Exploring Bioactive Fungal RiPPs: Advances, Challenges, and Future Prospects

Abstract

Fungal ribosomally synthesized and post-translationally modified peptides (RiPPs) represent a vital class of natural products with their potent biological activities, including anticancer, antitubulin, antinematode, and immunosuppressant effects. These bioactive fungal RiPPs play significant roles in chemical ecology and hold considerable therapeutic potential. The structural diversity of fungal RiPPs is particularly striking, driven by complex post-translational modifications of precursor peptides. Despite their importance, discovering fungal RiPPs has historically been challenging. However, the discovery of fungal RiPPs has been historically challenging, with only a limited number identified so far. These existing fungal RiPPs are primarily categorized into three groups: cycloamanides and borosins from basidiomycetes, and dikaritins from ascomycetes. Recent advancements in bioinformatics have uncovered the vast, untapped potential of fungi to produce RiPPs, offering new opportunities for their discovery. This perspective provides an overview of recent progress in the biosynthesis of fungal RiPPs and the genome-guided discovery of these compounds. We propose that integrating insights into the biosynthetic hallmarks of fungal RiPPs biosynthesis with cutting-edge gene manipulation techniques and bioinformatic approaches will accelerate the discovery of novel bioactive fungal RiPPs.

Introduction

Natural products have long been recognized as crucial sources for new drug discovery, with more than half of FDA-approved drugs derived from these origins.¹ These natural products hold significant potential for uncovering novel drug candidates and bioactive chemical templates. Fungi, in particular, offer a prolific source of diverse bioactive secondary metabolites, making them invaluable in producing unique chemicals for combating various diseases.² The complexity and diversity of fungal metabolites contribute to their richness as sources of potential therapeutics. Notable examples include penicillin, used against bacterial infections; lovastatin, for treating high cholesterol; and cyclosporins, as immunosuppressant medications.^3,4

Ribosomally synthesized and post-translationally modified peptides (RiPPs) are an important family of natural products known for their diverse chemical structures and potent biological activities.^5–7 While RiPPs are predominantly found in bacteria, a limited number have been identified in fungi. Many of these fungal RiPPs were initially misclassified as non-ribosomal peptides, largely due to their structural complexity and the sophisticated nature of their biosynthetic pathways. Notable examples of fungal RiPPs include cycloamanides, such as α-amanitin (1) and cycloamanide B (2); the borosins, exemplified by omphalotin A (3) from basidiomycetes; and dikaritins, represented by ustiloxins (4 and 5), phomopsin A (6), and victorin C (7) from ascomycetes (Figure 1).⁸ Despite the limited numbers, fungal RiPPs exhibit a wide array of biological functions and demonstrate remarkable chemical diversity.

Figure 1. Representative bioactive fungal RiPPs.

Basidomycetes-derived RiPPs are represented by cycloamanides including α-amanitin and cycloamanide B, and borosin including omphalotin A. Dikaritins are ascomycetes-derived RiPPs, exemplified by phomopsin A, ustiloxins, and victorin C.

Biosynthetically, RiPPs typically originate from precursor peptides composed of leader and core peptides. The intricate post-translational modifications of these precursor peptides result in the maturation of RiPPs, which significantly contributes to their structural diversity.⁵ However, our understanding of the biosynthesis of fungal RiPPs is limited, which has hindered the discovery of new fungal RiPPs. Despite this challenge, the expansion of fungal genomic databases and advancements in genetic analysis tools have revealed numerous biosynthetic gene clusters (BGCs) associated with fungal RiPPs. These findings underscore the immense potential of fungi to produce this largely unexplored class of natural products.⁹ Therefore, developing efficient strategies for identifying fungal RiPPs is of great importance.

This review explores the bioactivity and biological functions of the major classes of fungal RiPPs, emphasizing their essential roles in chemical ecology and therapeutics. Additionally, we provide an overview of recent research on fungal RiPP biosynthesis and efforts in genome-guided discovery, which lays the foundation for the identification and characterization of fungal RiPPs. Finally, we discuss the challenges and future perspectives in genome mining to facilitate the discovery of more fungal RiPPs.

Major classes of the known fungal RiPPs

Cycloamanides

Amatoxins and phallotoxins are bicyclic cycloamanides RiPPs, characterized by head-to-tail macrocyclization and the tryptathionine bridge between Trp and Cys. They are the primary toxins responsible for poisoning from mushrooms like Amanita phalloides, commonly known as the Deathcap fungus.¹⁰ α-Amanitin (1), one of the most toxic amatoxins, shows high toxicity to insects, nematodes, and mammals (Figure 1), with a median lethal dose of in humans of approximately 0.1 mg/kg.¹⁰ It is widely hypothesized that these toxins evolved as a defense mechanism against predation by microbes, invertebrates, and vertebrates, although it still has not been fully verified by chemical ecology studies.¹¹ α-Amanitin specifically inhibits RNA polymerase II, leading to a dramatic decrease in transcription and protein synthesis.^12,13 The co-crystallization of α-amanitin and RNA polymerase II reveals that amanitin binds adjacent sites of the elongation complex and contacts the trigger loop and bridge helix.^14,15 Despite the severity and potentially fatal outcome of poisoning from amanitin-containing mushrooms, no standardized treatment regimens are currently available. Over the past decades, several clinical drugs, including silybin and penicillin, have demonstrated potent therapeutic efficacy against human amatoxin poisoning.¹⁶ Additionally, through virtual docking, polymyxin B, a potential RNA polymerase II inhibitor, has been shown to alleviate amanitin toxicity.¹⁷ Recently, by a genome-wide CRISPR screen, STT3B, a key component in the N-glycan biosynthesis pathway has been validated to play a crucial role in α-amanitin toxicity.¹⁸ An STT3B inhibitor, ICG, has been shown to effectively block the toxic effects of α-amanitin, leading to a significant increase in animal survival. The potent cytotoxicity of amanitin makes it challenging to be applied in the clinic to treat cancer, therefore, amanitin-based antibody-drug conjugates (ADCs) have been developed to facilitate its efficacy and safety in treatment.¹⁹ For example, ToxiMab, a covalent conjugate of PankoMab and β-amanitin shows potent and specific cytotoxicity toward T-47D tumor cells.²⁰ The conjugation of a human anti-CD117 antibody with amanitin affords targeted conditioning and anti-leukemia ADC.^21,22 Furthermore, amanitin-based ADCs have been demonstrated to overcome therapy resistance and are effective against dormant tumor cells.¹⁹ Although phallotoxins, such as phallacidin, share high structural similarity with amatoxins, they function differently. Unlike amanitins, phallotoxins can bind and stabilize F-actin, preventing the depolymerization of actin fibers.²³ Furthermore, phallotoxins are non-poisonous when ingested orally due to poor absorption.²⁴ Additionally, monocyclic cycloamanides RiPPs have also been found in Amanita. Among those, cycloamanides A-F, represented by cycloamanide B (2) shown in Figure 1 from A. phalloides displayed immunosuppressive activity.

Until 2007, two genes AMA1 and PHA1, identified in Amanita bisporigera, directly encode α-amanitin and phallacidin, indicating that these compounds are synthesized on ribosomes rather than by nonribosomal peptide synthetases.²⁴ In this study, the prolyl oligopeptidase was proposed to participate in processing the proproteins of the Amanita toxins. Subsequently, Luo et al. characterized a prolyl oligopeptidase B (POPB) from Galerina marginata, named GmPOPB which can remove the N-terminal leader peptides at the first Pro to release the C-terminal propeptide. Additionally, it catalyzes the subsequent transpeptidation at the second Pro to produce the head-to-tail octacyclic peptide (Figure 2A). Moreover, amanitin and AbPOPB, a POPB from A. bisporigera, showed a high degree of colocalization, indicating that the biosynthesis and accumulation of the amatoxin occur in the same cells revealed by confocal immunomicroscopy.²⁵ This finding also indicates that amanitin likely accumulates in specialized storage vacuoles. Cystidia, commonly located on the surface of gills in basidiomycetes and characterized by their distinctive morphologies, are proposed to function as storage cells to mitigate the toxicity of amatoxins, although their precise role and relationship to these toxins remain unconfirmed.¹¹ Accordingly, POPB is indispensable for the maturation of cycloamanides, marking a distinct feature of their biosynthesis. In 2020, Luo et al. reported that a cytochrome P450 GmP450–29 can catalyze hydroxylations of proline and isoleucine. Additionally, a flavin monooxygenase family protein GmFMO1 is involved in the production of α-amanitin (Figure 2A). The deletion of GmFMO1 abolished the production of amanitin and led to the accumulation of three new compounds, each corresponding to α-amanitin with one, two, or three fewer oxygen atoms, respectively.²⁶ However, due to the difficulty of culturing GmFMO1 deletion mutant, these compounds have not yet been isolated and identified, necessitating further investigation.

Figure 2. Biosynthetic pathway of basidiomycetes-derived RiPPs.

Proposed biosynthetic pathways of amanitin (A) and omphalotins (B).

Borosins

Borosins are a family of head-to-tail cyclized, heavily backbone α-N-methylated peptides. Omphalotins, the most well-studied class of the borosin family, are modified cyclic dodecapeptides with a core peptide sequence of WVIVVGVIGVIG. Omphalotin A (3) displays a remarkable selective nematocidal activity towards the plant pathogenic nematode Meloidogyne incognita, with no obvious toxicity against other nematode species.^27,28 Its potency surpasses that of the bacterial nematotoxin ivermectin, and it shows no evident toxicity towards mammalian cells, plants, insects, fungi, or bacteria, suggesting its promising in nematode control.^27,29 Recently, two omphalotin analogs, lentinulin A from Lentinula edodes and dendrothelin A from Dendrothele bispora, were discovered, featuring core sequences of WIIVVGVVGVVG and WVIVTGIVGVIG, respectively. Lentinulin A shows nematocidal activity against M. incognita comparable to omphalotin A, whereas dendrothelin A displays significantly lower activity, indicating that the mutation of Val to Thr in dendrothelin A may reduce its nematotoxicity.²⁸ These findings highlight the critical role of core peptide sequences in determining both the bioactivity and selectivity of borosins, likely due to their influence on target accessibility or peptide stability across different nematode species. Despite these insights, the precise mode of action of borosins remains unclear and warrants further investigation.

Although α-N-methylation is typically associated with non-ribosomal peptides, the release of genomic data from omphalotin-producing organisms has revealed that these compounds are, in fact, ribosomal peptides.^30,31 The multiple α-N-methylated backbone is biosynthesized by an autocatalytic enzyme OphA (also known as OphMA) fusing the precursor peptide and a methyltransferase domain (Figure 2B).³¹ The structure of OphA has been determined, revealing that two precursor peptides interlock as a homodimer. Subsequent mutagenesis studies indicate that OphA facilitates the activation of the amide bonds for methylation by the removal of the amide proton and stabilization of the resulting negative charge.³² Guided by the crystal structure of OphA, the rationalized design for genetic mutagenesis offers omphalotin variants, highlighting the substrate promiscuity of this enzyme and its potential as a catalyst.³³ The methylated OphA is subsequently processed and cyclized by a prolyl oligopeptidase OphP to release the omphalotin A.³⁴ Following hydroxylations and acetylations were proposed to be catalyzed by two additional P450s (OphB1/OphM1 and OphB2/OphM2) and an acetyltransferase (OphC) to provide omphalotin analogs including omphalotins B, C, and D (8-10) (Figure 2B).^34,35

Dikaritins

Dikaritins play essential roles in chemical ecology, as they are predominantly produced by plant pathogens and closely linked to their pathogenicity. Phomopsin A (4), a cyclic hexapeptide mycotoxin produced by the fungal pathogen Phomopsis leptostromiformis, infects lupins and causes lupinosis disease in livestock.^36,37 Ustiloxins A (5) and B (6), sharing high structural similarity with phomopsins, are circularized tetrapeptide and an additional sulfur-containing side chain, first isolated from the plant pathogen Ustilaginoidea virens. They can infect the spikelets of rice, leading to the rice false smut disease, one of the most destructive rice diseases.^38–40 Ustiloxins acting as phytotoxins can inhibit the growth of rice, wheat, and maize.³⁸ Additionally, both phomopsins and ustiloxins are antimitotic agents and show cytotoxicity against cancer cell lines.^38,41,42 Both them can potently inhibit tubulin polymerization and suppress mitosis, which accounts for their bioactivity.⁴³ Studies on the structure-activity relationship of ustiloxins have shown that the free phenolic hydroxyl group and the S-configuration of the methylamino group are essential for bioactivity.^44,45 Additionally, victorin is a mixture of highly modified heterodetic cyclic hexapeptides with a distinct five-member ring with victorin C (7) being the predominant component. Victorin is the host-specific toxin of Cochliobolus victoriae, which causes the Victoria blight of oats.^46–48 The contamination of food crops with these compounds poses a significant safety risk and financial burden. Unlike phomopsins and ustiloxins, the phytotoxicity of victorin is primarily attributed to the inhibition of plant immunity regulatory protein, thioredoxin TRX-h5, which activates LOV1 and triggers a resistance-like response, leading to disease susceptibility.^49–51

In contrast to basidiomycetes-derived RiPPs, the precursor peptides of dikaritins derived from ascomycetes generally contain the N-terminal signal peptide, followed by multiple perfect or imperfect repeats that each contain a core peptide. Each repeat is flanked dibasic motif such as Lys-Arg (KR) motif, which was proposed to be cleaved by Golgi-specific protease Kexin endopeptidases.^52,53 Another hallmark of dikaritins is that they all possess ether bonds for their cyclization (Figure 1), rather than the head-to-tail cyclization via amide bonds in basidiomycetes-derived RiPPs. Unlike bacteria and plants RiPPs whose ether bonds are primarily facilitated by radical SAM enzymes⁵⁴ and BURP domains^55,56 respectively, fungi-specific DUF3328 oxidases featuring the conserved HXXHC are typically found in the fungal RiPPs biosynthesis to form the cyclization-related ether bonds.^9,57

The biosynthetic pathway of ustiloxins was elucidated by Ye et al. in 2016 through gene inactivation in another producer of ustiloxins Aspergillus flavus, heterologous expression in Aspergillus oryzae, and in vitro functional analyses.⁵⁸ After the cleavage of 16-repeats precursor peptide UstA to single core peptide with sequence YAIG, a single hydroxylation at benzylic position and the macrocyclization via an ether bond are catalyzed by two DUF3328 protein UstYa and UstYb along with a tyrosinase UstQ to produce N-demethylustiloxin F (11) (Figure 3A). The following N-methylation by methyltransferase UstM and the C-S bond formation by adding cysteine to the aromatic ring with a cytochrome P450 UstC provide S-deoxyustiloxin H (12) (Figure 3A).^58,59 The oxidation of thioether to sulfoxide and decarboxylative N-oxidation mediated by two flavoproteins UstF1 and UstF2, respectively offer oxime intermediate 13 which can be hydrolyzed under acid conditions to form the aldehyde intermediate 14. Finally, a pyridoxal phosphate (PLP)-dependent enzyme UstD catalyzes the decarboxylation of L-aspartic acid to form an enamine unit and subsequent condensation with aldehyde generates the final product ustiloxin B (6) (Figure 3A). Recently, through the direct evolution of UstD, this distinct decarboxylative aldol reaction was applied in asymmetric chemoenzymatic synthesis of γ-hydroxy amino acid.⁶⁰

Figure 3. Biosynthesis of ascomycetes-derived RiPPs.

Proposed biosynthetic pathways of ustiloxins (A), phomopsin (B), and victorin C (C).

In 2016, Ding et al. first identified phomopsins as fungal RiPPs and characterized their BGC in the producer P. leptostromiformis⁶¹. Subsequent research by Sogahata et al. further elucidated the biosynthetic pathway of phomopsins⁵⁷. The precursor peptide PhomA contains five repeats of short peptides with the core peptide sequence YVIPID (Figure 3B). Similar with UstQ, PhomQ2 is likely involved in the macrocyclization process, while the other tyrosinase, PhomQ1, functions as a halogenase. The catalytic roles of five DUF3328 proteins, PhomYa-Ye, are proposed to be oxidative cyclization, hydroxylation, and C=C double bond formation in the biosynthesis of phomopsins as illustrated in Figure 3B.^8,57 Additionally, a methyltransferase, PhomM, was verified to realize the α-N-methylation of the Tyr-derived residue to provide phomopsin E.⁶¹ However, the timing and the order of proteolytic release of core peptide regions and post-translationally modification are still unknown.

Until 2020, the BGC of victorins was demonstrated by Kessler et al. from the producing fungus and confirmed the ribosomal origin of victorins.⁴⁸ The characterization of their BGC was challenging for researchers, due to three copies of the precursor peptide-encoding gene, vicA, located in two genomic loci. Additionally, VicA exhibits variable numbers of core peptide repeats containing the GLKLAF core sequence. Based on gene deletion studies, a DUF3328 enzyme, VicYb, was proposed to participate in the ether bond formation in the cyclization process. Interestingly, the copper-amine-oxidase VicK catalyzes oxidative deamination, transforming HV-toxin M (15) with the unmodified N-terminal glycine into the corresponding compound 16, which contains a glyoxylate residue that is indispensable for the bioactivity of victorins and is interconvertible with victorin C (7) (Figure 3C).⁴⁸ However, enzymes responsible for other modification steps, such as chlorination and the formation of the five-member ring via the dearomatization and ring contraction of the benzene ring in Phe residues remain elusive.

Recent strategies of genome-guided identification of fungal RiPPs

With the development of bioinformatic tools and MS-based metabolism analysis, combined with the current knowledge of their biosynthesis, there is now an opportunity to identify fungal RiPPs through genome mining. Bioinformatic analysis revealed that genomes of both A. phalloides and A. bisporigera contain approximately 30 candidate RiPPs-related genes, most of which are predicted to encode unknown cycloamanide family peptides. Notably, 20 of these genes in A. bisporigera were found to be expressed, suggesting a significant potential for the production of additional cycloamanides by this genus.⁶² Furthermore, the analysis identified the genes encoding cycloamanide B (2), along with two cyclic peptides—cycloamanides E and F—with structures cyclo(SFFFPVP) and cyclo(IVGILGLP), respectively, which were detected using LC-MS/MS in A. phalloides. Moreover, Vignolle et al. combined the existing bioinformatic tools including antiSMASH and RiPPMiner with RNASeq data to find fungal RiPPs effectively. The analysis revealed an abundance of cycloamanide-like precursor peptide genes in the Trichoderma genome, indicating the potential of Trichoderma to produce cycloamanide RiPPs.⁶³ Quijano et al. utilized the methyltransferase domain of OphMA as a probe for genome mining to expand the borosin family. By combining bioinformatic analysis with heterologous expression of substrate-fused α-N-methyltransferases in E. coli, they identified over 50 putative borosin pathways and functionally characterized 11 new α-N-methylating catalysts.⁶⁴ In this study, gymnopeptides produced by the basidiomycete Gymnopus fusipes were first identified to be biosynthesized by the borosin RiPP pathway. Recently, Lee et al. developed a new RODEO module, which resulted in the identification of a native substrate of shewasin A, a previously known aspartyl peptidase with no identified substrates.⁶⁵

The discovery of dikaritins mainly relies on the two distinct properties involved in their biosynthetic pathway: the repeats in precursor peptides and the presence of DUF3328 enzymes. GigA, with high expression in endophyte-infected grass-associated Epichloë spp. features varying numbers of imperfect repeats and kexin protease cleavage sites across different strains. The metabolic profile of gigA deletion mutant let Johnson et al. identify five major products, known as epichloëcyclins (17), featuring heptapeptide macrocyclic scaffold with an ether bond, revealed by MS/MS spectra (Figure 4A).⁶⁶ Recently, Zhang et al. further characterized the BCG of epichloëcyclins. A DUF3328 oxidase GigB and a methyltransferase GigC catalyze the cyclization N-methylation of epichloëcyclins, respectively (Figure 4A).⁶⁷ Furthermore, Nagano et al. developed a pipeline to mine the genomic database of Aspergillus species to find ustiloxin homologous gene clusters, revealing 94 additional ustiloxin-like precursor peptide genes.⁹ One of the BGCs from Aspergillus flavus, was found to mediate the biosynthesis of a novel cyclic peptide, asperipin-2a (18) (Figure 4B). In 2019, the biosynthesis of 18 was reconstituted in the heterologous host Aspergillus oryzae.⁶⁸ Besides precursor peptide AprA, DUF3328 enzyme AprY, a reductase AprR, and a transporter AprT are responsible for sequential oxidative macrocyclization, deamination, and transport, respectively to produce the bi-macrocyclic compound 18.⁶⁸

Figure 4.

Genome-guided discovery and biosynthesis of epichloëcyclin B (A) and asperipin-2a (B).

Challenges and perspective for fungal RiPPs discovery

Despite the great efforts that have been made, fungal RiPPs are still largely untapped, suggesting that the identification of fungal RiPPs has been a significant challenge. Historically, most known RiPPs were initially misannotated as non-ribosomal peptides (NRPs), due to the extensive overlap in post-translational modifications between these two peptide types, such as head-to-tail cyclization and N-methylation.^69,70 This misclassification implies that some fungal RiPPs may have been overlooked or are still difficult to detect, even when present. However, unlike NRPs, fungal RiPPs do not typically contain non-proteinogenic amino acids in their peptide backbones.⁶⁹ Furthermore, ascomycete-derived RiPPs are macrocyclized via ether bonds, whereas fungal NRPs are usually cyclized by amide or ester bonds. These distinctions may provide a means of differentiating between the two peptide types. Although progress in elucidating some key hallmarks of fungal RiPP biosynthesis, substantial gaps remain in our knowledge. Notably, some enzymes involved in post-translational modifications, such as those responsible for cross-bridge formation in amanitins and the condensation of the benzene ring in victorins, are still unidentified. Additionally, the molecular mechanisms of critical tailoring enzymes, particularly DUF3328 oxidases with diverse catalytic functions including hydroxylation, C=C bond formation, and chlorination,^57,58,71 remain poorly understood, significantly impeding the discovery of new fungal RiPPs.

Advancements in genome mining tools and techniques over the past decade have greatly enhanced the ability to predict and identify new RiPPs. Beyond traditional methods like basic local alignment search tool (BLAST), newer approaches involve machine learning and neural network algorithms, which allow for the prediction of the new RiPPs with unconventional BGCs.⁷² Several RiPP-specific genome-mining tools such as BAGEL3⁷³, RODEO⁷⁴, RiPPPRISM⁷⁵, and fungal RiPPs specific fRIPPA⁸ have been developed to expand the library of RiPPs.⁶⁴ Furthermore, mass spectrometry has also emerged as a powerful technique for identifying and characterizing metabolites in natural product research.^72,74 Incorporating MS-based methods into the discovery process enhances the accuracy and efficiency of RiPP identification.⁵⁵ Tools like RiPPQuest⁷⁶ and PepSAVI-MS⁷⁷ integrate MS data with genome mining, enabling the correlation of genetic information with metabolite profiles

Although the bioinformatic analysis of the fungal genome database has revealed the widespread presence of RiPP BGCs across fungal genomes, indicating a huge potential for fungi to produce these compounds, these BGCs are either expressed at low levels or not expressed under laboratory conditions, leading to their undetectability. Therefore, developing effective strategies to activate these potential fungal RiPP BGCs is also crucial for advancing the discovery of fungal RiPPs. Several classical methods are available to access the products of silent BGCs, including OSMAC (One Strain Many Compounds), heterologous expression, promoter manipulation, and so on.^78,79 Recently, CRISPR-Cas9-based technologies, including CRISPR base editing and CRISPR activation, have provided opportunities to efficiently manipulate regulatory genes or biosynthetic genes for the activation of BGCs.^80–83

The expanding understanding of RiPP biosynthesis is driving genome mining strategies to play an increasingly important role in discovering new RiPPs. Identifying novel fungal RiPPs and elucidating the details of their biosynthesis will enhance our understanding of these compounds, creating a positive feedback loop between biosynthesis studies, genome mining, and RiPP discovery.

Conclusion

Fungal RiPPs are unexplored treasures in nature products, representing a highly diverse array of chemicals with remarkable structural and biological activity diversity. They play an important role in chemical ecology and show great promise as potential drug leads in pharmaceutical applications. In this review, we highlight recent advances in the biosynthetic studies of fungal RiPPs and discuss the challenges associated with their discovery. We hope integrating bioinformatics, mass spectrometry, and biosynthetic investigation can enhance the identification of novel fungal RiPPs and deepen our understanding of their biosynthetic pathways.