Plants fight fungi using kiwellin proteins

Fungal infection can affect crop yield. A plant protein found to counter fungal-induced interference with host metabolism illuminates antifungal defences and mechanisms that inhibit metabolic enzymes.
Mary C. Wildermuth is in the Department of Plant & Microbial Biology, University of California, Berkeley, Berkeley, California 94720, USA.

Search for this author in:

Organisms, such as fungi, that cause disease in plants often secrete proteins that aid growth and reproduction in the host. These are termed effector proteins, and some are deregulated metabolic enzymes that manipulate key metabolic pathways in plants. Writing in Nature, Han et al.1 reveal that a protein in maize (corn) blocks the enzymatic activity of a fungal effector enzyme, thereby thwarting the effector’s ability to influence maize metabolism in a way that limits the plant’s defence response.

The authors studied infection of maize by the fungus Ustilago maydis, which can cause corn smut disease and results in substantial crop loss worldwide. The enzyme chorismate mutase (Cmu1), which catalyses the molecular conversion of chorismate to prephenate, is a known effector protein of this fungus2. Han and colleagues engineered a tagged version of Cmu1 and used a technique called co-immunoprecipitation to try to identify whether any plant proteins interact with Cmu1 in maize leaves infected with U. maydis. This revealed a maize protein, called ZmKWL1 by the authors, that binds to Cmu1.

Han et al. determined that ZmKWL1 is a member of a family of proteins called kiwellins. Of the 20 kiwellin proteins found in maize, only ZmKWL1 is highly expressed in response to U. maydis infection3. The authors found that only ZmKWL1, of the four maize kiwellins they tested, interacts with Cmu1 in vitro. Furthermore, ZmKWL1 exclusively bound and inhibited purified U. maydis Cmu1 in vitro, whereas maize versions of chorismate mutase were not affected by ZmKWL1. The specificity of this interaction is remarkable, given the structural similarity between the fungal and maize enzymes.

Little was known previously about how kiwellins function. They are highly expressed in kiwi fruit and can trigger allergic responses in humans4,5. Han et al. found that kiwellin-encoding sequences are present both in non-seed plants, such as mosses, and in seed-producing plants, such as conifers and flowering plants. However, kiwellins are not present universally, and were not identified in the Brassicaceae family of plants, which includes the model plant Arabidopsis thaliana. Although the genomes of many species, including mosses, encode only one kiwellin, gene analysis by Han et al. suggests that some lineage-specific increases in kiwellins occurred as plants evolved, probably through gene duplication.

ZmKWL1 is part of small subgroup of nearly identical kiwellins found only in cereal plants. It is tempting, therefore, to speculate that this subgroup of kiwellins binds to the same target. Anecdotal evidence that requires additional experimental verification supports this hypothesis. For example, ZmKWL1 and its most closely related kiwellin protein Sb01g018600, from the cereal sorghum, have amino-acid sequences that are 87% identical. U. maydis does not infect sorghum, so U. maydis Cmu1 is not the target of sorghum kiwellin. However, another smut-causing fungus, Sporisorium reilianum, infects both sorghum and maize. This fungus encodes a secreted version of chorismate mutase (called sr16064)6 that might be the target of Sb01g018600.

Han and colleagues also provide evidence suggesting that other subgroups of kiwellins have evolved to recognize distinct effector proteins. Apart from ZmKWL1, maize kiwellins did not bind to Cmu1 in vitro, raising the possibility that some of these other kiwellins recognize other fungal proteins. The authors also noted previous studies of potatoes7,8 and husk tomato plants9 in which some kiwellins were highly expressed in response to organisms that cause disease in these plants.

It is intriguing that kiwellin-encoding genes are also found in genomes of some fungi that infect plants, including U. maydis. Han and colleagues do not speculate about the possible origins or functions of these genes. Perhaps fungal kiwellins were acquired by gene transfer from a cereal host plant, given that they are most similar to a large group of kiwellins in cereals. Could these fungal kiwellins counter the inhibition of their effectors by the plant kiwellins? Future experiments should investigate this possibility.

To determine what features of kiwellins might allow them to form strong and specific interactions with effector proteins, Han and colleagues used X-ray crystallography to generate structural models of ZmKWL1. This revealed that, like kiwellins in kiwi plants4,5, ZmKWL1 has a central ‘β-barrel’ domain that is stabilized by numerous connections called disulfide bridges. This type of arrangement is evolutionarily conserved in plant-secreted defence proteins commonly referred to as pathogenesis-related 4 family proteins10,11, and also found in fungal secreted proteins called cerato-platanins12 that modulate fungal interactions with hosts such as plants. However, plant kiwellins also contain a structure called an anti-parallel β-sheet, comprised of two β-strands and multiple surface-exposed loops. Han and colleagues’ structural studies of U. maydis Cmu1 in complex with ZmKWL1 shows extensive interaction between the proteins. The interactions form mainly between ZmKWL1 amino-acid residues in the anti-parallel β-sheet and Cmu1 residues in an extensive loop region unique to the fungal enzyme.

Han et al. report that ZmKWL1 inhibits Cmu1 by affecting the enzyme’s catalysis and not by competing for binding of its substrate molecule, chorismate. When the authors prevented the expression of ZmKWL1 in maize, plant infections with U. maydis were more severe than in maize that expressed ZmKWL1. Presumably, this is because the absence of ZmKWL1 enabled U. maydis Cmu1 to convert chorismate to prephenate, thereby limiting the availability of chorismate for synthesis of the plant defence hormone salicylic acid (Fig. 1). Previous work2 showed that deletion of the gene encoding Cmu1 in U. maydis caused an increase in salicylic acid production and associated plant-defence responses and a reduction in the success of fungal infection, compared with infection by strains that had Cmu1.

Figure 1 | A battle between plant and fungal enzymes for control of plant metabolism. Han et al.1 studied the infection of maize (corn) by the disease-causing fungus Ustilago maydis. They report that a plant protein that they term ZmKWL1 can block the action of a fungal enzyme called Cmu1 in maize cells. Plant enzymes are shown in red, and fungal enzymes in yellow. Cmu1 belongs to a family of chorismate mutase (CM) enzymes that is also found in maize, and catalyses the conversion of chorismate to prephenate as part of a pathway of amino-acid synthesis. This reaction prevents chorismate from contributing to a pathway that generates the plant defence molecule salicylic acid through the action of an isochorismate synthase (ICS) enzyme18 and another unknown plant enzyme. ZmKWL1, a member of a family of proteins called kiwellins, prevents Cmu1 from subverting plant metabolism to limit the production of salicylic acid. Ustilago maydis also encodes an isochorismatase enzyme, which is predicted to be secreted. This is a type of enzyme that can convert isochorismate to 2,3-dihydroxybenzoate, although whether the fungal enzyme functions in this way in plants is unknown. If this enzyme affects the availability of molecules in the salicylic-acid-generating pathway, it might be targeted by another kiwellin protein.

Chorismate is a key metabolite molecule in plants, and enzymes that use it compete to co-opt the molecule into their respective biosynthetic pathways, which generate molecules that include amino acids, hormones, vitamins and plant cell-wall components13. Enzymes that use chorismate have similar structures and reaction mechanisms14,15. The ability of ZmKWL1 to inhibit Cmu1 with such specificity raises the question of whether other kiwellins in plants have evolved to specifically inhibit related enzymes, including isochorismatases. Plant-infecting fungi can secrete isochorismatases to limit the availability of the molecule isochorismate16. This molecule can be converted to salicylic acid (Fig. 1). U. maydis encodes an isochorismatase that is predicted to be secreted. The use of multiple effectors to target the pathway that generates salicylic acid would have a major effect on the ability of plants to mount a defence response. If distinct kiwellins evolved to inhibit such fungal effectors, it might enable plants to generate sufficient salicylic acid to induce the robust defences needed to limit fungal infection.

Han and colleagues have set the stage for the identification of other kiwellin effector targets. Detailed future analyses of kiwellin evolution and function might help to reveal the full range of roles of these proteins. Perhaps naturally occurring or engineered kiwellins that specifically inhibit a range of enzymes that use chorismate or isochorismate could be developed to enhance agricultural productivity. Furthermore, kiwellins could be used to manipulate chorismate metabolism to enhance the production of a variety of commercial chorismate-derived products17.

Kiwellins might also have the potential to be developed as antimicrobial agents for the treatment of human disease. Certain bacteria, including the bacterium that causes tuberculosis, use chorismate to make molecules that they require for infection14. The human genome encodes neither chorismate-using enzymes nor kiwellins; therefore, kiwellin-based inhibition of these microbial targets should be investigated. The range of metabolic proteins bound by kiwellins probably extends beyond enzymes that use chorismate, and it will be exciting to uncover the full versatility of those proteins.

Nature 565, 575-577 (2019)

doi: 10.1038/d41586-019-00092-2


  1. 1.

    Han, X. et al. Nature 565, 650–653 (2019).

  2. 2.

    Djamei, A. et al. Nature 478, 395–398 (2011).

  3. 3.

    Lanver, D. et al. Plant Cell 30, 300–323 (2018).

  4. 4.

    Hamiaux, C. et al. J. Struct. Biol. 187, 276–281 (2014).

  5. 5.

    Offermann, L. R. et al. J. Agric. Food Chem. 63, 6567–6576 (2015).

  6. 6.

    Schirawski, J. et al. Science 330, 1546–1548 (2010).

  7. 7.

    Draffehn, A. M. et al. Front. Plant Sci. 4, 423 (2013).

  8. 8.

    Mosquera, T. et al. PLoS ONE 11, e0156254 (2016).

  9. 9.

    Quitana-Camargo, M. et al. Acta Physiol. Plant. 37, 29 (2015).

  10. 10.

    Huet, J. et al. Acta Cryst. D 69, 2017–2026 (2013).

  11. 11.

    Pereira Menezes, S. et al. BMC Plant Biol. 14, 161–182 (2014).

  12. 12.

    De Oliveira, A. L. et al. J. Biol. Chem. 286, 17560–17568 (2011).

  13. 13.

    Maeda, H. & Dudareva, N. Annu. Rev. Plant Biol. 63, 73–105 (2012).

  14. 14.

    Shelton, C. L. & Lamb, A. L. Trends Biochem. Sci. 43, 342–357 (2018).

  15. 15.

    Lamb, A. L. Biochemistry 50, 7476–7483 (2011).

  16. 16.

    Liu, T. et al. Nature Commun. 5, 4686 (2014).

  17. 17.

    Avereschm, N. J. H & Krömer, J. O. Front. Bioeng. Biotechnol. 26, 32–50 (2018).

  18. 18.

    Wildermuth, M. C., Dewdeny, J., Wu, G. & Ausubel, F. M. Nature 414, 562–565 (2001).

Download references

Nature Briefing

An essential round-up of science news, opinion and analysis, delivered to your inbox every weekday.