A team led by Jian-Feng Li at Sun Yat-sen University in Guangzhou, working with collaborators at South China Agricultural University, the Chinese University of Hong Kong and the University of Macau, identified two paralogous chitinases in the model plant Arabidopsis thaliana that they named MIC1 and MIC2. Both genes are induced by chitin and by other microbe-associated molecular patterns, including the bacterial flagellin-derived peptide flg22 and the elongation factor Tu-derived peptide elf18. Both proteins are secreted into the apoplast, the fluid-filled space outside the plant cell membrane where fungal contact first occurs, and both possess direct antifungal activity in vitro, breaking down chitin in the fungal cell wall itself. On the surface, the two enzymes look like redundant copies of the same tool.

The resemblance is deceptive. When the researchers generated CRISPR/Cas9 knockout mutants, the two paralogs pulled in opposite directions. Plants lacking MIC1 became markedly more resistant to the grey mould pathogen Botrytis cinerea and to the vascular wilt fungus Verticillium dahliae, while plants lacking MIC2 became more susceptible. The team also showed that the single mutants and the mic1 mic2 double mutant grow normally under standard conditions, which means the antagonistic effects of these enzymes are specific to immune outcomes rather than general developmental defects. The picture that emerged is of two enzymes with the same substrate but opposing roles in tuning the immune response.

The key to this antagonism lies in the products of chitin digestion. Chitin is a polymer of N-acetylglucosamine units, conventionally abbreviated NAG, and the length of the oligomers released by hydrolysis determines their biological meaning. Using high-performance liquid chromatography to track the fate of defined chitin oligomers, the researchers found that MIC2 cleaves long chitin chains into intermediate fragments, principally NAG4 to NAG6, which are precisely the sizes that plant chitin receptors recognize as immunogenic elicitors. These fragments activate mitogen-activated protein kinase signalling, a canonical early readout of chitin-triggered immunity, and pretreatment with NAG4 primed Arabidopsis leaves for enhanced resistance against B. cinerea. MIC2, in other words, is an alarm bell: it chews the fungal wall in a way that generates the molecular signals that rally the plant’s defenses.

MIC1 does the opposite. Rather than stopping at immunogenic intermediates, MIC1 processes chitin all the way down to NAG2 dimers and NAG1 monomers, fragments that are too short to activate receptor signalling. In seedling assays, exogenous chitin normally inhibits root growth, a visible manifestation of constitutive chitin signalling, and MIC1 is the enzyme responsible for preventing this self-inflicted damage. When the researchers tested additional Arabidopsis chitinases from other clades, they found that the MIC1-like degradation pattern, exemplified by the clade IV enzyme chitIV1, is functionally conserved, and that deleting whole clusters of MIC1-like genes reproduced the chitin-hypersensitive growth phenotype. MIC1 and its relatives therefore act as a molecular off-switch, extinguishing the growth-inhibitory signal once the immune alarm has sounded.

What makes one enzyme stop at immunogenic fragments while its near-identical paralog grinds chitin to dust? The answer, uncovered through a combination of mass spectrometry, PNGase digestion assays and site-directed mutagenesis, is N-glycosylation, the attachment of sugar chains to asparagine residues of secreted proteins. Both MIC1 and MIC2 carry N-glycans, and these modifications are not decorative: glycosylation-deficient variants of either protein failed to be secreted efficiently into the apoplast, and a glycosylation-deficient MIC2 could no longer complement the enhanced susceptibility of mic1 mic2 double mutants to B. cinerea. Glycosylation is thus required for these enzymes to reach their site of action at all.

But one glycan does far more than ensure secretion. Mass spectrometry mapped three N-glycosylation sites on MIC1 and two on MIC2, and the decisive difference came down to a single residue: asparagine 137 in MIC1. When the researchers swapped variable segments between the two proteins to create chimeric enzymes, the exchange that transferred the MIC1 region containing Asn137 into MIC2 converted MIC2 into a MIC1-like enzyme, digesting NAG7 into more dimers and monomers and fewer immunogenic pentamers and tetramers. Conversely, the corresponding chimeric version of MIC1 lost its characteristic band shift and behaved more like MIC2. In complementation tests, the chimeric MIC2 carrying the MIC1 glycosylation region partially suppressed chitin-induced growth inhibition, whereas wild-type MIC2 did not, and it slightly increased susceptibility to B. cinerea instead of restoring resistance. Removing the Asn137 glycan from MIC1 converted it into a MIC2-like enzyme, demonstrating that this one sugar attachment dictates the chitinolytic specificity of the entire protein.

Structural modelling with AlphaFold showed that MIC1 and MIC2 are highly similar in sequence and fold, with their catalytic glutamates aligned and their glycosylation sites positioned on surface loops, including the Loop III region implicated in substrate handling in GH19 family chitinases. The Asn137 glycan apparently reshapes the local architecture of the catalytic cleft in a way that changes how many NAG units the enzyme removes per cleavage event. This is a remarkable example of post-translational modification acting as a functional diversification code: two genes born of duplication, nearly interchangeable in sequence, are sorted into opposing immune roles by a single covalently attached oligosaccharide. The authors describe this as an N-glycosylation code governing the functional specialization of paralogous chitinases.

The implications extend well beyond Arabidopsis. A bioinformatic survey revealed that MIC1-like chitinases are widespread across the plant kingdom, and the Asn137-equivalent residue is highly conserved among MIC1 orthologs from different Arabidopsis ecotypes. The team turned to rice, the world’s most important staple crop, and generated CRISPR knockout mutants of the rice ortholog OsMIC1. The osmic1 mutants showed enhanced resistance to rice blast and to bacterial blight, the latter an intriguing bonus suggesting the mechanism is not strictly antifungal, while their grain length, grain width and thousand-grain weight remained indistinguishable from wild-type plants. Like their Arabidopsis counterparts, however, osmic1 seedlings did exhibit chitin-induced root growth inhibition, confirming that the dampening function of OsMIC1 is conserved.

For breeders, this is the rare kind of result that points directly at a molecular target. Fungal pathogens destroy a substantial fraction of global crop yields each year, and conventional resistance breeding often trades yield for defense. The MIC1/MIC2 system suggests a third path: tune the enzymatic processing of a universal pathogen-derived signal so that immunity is amplified without the growth penalty that constitutive immune signalling imposes. Editing or selecting variants of MIC1-like chitinases, or manipulating the glycosylation machinery that controls them, could in principle produce crops that detect and respond to fungal attack more vigorously while growing as if no pathogen were present. A single sugar on a single asparagine, it turns out, may hold the balance between feast and famine in the plant kingdom.

Subject of Research: N-glycosylation-dependent functional divergence of paralogous plant chitinases balancing immunity and growth

Article Title: N-glycosylation code diversifies paralogous chitinases to balance plant immunity and growth

Article References: Li, C., Lin, X.-X., Tian, L., Sun, W., Liu, J., Liu, Q., Du, Y., Jiang, X., Liang, Z., Deng, Y. Z., Au, S. W. N., Wan, J.-B., & Li, J.-F. (2026). N-glycosylation code diversifies paralogous chitinases to balance plant immunity and growth. Nature Plants. https://doi.org/10.1038/s41477-026-02420-7

Image Credits: AI Generated

DOI: 10.1038/s41477-026-02420-7

Keywords: chitinases, N-glycosylation, plant immunity, chitin signalling, Arabidopsis thaliana, MIC1, MIC2, growth-defense trade-off, fungal resistance, rice breeding, apoplast, CRISPR

Fusarium wilt of watermelon is a devastating vascular disease. The fungus enters through the roots, threads its way into the water-conducting xylem vessels, and effectively strangles the plant from within, disrupting water transport until leaves wilt and the plant dies. What makes the disease especially difficult to control is the pathogen’s biology: the fungus can survive in soil for many years, new races capable of overcoming resistant cultivars continue to emerge, and effective field treatments remain scarce. Because Fusarium oxysporum as a species can infect more than 150 plants, including cotton, tomato, melon, and banana, understanding how its virulence works has implications far beyond the watermelon patch.

The Zhejiang team, led by Fengming Song, focused on a class of molecules known as effector proteins. Fungal pathogens secrete these proteins in large numbers during infection to manipulate or suppress plant immunity. Effectors are broadly divided into two categories based on where they operate: apoplastic effectors, which act in the space outside plant cell membranes, and cytoplasmic effectors, which function inside host cells. In the tomato-infecting form of the fungus, researchers previously identified a family of Secreted In Xylem, or SIX, proteins, many of which proved pivotal to disease. But for the watermelon-infecting strain, the functional repertoire of effectors remained largely unexplored.

The protein at the center of the new study, FonCHRD, caught the researchers’ attention because it carries a chordin, or CHRD, domain whose function in fungi was unknown. Chordin is famous in developmental biology: it was first isolated from the Spemann organizer of the frog embryo Xenopus laevis, where it antagonizes bone morphogenetic proteins and helps pattern the body axis. CHRD domain proteins also appear in green algae, where they are linked to stress responses, and have been implicated as potential biomarkers in human cancers. In the tomato wilt fungus, a CHRD-containing protein had been flagged as a likely pathogenicity effector, but nobody had experimentally pinned down what such a protein actually does in a plant-pathogenic fungus.

FonCHRD turned out to be a compact protein of 193 amino acids with a 16-amino-acid signal peptide at its N-terminus, the molecular postal address that directs proteins into the conventional secretion pathway. The team confirmed this experimentally using a yeast secretion trap assay: yeast cells engineered to carry the FonCHRD signal peptide could grow on selective medium and reduce a colorless compound to a red precipitate, proving the peptide genuinely drives secretion. Bioinformatic tools classified the protein as an apoplastic or cytoplasmic effector, and structural modeling showed its three-dimensional fold closely resembles known CHRD domain proteins. Notably, CHRD-containing proteins also occur in other major plant pathogens, including the rice blast fungus Magnaporthe oryzae, Botrytis cinerea, and Verticillium dahliae, suggesting an evolutionarily conserved role.

Expression profiling added a telling clue. FonCHRD was most active in macroconidia, the banana-shaped spores that spread the fungus, and its expression surged roughly thirtyfold when spores were exposed to watermelon root tissue, peaking at 36 hours. That pattern, a gene quiet in resting mycelium but switched on by host contact, is a hallmark of effectors deployed at the earliest stages of infection. To test function, the researchers deleted the gene using targeted gene replacement and then restored it in a complemented strain. The deletion mutant, designated ΔFonCHRD, grew more slowly on culture medium, with colony diameters reduced by 44 percent on potato dextrose agar and 26 percent on minimal medium. It also produced about a third fewer macroconidia, and those spores were visibly abnormal: curled, shorter, and carrying fewer cross-walls, or septa, than wild-type spores. Intriguingly, spore germination was unaffected, and the mutant showed no heightened sensitivity to cell wall, oxidative, or salt stress, indicating the protein’s role is selective rather than a general stress function.

The pathogenicity results were dramatic. Twenty-one days after inoculation, 71 percent of watermelon plants infected with the wild-type fungus had severely wilted or died, while only 7 percent of plants exposed to the deletion mutant had died, translating to a 51 percent drop in disease severity. Plants inoculated with the mutant also began dying three to six days later than those facing the wild type, and at 27 days, all wild-type-infected plants were dead while only 43 percent of mutant-infected plants had succumbed. Reintroducing the gene restored full virulence, confirming the effect was due to FonCHRD itself. Quantitative PCR revealed that fungal biomass inside infected roots was reduced by 56 percent and inside stems by a striking 85 percent, yet the mutant penetrated cellophane membranes as efficiently as the wild type. The picture that emerges is that FonCHRD does not help the fungus break into the plant; it helps the fungus spread once inside, particularly as it moves from root to stem through the vascular system.

To see where the protein acts in plant cells, the team tagged FonCHRD with green fluorescent protein and expressed it transiently in leaves of Nicotiana benthamiana, a widely used experimental plant. The protein appeared in the cytoplasm, nucleus, and plasma membrane, and formed punctate spots whose significance remains unclear. When the researchers induced plasmolysis with mannitol to shrink the cell membrane away from the wall, fluorescent signal accumulated in the apoplastic space, confirming that FonCHRD is secreted outside plant cells and that its signal peptide directs this targeting. A variant lacking the signal peptide failed to reach the apoplast, while a hybrid variant in which the native peptide was replaced with a plant apoplastic targeting sequence restored apoplastic localization.

The most striking results came from immune suppression assays. INF1, an elicitor from the potato late blight pathogen, and BAX, a pro-apoptotic mouse protein, both reliably trigger programmed cell death in Nicotiana leaves, a defense reaction analogous to the hypersensitive response plants use to wall off invaders. When FonCHRD was co-expressed with either inducer, the cell death response disappeared or was markedly alleviated. The protein also dampened the expression of six defense-related genes, including pathogenesis-related genes NbPR1 and NbPR2, that INF1 and BAX normally switch on. Remarkably, the signal peptide was dispensable for this suppression: even the variant lacking it blocked cell death, implying that FonCHRD acts inside plant cells to interfere with immunity, while the peptide serves mainly to route the protein to the apoplast. This dual behavior mirrors other Fusarium effectors such as FoSSP17 and FoCupin1, which also suppress cell death independently of their signal peptides.

What makes FonCHRD unusual among Fusarium effectors is that it is not merely a virulence accessory. Most known effectors, such as FolAsp, FoRnt2, and FoSSP17, are dispensable for the fungus’s basic biology and live on mobile pathogenicity chromosomes. FonCHRD, by contrast, is woven into vegetative growth, asexual reproduction, and spore morphology, making it a pleiotropic factor whose loss cripples the organism in multiple ways at once. That combination of housekeeping and warfare roles suggests the sole CHRD domain protein in the fungal genome may have been recruited from an ancestral developmental function into pathogenesis, echoing chordin’s deep role in animal development. The authors caution that the reduced virulence of the mutant could partly reflect its impaired growth, and that the protein’s precise biochemical activity and its interaction partners in both fungus and host remain to be identified. Still, the study delivers a clear strategic insight: a pathogen that must both build its own body and dismantle the plant’s defenses depends on a single secreted protein to do both. Targeting such dual-purpose molecules, whether through resistance breeding, engineered decoys, or small molecules that block their secretion or folding, could offer a way to blunt Fusarium wilt that the fungus would find hard to evade, since losing the protein costs it far more than a single weapon.

Subject of Research: Functional characterization of the secreted effector protein FonCHRD in the watermelon Fusarium wilt fungus Fusarium oxysporum f. sp. niveum

Article Title: The secreted protein FonCHRD is essential for vegetative growth, asexual reproduction, and pathogenicity in watermelon Fusarium wilt fungus

Article References: Lou, J., Wang, J., Zeng, S., Xiong, X., Guo, M., Li, D., & Song, F. (2024). The secreted protein FonCHRD is essential for vegetative growth, asexual reproduction, and pathogenicity in watermelon Fusarium wilt fungus. Crop Health, 2(1), Article 16. https://doi.org/10.1007/s44297-024-00036-x

Image Credits: AI Generated

DOI: 10.1007/s44297-024-00036-x

Keywords: FonCHRD, Fusarium oxysporum f. sp. niveum, watermelon Fusarium wilt, effector protein, plant immunity, chordin domain, pathogenicity, programmed cell death, apoplast, macroconidia, vascular wilt, secreted protein

Phytocytokines, the plant counterparts of animal cytokines, are small secreted peptides that coordinate defense responses when pathogens or herbivores wound plant tissue. They act like molecular alarms: once released, they bind to cell-surface receptor kinases and trigger a cascade of intracellular events, including calcium influx, the production of reactive oxygen species, and the activation of defense genes. CTNIP peptides belong to a family of these signals that has intrigued structural biologists because their sequences do not resemble the classical cysteine-rich peptide motifs that dominate the phytocytokine world. Without an obvious templating pattern, the question of how CTNIP4 achieves a stable, recognizable structure has lingered since the family was first connected to immune signaling.

The new work answers that question with a combination of structural biology, biochemistry, and genetics. Using nuclear magnetic resonance spectroscopy, the team determined the solution structure of mature CTNIP4 and found that it adopts a compact, well-defined fold anchored by an intramolecular disulfide bond, a covalent bridge formed between two cysteine residues within the same peptide chain. Rather than serving as an interchangeable decoration, this bond acts as the load-bearing element of the structure, pinning the peptide into a looped architecture that presents its receptor-binding face in an exact orientation. When either of the two participating cysteines was mutated to a non-reactive residue, the peptide lost its ordered conformation and, critically, its ability to be recognized.

The researchers then turned to the other half of the pairing: HSL3, a receptor-like kinase embedded in the plasma membrane. Receptor-like kinases form the frontline of plant perception, typically comprising an extracellular domain that captures ligands, a single membrane-spanning helix, and an intracellular kinase domain that relays the message. By mapping the interaction surface, the study showed that HSL3 recognizes not merely the chemical identity of CTNIP4 but its geometry. Binding assays demonstrated that the oxidized, disulfide-bonded form of the peptide engages the extracellular domain with high affinity, whereas the reduced or mutant forms bind poorly or not at all. In other words, the disulfide bond functions as a quality-control stamp, ensuring that only properly folded peptide molecules can sound the alarm.

This mechanism has implications that reach beyond a single ligand-receptor pair. Small peptides are intrinsically floppy; without stabilizing features, they sample many conformations in solution, and a receptor must either tolerate that flexibility or demand a specific shape. The CTNIP4-HSL3 system illustrates the second strategy, sometimes described as conformational selection with a structural gate: the receptor waits for, and exclusively accepts, the correctly folded state. The disulfide bond effectively reduces the search space, pre-organizing the peptide so that entropic costs of binding are minimized. For the plant, this design offers both sensitivity and specificity, allowing the immune system to respond rapidly to genuine danger signals while ignoring misfolded or degraded fragments that could otherwise trigger wasteful false alarms.

Genetic evidence reinforced the structural picture. Plants in which CTNIP4 or HSL3 function was disrupted showed blunted defense responses, and plants expressing mutant versions of CTNIP4 lacking the disulfide-forming cysteines failed to complement those defects, despite accumulating the peptide to normal levels. The deficiency was not in production but in presentation: the peptide was made, secreted, and present, yet structurally illegible to its receptor. This separation between abundance and functionality echoes a growing theme in peptide biology, where post-translational modifications and folding states, not raw expression levels, determine biological activity. It also suggests that breeding or engineering efforts aimed at boosting immune peptides must account for the machinery, such as the secretory pathway’s oxidative folding environment, that allows them to fold correctly in the first place.

The study also clarifies the evolutionary logic of the CTNIP family. Sequence comparisons across species reveal that the cysteine residues are among the most conserved positions, even as the surrounding amino acids vary considerably. That conservation pattern makes sense now: the cysteines are under selection because the fold they create is the substrate of recognition, while other positions can drift as long as the overall architecture and receptor-contacting residues are preserved. Such structural constraints explain how a family of peptides can diversify into multiple members with distinct receptor partners, expanding the vocabulary of plant immune communication without breaking the existing grammar. Each new pair of cysteines is a scaffold; the surface painted on that scaffold determines who listens.

From a biotechnological standpoint, the results arrive at a moment of intense interest in peptide-mediated immunity. Researchers are exploring synthetic phytocytokines as disease-resistance tools, either by applying peptides directly to crops or by engineering plants to produce enhanced versions. The CTNIP4-HSL3 work supplies a design principle for such efforts: any engineered variant must respect the disulfide architecture that defines the fold. It also suggests opportunities, since the disulfide-bonded loop is a discrete, portable module that could in principle be grafted onto other peptide scaffolds to create novel ligands, or targeted by small molecules that lock the peptide into active or inactive conformations. Structural knowledge of this kind turns peptide immunity from a black box into an engineerable system.

There are also lessons for understanding receptor kinase signaling more broadly. HSL3 belongs to a large superfamily of receptors, many of which bind small, post-translationally modified peptides through extracellular domains whose ligand-binding mechanisms remain unresolved. Demonstrating that a single covalent bond can be the decisive determinant of recognition provides a template hypothesis for other orphan receptors: when a peptide ligand seems too short or too featureless to encode specificity, look for disulfides, hydroxyprolines, tyrosine sulfations, or other modifications that impose order. The new study adds weight to the idea that the extracellular space of the plant cell is a chemically decorated landscape where modifications are not embellishments but the very language of communication.

As with any strong structural study, questions remain. How exactly does the HSL3 kinase domain translate ligand binding at the cell surface into intracellular phosphorylation events, and does CTNIP4 require a co-receptor, as many immune peptides do? What proteases liberate CTNIP4 from its precursor in vivo, and do those processing steps couple directly to the oxidizing environment that permits disulfide formation? Answering these will require further structural work on full-length receptor complexes and careful cell biology of the apoplast. But the core achievement stands: the study identifies the disulfide bond of CTNIP4 as the sculptor of its fold and, through that fold, the gatekeeper of its recognition by HSL3. In doing so, it reveals how plants squeeze remarkable information density into molecules barely a dozen amino acids long, and it hands researchers a blueprint for reading, and perhaps rewriting, the molecular conversations that keep crops alive.

Subject of Research: Structural basis of CTNIP4 phytocytokine folding via a disulfide bond for recognition by the receptor kinase HSL3

Article Title: A disulfide bond sculpts the CTNIP4 phytocytokine fold for recognition by the receptor kinase HSL3

Article References: Jiménez-Sandoval, P., Johanndrees, O., Snoeck, S., Harshith, C. Y., Omary, M., Broyart, C., Rhodes, J., Bender, K. W., Zipfel, C., & Santiago, J. (2026). A disulfide bond sculpts the CTNIP4 phytocytokine fold for recognition by the receptor kinase HSL3. Nature Plants, 12(9), 1688-1697. https://doi.org/10.1038/s41477-026-02380-y

Image Credits: AI Generated

DOI: 10.1038/s41477-026-02380-y

Keywords: CTNIP4, HSL3, phytocytokine, disulfide bond, receptor-like kinase, plant immunity, peptide folding, structural biology, ligand recognition, defense signaling, apoplast, secreted peptides