henhaobio
Sign in

Nature PlantsPlant BiologyPeer-reviewed journal article

Reverse engineering of the pattern recognition receptor FLS2 reveals key design principles of broader recognition spectra against evading flg22 epitopes

S. H. ZhangiD, Songyuan LiuiD, Hung-Fei Lai, Kyle W. BenderiD, Gijeong KimiD, Amedeo CaflischiD, Cyril ZipfeliD

PDFDOIPreprint
In-depth full-text digest available · Jump to digest ↓

Abstract

In the ongoing plant-pathogen arms race, plants employ pattern recognition receptors (PRRs) to recognize pathogen-associated molecular patterns (PAMPs), while in successful pathogens, PAMPs can evolve to evade detection. Engineering PRRs to recognize evading PAMPs could potentially generate broad-spectrum and durable disease resistance. In this study, we reverse-engineered two natural FLAGELLIN SENSING 2 (FLS2) variants, VrFLS2XL and GmFLS2b, with extended recognition specificities towards evading flg22 variants. We identified minimal gain-of-function residues enabling blind FLS2s to recognize otherwise evading flg22 variants. We uncovered two strategies: (i) enhancing FLS2-flg22 interaction around flg22s key evasion sites, and (ii) strengthening direct interaction between FLS2 and its co-receptor BAK1 to overcome weak agonistic and antagonistic flg22s, respectively. Additionally, we leveraged polymorphisms that enhance recognition through unknown mechanisms to engineer superior recognition capability. These findings offer basic design principles for PRRs with broader recognition spectra, paving the way for PRR engineering using precise gene-editing to increase disease resistance in crops.

In-depth summary

Claude Sonnet 4.6 · xhigh effort · 8 Nov 2026 · from the full text of version 1 (13 Oct 2024) · CC BY 4.0

Minimal residue sets were mapped that let blind grape and soybean FLS2 receptors recognize evading flg22 peptides, giving two design principles and an engineered GmFLS2 variant with stronger flg22Rso responses.

The paper says the design rules for editing native pattern recognition receptors (PRRs) to detect evading epitopes are poorly understood, and earlier FLS2 studies gave no gain-of-function receptor or mapped too coarsely to guide precise engineering. Here two natural FLS2 variants are reverse-engineered to minimal gain-of-function residues, yielding two principles keyed to how the flg22 variant evades.

Key results

  1. ShownThe LRR12-19 region of VrFLS2XL is necessary and sufficient for recognizing flg22Atum, which evades VvFLS2 mainly by loss of binding.

    VvFLS2 carrying LRR12-19 of VrFLS2XL (chimera '0100') gained flg22Atum-induced ROS; VrFLS2XL with VvFLS2 LRR12-19 (chimera '1011') lost it. In the GUS infection assay, '0100' restricted A. tumefaciens similarly to VrFLS2XL. · LRR12-19 region of VrFLS2XL is mainly…

  2. ShownAbout ten residue changes in LRR12-19 are said to be required for VvFLS2 to recognize flg22Atum (Fig. 2f), while the full 'comp.Med' set that would be introduced into grape comprises 14 residue changes (18 nucleotide substitutions); buried residues matter too.

    'comp.Med' showed only a minimal drop in ROS recognition relative to 'comp.Max' and chimera '0100', and its resistance to A. tumefaciens was comparable to or better than VrFLS2XL, while 'comp.Min' lost significant ROS recognition and resistance (not different from VvFLS2). Mutating buried residues 391, 413 and 439 together abolished flg22Atum recognition. · VrFLS2XL residues critical for flg22Atum…; Discussion

  3. ShownTwo residues (Q600H and S603A) in the LRR20-EJM region of GmFLS2a are enough for it to recognize the antagonistic flg22Rso, with ROS, calcium and MAPK responses.

    The double mutant gave stable flg22Rso-induced ROS not significantly weaker than GmFLS2b, with no change in protein abundance. Neither single mutation worked consistently; S603A alone gave no response. Calcium influx and MAPK activation were induced by flg22Rso. Thirty shuffled variants identified these two residues. · LRR20-EJM region of GmFLS2b is mainly…; DNA shuffling identifies minimal…

  4. ShownDNA shuffling also exposed polymorphisms outside the flg22 and BAK1 interfaces (including a C-terminal capping/juxtamembrane group) that enhance flg22Rso recognition; combining enhancers gave 'superGmFLS2'.

    'superGmFLS2' produced more than 3-fold higher total flg22Rso-induced ROS than wild-type GmFLS2b, with unchanged expression. The calcium influx did not rise significantly, and resistance to R. solanacearum was not tested. · DNA shuffling identifies minimal…

  5. SuggestedThe principles transfer to other FLS2 homologs only in line with the evasion mechanism: LRR12-19 transplantation works for the weak agonist; LRR20-EJM transplantation works where flg22Rso is a strong antagonist.

    LRR12-19 of VrFLS2XL gave AtFLS2, SlFLS2, GmFLS2a and GmFLS2b flg22Atum recognition, but NbFLS2 lost flg22Pa detection. LRR20-EJM of GmFLS2b enabled SlFLS2 and NbFLS2, not AtFLS2, VvFLS2 or VrFLS2XL. Among homologs tested in Extended Data Fig. 5, flg22Rso is a strong antagonist for SlFLS2 and NbFLS2, with minimal effect on AtFLS2, VvFLS2, VrFLS2XL. · The broad applicability of FLS2 engineering…

Methods, data and reproducibility

Natural FLS2 variants were dissected with domain-swapped chimeras, AlphaFold3-guided residue sets and an optimized DNA shuffling library. Variants were scored for peptide-induced responses in Nicotiana benthamiana fls2 mutant leaves (ROS, calcium, MAPK, GUS infection proxy).

  • Chimeric and point-mutant FLS2 variants (VvFLS2/VrFLS2XL; GmFLS2a/b) tested by transient expression and ROS assays · Nicotiana benthamiana (fls2 mutant host); grape and soybean receptor sources · n = at least eight leaf discs per treatment; at least three leaves from three plants
  • Shuffled GmFLS2a/b LRR20-EJM variants analysed for genotype-phenotype association · Nicotiana benthamiana (host); soybean receptor sources · n = thirty variants
  • Other FLS2 homologs (AtFLS2, SlFLS2, NbFLS2) used for transplantation tests · Arabidopsis thaliana, Solanum lycopersicum, N. benthamiana
Methods & tools
AlphaFold3, Mapiya, PyMOL, Jalview, Golden Gate cloning, SacB counterselection, JBS DNA-Shuffling Kit, Welch's t-test, Kruskal-Wallis with Dunn's test, luminol ROS assay, aequorin calcium assay
Reproducibility gaps
No code or data repository URLs are given; DNA sequences, primers and shuffled variant details are in Supplementary Tables 1-3, and some plasmid templates were gifts from other labs.
Limitations (6)
  • Screening throughput of the DNA shuffling and variant-testing workflow is low and remains a bottleneck. · stated by the authors
  • DNA shuffling needs highly homologous parental fragments with polymorphic sites that are not tightly clustered, so it suits GmFLS2a/b LRR20-EJM but not VvFLS2/VrFLS2XL LRR12-19. · stated by the authors
  • Whether superGmFLS2 confers stronger resistance to R. solanacearum than GmFLS2b is untested, and the LRR20-EJM strategy works only for FLS2s to which flg22Rso is an antagonist. · stated by the authors
  • The mechanisms behind the 'dark matter' polymorphisms and the novel specificity gains remain unknown, and extending the principle to antagonistic variants beyond flg22Rso is untested. · stated by the authors
  • All receptor tests use transient expression in N. benthamiana leaves with synthetic peptides; resistance was assessed only for A. tumefaciens (GUS proxy), and no engineered native crop gene was tested. · noted in this summary
  • Residue effects are inferred from AlphaFold3 models and ROS readouts without experimentally resolved structures or binding measurements for the engineered variants. · noted in this summary

Read this if you want mapped minimal residue changes and design rules for engineering FLS2 to recognize evading flg22 variants.

Citations
37
FWCI
17.27
Field percentile
top 1%

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISLi '25Zhong '26Xiao '26Ma '26Roberts '20Yang '25
80% semantic matchPlant Biology
Tianrun Li, Esteban Jarquin Bolaños, Danielle M. Stevens+3 more·Nature Plants·Jul 28, 2025· 36 citations