A dominant-negative avirulence effector of the barley powdery mildew fungus provides mechanistic insight into barley MLA immune receptor activation

Abstract Nucleotide-binding leucine-rich repeat receptors (NLRs) recognize pathogen effectors to mediate plant disease resistance often involving host cell death. Effectors escape NLR recognition through polymorphisms, allowing the pathogen to proliferate on previously resistant host plants. The powdery mildew effector AVRA13-1 is recognized by the barley NLR MLA13 and activates host cell death. We demonstrate here that a virulent form of AVRA13, called AVRA13-V2, escapes MLA13 recognition by substituting a serine for a leucine residue at the C-terminus. Counterintuitively, this substitution in AVRA13-V2 resulted in an enhanced MLA13 association and prevented the detection of AVRA13-1 by MLA13. Therefore, AVRA13-V2 is a dominant-negative form of AVRA13 and has probably contributed to the breakdown of Mla13 resistance. Despite this dominant-negative activity, AVRA13-V2 failed to suppress host cell death mediated by the MLA13 autoactive MHD variant. Neither AVRA13-1 nor AVRA13-V2 interacted with the MLA13 autoactive variant, implying that the binding moiety in MLA13 that mediates association with AVRA13-1 is altered after receptor activation. We also show that mutations in the MLA13 coiled-coil domain, which were thought to impair Ca2+ channel activity and NLR function, instead resulted in MLA13 autoactive cell death. Our results constitute an important step to define intermediate receptor conformations during NLR activation.


Introduction
During infection of their host, pathogens secrete numerous virulence factors that act extracellularly or inside host cells.These so-called effectors manipulate the host's physiology in favour of the pathogen.Disease resistance of a plant against a pathogen is often mediated by nucleotide-binding leucinerich repeat receptors (NLRs) (Maekawa et al., 2011b;Jones et al., 2016).NLRs recognize effectors directly or by indirectly detecting effector-mediated alterations of host targets (Cesari, 2018).Effector-mediated NLR activation is often linked to localized host cell death (Dodds and Rathjen, 2010;Saur and Hückelhoven, 2021;Maekawa et al., 2022), and recognized effectors are called avirulence (AVR) effectors.Diversification of AVR genes can lead to loss of recognition by the respective NLR, resulting in pathogen virulence and breakdown of disease resistance (Märkle et al., 2022).In the case of direct AVR recognition, the NLR can usually no longer bind the diversified effector proteins (Saur et al., 2021).
NLRs are multidomain proteins with a central nucleotidebinding (NB) domain and C-terminal leucine-rich repeats (LRRs).At the N-terminus, most NLRs encode either a Toll/ interleukin-1 receptor-like (TIR) or a coiled-coil (CC) domain, classifying the majority of NLRs into either TIR-type NLRs (TNLs) or CC-type NLRs (CNLs) (Shao et al., 2016).A subgroup of CNLs (also called RPW8-like NLRs or RNLs) are the helper NLRs NRG1 (N REQUIREMENT GENE 1) and ADR1 (ACTIVATED DISEASE RESISTANCE GENE 1) that are required for TNL-mediated disease resistance (Saile et al., 2020).The N-terminal CC and TIR domains mediate NLR signal emission upon NLR activation (Swiderski et al., 2009;Bernoux et al., 2011;Collier et al., 2011;Maekawa et al., 2011a;Williams et al., 2014).In inactive receptors, CC and TIR domains are locked in inactive conformations, and this autoinhibition is mediated by interdomain interactions between the N-terminal domains and the NB and LRR domains (Burdett et al., 2019;Saur et al., 2021;Tamborski et al., 2023).Although structural information on intermediate forms between inactive and active NLRs is limited to the structure of the Arabidopsis thaliana CNL ZAR1 (HOPZ-ACTIVATED RESISTANCE 1) (Wang et al., 2019b), NLR activation appears to be a multistep process (Förderer et al., 2022b).The first activation step is ligand binding, which induces a steric clash between the LRR and the NB domains.The resulting open conformation of the NB domain then allows ADP (inactive) to ATP (active) exchange, which in turn induces allosteric changes to release the conformational autoinhibition of the CC or TIR domains.This induces NLR oligomerization, and these NLR oligomers are referred to as resistosomes (Förderer et al., 2022b).Certain amino acid replacements within the conserved MHD motif of the NB domain mimic ATP binding and thus result in an active NLR conformation (Dinesh-Kumar and Baker, 2000;Bendahmane et al., 2002;Gao et al., 2011;Bai et al., 2012;Ntoukakis et al., 2013;Roberts et al., 2013;Nishimura et al., 2017).The N-terminal portion of the LRR domain in CNLs also contributes to receptor autoregulation through interactions with CC and NB domains, and amino acid exchanges at these sites can affect NLR autoactivity (Rairdan and Moffett, 2006;Slootweg et al., 2013;Burdett et al., 2019;Förderer et al., 2022a;Tamborski et al., 2023).For receptor activation via direct effector recognition, amino acids in the LRR can also function as effector contact sites and can define the specificity of effector recognition (Jia et al., 2000;Shen et al., 2003;Dodds et al., 2006;Bauer et al., 2021).Upon direct effector recognition by the LRR or other integrated domains, effector binding correlates directly with NLR signal activation, and studies on the Magnaporthe oryzae effectors AvrPik and AVR-Pia and the rice NLRs Pik and RGA5 (RESISTANCE GENE ANALOG5), respectively, argue for an affinity threshold between receptor and effector for activation of NLR immune signalling and pathogen resistance (Ortiz et al., 2017;de la Concepcion et al., 2018).
While the mechanisms underlying the restriction of pathogen growth by resistosomes is not fully elucidated, recent cryo-EM structures of multiple resistosomes (Wang et al., 2019a, b;Ma et al., 2020;Martin et al., 2020;Förderer et al., 2022a) revealed fundamental differences in TNL and CNL signalling: the pentameric resistosomes of A. thaliana ZAR1 CNL and wheat Sr35 CNL have Ca 2+ channel activity (Bi et al., 2021;Förderer et al., 2022a).The funnel-shaped ZAR1 cation channel is formed by the N-terminal CC domain α1-helix of the ZAR1 resistosome (Wang et al., 2019a, b).Substitutions of negatively charged amino acids to alanine in the inner lining of the funnel abolish Ca 2+ channel and cell death activity and ZAR1-mediated resistance (Wang et al., 2019b;Bi et al., 2021).The α1-helix of the wheat Sr35 resistosome is not well resolved and Sr35 α1-helix amino acid exchanges equivalent to those in ZAR1 do not affect Sr35 resistosome channel and cell death activity (Förderer et al., 2022a;Zhao et al., 2022), suggesting differences in Ca 2+ signalling functions between ZAR1 and Sr35.Effector binding to the TNLs RPP1 (RECOGNITION OF PERONOSPORA PARASITICA 1) and ROQ1 (RECOGNITION OF XopQ 1) from A. thaliana and Nicotiana benthamiana, respectively, induces the formation of homotetrameric complexes, stimulating TIR enzyme activity.The resistosome TIR enzyme, but also TIR-only proteins, produce a variety of nucleotide-based second messenger molecules (Horsefield et al., 2019;Wan et al., 2019;Huang et al., 2022;Jia et al., 2022;Yu et al., 2022), some of which serve as ligands to activate the EDS1 protein family plus the signalling/helper CNLs ADR1 or NRG1 (Lapin et al., 2019;Huang et al., 2022;Jia et al., 2022).ADR1 and NRG1 can also function as calcium ion-permeable channels (Jacob et al., 2021) and, as such, disruption of Ca 2+ homeostasis appears to be central in CNL and TNL resistosome signalling.
The polymorphic barley Mildew locus A (Mla) encodes allelic variants of CNLs (MLA NLRs), each conferring isolatespecific disease resistance to the barley powdery mildew fungus Blumeria graminis f. sp.hordei (Bgh) (Moseman and Schaller, 1960;Glawe, 2008;Seeholzer et al., 2010;Maekawa et al., 2019).Some barley MLAs and Mla homologues of other cereals confer additional resistance to isolates of unrelated fungal pathogens (Periyannan et al., 2013;Mago et al., 2015;Chen et al., 2017;Bettgenhaeuser et al., 2021;Brabham et al., 2022, Preprint;Ortiz et al., 2022).The Bgh effectors recognized by barley MLAs are known as AVR A effectors (Jorgensen, 1994), and diversified variants that have escaped Mla recognition are designated as AVR A -V variants (Lu et al., 2016).To date, fulllength structures of inactive or effector-activated MLAs are not available, but protein interaction assays suggest a direct interaction between at least some MLA NLRs and matching AVR A effectors (Saur et al., 2019a).Most amino acids under positive selection of Mla resistance specificities map to the predicted solvent-exposed sites of the LRR, suggesting that these serve as AVR A contact residues (Seeholzer et al., 2010;Maekawa et al., 2019), but interaction between effectors and MLA LRR domain deletion constructs could not be shown.Most of the known Bgh AVR A effectors are unrelated in sequence, but share a common fold reminiscent of RNases lacking catalytic residues (Bauer et al., 2021).
Because receptor-effector interaction is commonly linked to receptor activation, we aimed here to investigate the seeming paradox of MLA13 inactivity despite stable AVR A13 -V2-MLA13 association.By applying proximity-dependent protein labelling (BioID), yeast two-hybrid (Y2H) interaction assays, and structural prediction (Alphafold2) in combination with in planta expression of AVR A13 effector variants, we demonstrate that a single surface-exposed amino acid at the C-terminus of AVR A13 effectors determines the association with and activation of MLA13.Our data also reveal that AVR A13 -V2 acts as a dominant-negative effector on MLA13-mediated cell death.This proposes that breakdown of Mla13-mediated resistance can be explained by Bgh isolates carrying dominant-negative AVR A13 -V2.We also demonstrate that amino acid exchanges in the MLA13 NB and LRR domains compromise effector binding.In turn, amino acid changes in the MLA13 CC domain predicted to disrupt cation channel activity do not affect MLA13-mediated cell death.Nevertheless, inhibition of Ca 2+ and other cation channels by LaCl 3 impaired MLA13mediated cell death of barley protoplasts.Collectively, these results provide insights and tools for understanding the conformational changes NLRs undergo during effector-mediated NLR resistosome activation.

Plant and fungal materials and growth conditions
Golden Promise and near sssisogenic lines (NILs) of the barley cultivar Manchuria were grown at 19 °C, 70% relative humidity, and under a 16 h photoperiod.Nicotiana benthamiana plants were grown under standard greenhouse conditions with a 16 h photoperiod.Maintenance of Bgh isolates was carried out as described previously (Lu et al., 2016).
Transient gene expression by Agrobacterium-mediated transformation of N. benthamiana leaves Agrobacterium tumefaciens GV3101:pMP90K were freshly transformed with the respective constructs of interest and grown from single colonies in liquid Luria broth medium containing appropriate antibiotics for ~24 h at 28 °C to an OD 600 not higher than 1.5.Bacterial cells were harvested by centrifugation at 2500 g for 15 min followed by resuspension in infiltration medium (10 mM MES, pH 5.6, 10 mM MgCl 2 , and 200 µM acetosyringone) to a final OD 600 =1.Cultures were incubated for 2-4 h at 28 °C with 180 rpm shaking before infiltration into leaves from 3-to 5-week-old N. benthamiana plants.For co-expression of multiple constructs, Agrobacteria carrying the genes of interest were mixed equally unless indicated otherwise.Cell death was assessed 1-5 d post-infiltration as indicated, and tissue for immunodetection analysis was harvested 1-2 d post-infiltration as indicated.

Protein extraction from N. benthamiana leaf tissue for protein detection by immunoblotting
Frozen leaf material was ground to a fine powder using pre-cooled adapters in a bead beater (Retsch) and thawed in cold plant protein extraction buffer [150 mM Tris-HCl, pH 7.5, 150 mM NaCl, 10 mM EDTA, 10% (v/v) glycerol, 5 mM DTT, 2% (v/v) plant protease inhibitor cocktail (Sigma), 1 mM phenylmethylsulfonyl fluoride (PMSF), and 0.5% (v/v) IGEPAL] at a ratio of 50 mg fresh tissue/150 µl of extraction buffer.Extracts were centrifuged twice at 15 000 g for 10 min at 4 °C.For SDS-PAGE, extracts were diluted 4:1 with 4× SDS loading buffer and heated to 85 °C for 10-15 min before again removing insoluble material by centrifugation at 15 000 g for 5 min.For pull-down of monomeric yellow fluorescent protein (mYFP)-tagged proteins, green fluorescent protein (GFP)-Trap-MA (Chromotek) beads were incubated in equilibration buffer (Saur et al., 2015) for 1 h at 4 °C and subsequently mixed with 1 ml of protein extracts for 2-3 h at 4 °C with slow but constant rotation.

Proximity-dependent protein labelling of proteins transiently expressed in N. benthamiana leaves
Pull-down of biotinylated proteins was performed by following published protocols (Conlan et al., 2018) with the alteration that free biotin was not removed before adding streptavidin to protein extracts.Instead, we infiltrated (Shi et al., 2023) a 10 µM biotin solution to the plant tissue (instead of a 75 µM solution; Conlan et al., 2018).We followed a sequence of infiltrations to minimize MLA-mediated cell death of N. benhamiana leaf tissue: Agrobacterium tumefaciens GV3101::pMP90K carrying 35S:Mla-4Myc constructs were grown from glycerol stocks and infiltrated (day 1).At 24 h post-infiltration of the Mla constructs, Agrobacteria freshly transformed with 35S:AVR a13 -BirA-4Myc constructs or the empty vector (EV) were infiltrated as indicated (day 2).Free biotin (10 µM) in infiltration buffer lacking acetosyringone was infiltrated at 24 h after the second infiltration and 48 h after the first infiltration (day 3).Tissue for streptavidin-based precipitation of biotinylated proteins was harvested 24 h post-infiltration of free biotin.Frozen leaf material was ground to a fine powder using pre-cooled adapters in a bead beater (Retsch) and thawed in cold plant denaturing extraction buffer [150 mM Tris-HCl, pH 7.5, 150 mM NaCl, 10 mM EDTA, 5% (v/v) glycerol, 5 mM DTT, 1% (v/v) plant protease inhibitor cocktail (Sigma), 1 mM NaF, 1 mM sodium orthovanadate, 1 mM PMSF, 1% Triton X-100, and 0.5 % (w/v) SDS] at a ratio of 300 mg fresh tissue/2 ml of denaturing extraction buffer.Extracts were incubated rotating at 4°C for 30 min before the removal of insoluble material by centrifugation at 21 000 g for 30 min at 4 °C.Streptavidin-coated Dynabeads (100 μl per sample, MyOne streptavidin C1, Thermo Fisher) were incubated in wash buffer [150 mM Tris-HCl, pH 7.5, 150 mM NaCl, 10 mM EDTA, 5% (v/v) glycerol, 1% (v/v) plant protease inhibitor cocktail (Sigma)] containing 1% BSA for 1 h at 4 °C and subsequently mixed with 2 ml of protein extracts for 3 h at 4 °C with slow but constant rotation.Then, conjugated streptavidin beads were washed four times in 1 ml of cold wash buffer before interacting proteins were stripped from the beads by heating to 85 °C for 10-15 min in 50 μl of 4× SDS loading buffer.From these 50 µl, a 30 µl aliquot was loaded on 9% SDS-PAGE gels.Proteins were blotted onto a PVDF membrane and probed with anti-Myc (abcam ab9106) followed by anti-rabbit IgG-HRP (Santa Cruz Biotechnology sc-2313) secondary antibodies.Myctagged proteins were detected by the HRP activity on SuperSignal West Femto Maximum Sensitivity Substrate (Thermo Fisher 34095) using a Gel Doc™ XR+ Gel Documentation System (Bio-Rad).

Transient gene expression and cell death assay in barley protoplasts
Assessment of protoplast cell death using a luciferase (LUC) activity as a proxy for cell viability was performed as described (Saur et al., 2019b).Briefly, Mla cDNA and AVR a cDNAs lacking the respective signal peptide were expressed from the Zea mays ubiquitin promotor in protoplasts isolated from barley cultivar Golden Promise, Manchuria CI 2330, and cultivar Manchuria Mla13 NIL CI 16155.For this, the epidermis of the secondary (Golden Promise) or primary (Manchuria) leaves from 7-to 8-day-old plants was removed before leaves were immersed in the enzyme solution.A total volume of 30 µl of water containing the LUC reporter and other constructs was transfected as indicated into 300 µl of barley protoplasts at a concentration of 3.5×10 5 protoplasts ml -1 solution.Protoplasts were recovered in regeneration buffer supplemented with LaCl 3 as indicated.About 16 h after transfection, protoplasts were collected by centrifugation at 1000 g, the supernatant was discarded, and 200 µl of 2× cell culture lysis buffer were added (Promega, E1531).LUC activity was determined by mixing 50 µl of protoplast lysate with 50 µl of LUC substrate (Promega, E1501) in a white 96-well plate, and light emission was measured at 1 s per well using a microplate luminometer (Centro, LB960).

Protein extraction from barley protoplasts, and fusion protein detection by immunoblotting
To determine the effect of LaCl 3 treatment on AVR A13 protein, for each LaCl 3 treatment, 300 µg of the AVR a13 -V2-mYFP effector construct or an EV was transfected into 3 ml of barley protoplasts cultivar Manchuria CI 2330 at a concentration of 5×10 5 protoplasts ml -1 solution.Protoplasts were recovered in regeneration buffer supplemented with the LaCl 3 to the final concentrations indicated.About 16 h posttransfection, protoplasts were collected by centrifugation at 1000 g, the supernatant was discarded, and protoplast pellets were frozen in liquid nitrogen.Total protein was extracted by the addition of 100 µl of cold plant protein extraction buffer [200 mM Tris-HCl, pH 7.5, 150 mM NaCl, 10 mM EDTA, 10% (v/v) glycerol, 12 mM DTT, 2% (v/v) plant protease inhibitor cocktail (Sigma), and 1% (v/v) IGEPAL] to each protoplast pellet.Extracts were centrifuged at 15 000 g for 5 min at 4 °C.For SDS-PAGE, extracts were diluted 4:1 with 4× SDS loading buffer and heated to 85 °C for 10-15 min before removing insoluble material by centrifugation at top speed for 5 min.Samples were separated by 10% SDS-PAGE, blotted onto a PVDF membrane, and probed with anti-GFP (Santa Cruz Biotechnology sc-8334 or abcam ab6556) followed by anti-rabbit IgG-HRP (Santa Cruz Biotechnology sc-2313) secondary antibodies.mYFP-tagged proteins were detected by the HRP activity on SuperSignal West Femto Maximum Sensitivity Substrate (Thermo Fisher 34095) using a Gel Doc™ XR+ Gel Documentation System (Bio-Rad).
For protein detection, yeast strains were grown to OD 600 =1 in SD-UHW/Gal/Raf liquid medium at 30 °C and 200 rpm shaking, and proteins were extracted using 200 mM NaOH (NaOH method) (Zhang et al., 2011).Total protein samples were separated by 9% or 12% SDS-PAGE, blotted onto a PVDF membrane, and probed with anti-HA (Merck, clone 3F10) or anti-LexA (Santa Cruz Biotechnology, sc7544) primary antibodies followed by anti-rat (Santa Cruz Biotechnology, sc2065) or anti-mouse IgG-HRP (Santa Cruz Biotechnology, sc2005) secondary antibodies as appropriate.HA and LexA fusion proteins were detected by HRP activity on SuperSignal West Femto Maximum Sensitivity Substrate (Thermo Fisher 34095) using a Gel Doc™ XR+ Gel Documentation System (Bio-Rad).

The C-terminus of AVR A13 effectors determines interaction with and activation of MLA13
The C-terminally located polymorphisms between genes encoding avirulent AVR A13 -1 effector and virulent AVR A13 -V1 or AVR A13 -V2 variants (Fig. 1A) indicate a role for the AVR A13 -1 C-terminus in the interaction with and activation of MLA13.Previously, no avirulence activity could be detected for AVR A13 -V1, but this could be attributed to its protein instability upon transient expression in planta (Lu et al., 2016;Saur et al., 2019a).Here we aimed to stabilize AVR A13 -V1 protein to retest the association patterns of the AVR A13 variants with MLA13 in planta.To this end, we fused the three effector variants to a biotin ligase (BirA), and indeed this fusion allowed immunodetection of the AVR A13 -V1 at levels comparable with the other two variants in N. benthamiana leaves (Supplementary Fig. S1).We also confirmed the functionality of the tagged proteins by demonstrating MLA13-specified cell death induced by AVR A13 -1-BirA-4×Myc.In turn, AVR A13 -V1-BirA-4×Myc and AVR A13 -V2-BirA-4×Myc did not activate MLA13-specified cell death in these assays (Supplementary Fig. S1).We detected biotinylated MLA13, but not MLA1 or MLA7 protein, in samples expressing Mla13-4Myc together with AVR a13 -1-BirA or AVR a13 -V2-BirA, but not AVR a13 -V1-BirA after biotin treatment followed by a streptavidin pulldown (Supplementary Fig. S1B).Given that AVR A13 -V1 lacks the 42 C-terminal amino acids of AVR A13 -1 (Fig. 1A), the data provide experimental evidence that the C-terminal half of AVR A13 is needed for the association and activation of the MLA13 receptor.
MLA13 interacts more efficiently with AVR A13 -V2 than with AVR A13 -1, and this enhanced association correlates with the inability to induce MLA13-mediated cell death (Saur et al., 2019a).We therefore tested the association of AVR A13 -1 ∆SPE and the AVR A13 -1/AVR A13 -V2 hybrid variants with MLA13.Protein stability of AVR A13 hybrid variants varies in planta, which makes the assessment of quantitative differences difficult (Fig. 1C).We therefore used a Y2H assay drop-out series to evaluate putative quantitative differences.We fused Mla N-terminally to the LexA-binding domain sequence (BD-Mla13) and the AVR a13 variants to the B42 activation domain (AD-AVR a13 ) and determined yeast growth in the absence of leucine as a proxy for protein interaction.Yeasts co-expressing BD-Mla13 with AD-AVR a13 -1 and AD-AVR a13 -1 ∆SPE grew less in the dilution series than yeasts carrying AD-AVR a13 -V2 or any of the AD-AVR a13 hybrid constructs (Fig. 1D).No growth was detected when BD-Mla13 was co-expressed with AD-AVR a13 -V1 or when it was replaced by BD-Mla1 (Fig. 1D; Supplementary Fig. S2).The data imply that L 119 of AVR A13 -V2 (Fig. 1A) is responsible for the enhanced interaction with MLA13.The corresponding residue in AVR A13 -1 is a serine.We generated structural predictions of the AVR A13 variants [lacking the respective signal peptides (SPs)] using AlphaFold2 (pLDDT overall =89, pLDDT L/ S119 >80) and found that indeed both, L 119 of AVR A13 -V2 ∆SP and S 119 of AVR A13 -1 ∆SP appear to be surface-exposed in these structural models, suggesting that they are accessible for binding to MLA13 (Supplementary Fig. S2C).

AVR A13 -V2 can act as a dominant-negative effector on MLA13-mediated cell death
The enhanced association between MLA13 and AVR A13 -V2 could affect Mla13 disease resistance and the activity of other MLA NLRs.To test this, we measured AVR A -induced MLA-mediated cell death in the presence of AVR A13 -V2.Co-expression of Mla13-4×Myc with AVR a13 -1-mYFP and an EV in N. benthamiana leaves resulted in cell death within 50-72 h, and this response was not detectable when EV was exchanged for AVR a13 -V2-mYFP (Fig. 2A).We also tested whether AVR a13 -V2-4×Myc affects cell death mediated by Mla1-3×HA and AVR a1 -mYFP or by Mla7-3×HA and AVR a7 -2-mYFP.We assessed the severity of cell death on a scale from 0 to 3 and found that AVR A13 -1-and MLA13mediated cell death was abrogated by the co-expression of AVR A13 -V2.In contrast, Mla1 and AVR a1 or Mla7 and AVR a7 -2 were not affected by AVR a13 -V2 (Fig. 2B).The specific inhibitory effect of AVR a13 -V2 on the MLA13 receptor (Fig. 2B) is not due to low MLA13 or AVR A13 -1 protein stability in the AVR a13 -V2-expressing samples (Fig. 2C).Importantly, AVR A13 -V1 had no inhibitory effect on cell death mediated by co-expression of Mla13 and AVR a13 -1, even when AVR A13 -V1 protein was stabilized by C-terminal fusion with the BirA tag (Supplementary Fig. S1C).Using a protoplastbased assay that relies on LUC activity as a proxy of cell viability (Saur et al., 2019b), we also determined if AVRA 13 -V2 inhibits MLA13 cell death in homologous barley.For this, we co-transfected protoplasts of the barley cultivar Golden Promise with the LUC reporter gene, Mla1, Mla7, or Mla13, and either EV or the matching AVR a variant.In addition, EV or plasmids encoding AVR a13 -V1 or AVR a13 -V2 genes were co-expressed.In comparison with the protoplasts transfected with Mla variants and EV plasmids, which served as control samples (relative luciferase activity=1), we detected strongly reduced LUC activity in the presence of the matching AVR A variants (Fig. 2D).Co-transfection with genes encoding AVR a13 -V1 instead of EV did not affect relative LUC activity.However, when EV was exchanged with AVR a13 -V2, the reduction of LUC activity induced by co-expression of AVR a13 -1 and Mla13 was abolished but AVR a13 -V2 expression had no significant effect on the cell death induced by samples expressing Mla1 and AVR a1 or Mla7 and AVR a7 -2 (Fig. 2D).Together, or data suggest that AVR A13-V2 has a dominantnegative effect on cell death activity specifically mediated by MLA13 and that this is accompanied by enhanced interaction of the proteins.

Amino acid exchanges in the nucleotide-binding site of MLA13 compromise AVR A13 effector binding
Previous reports on flax TNL L6 suggest an equilibrium between inactive and active NLR conformations in the absence of pathogen effectors, but that binding of the matching effector stabilizes the active NLR conformation (Bernoux et al., 2016).We therefore hypothesized that avirulent AVR A13 -1 stabilizes the active ATP-bound oligomeric conformation of MLA13.Given that AVR A13 -V2 can inhibit MLA13-mediated cell death in co-expression assays (Fig. 2), we hypothesized that AVR A13 -V2 binds and stabilizes the inactive MLA13 receptor.To test this hypothesis, we applied the aforementioned Y2H approach to examine the interaction between naturally occurring AVR A13 variants and MLA13 variants carrying mutations in the NB domain that render the MLA receptor inactive (P-loop mutants that cannot bind ADP or ATP at the NB domain) or autoactive (MHD mutant mimicking ATP binding at the NB domain) (Bai et al., 2012).In the Y2H assay, yeast expressing BD-MLA13 together with AD-AVR A13 -1 or AD-AVR A13 -V2, but not AD-AVR A13 -V1, grew as expected.None of the yeast samples co-expressing BD-MLA13 D502V or BD-MLA13 K207R together with any AVR A13 variants grew in the absence of leucine, although all proteins were stably detectable (Fig. 3A, B).We also wondered if similar results can be observed for other cereal CNLs, and therefore determined the interaction of autoactive Sr50 with AvrSr50.MLA13 and Sr50 are homologous genes and share 78% amino acid identity, whereas the effector genes are not related.Indeed, we observed similar results for the Mla homologue Sr50, although we detected growth of yeast expressing AD-AvrSr50 with the MHD variant Sr50 D498V fused N-terminally to the B42 BD.However, this interaction was consistently weaker when compared with samples co-transformed with wild-type Sr50.When AD-AvrSr50 was replaced by AD-AvrSr50 QCMJC , a variant lacking avirulence activity, no interaction was detected (Supplementary Fig. S3A, B).
AVR A13 -V2 binds specifically and strongly to wild-type MLA13 and can inhibit MLA13-specified cell death, suggesting a direct link between effector binding and cell death inhibition for this association.However, AVR A13 -V2 cannot bind autoactive MLA13 D502V in the Y2H assay (Fig. 3A) and we therefore speculate that it cannot inhibit MLA13 D502V -mediated cell death.Indeed, co-overexpression of AVR a13 -V2 or AVR A13 -V1 had no effect on MLA13 D502V -induced cell death observed as early as 2 d post-infiltration of the respective constructs in N. benthamiana leaves (Fig. 3C).Four to five days after infiltration of leaves with Agrobacteria carrying 35S:Mla13 at OD 600 =1, we also detected effector-independent cell death mediated by wild-type MLA13 (MLA13 autoactivity).This average cell death score of 2 was significantly impaired in samples co-overexpressing AVR a13 -V2 (average cell death score=0.5)but not AVR a13 -V1.Co-expression of AVR a13 -V2 had no effect on the protein levels of any of the MLA13 variants used (Fig. 3D).Of note, cell death mediated by overexpression of the MLA13 CC domain (MLA13 CC , amino acids 1-160) was not affected by AVR A13 -V2 (Supplementary Fig. S3C, D).

Different affinities between MLA13 mutant variants and AVR A13 effectors
The lack of AVR A13 interaction with both inactive and active CNL MLA13 mutant variants was unexpected, as it contrasts with previous reports on flax TNL L6 and its matching effector AvrL567 (Bernoux et al., 2016).We therefore investigated whether this lack of effector-receptor association could be generalized to other putatively inactive or autoactive MLA13 variants (Fig. 4A).We chose the MHD mutant variant H501G, whose autoactivity in MLA10 appears to be less pronounced than that of D502V (Bai et al., 2012).Receptor autoactivity was also previously reported for MLA10 F99E (mutation in the CC domain) (Bai et al., 2012).We also ssincluded the D284A mutant (mutation in the Walker A motif of the NB site, Fig. 4A) because the corresponding variant in the A. thaliana CNL RPM1 (RESISTANCE TO P. SYRINGAE PV MACULICOLA 1) leads to RPM1 autoactivity (Gao et al., 2011).By substituting negatively charged residues in the first α-helix of MLA13 with alanine (MLA13 D2A_E17A ), we aimed to generate an MLA13 resistosome that is structurally intact but impaired in immune signalling via Ca 2+ influx (Wang et al., 2019a, b;Bi et al., 2021).This hypothesis is based on the observation that the replacement of negatively charged amino acids in the ZAR1 α1-helix abrogates Ca 2+ influx and impairs cell death activity and ZAR1 disease resistance, but not formation and membrane association of the ZAR1 resistosome (Wang et al., 2019a, b;Bi et al., 2021).The S902F_F935I substitutions affect residues in the 14th and 15th LRRs of MLA13 (Fig. 4A), and the corresponding receptor is not expected to detect AVR A13 -1 as it is encoded by the barley line SxGP DH-47 (cross of cultivars SusPtrit and Golden Promise), which is fully susceptible to Bgh isolates carrying avirulent AVR a13 (Bettgenhaeuser et al., 2021).We first tested our assumption that the MLA13 mutants exhibit altered cell death activities (inactive/autoactive).We expressed the corresponding gene constructs in N. benthamiana leaves and determined cell death in the presence and absence of AVR A13 -1.As reported for other MLA variants (Bai et al., 2012), MLA13 H501G and MLA13 F99E showed effector-independent cell death activity in this assay.MLA13 D284A and SusPtritis MLA13 S902F_F935I receptor variants are unable to trigger host cell death when expressed together with AVR a13 -1.In turn, expression of MLA13 D2A_E17A , which is thought to be impaired in Ca 2+ and cell death signalling (Bi et al., 2021), resulted in effector-independent cell death in N. benthamiana leaves within 2 d post-infiltration (Fig. 4B).All MLA13 variants are detectable as fusion proteins (Fig. 4C).
We next determined the ability of AVR A13 -V2 to bind the aforementioned MLA13 variants in a Y2H assay.Again, MLA13 D502V and MLA13 K207R variants served as negative controls.Yeast samples expressing AD_AVR a13 -V2 together with wild-type BD-Mla13 grew to a dilution of OD 600 =0.001 quantitatively less when wild-type MLA13 was replaced with MLA13 D2A_E17A or MLA13 F99E .Samples transformed with AD_AVR a13 -V2 and MLA13 D284A , MLA13 K207R , or MLA13 S902F_F935I showed no growth in the absence of leucine (Fig. 4D) although these MLA13 variants are stably expressed in yeast (Fig. 4E).The MLA F 99 residue is not conserved in other CNLs and, therefore, the currently available CNL resistosome structures of ZAR1 and Sr35 cannot give functional insight into the role of this residue.However, the ZAR1 resistosome structures postulate that upon ligand binding, the release of the α1-helix in CNLs is an important conformational change that occurs immediately before resistosome formation (Wang et al., 2019a, b).We thus speculate that the autoactivity of MLA13 D2A_E17A is a result of mutation-induced α1-helix release.If this is the case, then this autoactivity cannot be inhibited by the dominant-negative AVR A13 -V2 ligand.Co-expression of AVR a13 -V2-mYFP with MLA13 D2A_E17A in N. benthamiana leaves indeed had no impact on the average cell death score, whereas autoactivity of wild-type MLA13 was again inhibited by co-expression of AVR a13 -V2-mYFP (Fig. 4F).

Activity of cation channels is required for MLA13 cell death
In ZAR1, the negatively charged residues on the inner lining of the ZAR1 resistosome funnel are required for Ca 2+ channel activity, and substitutions of these amino acids impaired ZAR1 signalling (Wang et al., 2019b;Bi et al., 2021).In contrast, such substitutions in Sr35 had no effect on cell death or channel activity (Förderer et al., 2022a), and the same appears to be true for MLA13 D2A_E17A (Fig. 4B).The data suggest that MLA13 does not require the negatively charged amino acids of the α1-helix in the CC domain for cell death signalling.We thus aimed to determine whether Ca 2+ channel activity is needed for MLA13-mediated cell death in barley by applying the potent cation channel inhibitor LaCl 3 .Toward this end, we expressed a LUC reporter together with AVR a13 -1 in barley mesophyll protoplasts, prepared from the Mla13-containing near-isogenic backcross line Manchuria (CI 16155), and measured LUC activity as an indicator of protoplast viability.Protoplasts from the cultivar Manchuria (CI 2330), which lack Mla13, served as control.With increasing LaCl 3 concentration, we observed a reduction in LUC activity by up to Growth of transformants was determined on selective growth medium containing raffinose and galactose as carbon sources but lacking uracil, histidine, and tryptophan (-UHW), and interaction of proteins was determined by leucine reporter activity reflected by growth of yeast on selective medium containing raffinose and galactose as carbon sources but lacking uracil, histidine, tryptophan, and leucine (-UHWL).Figures shown are representatives of at least three experiments, and pictures were taken 6-8 d after drop-out.(E) Protein levels of BD-MLA13 variants and AD-AVR A13 -V2 corresponding to yeast of (D).Yeast transformants were grown in raffinose-and galactose-containing selective medium lacking uracil, tryptophan, and histidine to OD 600 =1.Then, cells were harvested, total protein extracted, separated by gel electrophoresis, and western blots were probed with anti-LexA or anti-HA as indicated.CBB: Coomassie brilliant blue.(F) N. benthamiana leaves were co-transformed transiently with cDNAs of AVR a13 -V1, AVR a13 -V2, or empty vector (EV) together with constructs encoding the MLA13 variant as indicated and under the control of the 35S promoter sequence at a 2:1 ratio.Cell death was determined based on the cell death scale indicated.All values obtained in at least two independent experiments are indicated by dots, error bars=SD.Differences between samples were assessed by non-parametric Kruskal-Wallis and subsequent Dunn's tests for each MLA variant.Calculated P-values were as follows: MLA13, P=9.38E-07; MLA13 D2A_E17A , P=0.77.n.s.=no significant difference.
50% of CI 2330 protoplasts (20 µM LaCl 3 ), suggesting a detrimental impact of LaCl 3 treatment on protoplast viability independent of Mla13 or a reduction in LUC activity independent of cell death.Nonetheless, in the absence of LaCl 3 , LUC activity is on average >70% lower in Mla13 protoplasts transfected with the AVR a13 -1 construct than in protoplasts that do not express Mla13 (Fig. 5A).This difference in LUC activity between the two samples diminishes with increasing LaCl 3 concentration and is no longer significant in samples treated with 10 µM LaCl 3 .Although LUC activity decreases with increasing LaCl 3 concentrations, LaCl 3 treatment does not affect AVR A13 -1 protein stability in protoplasts of the cultivar Manchuria (Fig. 5B).Although we cannot exclude that LaCl 3 treatment affects Mla13 expression in barley line CI 16155, our data show that blocking the function of cation channels by LaCl 3 compromises MLA13-mediated cell death in barley leaf protoplasts.

Discussion
Functional studies of effector recognition by NLRs are important not only for a better understanding of plant disease resistance but also for dissecting the mechanisms pathogens employ to overcome NLR-mediated resistance.To address both aspects, we studied MLA13-mediated recognition of the barley powdery mildew AVRa A13 effector family with a particular focus on AVR A13 -V2, which originated from a Bgh isolate that has overcome Mla13 resistance.We demonstrate that AVR A13 -V2 can act as a dominant-negative effector on MLA13-mediated cell death.The concept of effector proteins suppressing the recognition of another effector by NLRs or other classes of resistance proteins has been described previously for multiple independent interactions.For example, the Leptosphaeria maculans effector AvrLm4-7 masks AvrLm3 recognition by the Arabidopsis TIR-containing protein RLM3 (Plissonneau et al., 2016), and the wheat powdery mildewencoded suppressor of avirulence SvrPm3a1/f1 gene negatively acts on wheat Pm3, which encode CNLs that recognize AvrPm3 variants from the wheat powdery mildew pathogen (Bourras et al., 2015).Similarly, Phytophthora infestans effector IPI-O1 (Avrblb1) elicits Rpi-blb1 resistance in wild potato, while the effector variant IPI-O4, that can also bind Rpi-blb1, functions to suppress this resistance elicitation (Chen et al., 2012).Our data demonstrate that the inhibitory function of AVR A13 -V2 on MLA13-mediated cell death is linked to enhanced association between AVR A13 -V2 and MLA13, and this in turn can prevent the detection of AVR A13 -1 by MLA13.

Mutations in the NB site of MLA13 abrogate association with its matching effector
The residues of the MLA LRR domains, which are under positive selection, may serve as effector contact residues (Seeholzer et al., 2010;Maekawa et al., 2019).Residues S 902 and P 935 in the 14th and 15th LRRs of MLA13 are exchanged for other amino acids in MLA13 encoded by a cultivar that has lost Mla13 resistance function (Bettgenhaeuser et al., 2021), and we showed here that these amino acid exchanges abrogate effector binding and activation of MLA13 (Fig. 4).Importantly, however, our data show that an intact, ADP-bound MLA13 receptor conformation is required for efficient effector-receptor association in yeast.Disruption of this intact conformation by mutations in the NB site of MLA13, which result in the so-called 'MHD' (mimicking ATP binding) and 'P-loop' (no binding of ADP/ATP) receptor versions (Supplementary Fig. S4) fully abrogated interaction with the matching AVR A13 effector variants in Y2H assay, probably because of spatial hindrance.One possible explanation for this hindrance is that residues of the MLA13 NB domain are engaged in the formation of an effectoraccessible conformation of the MLA LRR domain; that is, a site of effector entry (Förderer et al., 2022b) only provided by ADP-bound MLA13 (Supplementary Fig. S4).At this effector entry site of ADP-bound MLA13, the MLA13 NB domain may transiently contact the AVR A13 ligand, and this contact may be required for the steric clash that dislocates the NB domain for ADP to ATP exchange.In fact, one intermediate state structure of the ADP-bound ZAR1 monomer bound to the activating PBL2 ligand (PDB 6j5v) implies contact between the ZAR1 NB domain and the PBL2 ligand ultimately before the steric clash that allows effector-mediated ZAR1 resistosome formation, although association between these contact-forming residues cannot be detected in the active, ATP-bound ZAR1 resistosome (Wang et al., 2019b).An alternative hypothesis of our findings is a transient association between AVR A13 and MLA13, implying that conformational changes of MLA13 to the active oligomeric ATP-bound state lead to dislodging of AVR A13 effectors from the resistosome complex.However, this model is in contrast to the observation of all active NLR resistosome structures available to date, where each NLR monomer stably binds one activating ligand.The autoactive wheat CNL Sr50 MHD mutant was also impaired in AvrSr50 association when compared with wild-type Sr50 (Supplementary Fig. S3), but our data contrast with the example of enhanced association between the flax TNL L6 MHD version and its matching effector (Bernoux et al., 2016).Also, a disrupted P-loop does not hinder the CNL Rpi-amr3 binding to the matching Phytophthora effector in co-immunoprecipitation assays (Ahn et al., 2023).We therefore suggest different requirements for NB domains at the site of effector entry for individual NLRs.However, we cannot entirely exclude that this difference may be due to the initiation of yeast cell death upon expression of CNL MHD , whereas TNL MHD variants cannot induce cell death in yeast.However, the MLA13 MHD and Sr50 MHD protein levels are as stable as those of wild-type receptors, and yeast growth in the presence of leucine is similar between yeasts expressing the wild type and the MHD variants (Fig. 3B; Supplementary Fig. S3D).
Blocking TNL ROQ1-mediated cell death signalling in eds1 knockout lines in N. benthamiana was important for purification of the tetrameric ROQ1-effector resistosome (Martin et al., 2020).We and others have previously attempted to detect interaction between CNLs and their matching effector in planta by using NLR P-loop mutants to prevent NLR-mediated cell death.Our data here showing that MLA13 P-loop variants have lost the ability to bind matching effectors explains why these attempts were unsuccessful.
Amino acid exchanges in the MLA13 α1-helix deregulate autoinhibition but not Ca 2+ -dependent MLA13 cell death function Negatively charged residues in the α1-helix of NLR CC domains are thought to be required for Ca 2+ channel activity of CNL resistosomes (Förderer et al., 2022b).This was inferred from the observation that replacement of these residues with alanine abrogated ZAR1 Ca 2+ channel activity and ZAR1mediated resistance.We observed that the negatively charged residues MLA13 D2 and MLA13 E17 in the α1-helix are not required for MLA13-mediated cell death and that these amino acid exchanges instead lead to effector-independent cell death in N. benthamiana (Fig. 4).We speculate that in the absence of a matching effector, these negatively charged amino acids in MLA13 are required for burying the α1-helix and that this autorepression malfunctions in MLA13 D2A_E17A (i.e. the α1-helix is exposed and available for oligomerization; Supplementary Fig. S4).However, our data cannot clarify whether the hypothetical autoactive α1-helix conformation of MLA13 D2A_E17A allows the exchange of ADP to ATP or whether an ADP-bound NB domain is even capable of forming a functional oligomer (Supplementary Fig. S4).Notably, the MLA residues L 15 and L 19 , which are predicted to be essential for MLA membrane association by analogy with the ZAR1 resistosome, were previously shown to abrogate cell death activity (Bai et al., 2012), and the same was demonstrated for Sr35 (Förderer et al., 2022a).
The cell death autoactivity of MLA13 D2A_E17A contrasts with similar ZAR1 mutants, which abolish cell death, but the data are comparable with results reported for other CNLs, including wheat Sr35 (Adachi et al., 2019;Förderer et al., 2022a).Despite these differences, we demonstrate that MLA13dependent and AVR A13 -triggered cell death activity in barley protoplasts is impaired in the presence of the cation channel inhibitor LaCl 3 (Fig. 5), suggesting that cation transport across plant cell membranes by a putative MLA13 channel and/or other cation channels is also an important biochemical activity of the deduced MLA13 resistosome.Although the exact mechanism for cation transport in the putative MLA13 resistosome remains to be determined, our data align with reports on other CNLs that confer calcium channel-dependent cell death (Grant et al., 2000;Förderer et al., 2022a), and underline that perturbation of Ca 2+ homeostasis is a fundamental component of both TNL-and CNL-mediated cell death in plants (Jubic et al., 2019;Jacob et al., 2021;Saur et al., 2021;Förderer et al., 2022a).

A single effector residue can disrupt NLR activation
As LRR domains have the potential to bind a variety of proteinaceous ligands, engineering the LRR domains of NLRs to bind pathogen effectors that are not recognized by the natural immune system appears to be an attractive strategy for controlling plant diseases.Our data demonstrate that ligand binding per se is not sufficient for NLR activation and that the exchange of a single, potentially surface-exposed residue (S119L exchange between AVR A13 ∆SPE and AVR A13 TCML ) can abrogate NLR activation in planta despite enhanced interaction of MLA13 and L 119 -containing AVR A13 TCML in Y2H assays (Fig. 1).L 119 may mediate direct contact with MLA13 or change the conformation of the AVR A13 for enhanced interaction with MLA13.The dominant-acting interaction may directly allow AVR A13 -V2 to outcompete all AVR A13 -1 effectors for association with MLA13 and subsequent receptor activation.Alternatively, AVR A13 -V2 sequestration of some MLA13 monomers might be sufficient to disrupt putative MLA13 resistosome formation if a threshold of ligandactivated CNLs must be available for CNL resistosomes to be formed (Förderer et al., 2022b).The possibility that AVR A13 -V2 sequesters AVR A13 -1 from activation of MLA13 appears less likely because AVR A13 -V2 can also inhibit MLA13 autoactivity (Fig. 2).The contact residues responsible for the activation of MLA13 by AVR A13 are likely to be unique, despite the overall structural similarity of AVR A effectors and allelic, highly sequence-similar MLA receptors (Seeholzer et al., 2010;Bauer et al., 2021).This appears to be also true for the residues of AVR A13 -V2 that mediate MLA13 interaction, as neither the enhanced interaction, nor the dominant-negative effect of AVR A13 -V2 was detected when MLA13 was replaced by the highly sequence-similar MLA1 or MLA7 NLRs.The overall high sequence and predicted structural identity between AVR A13 -1 and AVR A13 -V2, as well as the identification of a single residue, L 119 of AVR A13 -V2, as the main driver of enhanced MLA13 interaction, suggest that the binding surfaces to the MLA13 receptor overlap.However, our data imply that AVR A13 -V2 locks MLA13 into an inactive, effector-bound state by preventing the receptor from transitioning to one of the conformational changes downstream of effector binding (Supplementary Fig. S5).AVR A13 -V2 cannot inhibit cell death signalling of MLA13 constitutive gain-of-function mutants with amino acid replacements in the CC domain despite interaction with MLA13 D2A_E17A (Fig. 4D).We therefore suggest that the inhibitory function of AVR A13 -V2, mediated by L 119 , affects conformational changes that take place before the release of the MLA13 α1-helix; that is, AVR A13 -V2 binding to MLA13 either fails to induce an interdomain steric clash in the receptor or blocks the transition to the steric clash-mediated open conformation, which allows exchange of ADP to ATP in the NB site of MLA13 (Supplementary Fig. S5).Alternatively, AVR A13 -V2 binding to MLA13 induces a steric clash, but AVR A13 -V2 association inhibits the release of the α1-helix from autorepression.As MLA13 MHD mutants are generally inaccessible to effector binding in Y2H assay (including binding to avirulent AVR A13 -1, Fig. 3), our data cannot clarify whether the loss of inhibitory function of AVR A13 -V2 on MLA13 cell death takes place before or after ADP exchange to ATP in wild-type MLA13.Collectively, we demonstrate that the stable interaction between AVR A13 -V2 and inactive MLA13 has the potential to define distinct conformations of intermediate states of CNL receptors.This knowledge is currently largely elusive for both animal and plant NLRs.Understanding such conformations will help ensure that future synthetic NLRs do not become locked into intermediate non-functional states.

Role of AVR A13 -V2 in the breakdown of Mla13 resistance in the European Bgh population
Evasion of NLR-mediated pathogen recognition is usually mediated by diversification of the pathogen's effector repertoire, including allelic variation of effector genes that results in abrogation of effector-NLR receptor associations.This model applies to the virulent variant AVR A13 -V1.However, AVR A13 -V2 not only interacts strongly with MLA13, but also inhibits MLA13 cell death signalling in a dominant manner (Fig. 2).This raises the possibility that Bgh AVR A13 -V2 facilitates dispersal of virulence in Bgh populations that are genetically avirulent on Mla13.In the European Bgh population, the virulence frequency on Mla13 increased from 0.2% in the 1980s to as high as 60% in 1995 (Gacek, 1987;Jørgensen and Hovmøller, 1987;Hovmøller et al., 2000), suggesting a major shift in genetic variation of AVRa 13 on a continental scale.In contrast, only 7% of Bgh isolates in a global strain collection carry virulent AVR A13 variants (Rsaliyev et al., 2017;Saur et al., 2019a).In addition, AVR a13 /BGH_20990 has a very low frequency of non-synonymous polymorphisms in tested Bgh populations (0.9 non-synonymous single nucleotide polymorphisms/100 bp coding sequence), indicating an overall low genetic diversity of AVR a13 (Saur et al., 2019a).Our data demonstrate a dominant-negative activity of AVR A13 -V2 on MLA13, therefore suggesting that the breakdown of Mla13 resistance was caused by direct manipulation of the receptor activation mechanism rather than by evasion of MLA13 recognition.

Supplementary data
The following supplementary data are available at JXB online.

Fig. 1 .
Fig. 1.The C-terminus of AVR A13 effectors controls interaction with and activation of MLA13.(A) Amino acid alignment of AVR A13 variants analysed for interaction with MLA13 and activation of MLA13-mediated cell death.Signal peptide (SP) residues are underlined; amino acids in blue and pink highlight the amino acid variation between AVR A13 -V2 and AVR A13 -1, respectively.(B and C) Nicotiana benthamiana leaves were transformed transiently with 35S:Mla13-4Myc (pGWB517) with one of the AVR a13 variants lacking SPs cloned between the 35S promoter and a C-terminal mYFP sequence or empty vector (EV).(B) Cell death was determined 3 d post-transformation, and figures shown are representatives of at least nine independent leaves from at least three independent plants.(C) Protein stability of the AVR A13 variants fused to mYFP corresponding to constructs of (B).Leaf tissue was harvested 2 d post-infiltration.Total protein was extracted, separated by gel electrophoresis, and probed by anti-GFP.(D and E) Yeast cells were co-transformed with Mla13 fused N-terminally to the LexA-binding domain (BD) sequence and AVR a13 variants lacking SPs fused N-terminally to the B42 activation domain (AD) and 1×HA tag sequence as indicated.Growth of transformants was determined on selective growth medium containing raffinose and galactose as carbon sources but lacking uracil, histidine, and tryptophan (-UHW), and interaction of proteins was determined by leucine reporter activity reflected by growth of yeast on selective medium containing raffinose and galactose as carbon sources but lacking uracil, histidine, tryptophan, and leucine (-UHWL).Figures shown are representatives of at least three experiments, and pictures were taken 6-8 d after drop-out.(E) Protein levels of BD-MLA13 and AD-AVR A variants corresponding to yeast of (D).Yeast transformants were grown in raffinose-and galactose-containing selective medium lacking uracil, tryptophan, and histidine to OD 600 =1.Then, cells were harvested, total protein extracted, separated by gel electrophoresis, and western blots were probed with anti-LexA or anti-HA as indicated.CBB: Coomassie brilliant blue.

Fig. 2 .
Fig. 2. AVR A13 -V2 can act as dominant-negative effector on MLA13.Nicotiana benthamiana leaves were co-transformed transiently with cDNAs of Mla1 or Mla7 or MLA13 (pGWB vectors) with AVR a1 , AVR a7 -2, AVR a13 -1, or empty vector (EV) as indicated and either AVR a13 -V1, AVR A13 -V2, or EV fused to epitope tags as indicated.All constructs were expressed from the 35S promoter.(A and B) Cell death was determined 3-4 d post-transformation and (B) scored from 0 to 3 based on the cell death scale indicated.All values obtained in at least three independent experiments are indicated by dots; error bars=SE.Differences between samples were assessed by non-parametric Kruskal-Wallis and subsequent Dunn's tests for each MLA variant.Calculated P-values were as follows: Mla1, P=0.824; Mla7, P=0.551; and Mla13, P=1.00E-06.Samples marked by identical letters in the plots do not differ significantly (P<0.05) in the Tukey test for the corresponding MLA.(C) Protein levels corresponding to samples of (B).Leaf tissue was harvested 2 d postinfiltration.Total protein was extracted and recovered by GFP-Trap (AVR a1 and AVR a7 -2) separated by gel electrophoresis, and probed by anti-HA (MLAs), anti-Myc (AVR A13 -V2-4×Myc), or anti-GFP (AVR A1 -mYFP, AVR A7 -2-mYFP, and AVR A13 -1-mYFP) as indicated.CBB: Coomassie brilliant blue.(D) Barley protoplasts were transfected with pUBQ:luciferase (4.5 µg) and genes encoding Mla1, Mla7, or Mla13 and either an EV (reference sample) or AVR a1 , AVR a7 -2, or AVR a13 -1 lacking their respective signal peptides (SPs), respectively.Additionally, an EV or AVR a13 -V1 or AVR a13 -V2 lacking their respective SPs was co-expressed.The piPKb002 vector was used for all Mla and AVR a constructs and, for each transfection, 9 µg of Mla-containing vector and 4.5 µg of each AVR a -containing vector or EV were transfected.Luciferase activity was measured at 16 h post-transfection, and relative luciferase activity determined by setting the reference samples (Mla+EV) to 1. Differences between samples were assessed by non-parametric Kruskal-Wallis and subsequent Dunn tests for each MLA variant.Calculated P-values were as follows: Mla1: P=0.412; Mla7, P=0.683; and Mla13, P=1.9E-04.Samples marked by identical letters in the plots do not differ significantly (P<0.05) in the Dunn test for the corresponding MLA.n.s=not significant.

Fig. 3 .
Fig. 3. Amino acid exchanges in the nucleotide-binding site of MLA13 compromise AVR A13 effector binding.(A, B) Yeast cells were co-transformed with Mla13 wild type (wt) or mutant variants Mla13 D502V (MHD) or Mla13 K207R (P-loop) fused N-terminally to the LexA-binding domain (BD) sequence and AVR a13 variants lacking SPs fused N-terminally to the B42 activation domain (AD) and 1×HA tag sequence as indicated.(A) Growth of transformants was determined on selective growth medium containing raffinose and galactose as carbon sources but lacking uracil, histidine, and tryptophan (-UHW), and interaction of proteins was determined by leucine reporter activity reflected by growth of yeast on selective medium containing raffinose and galactose as carbon sources, but lacking uracil, histidine, tryptophan, and leucine (-UHWL).Figures shown are representatives of at least three experiments, and pictures were taken 6-8 d after drop-out.(B) Protein levels of BD-MLA13 variants and AD-AVR A variants corresponding to yeast of (A).Yeast transformants were grown in raffinose-and galactose-containing selective medium lacking uracil, tryptophan, and histidine to OD 600 =1.Cells were harvested, total protein extracted, separated by gel electrophoresis, and western blots were probed with anti-LexA or anti-HA as indicated.(C and D) Nicotiana benthamiana leaves were co-transformed transiently with cDNAs of AVR a13 -V1, AVR a13 -V2, or empty vector (EV) together with constructs encoding either MLA13 or MLA13 D502V (pAM-PAT vector) as indicated and under the control of the 35S promoter sequence at a 2:1 ratio.(C) Cell death was determined 2 d (MLA13 MHD) to 5 d (MLA13) post-transformation and scored from 0 to 3 based on the cell death scale indicated.All values obtained in at least three independent experiments are indicated by dots;| error bars=SD.Differences between samples were assessed by non-parametric Kruskal-Wallis and subsequent Dunn's tests for each MLA variant.Calculated P-values were as follows: MLA13, P=5E-05; MLA13 MHD, P=0.078.Samples marked by identical letters in the plots did not differ significantly (P<0.05) in the Dunn test for the corresponding MLA.(D) Protein levels corresponding to samples of (C).Leaf tissue was harvested 36 h post-infiltration.Total protein was extracted, separated by gel electrophoresis, and probed by anti-Myc (MLAs) or anti-GFP (AVR A13 -V2) western blotting as indicated.CBB: Coomassie brilliant blue.

Fig. 4 .
Fig. 4. Amino acid exchanges in the coiled-coil (CC) domain de-regulate MLA13 autoinhibition.(A) Amino acid changes in MLA13 mutant variants.The D2A_E17A and the F99E variants encode changes in the MLA13 CC domain, which spans from amino acid 1 to 160.The K207R, D284A, D502V, and H501G variants encode changes in the nucleotide-binding site (NB, amino acids 161-549).The S902F_F935I variant affects the leucine-rich repeats (LRRs, amino acids 550-942) which are followed by a short C-terminal amino acid sequence.(B and C) Nicotiana benthamiana leaves were transformed transiently with cDNAs of one of the Mla13 variants as indicated (pGWB517 vector) either with or without AVR a13 -1 lacking SPs and fused C-terminally to an mYFP sequence.All constructs are under the control of the 35S promotor.(B) Cell death was determined 3 d post-transformation; n≥9.(C) Protein stability of the MLA variants fused to 4×Myc corresponding to constructs of (B).Leaf tissue was harvested 2 d post-infiltration.Total protein was extracted, separated by gel electrophoresis, and probed by anti-Myc western blotting as indicated.(D and E) Yeast cells were co-transformed with Mla13 variants fused N-terminally to the LexA-binding domain (BD) sequence and AVR a13 -V2 lacking SPs fused N-terminally to the B42 activation domain (AD) and 1×HA tag sequence as indicated.Growth of transformants was determined on selective growth medium containing raffinose and galactose as carbon sources but lacking uracil, histidine, and tryptophan (-UHW), and interaction of proteins was determined by leucine reporter activity reflected by growth of yeast on selective medium containing raffinose and galactose as carbon sources but lacking uracil, histidine, tryptophan, and leucine (-UHWL).Figures shown are representatives of at least three experiments, and pictures were taken 6-8 d after drop-out.(E) Protein levels of BD-MLA13 variants and AD-AVR A13 -V2 corresponding to yeast of (D).Yeast transformants were grown in raffinose-and galactose-containing selective medium lacking uracil, tryptophan, and histidine to OD 600 =1.Then, cells were harvested, total protein extracted, separated by gel electrophoresis, and western blots were probed with anti-LexA or anti-HA as indicated.CBB: Coomassie brilliant blue.(F) N. benthamiana leaves were co-transformed transiently with cDNAs of AVR a13 -V1, AVR a13 -V2, or empty vector (EV) together with constructs encoding the MLA13 variant as indicated and under the control of the 35S promoter sequence at a 2:1 ratio.Cell death was determined based on the cell death scale indicated.All values obtained in at least two independent experiments are indicated by dots, error bars=SD.Differences between samples were assessed by non-parametric Kruskal-Wallis and subsequent Dunn's tests for each MLA variant.Calculated P-values were as follows: MLA13, P=9.38E-07; MLA13 D2A_E17A , P=0.77.n.s.=no significant difference.

Fig. 5 .
Fig. 5. Calcium channel activity is required for Mla13-mediated cell death in barley.(A) Barley protoplasts of lines CI 16155 (cultivar Manchuria Mla13) and CI2330 (Manchuria) were transfected with pUBQ:luciferase (6 µg) and piPKb002 containing AVR a13 -1 cDNA without signal peptide (5 µg) or a piPKb002 empty vector control (5 µg) and recovered in the presence of LaCl 3 at the concentrations indicated.Luciferase activity was determined 16 h post-transfection/addition of LaCl 3 as a proxy for cell death and normalized against the respective EV sample.Error bars=SE.Differences between samples were assessed using non-parametric Kruskal-Wallis and subsequent Dunn's post-hoc tests.P=6.179e-10.Samples marked by identical letters in the plot did not differ significantly (P<0.05) in Dunn's test.(B) Protoplasts derived from cultivar Manchuria CI2330 leaves transfected with pZmUBQ:AVR a13 -1-mYFP were harvested 16 h post-transfection/LaCl 3 treatment.Total protein was extracted, separated by gel electrophoresis, and western blots were probed with anti-GFP.CBB: Coomassie brilliant blue.
Fig. S1.Proximity-dependent protein labelling confirms the requirement of AVR A13 C-terminus for MLA13 interaction.Fig. S2.Specificity control to Fig. 1D and structural prediction models Fig. S3.Gain-of-function NLR mutants and their ability to bind matching avirulence effectors.