Abstract

Penicillium strains are renowned for producing diverse secondary metabolites with unique structures and promising bioactivities. Our chemical investigations, accompanied by fermentation media optimization, of a newly isolated fungus, Penicillium shentong XL-F41, led to the isolation of twelve compounds. Among these are two novel indole terpene alkaloids, shentonins A and B (1 and 2), and a new fatty acid 3. Shentonin A (1) is distinguished by an unusual methyl modification at the oxygen atom of the typical succinimide ring, a feature not seen in the structurally similar brocaeloid D. Additionally, shentonin A (1) exhibits a cis relationship between H-3 and H-4, as opposed to the trans configuration in brocaeloid D, suggesting a divergent enzymatic ring-expansion process in their respective fungi. Both shentonins A (1) and B (2) also feature a reduction of a carbonyl to a hydroxy group within the succinimide ring. All isolated compounds were subjected to antimicrobial evaluations, and compound 12 was found to have moderate inhibitory activity against Candia albicans. Moreover, genome sequencing of Penicillium shentong XL-F41 uncovered abundant silent biosynthetic gene clusters, indicating the need for future efforts to activate these clusters and unlock the full chemical potential of the fungus.

Introduction

Penicillium, a genus within the Ascomycota phylum, is a type of critical saprophytic fungus with over 400 strains identified in diverse environments such as mountains, oceans, and the human gut . After the first antibiotic mycophenolic acid originally isolated by Gosio in the 1890s , the important antibiotic penicillin was characterized more than one decade after Fleming discovered the antibacterial activity of a Penicillium extract, and since then, Penicillium has been an important target in drug development. Researchers have identified numerous compounds with anticancer properties, including mycophenolic acid, brefeldin A, and wortmannin , as well as compounds with antibacterial properties like xestodecalactones A–C, penicifurans A, and anthraquinone-citrinin . From 2010 to 2022, researchers have identified over 260 secondary metabolites from Penicillium , exhibiting not only antibacterial and anticancer activities but also potent antioxidant properties, inhibition of GSK-3β and α-glucosidase activities, and interaction with the pregnane X receptor (PXR). These compounds are categorized into polyketides, alkaloids, sterol derivatives, terpenoids, and macrolides, with polyketides and alkaloids comprising 40% and 32% of the total, respectively.

Alkaloids are a diverse group of compounds with multiple pharmacological activities, including anti-inflammatory, antibacterial, antiviral, insecticidal, and anticancer properties . Historically, most alkaloids were isolated from higher plants, with a significant number found in the Apocynaceae family. Notable examples such as vinblastine, vinorelbine, vincristine, and vindesine have gained prominence as effective anticancer drugs . Recent studies have revealed that certain fungi are also prolific sources of indole alkaloids, which are among the largest classes of nitrogen-containing secondary metabolites. Characterized by at least one indole moiety and derived from tryptophan or tryptamine, indole alkaloids are known for their diverse structures, electron-donating capabilities, and excellent biocompatibility, contributing to their potent antibacterial and anticancer activities . Over 4000 species producing indole alkaloids have been identified, and many of these compounds are now successfully employed in clinical applications.

Despite the extensive catalog of secondary metabolites discovered, the pace of new findings has decelerated. However, the advent of bioinformatics analysis tools has reinvigorated the search for fungal secondary metabolites. The estimated number of non-redundant clusters in Penicillium is around 25,000 , yet the number of isolated compounds is significantly lower, indicating the presence of many unexpressed gene clusters. This suggests a wealth of undiscovered compounds with potentially novel structures and significant biological activities. To stimulate the expression of biosynthetic gene clusters (BGCs), several methods can be utilized, for instance, epigenetic regulation, co-culture, precursor feeding, heterologous expression, and changing fermentation parameters .

In the present study, we focused on a newly identified Penicillium strain, Penicillium shentong XL-F41. To activate the BGCs of this strain, we employed a combination of elicitors in our fermentation media, including histone deacetylase inhibitors and DNA methyltransferase inhibitors. We developed two specialized media, XISR I and XISR III, which outperformed the traditional potato dextrose broth (PDB) in stimulating the production of a greater number of metabolite peaks, as shown in Figure 1. Scaled-up fermentation allowed us to isolate and characterize two new indole terpene alkaloids, shentonins A and B (1 and 2), a new fatty acid 3, and nine previously identified compounds 4–12, among which were gram quantities of curvularin analogs. Our bioactivity assays identified one compound, 12, with promising antimicrobial properties. Subsequent genome sequencing analysis pinpointed the likely BGCs associated with our isolated compounds and suggested a vast potential for the production of additional compounds, given the application of suitable activation techniques.

Figure 1: HPLC analysis of small-scale fermentation with different media. More details of media, XISR I and XISR III can be found in the methods section.

Results and Discussion Compound isolation and structure elucidation

To activate the silent BGCs in Penicillium shentong XL-F41, we conducted small-scale fermentations using various media. Analysis revealed that HPLC peaks, which correspond to fermentation products, showed a lower number and abundance in the PDB medium than in the XISR I and XISR III media, as illustrated in Figure 1. Consequently, we chose XISR I and XISR III media for further fermentation.

The fermentation broth was exhaustively extracted with ethanol, after which the ethanol extract was partitioned between EtOAc and H2O. The EtOAc fraction was chromatographed repeatedly over silica gel and reversed-phase high-performance liquid chromatography (RP-HPLC), resulting in the isolation of pure compounds 1–12 (Figure 2). According to literature reports of known compounds, some of them were identified as fusarindoles B (4) , dehydrocurvularin (5) , hydroxycurvularin (6) , curvularin (7) , curvulopyran (8) , (S)-6-(sec-butyl)-3-isobutylpyrazin-2(1H)-one (9) , 3,6-di-sec-butyl-2(1H)-pyrazinone (10) , daidzein (11) , and genistein (12) . Notably, compound 7, corresponding to the major peak in our optimized fermentation (Figure 1), was obtained at the gram level.

Figure 2: Chemical structures of compounds 1–12.

Compound 1 (shentonin A) was obtained as a light green solid with a chemical formula of C20H26N2O3, as determined by HRMS m/z 365.1828 [M + Na]+ (calcd for C20H26N2O3Na+, 365.1835) and HRMS m/z 341.1862 [M − H]− (calcd for C20H25N2O3, 341.1870). Spectroscopic analysis, including 1H NMR, 13C NMR (Table 1), and DEPT, revealed that compound 1 contains three methyl groups, one of which is oxygenated, four methines, three saturated non-protonated carbons, and two ketone carbonyl carbons (δC 175.94, δC 194.36). Its NMR data closely resemble those of brocaeloid D , with the notable addition of a methoxy group (δH 3.20/δC 53.92). HMBC correlations confirmed the presence of a reversed prenyl group and differentiated compound 1 from brocaeloid D by the substitution of a succinimide substructure at C-14 with a methine at C-16, indicated by the methoxy group. The position of the methoxy substituent was established by HMBC correlations, and the 13C NMR data suggested that compound 1 includes a 4-oxo-2,3-dihydro-(1H)-quinolin-3-yl fragment. The planar structure was established from HMBC correlations linking three different fragments.

Table 1: 1H and 13C data of compound 1 (recorded in CDCl3).

  δH mult (J in Hz) δC mult 1 4.66 (s) NH 2 3.09 (dd, 3.9, 1.0) 61.4, CH 3 2.92 (t, 7.6) 45.2, CH 4 – 194.3, qC 4a – 116.9, qC 5 7.68 (dd, 7.9, 1.6) 127.2, CH 6 6.61, m 114.9, CH 7 7.28 (d, 1.7) 136.0, CH 8 6.61, m 116.7, CH 8a – 149.9, qC 9 – 43.1, qC 10 5.65 (dd, 17.5, 10.8) 144.3, CH 11 5.02, m 114.3, CH2 12 0.97 (d, 7.2) 23.2, CH3 13 0.97 (d, 7.2) 23.5, CH3 14a 3.16, m 41.8, CH2 14b 3.90 (dd, 13.9, 8.8) – 15 – N 16 4.77 (dd, 6.1, 1.2) 89.7, CH 17a 1.99 (ddd,13.5,9.5) 24.3, CH2 17b 2.12, m – 18a 2.37 (ddd, 17.2, 9.9) 28.9, CH2 18b 2.51 (dt, 17.8, 9.2) – 19 – 175.9, qC 20 3.20, s 53.9, CH3

Compound 1 features three stereogenic centers at C-2, C-3, and C-16. The relative configuration of C-2 and C-3 was determined as (2R*,3R*) by 1H-1H NOESY correlations (Figure 3), while the relative configuration of C-16 remains unresolved due to the inapplicability of the NOESY experiment. To establish the absolute configuration of compound 1, electron circular dichroism (ECD) calculations were conducted using the time-dependent density functional theory (TDDFT) approach at a B3LYP/6-311G (d,p) (IEFPCM) level (Figure 4). Considering the uncertainty of the relative configuration of C-16, both ECD spectra of 2R,3R,16R-1 and 2R,3R,16S-1 were calculated, and compared with experimental ECD. Both calculated spectra displayed almost identical curves compared to the experimental one, which was suggestive of the 2R,3R absolute configuration for compound 1. The calculated spectrum of 2R,3R,16R-1 matched better with the experimental one, so the absolute configuration of 1 was tentatively determined to be 2R,3R,16R, namely shentonin A.

Figure 3: Key 2D NMR correlations of compounds 1–3.

Figure 4: Experimental and calculated ECD spectra at the CAM-B3LYP/6-311G(d,p) level of theory for compound 1.

Compound 2 (shentonin B) was isolated as a light green solid. Its chemical formula, C19H24N2O2, was confirmed by HRMS with m/z 335.1719 [M + Na]+ (calcd for C19H24N2O2Na+, 335.1730) and m/z 311.1755 [M − H]− (calcd for C19H23N2O2, 311.1765). Spectroscopic analysis using 1H NMR, 13C NMR, and DEPT (Table 2) indicated that compound 2 comprises two methyl groups, five methines, five saturated non-protonated carbons, and one ketone carbonyl carbon (δC 174.6). Its NMR profile is similar to brocaeloid C , with the distinction of an added succinimide substructure at N-15, where the ketone carbonyl carbon at C-16 is replaced by a hydroxy carbon. The isoprene group is consistent with that in compound 1. HMBC cross-peaks from H-10 to C-2 and H-12 to C2 connect the indole and isoprene units, while HMBC correlations from H-14 to C-4, C-16, and C-19, and from H-3 to C-4a and C-9, elucidate the connectivity of three fragments. These data collectively establish the planar structure of compound 2.

Table 2: 1H and 13C NMR data of compound 2 (recorded in CDCl3).

  δH mult (J in Hz) δC mult 1 7.90, s NH 2 – 140.0, qC 3 3.09 (ddd, 14.1, 9.3, 5.6) 23.9, CH2 3a 3.16 (ddd, 14.1, 9.5, 6.5) – 4 – 108.1, qC 4a – 129.6, qC 5 7.61 (dp, 7.8, 0.7) 118.3, CH 6 7.09 (ddd, 8.1, 7.0, 1.1) 119.6, CH 7 7.14 (ddd, 8.1, 7.0, 1.2) 121.7, CH 8 7.29 (dt, 8.0, 1.0) 110.6, CH 8a – 134.2, qC 9 – 39.1, qC 10 6.14 (dd, 17.4, 10.5) 146.0, CH 11 5.16 (d, 1.1) 112.2, CH2 11a 5.18 (dd, 2.5, 1.1) – 12 1.56 (d, 1.5) 27.7, CH3 13 – 27.8, CH3 14 3.51 (m) 41.5, CH2 14a 3.67 (ddd, 13.7, 9.5, 5.6) – 15 – N 16 4.98 (s) 84.3, CH 17 2.30 (ddd, 17.1, 10.1, 4.3) 28.9, CH2 17a 2.54 (ddd, 16.9, 9.7, 7.2) – 18 1.76 (dddd, 13.8, 9.7, 4.3, 2.4) 28.7, CH2 18a 2.19 (dddd, 13.7, 10.1, 7.3, 6.4) – 19 – 174.6, qC

Compound 3 was isolated as a transparent oily liquid, and its chemical formula, C16H28O4, was confirmed by LC–MS with m/z 283.2 [M − H]− (calcd for C16H27O4, 283.2) (Figure S25 in Supporting Information File 1). Spectroscopic analyses, including 1H NMR, 13C NMR, DEPT, HSQC, COSY, and HMBC (Table 3, Figure 3), identified compound 3 as a sixteen-carbon fatty acid. Notably, two methylene carbons overlapped in the 13C NMR spectrum. The COSY correlations facilitated the determination of the carbon chain fragments from C-11 to C-16 and C-2 to C-10, despite two methylene signals overlapping. The carboxyl group's position at C-1 was confirmed by HMBC correlations from H-2/3. Furthermore, HMBC cross-peaks from H-12 to C-10, H-11 to C-9, and H-10 to C-12 indicated that the fragments are connected through C-11 and C-10, establishing the structure of compound 3.

Table 3: 1H and 13C NMR data of compound 3 (recorded in CDCl3).

  δH mult (J in Hz) δC mult 1 – 177.2, qC 2 2.41, m; 1.96 (dp, 12.9, 9.3) 29.3, CH2 3 2.57, m 29.0, CH2 4 2.57, m 29.0, CH2 5 2.41, m; 1.96 (dp, 12.9, 9.3) 29.3, CH2 6 5.30, m 76.4, CH 7 5.47, m 127.5, CH 8 5.66 (dt, 11.0, 7.5) 134.0, CH 9 2.86 (dt, 15.3, 7.3) 26.3, CH2 9a 2.96 (dt, 13.5, 7.6) – 10 5.36, m 126.9, CH 11 5.47, m 131.0, CH 12 2.19 (q, 7.6) 23.7, CH2 13 1.49, m 36.5, CH2 14 3.53 (tt, 8.3, 4.4) 72.8, CH 15 1.49, m 30.5, CH2 16 0.94 (t, 7.5) 10.0, CH3 Biological activities

In our bioassays, we evaluated the inhibitory activity of all isolated compounds against a panel of microorganisms (Table 4), including Escherichia coli (ATCC 25922), Candida albicans (ATCC 76485), Staphylococcus aureus (ATCC 27154), Pseudomonas fulva (CGMCC 1.15147), and Enterobacter hormaechei (CGMCC 1.10608). The results indicated that compounds 3, 5, 6, 7, and 12 were active against Candida albicans. Notably, compound 12 showed particularly promising inhibitory activity against this fungal pathogen.

Table 4: Antimicrobial activity of compounds 1–12. Minimum inhibitory concentrations were shown in µg/mL.

No. Escherichia coli Candida albicans Staphylococcus aureus Pseudomonas Fulva Enterobacter hormaechei 1 – – >100 – – 3 – 50–100 – >100 >100 4 – >100 – – – 5 >100 25–50 >100 >100 >100 6 >100 25–50 >100 >100 >100 7 >100 64–128 >100 >100 >100 8 – – – >100 >100 9 >100 >100 – >100 >100 11 – – – – – 12 >100 12.5–25 >100 >100 >100 Genome sequencing analysis

The genome sequencing yielded 7,118,236 reads with an average read length of 1,858.7 bp. The assembled genome is 34,621,366 bp long, comprising 9 contigs with a mean contig length of 3,846,818.44 bp, and the longest contig is 5,975,444 bp. The genome's GC content is 46.43%. Annotation of the genome sequence of Penicillium shentong XL-F41 identified 11,235 coding sequences and 172 tRNA genes.

Upon utilizing the fungal version of antiSMASH 7.0 software for the analysis of the Penicillium shentong XL-F41 genome, we identified 46 BGCs. These include 13 NRPS-like fragments, 6 NRPS, 13 type I PKS, 2 PKS/NRPS hybrids, 1 NI-siderophore, 2 NRP-metallophore/NRPS hybrids, 1 fungal RiPP with POP or UstH peptidase types, 1 fungal-RiPP-like/T1PKS, 1 betalactone, 1 PKS type I/NRPS/indole hybrid, 1 fungal-RiPP-like/T1PKS hybrid, 1 NRP-metallophore/NRPS hybrid, NRPS-like/terpene/phosphonate hybrids, 3 terpenes, and 1 indole-related cluster (Table 5).

Table 5: Biosynthetic gene clusters of the Penicillium shentong XL-F41.

BGC Type Putative product 1.1 NRPS-like   1.2 NRPS-like   1.3 NI-siderophore   1.4 NPR-metallophore, NRPS   1.5