J1, cells were treated with compounds (1-3) at the indicated concentrations (final concentration obtained by serial dilutions from 100 µg/mL to 1.56 µg/mL in DMSO 1%). At J2, the plate and CellTiter Glo 2.0 reagent are equilibrated to room temperature for 20 min, and 100 µL of CellTiter Glo 2.0 reagent was added to each well. The plate is agitated for 2 min on an orbital shaker and then incubated at room temperature for 10 min. Mono-oxygenation of luciferin is catalyzed by luciferase in the presence of ATP, liberated by viable cells. Luminescence was measured in a microplate reader, Infinite 200Pro (TECAN), as described by Dillard et al. [10]. The results are expressed as a ratio compared to the control using the following formula: Pourcentage = 100 × [(compounds lum)/(control lum)]. All values are averages of three experiments performed in triplicate.
Compounds (1)(2)(3) were also evaluated against A549 and 16HBE cell lines, adenocarcinomic human alveolar basal epithelial cell line and human bronchial epithelial cell lines, respectively. Briefly, A549 cells were seeded in a 96-well plate at 1.5 × 10 4 cells/well with 100 µL DMEM cell culture media. In total, 16 HBE cells were seeded in a 96-well plate at 4 × 10 4 cells/well with 100 µL MEM cell culture media. Both cell lines were cultivated overnight into a humified incubator (37 • C and 5% of CO 2 ) for cell growth. On day 2, half of the cell culture media (50 µL) was removed from the plate and replaced with the same volume of fresh cell culture media containing various concentrations of the compounds (1-3, final concentration obtained by serial dilutions from 100 µg/mL to 1.56 µg/mL in DMSO 1%) and various concentrations of DMSO as solvent control (serial dilution from 10% to 0.1%) and cells alone as the negative control. On day 3, cell viability was carried out by two types of assays, ViaLight and ToxiLight-LONZA, based on luminescence measurement (Vialight and Toxilight). All values are averages of three experiments performed in triplicate.

Antimicrobial Assay
The antimicrobial effects of 1, 2, and 3 were evaluated on human pathogenic gramnegative and positive bacteria, Pseudomonas aeruginosa and Staphylococcus aureus, respectively, as reported earlier [11][12][13]. The methicillin-resistant clinical strain of S. aureus, N315 (13), and the multidrug-resistant clinical strain of P. aeruginosa, PAH [11] were used in the assay. Briefly, DMSO-dissolved azaphilones pigments were diluted in 50 µL of Luria broth (LB) media and transferred to a 96-well transparent plate. Subsequently, the volume was made up to 100 µL by adding 50 µL of bacterial culture in LB media to achieve bacterial cell numbers of 7.5 × 10 5 CFU/mL. The bacterial inoculum used for the assay was obtained from an overnight-grown pre-culture of the respective strains. After 24 h of growth (37 • C, at constant shaking), the optical density (OD) at an absorbance of 600 nm of treated and non-treated samples was evaluated using the multiple-reader Mithras2 LB 943 (Berthold Technology, Thoiry, France). In addition, 5 µL from each well of the 96-well plate was transferred to a solid LB agar plate. The following day, the overnight incubated (37 • C) LB agar plate was monitored for the antimicrobial effects of azaphilones, and an image was acquired.

Results
Penicillium hirayamae Udagawa (CBS 229.60) was discovered by S. Udagawa in 1959 and has been little studied so far [14]. Indeed, only three metabolites were described from this species, the azaphilones (-) sclerotiorin, (-) sclerotioramine, and rubrorotiorin [15,16]. From the holotype Eupenicillium hirayamae D.B. Scott & Stolk, only the secondary metabolite desoxyverrucarin E was described [17]. Accordingly, in order to isolate new azaphilones from fungi, Penicillium hirayamae U. was purchased, and its crude extract was analyzed by molecular network. This latter pointed out a specific cluster of azaphilones. Indeed, combined annotations from GNPS databases and SIRIUS version 5.5.5 highlighted the presence of various ions identified as azaphilones ( Figure 1). Among the annotated azaphilones, penazaphilone D [8], isochromophilone VI [7], and sclerketide E [6] were fully chemically characterized by 2D NMR after isolation from the Penicillium hirayamae crude extract, reinforcing the GNPS and SIRIUS predictions. Among the known compounds produced by P. hirayamae, only the sclerotiorin was identified from GNPS and Sirius ( Figure S1).
Non-identified ions of the cluster were thus targeted for the next isolation steps, which led to the isolation of three new azaphilones named penazaphilone J-L (1-3) (green nodes, Figure 1). It is noteworthy that grey nodes represent azaphilones present in too small quantities in the extract to be isolated in sufficient amounts and fully elucidated.
The 1 H-1 H COSY spectrum of 1 highlighted the presence of two main spin-coupling systems shown in bold lines ( Figure 2). The first one included the 2 olefinic methines, H9 and H10, as well as the methine H12 and H13, the methylene H14, along with the three methyl H15, H16, and H17 indicating the presence of a 3, 5-dimethyl-1,3-heptadiene moiety as a side chain. The second spin coupling system included the two methylene H1′ and H2′, the methine H3′ as well as the two methyl H5′ and H4′.
Key HMBC correlations from H1 to C3, C4a, C8, and C8a, as well as HMBC correlations from H4 to C4a, C5, C9, and C8a, from H18 to C6, and C8, indicated the presence of an isoquinoline 6,8 (2H, 7H)-dione moiey characteristic of an azaphilone skeleton. HMBC correlations from H-9 to C4 and H10 to C3, along with the 3, 5-dimethyl-1,3-heptadiene moiety spin system detected in 1 H-1 H COSY spectrum allowed to connect the side chain with C3. HMBC correlation between C1 and H1 and H3, together with the chemical shift of the methylene carbon C1 at δ C 53.7, suggested that C1 is linked to an N-atom ( Figure 2). Comparing the NMR data of compound (1) with penazaphilone A [8], we concluded that 1 is a sclerotioramine derivative displaying a decarboxylated leucine located at N2.
2). Comparing the NMR data of compound (1) with penazaphilone A [8], we concluded that 1 is a sclerotioramine derivative displaying a decarboxylated leucine located at N2. The geometry of the double bonds ∆ 9,10 and ∆ 11,12 was deduced based on the coupling constants between H9 and H10 (J = 15.5 Hz). The Overhauser effect (NOE) observed between H9 and H17, and the NOE between H10 and H12, suggesting the E-type ∆ 9,10 and ∆ 11,12 . The observed alternate Cotton effects of 1 on the ECD spectrum were all consistent with those of penazaphilones reported in the literature [8] ( Figure S7). Thus, the absolute configuration was assigned 7S, 13S accordingly, and compound (1) was thus assigned as penazaphilone J.
Compound (2) was obtained as a red amorphous powder and assigned a molecular formula of C29H32ClNO5 based on the HRESIMS peak at m/z 510.2057 [M+H] + . Comparison of 1 H NMR spectra (Table 1) of compounds 1 and 2 showed that both compounds share the same structural features, except the presence of additional aromatic signals at δH 6.66 and 6.96, along with the absence of a dimethyl moiety. The 13 C and HSQC spectra revealed the presence of 4 aromatic carbons along with a hydroxy phenol at 156.3 and a quaternary carbon at δC 126.7, suggesting thus the presence of a phenol on the aminated side chain. This was supported by the 1 H-1 H COSY correlations between the aromatic methine groups H4′ and H5′ and between H1′ and H2′ ( Figure 3). HMBC correlations between H2′ and C3′, C4′ as well as HMBC correlations between H1′, C2′ and C3′ confirmed the presence of a 4-ethyl phenol moiety. Finally, the HMBC correlation of H1′ with C1 and C4, along with the chemical shift of H1′ at δH 4.25, confirmed the presence of the 4-ethyl phenol side chain at N-2. The geometry of the double bonds ∆ 9,10 and ∆ 11,12 was deduced based on the coupling constants between H9 and H10 (J = 15.5 Hz). The Overhauser effect (NOE) observed between H9 and H17, and the NOE between H10 and H12, suggesting the E-type ∆ 9,10 and ∆ 11,12 . The observed alternate Cotton effects of 1 on the ECD spectrum were all consistent with those of penazaphilones reported in the literature [8] ( Figure S7). Thus, the absolute configuration was assigned 7S, 13S accordingly, and compound (1) was thus assigned as penazaphilone J.
Compound (2) was obtained as a red amorphous powder and assigned a molecular formula of C 29 H 32 ClNO 5 based on the HRESIMS peak at m/z 510.2057 [M+H] + . Comparison of 1 H NMR spectra (Table 1) of compounds 1 and 2 showed that both compounds share the same structural features, except the presence of additional aromatic signals at δ H 6.66 and 6.96, along with the absence of a dimethyl moiety. The 13 C and HSQC spectra revealed the presence of 4 aromatic carbons along with a hydroxy phenol at 156.3 and a quaternary carbon at δ C 126.7, suggesting thus the presence of a phenol on the aminated side chain. This was supported by the 1 H-1 H COSY correlations between the aromatic methine groups H4 and H5 and between H1 and H2 ( Figure 3). HMBC correlations between H2 and C3 , C4 as well as HMBC correlations between H1 , C2 and C3 confirmed the presence of a 4-ethyl phenol moiety. Finally, the HMBC correlation of H1 with C1 and C4, along with the chemical shift of H1 at δ H 4.25, confirmed the presence of the 4-ethyl phenol side chain at N-2.  The observed alternate Cotton effects of 2 on the ECD spectrum were all consistent with those of penazaphilones reported in the literature [8] ( Figure S15). Thus, the absolute configuration was assigned 7S, 13S accordingly, and compound (2) was thus assigned as penazaphilone K.
Compound (3) was obtained as a red amorphous powder and assigned a molecular formula of C48H56Cl2N2O10 based on the HRESIMS peak at m/z 891.3401 [M+H] + , indicating 21 unsaturation degrees and suggesting a dimer compound. This compound was consistent with the node at m/z: 446.1716 [M+2H] 2+ observed on the molecular networking. The IR spectrum of 3 featured typical absorption bands for alcohol (3437 cm −1 ), conjugated The observed alternate Cotton effects of 2 on the ECD spectrum were all consistent with those of penazaphilones reported in the literature [8] ( Figure S15). Thus, the absolute configuration was assigned 7S, 13S accordingly, and compound (2) was thus assigned as penazaphilone K.
Compound (3) was obtained as a red amorphous powder and assigned a molecular formula of C 48 H 56 Cl 2 N 2 O 10 based on the HRESIMS peak at m/z 891.3401 [M+H] + , indicating 21 unsaturation degrees and suggesting a dimer compound. This compound was consistent with the node at m/z: 446.1716 [M+2H] 2+ observed on the molecular networking. The IR spectrum of 3 featured typical absorption bands for alcohol (3437 cm −1 ), conjugated ketone (1703 cm −1 ), alkene (1587 cm −1 ), and aromatic amine (1252 cm −1 ). The UV spectrum displayed an absorption maximum of 371 nm. The 1 H NMR and 13 C NMR spectra were identified to have the same planar structures as the sclerotionamine dimer [6,8].

Discussion and Conclusions
By 2024, it is expected that the world color market used in the food industry will reach approximately USD 5.7 billion [18], and the contribution of the global food pigment market is projected to reach 1271.4 million USD by 2025. Unfortunately, the basic science and industry still have not overcome the high cost and low availability of natural pigments. Therefore, the demand for naturally produced biopigments will continue to rise. Microbial pigments in comparison with other natural pigments have several advantages, including an ease of scaling up and harvest, and the fact that they are not subjected to nature vagaries. Fungi are thus alternate sources of naturally derived, stable, and low-cost pigments for food industry applications [19]. One of the most promising classes of fungal pigments in research as industrial pigments are azaphilones, compounds that stand out for their yellow, orange, and red colors [5]. Azaphilones are fungal polyketides containing a highly oxygenated pyranoquinone bicyclic core and produced by numerous species of ascomyceteous and basidiomyceteous [5]. Azaphilones research is extremely important, with abundant literature. For instance, more than 600 new azaphilone derivatives displaying a wide scope of biological activities have been reported [5]. Orange azaphilones generally display a pyran oxygen-containing heterocycle, which is susceptible to aminophilic reactions with peptides, nucleic acids, or proteins. This exchange of oxygen for nitrogen changes the absorption of the pigment from orange to red and may alter its biological properties [20]. For instance, the readiness of fungi-derived red colorants for use in the food industry was discussed recently in an interesting paper by Dufossé [21].
Molecular network (MN) is a method that exploits the assumption that structurally related molecules produce similar fragmentation patterns; therefore, they should be related within a network [22]. In MN, MS/MS data are represented in a graphical form, where each node represents an ion with an associated fragmentation spectrum. The edge among the nodes represents similarities of the MS/MS spectra. Unknown but structurally related molecules can thus be highlighted, and successful dereplication can be obtained [22]. Identification of compounds can also be achieved thanks to the numerous natural products database present in the GNPS platform. A molecular network can thus be an efficient method to detect the presence of new azaphilones among complex fungal crude extract.
Herein, MN mapping of crude extract of P. hirayamae was performed and led to the detection of known alkaloids azaphilones such as penazaphilone D, penazaphilone F, isochromophilone VI and sclerketide E as well as new azaphilones belonging to the same cluster. These azaphilones were specifically sought by MS-guided isolation and led to the identification of three new compounds related to penazaphilone compounds named penazaphilones J-L. These compounds were fully described, and their absolute configuration was performed by ECD. To date, only nine penazaphilones have been previously reported [8]. Interestingly, most of them are devoid of cytotoxicity but display potent antimicrobial activities [23]. Penazaphilones J-L isolated in this framework did not display any cytotoxicity at 100 µg/mL on three different human cell lines used, as well as no antimicrobial on the bacterial strains tested. This study suggested thus that penazaphilones J-L may be potential fungal bio-pigments, and further investigation in the future would be necessary to consider their potential industrial development.
Furthermore, unlike Monascus species, which are not allowed to use azaphilones as food dyes in the USA and European countries due to the presence of co-production of toxic metabolites [24], no toxins were identified from the crude extract of P. hirayamae by MN. Therefore, this result suggests that P. hirayamae could be a potentially safe and promising fungal cell factory for the production of natural polyketide food colors.

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/colorants2010003/s1. Figure S1: Full molecular network of P. hirayamae crude extract. Cluster or azaphilones is highlighted in yellow (known azaphilones) and green (new azaphilones). Diamond indicated azaphilones identified in the GNPS database and red square those identified, thanks to Sirius 5.5.5.; Figure S2 Figure S23: Comparison of the experimental ECD spectrum of 1 and calculated ECD spectrum for the (2R, 7S, 7"S, 13S, 13S") stereoisomer. Figure S24 Absorption and emission spectra of penazaphilone L.