At the cellular level, fungal macroautophagy uses a conserved network of more than 40 ATG-family proteins to sequester cytoplasmic material in a double-membrane phagophore and deliver it to the fungal vacuole for degradation. In yeast and most filamentous fungi, autophagosome assembly is concentrated at a single phagophore assembly site, unlike the multiple ER-associated omegasomes described in mammalian cells. BAR-domain proteins help generate the sharply curved isolation membrane, while branched actin forms a cage around the assembly site that restricts uncontrolled membrane diffusion. After autophagosomes close, molecular motors transport them along microtubules to the vacuole, a particularly demanding task in polarized hyphae where autophagosomes may travel dozens of micrometers.

Membrane lipid signaling and vesicle-trafficking systems add further control over maturation and vacuolar fusion. In M. oryzae, the oxysterol-binding protein-related proteins MoOrp1 and MoOrp2 interact with MoAtg8, localize to the phagophore assembly site and enrich phosphatidylinositol-4-phosphate on autophagosomal membranes. A double gene knockout blocked autophagic flux and was associated with loss of appressorial turgor and host penetration. In Fusarium graminearum, TRAPPIII interacts with the autophagy transmembrane protein FgAtg9 through FgTrs85 and TRAPPC13 while also activating FgRab1 as a guanine nucleotide exchange factor, linking autophagosome biogenesis with ER-to-Golgi and endosome-to-Golgi transport during reproduction and infection.

Trafficking also helps determine which proteases reach the vacuole. The retromer protein MoVps35 in M. oryzae binds the serine protease MoPrb1 on endosomes and directs it into the vacuole, where MoPrb1 assists in localizing MoPep4. Genetic analyses assigned distinct functions to these proteases: MoPrb1 supports degradation of macroautophagy cargo associated with infection structures, and its mutation was linked to failed appressorium development, whereas MoPep4 regulates microautophagy, including peroxisomal degradation, without being required for virulence under the tested conditions. These results indicate that vesicle sorting can differentiate among autophagic pathways.

Selective autophagy remodels organelles and storage reserves in ways that support fungal growth and virulence. Quantitative proteome-turnover analysis in Komagataella phaffii found that changing carbon and nitrogen sources remodeled up to 90% of the pre-existing proteome, with Atg9-dependent bulk autophagy operating alongside cargo-specific pathways. Peroxisomes were cleared through Atg30- and Atg11-dependent pexophagy, while mitochondria underwent slower Atg32-dependent mitophagy. In M. oryzae, the MoDnm1–MoMdv1–MoFis1 complex regulates mitochondrial and peroxisomal fission; disruption impaired mitophagic clearance, conidiation, appressorium-mediated penetration and full virulence.

For lipid droplets in Metarhizium robertsii and Beauveria bassiana, morphological observations indicate partial engulfment by isolation membranes or sequestration through microlipophagy. Deletion of ATG8 homologs was associated with impaired droplet use, conidial germination and host infectivity. Autophagy is also regulated through RNA modifications, transcription and protein modification. In M. oryzae, the m5C methyltransferase Ncl1 modifies specific sites in atg5 and the 3′ untranslated region of atg16, increasing transcript stability and supporting Atg5–Atg16 translation; Pmk1-MAPK phosphorylation and Smt3-dependent SUMOylation also stabilize Ncl1.

The clearest direct link between external stress sensing and autophagy described by the review occurs during appressorium development. The cell-wall-integrity MAP kinase MoMps1 directly phosphorylates the autophagy protease MoAtg4, which processes Atg8 before its lipid modification, and the reported interaction can adjust autophagic flux independently of TOR signaling. The Wee1-family kinase MoSwe1 also coordinates cell-cycle arrest, autophagy and localized cell-wall remodeling during penetration-peg formation. Deletions of ATG genes in Cryptococcus neoformans and Aspergillus fumigatus have also been associated with hypersensitivity to Congo red and echinocandins, supporting a broader association between autophagy and cell-wall stress, although the direct biochemical evidence for MoMps1–MoAtg4 signaling is concentrated in M. oryzae.

At the developmental and host-interaction scale, the review links autophagy to programmed conidial death in M. oryzae. As an appressorium develops, maternal spore cells undergo extensive autophagy, transfer degraded material toward the appressorium and eventually collapse; core ATG mutations can prevent this death and associated transfer. The review also describes lipid peroxidation and ferroportin-1-related iron release as coupled processes, but much of this evidence comes from one fungal model. In C. neoformans, loss of Sgl1 was associated with blocked GFP-Atg8 trafficking, accumulation and extracellular release of ergosteryl-glucoside derivatives, and Mincle-dependent inflammatory responses that cleared the pathogen from lungs in the reported mouse infection model. These experimental results do not establish a clinical intervention.

The review presents Cordyceps militaris as a model for linking host manipulation to multicellular development. During infection, the horizontally transferred trehalase CmTreH1 depletes host trehalose, while the review reports induction of the orexigenic neuropeptide HemaP, hyperphagia and formation of a lipid- and glycogen-rich “fatty pupa.” After host takeover, bulk autophagy, lipophagy and glycophagy are proposed to mobilize nutrients for primordium and fruiting-body growth. Direct functional characterization of core autophagy genes in C. militaris has only begun, so the analogy with the better-characterized FOX-factor program in Sclerotinia sclerotiorum remains a hypothesis. Separately, the review describes Ebselen derivatives reversibly inhibiting the Atg4 catalytic cysteine and reports activity in plant and human fungal models; these findings are preclinical and do not establish a clinical antifungal treatment.

Subject of Research: Review of autophagy in fungal development, pathogenicity and host interactions

Article Title: Autophagy as a multi-scale architect of fungal development and pathogenicity: membrane dynamics, multilayer regulation, and cell wall integrity crosstalk

Article References: Wang, F., Li, J., Chen, H., Zheng, Q., Wei, T., Feng, K., & Chen, B.-X. (2026). Autophagy as a multi-scale architect of fungal development and pathogenicity: membrane dynamics, multilayer regulation, and cell wall integrity crosstalk. IMA Fungus, 17, Article e196033. https://doi.org/10.3897/imafungus.17.196033

Image Credits: AI Generated

DOI: 10.3897/imafungus.17.196033

Keywords: fungal autophagy, membrane remodeling, cell wall integrity, fungal pathogenesis, Magnaporthe oryzae, Cryptococcus neoformans, organelle turnover, antifungal targets, Cordyceps militaris, vesicle trafficking, Autophagy, multi-scale

The finding matters because A. fumigatus is not an obscure laboratory curiosity. It is the dominant cause of invasive aspergillosis, a life-threatening infection of the lungs and other organs that primarily strikes people with weakened immune systems, including transplant recipients, chemotherapy patients, and those with severe influenza or COVID-19. Current antifungal drugs target a narrow set of conserved cellular components, and resistance is an escalating clinical problem. Regulators that exist only in pathogenic species, and not in humans, represent some of the most attractive drug targets imaginable, because inhibiting them could cripple the fungus while sparing the patient. The problem has always been finding them, since pathogen-specific elements by definition lack recognizable counterparts in well-studied model organisms, and standard annotation pipelines tend to dismiss anything without a known homolog.

The team’s strategy began with a large-scale re-annotation of the A. fumigatus genome, focusing on a set of 423 loci that had previously been labeled as protein-coding genes restricted to the Fumigati section, the fungal group to which this pathogen belongs. These so-called section-specific genes had long been suspected of contributing to pathogenicity precisely because they appear nowhere outside this lineage. But when the researchers interrogated them with transcriptome-guided evidence, examining the actual RNA molecules produced from each locus and the molecular signatures that distinguish coding from non-coding transcripts, a different picture emerged. Of the 423 loci, 176 turned out to be misannotated: they do not encode proteins at all but instead produce long non-coding RNAs, a class of regulatory molecules that function as RNA rather than being translated into amino acid chains.

That level of error is remarkable. Nearly half of the genes in this pathogen-specific category had been assigned the wrong molecular identity, meaning that years of functional studies, comparative genomics analyses, and drug-discovery efforts may have been looking at these loci through the wrong lens. Long non-coding RNAs, or lncRNAs, are well known in mammalian biology as orchestrators of gene expression, chromosome structure, and development, but in fungi they have been studied far less intensively, and their role in microbial pathogenicity has remained almost entirely unexplored. The re-annotation demonstrated that in A. fumigatus, this overlooked regulatory layer is not a minor footnote but a substantial component of the species-specific genome.

Identifying the misannotated loci was only the first step. The researchers then carried out a functional screen, systematically deleting or disrupting individual lncRNA genes and testing the resulting mutant strains for defects in growth and, critically, for the ability to cause disease. From this screen, two lncRNAs emerged with particularly dramatic phenotypes. The team named them SSP7 and SSP8, and in both cases, loss of the RNA molecule severely impaired hyphal growth, the filamentous expansion that A. fumigatus uses to colonize tissue. More importantly, mutants lacking either lncRNA showed markedly reduced virulence in a murine infection model, the standard preclinical test for fungal pathogenicity. In other words, these are not dispensable transcripts with subtle effects; they are required for the fungus to behave as a pathogen.

The mechanistic follow-up work on SSP7 revealed something genuinely novel about how fungi can regulate their biology. The researchers found that SSP7 binds directly to the coding region of the messenger RNA produced by a gene called pprA, which encodes a mitochondrial protein involved in energy production. This binding is not decorative: by associating with the pprA transcript, SSP7 promotes the recruitment of that mRNA to polysomes, the clusters of ribosomes that carry out protein synthesis. With SSP7 present, pprA mRNA is efficiently translated into protein; without it, translation falters. The consequence is a decline in mitochondrial function, which in a fast-growing filamentous fungus is catastrophic, because hyphal extension demands enormous amounts of energy. SSP7 thus acts as a translational enhancer, an RNA molecule that fine-tunes how much protein is made from a target message rather than how much of the message exists.

This mode of regulation is conceptually important. In eukaryotic cells, RNA-binding lncRNAs that stimulate the translation of specific mRNAs by engaging their coding sequences are rare, and their involvement in microbial virulence had not been demonstrated before. The authors emphasize that this is the first time pathogen-specific lncRNAs have been shown to modulate conserved cellular processes, in this case mitochondrial respiration and protein synthesis, to generate traits that are distinctive to a pathogenic fungus. The core machinery, ribosomes, mitochondria, and the genes that build them, is ancient and shared across nearly all eukaryotes. What SSP7 and SSP8 appear to do is layer pathogen-specific control on top of that conserved machinery, adjusting its output in ways that suit the lifestyle of an invasive mold.

SSP8 tells a complementary story. Rather than acting on mitochondrial translation, this lncRNA contributes to amino acid homeostasis, the maintenance of a balanced intracellular pool of the building blocks that fungi need to construct proteins and sustain metabolism. Disrupting SSP8 throws that balance off, and the resulting metabolic stress compounds the growth defects observed in the mutant strains. Together, the two lncRNAs illustrate that a single class of non-coding molecules can touch multiple arms of cellular physiology, from energy metabolism to nutrient management, and that these touchpoints converge on the two phenotypes that matter most clinically: how fast the fungus grows and how sick it makes its host.

The broader implications extend well beyond aspergillosis. Genome annotation is the foundation on which nearly all modern molecular biology rests, and this study is a pointed reminder that automated pipelines systematically under-detect non-coding genes, especially in lineages where they lack evolutionary conservation. If nearly half of the section-specific protein-coding annotations in one fungal pathogen were wrong, comparable errors almost certainly lurk in the genomes of other medically important fungi, including Candida, Cryptococcus, and the emerging multidrug-resistant Candida auris. Re-annotation efforts guided by transcriptomic evidence could therefore unlock a hidden reservoir of species-specific regulators across the fungal tree, many of which may prove to be virulence factors or drug targets.

For the immediate future, SSP7 and SSP8 themselves offer a proof of concept. Both molecules are required for full virulence, both operate through mechanisms that appear absent from human cells, and both belong to genomic neighborhoods that exist only in a narrow fungal lineage. A therapeutic that interferes with an lncRNA-mRNA interaction, or that destabilizes a pathogen-specific transcript, would represent an entirely new class of antifungal strategy at a moment when the clinical need for such strategies has never been greater. The study also reframes how scientists think about pathogenicity itself: rather than requiring wholly new genes, a deadly fungus may simply deploy species-specific RNA regulators to rewire ancient machinery in its own favor. In uncovering that layer, this work transforms a genome-annotation cleanup exercise into a discovery with direct relevance to one of medicine’s most stubborn infectious disease problems.

Subject of Research: Pathogen-specific long non-coding RNAs regulating growth and virulence in the human fungal pathogen Aspergillus fumigatus

Article Title: Large-scale genome re-annotation uncovers pathogen-specific lncRNA regulators of fungal growth and virulence

Article References: Gao, R., Liu, Z., Tang, H., Wang, N., Xue, X., Lu, L., & Zhang, Y. (2026). Large-scale genome re-annotation uncovers pathogen-specific lncRNA regulators of fungal growth and virulence. PLOS Pathogens, 22(10), e1014681. https://doi.org/10.1371/journal.ppat.1014681

Image Credits: AI Generated

DOI: 10.1371/journal.ppat.1014681

Keywords: Aspergillus fumigatus, long non-coding RNA, genome re-annotation, fungal pathogenicity, virulence, translational regulation, mitochondrial function, hyphal growth, invasive aspergillosis, PLOS Pathogens, functional genomics, antifungal targets

Lsm8 is the defining subunit of the nuclear Lsm2-8 complex, a heptameric ring of Like-Sm proteins that binds the 3-prime end of U6 small nuclear RNA, the catalytic core of the spliceosome. The spliceosome is the macromolecular machine that removes introns from pre-messenger RNA, and without properly assembled Lsm2-8, U6 is left unprotected and splicing cycles falter. Sequence analysis showed that the Fusarium Lsm8 protein shares roughly half its identity with its yeast counterpart and about 65 percent with the human version, underscoring how ancient this module is. Although Lsm8 lacks a predicted nuclear localization signal, a fluorescent Lsm8-GFP fusion co-localized precisely with a histone marker in fungal nuclei, confirming where the protein operates.

To map the complex’s architecture, the team deployed a battery of interaction assays. Yeast two-hybrid tests revealed direct contacts between Lsm8 and Lsm2 and Lsm3, bimolecular fluorescence complementation captured a nuclear interaction between Lsm8 and Lsm4, and co-immunoprecipitation validated all of these associations in living fungal cells. Most decisively, affinity purification followed by mass spectrometry, using Lsm8-GFP as bait, co-purified all six remaining Lsm subunits. The picture that emerged is of Lsm8 as a central architectural keystone. When the researchers deleted the LSM8 gene, GFP-tagged Lsm2, Lsm3 and Lsm4 abandoned the nucleus entirely and diffused throughout the cytoplasm, demonstrating that Lsm8 is indispensable for both assembling the ring and escorting it to its workplace.

The consequences of losing this keystone were dramatic. Deletion mutants grew slowly on every medium tested, from rich potato dextrose agar to minimal medium. Transcriptome sequencing revealed massive reprogramming: 1,710 genes were significantly upregulated and 1,796 downregulated. The upregulated set was enriched in ribosome biogenesis, spliceosome function and RNA degradation, which the authors interpret as a compensatory feedback response to the missing splicing machinery. The downregulated genes clustered in carbon and nitrogen metabolism, including glycolysis and the pentose phosphate pathway, providing a molecular explanation for the growth collapse. Quantitative PCR confirmed the sequencing results for splicing factors such as PRP4 and PRP5, degradation factors such as RRP6 and XRN2, and glycolytic genes such as FBA1 and ENO1.

The most striking defect, however, lay in splicing itself. Across 13,982 detected introns, the median intron retention rate jumped from 0.011 in the wild type to 0.089 in the mutant, a highly significant eight-fold increase. In total, 2,344 introns from 1,821 genes showed significantly elevated retention, while only 45 introns became better spliced. Crucially, most of the affected genes, 1,270 of them, showed no change in overall transcript abundance, meaning the damage was purely post-transcriptional: transcripts were made normally but never spliced correctly. Retained introns disrupt reading frames, producing truncated or non-functional proteins, or triggering nonsense-mediated decay. The affected genes were enriched in chromatin remodeling, RNA polymerase activity, nucleocytoplasmic transport and DNA repair, processes essential to cellular integrity.

The splicing failures mapped directly onto the fungus’s environmental vulnerabilities. When challenged with a panel of eight stress agents, the mutant proved hypersensitive to all of them, with the strongest effects under osmotic stress from salt and the fungicide fludioxonil, which works by hyperactivating the osmotic stress pathway. The team traced this to the high osmolarity glycerol, or HOG, MAPK cascade. HOG1 carries eight introns, and nearly all of them were inefficiently spliced in the mutant, with defects worsening under stress. Total HOG1 mRNA levels were unchanged, but Hog1 protein levels dropped significantly, showing that the defect lies in protein production rather than transcription. Intriguingly, the residual Hog1 that remained was hyperphosphorylated, and intracellular glycerol accumulated to roughly three times wild-type levels, suggesting distorted pathway feedback. The downstream transcription factor ATF1 also showed splicing abnormalities.

Reproduction fared no better. The mutant produced fewer conidia, the asexual spores that spread infection, and those spores were shorter and abnormally septated. Sexual development was abolished outright: on carrot agar under black light, the mutant formed no perithecia at all, the fruiting bodies that generate overwintering ascospores. The molecular culprits were again splicing defects, this time in two developmental regulators. GPA1, encoding a heterotrimeric G protein alpha subunit, retained all three of its introns at elevated rates, and STE12, a transcription factor essential for sexual reproduction, retained both of its introns. As with HOG1, total mRNA levels were largely unaffected, reinforcing the theme that Lsm8’s influence operates after transcription.

Virulence and mycotoxin synthesis collapsed in parallel. Infection assays on wheat heads, wheat coleoptiles and corn silks showed the mutant was nearly nonpathogenic, and deoxynivalenol production fell sharply. The TRI gene cluster that manufactures the toxin was hit twice: key genes including TRI1, TRI4, TRI5, TRI6, TRI10 and TRI101 were transcriptionally downregulated, and the splicing efficiency of TRI1, TRI4 and TRI5 transcripts was significantly reduced. Microscopy of a Tri1-GFP reporter delivered perhaps the most visually compelling result: in the wild type, Tri1 aggregates into the spherical toxisomes, specialized compartments where deoxynivalenol is synthesized, but in the mutant almost no toxisomes formed and Tri1-GFP protein levels plummeted. Lsm8 thus guards the fungus’s chemical arsenal at every regulatory level simultaneously.

Finally, the study answered a long-standing question about what happens to all those defective transcripts. In eukaryotes, aberrant RNAs are normally cleared by nonsense-mediated decay, the RNA exosome, which degrades transcripts in the 3-prime to 5-prime direction, or by Xrn enzymes working in the opposite direction. Mass spectrometry showed that Lsm8 physically interacts with core exosome subunits including Rrp45, Rrp4 and Dis3, and the mutant compensated by upregulating exosome components. Genetic dissection proved decisive: intron-retained HOG1 transcripts accumulated massively in an exosome-subunit deletion mutant and even more in an Lsm8-exosome double mutant, while accumulation in an Xrn1 mutant was comparatively modest. The exosome, the authors conclude, is the predominant degrader of intron-retained transcripts arising from Lsm2-8 dysfunction, with Xrn1 playing a supporting role. Because both the Lsm2-8 complex and the exosome are conserved across eukaryotes, from yeast and plants to worms and humans, this coupled splicing-and-surveillance module appears to be a fundamental feature of eukaryotic biology. In Candida albicans, a human fungal pathogen, Lsm loss already impairs filamentation and pathogenicity, suggesting the Fusarium findings generalize. Targeting this conserved module, the researchers argue, could yield antifungal strategies that strike at core RNA-processing machinery shared by diverse pathogens, simultaneously curbing crop disease and mycotoxin contamination while sparing the host, whose own cells process RNA through homologous but distinct regulatory contexts.

Subject of Research: RNA splicing fidelity and the Lsm8-exosome surveillance module in fungal stress adaptation and virulence

Article Title: A conserved Lsm8–exosome module maintains RNA splicing fidelity to control fungal stress adaptation and virulence

Article References: Ren, Y., Cheng, H., Han, X., Guo, M., Xu, C., Yan, J., Ge, Z., Ma, Z., & Chen, Y. (2026). A conserved Lsm8–exosome module maintains RNA splicing fidelity to control fungal stress adaptation and virulence. Stress Biology, 6(1), Article 14. https://doi.org/10.1007/s44154-026-00285-6

Image Credits: AI Generated

DOI: 10.1007/s44154-026-00285-6

Keywords: Fusarium graminearum, Lsm8, Lsm2-8 complex, RNA splicing, intron retention, RNA exosome, deoxynivalenol, mycotoxin biosynthesis, HOG-MAPK pathway, fungal virulence, post-transcriptional regulation, antifungal targets