Fungi harbour a vast superfamily of antifungal proteins that may act like secondary metabolites

A new study has catalogued nearly 500 small, cysteine-stabilised proteins from filamentous fungi and proposes that they form a single evolutionary superfamily with properties reminiscent of fungal secondary metabolites. The research, published on 29 August 2026 in the open-access journal Fungal Biology and Biotechnology, was carried out by Claudia Feurstein, Lisa Bachmann, Lena Heber, Birgit Dobbert and Sascha Jung. Using database searches, phylogenetics and structural modelling, the team identified 165 AFP-like, 102 BP-like and 219 NFAP2-like proteins, all of which were encoded exclusively by fungi. To the authors’ knowledge, this is the first study to show that this group of bioactive proteins is shared between the two fungal divisions Ascomycetes and Basidiomycetes.

The work began with an experimental discovery. While optimising the purification of the well-known antifungal protein AgAFP from Aspergillus giganteus, the researchers isolated a previously uncharacterised protein from the culture medium. Mass spectrometry identified three peptides matching bubble protein family members, including the highly conserved sequence HFCGCDR. Searching the genome of A. giganteus strain CBS 515.65 with this fragment revealed the corresponding gene, agbp, on scaffold 69, and the mature protein was named AgBP. Its predicted structure shows a knottin-fold stabilised by four disulfide bridges, consistent with other bubble proteins. Notably, the agbp gene was 100% conserved across five A. giganteus strains tested, whereas the agafp gene showed several nucleic acid and amino acid deviations, hinting at different mutation rates for the two genes.

In laboratory assays, AgBP showed a relatively narrow spectrum of growth inhibition. It suppressed the filamentous fungi Aspergillus niger with a minimal inhibitory concentration of roughly 50 micrograms per millilitre and Fusarium sporotrichioides at about 6.25 micrograms per millilitre, and inhibited the yeast Pichia pastoris, now called Komagataella phaffii, at approximately 1.6 micrograms per millilitre. Growth of bacteria such as Escherichia coli, Bacillus subtilis and Enterococcus species, as well as the yeasts Saccharomyces cerevisiae and Candida albicans and the fungus Penicillium chrysogenum, was not inhibited even at concentrations up to 100 micrograms per millilitre. The authors stress that this screening provides only a preliminary overview, and that more organisms and experimental conditions must be tested to evaluate the antimicrobial character of AgBP fully. A paradoxical effect on growth was observed with F. sporotrichioides and, to a lesser extent, with P. pastoris.

To map the wider reservoir of these proteins, the team used NCBI’s PSI-BLAST with the pre-pro-protein sequences of AgAFP, PbBP from Penicillium brevicompactum and NFAP2 from Neosartorya fischeri, now renamed Aspergillus fischeri, as queries. After applying filter criteria based on length, cysteine content and the presence of gamma-core motifs, the searches returned 165 AFP, 102 BP and 219 NFAP2 homologues. Six AFP family members and all 219 NFAP2 family members had not yet been assigned to a protein family in the InterPro database, because the NFAP2 family has yet to be formally defined there. All 486 proteins were found exclusively in fungi, and no yeast strains appeared among the encoding organisms.

The taxonomic distribution was strikingly uneven. The classes Eurotiomycetes, accounting for 35.6% of encoding organisms, and Sordariomycetes, with 46.5%, within the phylum Ascomycetes, and the class Agaricomycetes, with 14.0%, within the Basidiomycetes were the most abundant. At the genus level, Aspergillus contributed 18.3% of BPF-encoding fungi, Penicillium 15.8% and Claviceps 10.1%. Aspergillus and Penicillium dominated the AFP and BP families, whereas the main genera encoding NFAP2 members were Claviceps, with 22.4%, and Hypoxylon, with 9.6%; Penicillium encoded no NFAP2 member at all. Most strains, 90.0% or 325 organisms, encoded only one member of the three families, 9.1% or 33 encoded members of two, and just 0.8% or three organisms encoded members of all three. Only Apiospora kogelbergensis strain CBS 113332, Aspergillus fischeri strain NRRL 181 and Aspergillus hancockii strain FRR 3425 carried all three families.

Phylogenetic analyses of both precursor and mature protein sequences, built with maximum-likelihood methods and rooted with a radish defensin, showed a neat separation of the AFP, BP and NFAP2 families. The authors suggest two major speciation events: one that separated the BP-like clade from the AFP-NFAP2-like clade, and a second that differentiated AFP-like from NFAP2-like proteins, with AFP family members possibly representing the origin of the currently known families through gene duplication and subsequent change of function. Comparison of the two trees indicated that the mature proteins and their pre-pro-regions evolved at different rates. Subgroups within each family could be distinguished, for example AFP proteins with six versus eight conserved cysteines, yet all 486 proteins were clearly assigned to their respective families, underscoring conserved differences in overall primary structure.

When the researchers compared the encoding organisms’ genomes using the UFCG pipeline, the picture changed. The spread of the three protein families across the fungal species tree was patchy and showed phylogenetic incongruence with the protein tree, a pattern the authors interpret as a possible signature of horizontal gene transfer. Within individual genera, evolutionarily younger species encoded more BPF members than older species in Penicillium and Fusarium, while Aspergillus appeared limited to either an AFP or a BP member over time, and BP-encoding species appeared later than AFP-encoding species in all genera examined. Three genera, Hypoxylon, Schizophyllum and Claviceps, exclusively encoded NFAP2 members. The authors are careful to note that horizontal gene transfer remains a hypothesis: further research, including analyses of GC content, synteny and codon usage, is required, and genomic context analysis has not yet been performed.

Tertiary structure comparison added a further layer of evidence. Using USalign and multidimensional scaling of structural distances, the team found that members of the AFP, BP and NFAP2 families each cluster within their own family and that all three families clearly separate from small, cysteine-stabilised antifungal proteins of non-fungal origin. This structural distinctness, combined with the phylogenetic separation, supports treating the three families as distinct units within one fungal superfamily, which the authors named BPF, for bioactive proteins from fungi. They distinguish this group from the recently defined Fungal-Derived Bioactive Peptides, which comprise peptides of only 2 to 30 residues with molecular weights below 5 kilodaltons released by fermenting food sources with fungal strains.

The secondary metabolite question is deliberately framed as a hypothesis rather than a conclusion. Against the criteria set out by Keller and colleagues, BPF members satisfy several: they are bioactive, show a restricted taxonomic distribution, and, for some verified members, are not primarily required for survival. They are also ribosomally synthesised and encoded by single genes, as shown for the verified members. However, other criteria remain inconclusive. No study has yet shown that BPF member genes are organised in gene clusters, which by definition require at least two genes, and there is no direct evidence of targeted enzymatic processing cascades or dedicated processing enzymes, so the proteins cannot yet be unambiguously categorised as RiPPs. Expression patterns are also mixed: many members, including AnAFP, PAF and NFAP2, show late expression under nutrient limitation or in older cultures, but counterexamples exist, such as PAFB from Penicillium chrysogenum, which is expressed under nutrient excess. The authors therefore suggest that BPF members could be considered secondary metabolite-like proteins rather than definitive secondary metabolites.

The practical significance lies in biotechnology. Fungi are leveraged to produce organic acids, platform chemicals, enzymes and therapeutics, and several BPF-encoding genera, including Aspergillus, Penicillium, Trichoderma and Schizophyllum, are economically relevant. Because some family members have documented roles in carbon metabolism, autophagy, cell wall remodelling and asexual sporulation, the authors argue that their presence in production strains could interfere with fermentation performance and should be considered during genetic and metabolic optimisation. They also caution that the dataset includes pathogenic fungi such as Fusarium graminearum, Beauveria bassiana and Claviceps purpurea, so BPF expression during virulence warrants future study. With only a handful of the 486 predicted members investigated in the laboratory so far, and the expected number of members set to grow as more fungal genomes are sequenced, the superfamily’s true functional breadth, and its relevance for antifungal drug discovery and industrial fermentation, remains an open question that the authors’ hypotheses now set out to test.

Subject of Research: Fungal antifungal protein families and their proposed classification as a superfamily of secondary metabolite-like proteins

Article Title: A superfamily of bioactive proteins from fungi – are these secondary metabolites?

Article References: Feurstein, C., Bachmann, L., Heber, L., Dobbert, B., & Jung, S. (2026). A superfamily of bioactive proteins from fungi – are these secondary metabolites?. Fungal Biology and Biotechnology, 13(1), Article 15. https://doi.org/10.1186/s40694-026-00219-x

Image Credits: AI Generated

DOI: 10.1186/s40694-026-00219-x

Keywords: antifungal proteins, fungi, secondary metabolites, BPF superfamily, Aspergillus giganteus, horizontal gene transfer, phylogenetics, protein structure, biotechnology, Ascomycetes, Basidiomycetes, RiPPs