Barley guanine nucleotide exchange factor HvGEF14 is an activator of the susceptibility factor HvRACB and supports host cell entry by Blumeria graminis f. sp. hordei.

In barley (Hordeum vulgare), signalling rat sarcoma homolog (RHO) of plants guanosine triphosphate hydrolases (ROP GTPases) support the penetration success of Blumeria graminis f. sp. hordei but little is known about ROP activation. Guanine nucleotide exchange factors (GEFs) facilitate the exchange of ROP-bound GDP for GTP and thereby turn ROPs into a signalling-activated ROP-GTP state. Plants possess a unique class of GEFs harbouring a plant-specific ROP nucleotide exchanger domain (PRONE). Here, we performed phylogenetic analyses and annotated barley PRONE-GEFs. The leaf epidermal-expressed PRONE-GEF HvGEF14 undergoes a transcriptional down-regulation on inoculation with B. graminis f. sp. hordei and directly interacts with the ROP GTPase and susceptibility factor HvRACB in yeast and in planta. Overexpression of activated HvRACB or of HvGEF14 led to the recruitment of ROP downstream interactor HvRIC171 to the cell periphery. HvGEF14 further supported direct interaction of HvRACB with a HvRACB-GTP-binding CRIB (Cdc42/Rac Interactive Binding motif) domain-containing HvRIC171 truncation. Finally, the overexpression of HvGEF14 caused enhanced susceptibility to fungal entry, while HvGEF14 RNAi provoked a trend to more penetration resistance. HvGEF14 might therefore play a role in the activation of HvRACB in barley epidermal cells during fungal penetration.

INTRODUCTION

ROPs (rat sarcoma homolog [RHO] of plants) are small monomeric GTPases that function as signalling hubs in cell polarity processes that involve cytoskeleton reorganization (Mucha et al., 2011). Pollen tube growth, the development of epidermal pavement cells and root hairs, but also processes that are important during plant–microbe interactions are examples of ROP‐regulated processes (Engelhardt et al., 2020; Zheng & Yang, 2000). ROPs are considered molecular switches due to their ability to shuttle between a signalling‐inactive, guanosine diphosphate (GDP)‐bound state and a signalling‐activated guanosine triphosphate (GTP)‐bound state (Bloch & Yalovsky, 2013). An interaction with downstream signalling partners, and therefore signal transduction, only occurs in the GTP‐bound state (Nagawa et al., 2010). To locally interact with downstream signalling partners, ROPs further require to be membrane‐associated, which is achieved by posttranslational lipid modifications and electrostatic lipid interaction (Winge et al., 2000; Yalovsky, 2015).

Barley (Hordeum vulgare [Hv]) contains six ROPs: HvRACB, HvRACD, HvRAC1, HvRAC3, HvROP4, and HvROP6 (Schultheiss et al., 2003). Studies using barley plants overexpressing constitutively activated (CA) variants of HvRACB, HvRAC1, or HvRAC3 indicated a role for these ROPs in plant development. Additionally, these plants support either enhanced or reduced susceptibility to penetration by fungal leaf pathogens such as Blumeria graminis f. sp. hordei (Bgh) and the rice blast fungus Magnaporthe oryzae (Pathuri et al., 2008; Schultheiss et al., 2005). HvRACB in particular has been studied for its function in susceptibility to penetration and accommodation of haustoria from Bgh. In addition, a function of HvRACB in polar epidermal cell development has been shown. HvRACB is hence considered a key developmental protein that is co‐opted by Bgh during pathogenesis (Engelhardt et al., 2020).

Due to the vast number of signalling processes ROPs are involved in, it is apparent that a tight regulation of these molecular switches is required to fine‐tune the cellular mechanisms following ROP activation. This regulation is controlled by three classes of regulatory proteins: guanine nucleotide exchange factors (GEF), GTPase activating proteins (GAP), and guanine nucleotide dissociation inhibitors (GDI) (Nagawa et al., 2010; Vetter & Wittinghofer, 2001; Zheng & Yang, 2000). Regarding ROP signalling, GEFs are ROP‐signalling activating proteins that interact with ROPs to induce a conformational change facilitating the exchange of GDP by GTP (Berken et al., 2005; Thomas et al., 2007, 2009).

Plants evolved a specific class of GEFs with a highly conserved plant‐specific Rac/ROP nucleotide exchanger (PRONE) domain (Berken et al., 2005; Gu et al., 2006). The family of PRONE‐GEFs in the model plant Arabidopsis thaliana consists of 14 members that have been studied for their function in various polar growth processes, plant development (Chang et al., 2013; Chen et al., 2011; Gu et al., 2006; Huang et al., 2018), and immunity (Qu et al., 2017). So far, the PRONE domain is the only identified conserved part of these GEFs. It contains several interfaces that physically interact with ROPs in a heterotetrameric complex of two ROPs with two PRONE‐GEFs. After binding to ROPs, the PRONE domain facilitates a structural rearrangement, which leads to GDP release. Nucleotide‐free ROPs and GEFs can further interact as a stable complex (Berken et al., 2005; Chang et al., 2013; Gu et al., 2006; Thomas et al., 2007). The PRONE domain is flanked by N‐ and C‐terminal variable regions that probably possess regulatory functions. The C‐terminal stretch can be subject to phosphorylation by upstream receptor‐like kinases (RLK) that have been implicated in developmental as well as immunity‐related signalling pathways (Fehér & Lajkó, 2015). One example of RLK‐GEF‐ROP signalling involves the RLK FERONIA, which interacts directly with AtGEF14 in A. thaliana. Further downstream, AtGEF14 interacts with AtROP6, which functions in the polar growth of epidermal cells (Lin et al., 2022) and together with AtRIC1 facilitates microtubule organization to deal with mechanical stress (Tang et al., 2022). Furthermore, AtGEF14 localizes to the apical region of pollen tubes and interacts with AtROP1, a regulator of pollen tube growth (Gu et al., 2006). In roots, AtGEF14 may be also involved in polar growth processes. AtGEF14 accumulates at the root hair initiation site and is replaced by other GEFs during the root hair initiation and elongation phase (Denninger et al., 2019). We know little about the function of PRONE‐GEFs in the interaction of plants with fungal pathogens. In rice, OsGEF1 interacts with the RLK OsCERK1, a central part of the rice chitin receptor complex, and activates OsRAC1 for its function in defence against M. oryzae (Akamatsu et al., 2013).

Susceptibility factors, such as the ROP GTPase HvRACB, have become increasingly recognized as potential targets for breeding but often little is known about their mode of activation and corresponding molecular environment. This work investigates barley PRONE‐GEFs, focusing on HvGEF14 as an interactor of HvRACB. We report that HvGEF14 is expressed in the leaf epidermis and downregulated after inoculation with Bgh. HvGEF14 can bind to HvRACB in yeast and in planta, and may be involved in the activation of this susceptibility‐associated barley ROP in leaf epidermal cells.

RESULTS

Phylogenetic analysis reveals three distinct clades of barley PRONE‐GEFs

To investigate GEFs in barley, we concentrated on the PRONE domain‐containing class of exchange factors (Berken et al., 2005). We identified 11 PRONE domain‐encoding genes in H. vulgare 'Morex' genome version 3 (Mascher et al., 2021) and aligned full‐length primary amino acid sequences of all barley PRONE‐GEFs with 11 PRONE‐GEFs of Oryza sativa and 14 PRONE‐GEFs of A. thaliana. Because the O. sativa PRONE‐GEF annotation has been incomplete so far, we first constructed a maximum‐likelihood phylogenetic tree in which all O. sativa PRONE‐GEFs were annotated according to their primary sequence similarity to A. thaliana PRONE‐GEFs using the nomenclature by Berken et al. (2005) (Figure S1). Subsequently, another calculation was performed to obtain the phylogenetic tree with all three species (Figure 1, based on the MUSCLE alignment in Figure S2). H. vulgare PRONE‐GEFs were then annotated according to the closely related O. sativa PRONE‐GEFs to determine a nomenclature that is consistent for grasses. In this way, the 11 barley PRONE‐GEFs were named HvGEFs 1, 3a, 3b, 3c, 7a, 7b, 9a, 9b, 9c, 10, and 14. The phylogenetic tree shows three distinct clades, with clade III containing only PRONE‐GEF14 proteins from all three species (Figure 1). Notably, in clades I and II, PRONE‐GEFs from monocot species show plant clade‐specific clustering with high confidence. This suggests lineage‐specific proliferation of PRONE‐GEF genes after the separation of monocots from dicots.

FIGURE 1

PRONE‐GEFs cluster in three distinctive clades. Phylogenetic analysis of PRONE‐GEFs of Arabidopsis thaliana (At), Oryza sativa (Os), and Hordeum vulgare (Hv) based on MUSCLE alignment. The maximum‐likelihood tree was calculated based on available amino acid sequences (NCBI and barley Morex genome version 3) and annotation of barley PRONE‐GEFs was performed based on this tree.

The three PRONE‐GEF14 sequences differ not only in their N‐ and C‐terminal regions from the other PRONE‐GEFs but present higher levels of amino acid variations in the PRONE domain itself when compared to all other PRONE‐GEFs (Figures S1 and S2, and Table S1).

Despite its unique position in the phylogenetic tree, the overall design of the HvGEF14 PRONE domain is conserved with its three subdomains (Figure 2). The alignment of the 11 barley PRONE‐GEFs highlights the predicted PRONE domains with varying length of 344 amino acids (HvGEF10) to 379 amino acids (HvGEF3a) (Figure S1 and Table S2). HvGEF3a has no variable C‐terminal region beyond the PRONE domain, and the C‐termini of HvGEF3b and HvGEF3c are comparably short (Figures 2 and S1, and Table S2). HvGEF9a is predicted to contain only a short variable N‐terminus of 43 amino acids. It is evident from the alignment that, regarding the primary sequence, HvGEF14 substantially differs most from all other barley PRONE‐GEFs. In addition, the variability in the primary sequence of HvGEF14 is highest when compared to the consensus sequence of all 11 Hv‐GEFs (Figure S1 and Table S1). However, important residues for GEF‐GEF homodimerization (Thomas et al., 2007), such as phenylalanine 133 and leucine 138 (in HvGEF14), are conserved and we found that HvGEF14 interacted with itself or its PRONE domain in yeast two‐hybrid (Y2H) assays (Figure S3). Known residues involved in GEF‐ROP interaction (N161, Q206, E215, M217, W275, W276, L434, and R460; Thomas et al., 2007) are also conserved in HvGEF14. Furthermore, serine 394 in the HvGEF14 PRONE domain is a predicted phosphorylation site based on mass spectrometry analysis of phosphorylation sites in A. thaliana GEF14 (Mergner et al., 2020) (Figure 2).

FIGURE 2

Hordeum vulgare (Hv) PRONE‐GEF MUSCLE alignment with annotation of PRONE domain (according to NCBI prediction) in grey, and GEF–GEF interaction residues and ROP–GEF interaction residues highlighted in purple. Predicted phosphorylation site in HvGEF14 highlighted in turquoise. Protein IDs according to NCBI accessions and annotation based on MUSCLE alignment with Oryza sativa PRONE‐GEFs (see Figure S1).

is expressed in epidermal cells and downregulated after Bgh inoculation

To understand the potential function of barley PRONE‐GEFs, gene expression patterns in different tissues were investigated. An initial in silico expression analysis of the eight barley PRONE‐GEFs was performed by consulting the RNA sequencing (RNA‐Seq) database provided by the James Hutton Institute (https://ics.hutton.ac.uk/barleyGenes). The data show RNA fragments per reads per kilobase million (FPKM) of the barley PRONE‐GEFs in distinct tissues. HvGEF3a, for example, is mainly expressed in the developing embryo. Other barley GEFs, such as HvGEF1, show a broader expression pattern, with the highest number of RNA fragments detected in embryos, root, and grains. HvGEF14 is the most ubiquitously expressed barley PRONE‐GEF, with the highest fragment counts in almost all tissues except in senescing leaves. Interestingly, HvGEF1 and HvGEF14 are the only GEF genes expressed in seedling shoots and epidermal peels, with HvGEF14 showing the highest fragment counts in these two tissues (Table S3). The leaf epidermis of barley provides an important interface for plant–pathogen interaction. Because the barley ROP HvRACB has been shown to play a crucial role in epidermis development and susceptibility to Bgh, we concentrated on PRONE‐GEFs, which might be of importance to signal transduction in the epidermis. We confirmed the gene expression of HvGEF14 via reverse transcription‐quantitative PCR (RT‐qPCR) in leaf and leaf epidermis from 7‐day‐old leaves of the barley cultivar Golden Promise. In three independent biological replicates, we found higher transcript levels of HvGEF14 in the epidermis when compared to whole leaves (Figure 3a). Notably, in three independent biological experiments, the HvGEF14 expression level in the epidermis decreased after inoculation with the biotrophic powdery mildew fungus Bgh compared to epidermal peels from unchallenged leaves (Figure 3b).

FIGURE 3

HvGEF14 shows increased expression in epidermal peels and downregulation after inoculation with fungal spores. HvGEF14 is expressed in barley epidermis (a) and downregulated after Bgh inoculation (1 day postinoculation) (b) in three independent repetitions (plants harvested on different days). Foldchange expression was calculated with primer efficiency correction and the 2−ΔΔ method by Livak and Schmittgen (2001) and normalized to transcript levels in whole leaves (a) and noninoculated epidermis (b).

interacts with

So far, nothing is known about potential PRONE‐GEF‐mediated activation of barley ROPs. To check if HvGEF14 could function as an HvRACB‐activating PRONE‐GEF, we analysed the direct protein–protein interaction between HvRACB and HvGEF14 in yeast and in planta. Plant ROPs can be mutagenized in the GTPase domain (e.g., HvRACB‐G15V) to render the ROP constitutively activated (CA) (Schultheiss et al., 2003). Correspondingly, the HvRACB‐T20N substitution results in a dominant negative (DN), signalling‐inactive conformation. A third mutation (HvRACB‐D121N) leads to lower nucleotide affinity and potentially increases GEF‐binding affinity (Akamatsu et al., 2013; Berken et al., 2005; Cool et al., 1999). Interestingly, we found that full‐length HvGEF14 and the HvGEF14 PRONE domain (amino acids 124–485) directly interacted in yeast with wild‐type (WT) HvRACB, CA HvRACB‐G15V, and the low nucleotide affinity version HvRACB‐D121N, but not with the DN HvRACB‐T20N variant (Figure 4a, see also Figure S4 for the full drop‐out plates). All HvRACB variants were truncated at the HvRACB CSIL motif to inhibit prenylation and membrane association in yeast and hence facilitate protein accumulation in yeast nuclei. To substantiate the results, we performed western blotting, which confirmed protein stability in yeast (Figure S5). In addition, similar results were obtained in Y2H assays using the type II ROP HvRAC1 (Figures S6 and S7).

FIGURE 4

HvGEF14 interacts with barley ROP variants in yeast and in planta. (a) HvRACB wild type (WT), G15V constitutively activated (CA), D121N low nucleotide affinity, and T20N dominant negative (DN) variants were tested as bait against prey constructs HvGEF14 full length, HvGEF14 amino acids 124–485 (PRONE domain), or empty vector (EV). Interaction of proteins shown on medium containing amino acid mix without leucine (L), tryptophan (W), and histidine (H), (−L−W−H) in two dilutions (factor 10−1) to identify growth of single yeast colonies. Representative image of three experiments with the same result. Successful yeast transformation was confirmed with selective medium (amino acid mix without leucine (L) and tryptophan (W) (−L−W). Dropout was performed on one −L−W and one −L−W−H plate and images were cropped during figure preparation for better visibility. Original images can be found in Figure S4. (b) Representative images of barley epidermis cells measured in FRET‐FLIM showing false colour representation of meGFP lifetime as indicated by the colour bar at the bottom. Images were saved from PicoQuant SymPhoTime 64 software after lifetime fitting. (c) meGFP‐tagged HvGEF14 full length interacts with mCherry‐tagged HvRACB WT, G15V, and D121N but not with T20N RACB in FRET‐FLIM experiments of barley epidermis cells. Mean GFP lifetime is indicated with orange bars and n = total number of cells observed in three independent experiments. Kruskal–Wallis p value = 4.3e−14, pairwise comparison was performed via Wilcoxon test and Bonferroni adjustment for multiple testing (Rstudio, v. 1.2.5033).

To test direct protein–protein interaction between HvGEF14 and HvRACB in planta, we measured Förster resonance energy transfer by fluorescence lifetime imaging (FRET‐FLIM) in transiently transformed barley epidermal cells (Figure 4b,c). The monomeric enhanced green fluorescent protein (eGFP) fusion meGFP‐HvGEF14 served as a FRET‐donor and was used as a potential interaction partner in all combinations tested. As FRET acceptor, different variants of mCherry‐HvRACB (WT, G15V, D121N, T20N) and cytosolic mCherry were used. All measurements took place at the cell periphery at the equatorial cell plane (Figure 4b).

meGFP‐HvGEF14 showed a significant lifetime reduction in epidermal cells when transiently co‐expressed with mCherry‐HvRACB WT to 2.36 ns on average compared to the negative control (free mCherry) recorded at 2.57 ns. This shows that HvGEF14 interacts directly with WT HvRACB in planta (Figure 4c). The transient co‐expression of meGFP‐HvGEF14 with mCherry‐CA HvRACB‐G15V also resulted in a significantly reduced GFP‐lifetime of 2.32 ns on average. This reflects the interaction assays in yeast and provides further evidence that HvGEF14 can also interact with the CA variant HvRACB‐G15V in planta (Figure 4c). In contrast to that, co‐expression of mCherry‐HvRACB‐D121N moderately decreased the meGFP‐HvGEF14 lifetime to 2.45 ns on average, which was not significantly different from the negative control (Figure 4c). As observed in Y2H assays, there was no measurable interaction between HvGEF14 and DN HvRACB‐T20N in planta (Figure 4b,c). In total, the FRET‐FLIM experiments suggest a direct interaction between HvGEF14 full length and HvRACB WT and CA HvRACB‐G15V in planta.

overexpression leads to activation of barley ROPs

ROP interacting proteins often display a change of subcellular localization in the presence of an activated ROP. This change in localization is considered evidence for local ROP activity because those interactors preferably interact with GTP‐loaded ROPs (Li et al., 2020; McCollum et al., 2020; Schultheiss et al., 2008). In transiently transformed epidermal cells, CA HvRACB‐G15V is partially located at the cell periphery, which depends on its C‐terminal CSIL prenylation motif (Schultheiss et al., 2003, 2008). HvRIC171, a barley scaffold protein that directly interacts with CA but not DN barley ROP variants (Schultheiss et al., 2008), was recruited from the cytoplasm to the cell periphery and plasma membrane in the presence of co‐expressed CA HvRACB‐G15V but not DN HvRACB‐T20N (Figure 5c). The plasma membrane recruitment of HvRIC171 is therefore considered to be HvRACB activation‐dependent (Schultheiss et al., 2008). Based on this, we monitored the localization of mCherry‐HvRIC171 in barley epidermis cells in the presence or absence of co‐expressed HvGEF14 to analyse the activation potential of HvGEF14 towards HvROPs. mCherry‐HvRIC171 fluorescence significantly increased at the cell periphery when either CA HvRACB‐G15V or HvGEF14 was present compared to the empty vector (EV) or DN HvRACB‐T20N controls (Figure 5a–c). This was evident from an increase in normalized fluorescent signal intensity at the cell periphery in the equatorial plane of the cell. Additionally, mCherry‐HvRIC171 signals appeared very irregular in the cell periphery of control cells, whereas mCherry‐HvRIC171 more evenly labelled the cell periphery in cells co‐expressing HvGEF14 (Figure 5a, lower panel).

FIGURE 5

HvRACB can be activated by HvGEF14. The downstream interactor of activated ROPs in barley, mCherry‐HvRIC171, is recruited to the cell periphery when HvGEF14 is co‐expressed. (a) Representative images of three biological replicates show z‐stack and equatorial plane of mCherry‐HvRIC171 and cytosolic GFP fluorescence of transiently transformed barley epidermis cells. (b, c) Quantification of periphery mCherry fluorescence intensity normalized to whole cell mCherry and GFP fluorescence intensity measured via Fiji. Statistical analysis performed in Rstudio via Kruskal–Wallis after testing for distribution of data (Rstudio v. 1.2.5033). (d) meGFP‐HvRACB interacts with HvCRIB46‐mCherry, a 46 amino acid fragment of HvRIC171, when mutated (G15V) to a constitutively activated variant or in the presence of HvGEF14 during FRET‐FLIM measurements in Nicotiana benthamiana. GST‐mCherry used as negative control. Summary of three independent repetitions indicated in shades of blue. Kruskal–Wallis comparison of means performed in Rstudio (v. 1.2.5033) after testing for distribution of data.

CRIB (Cdc42/Rac interactive binding motif) domains of RHO‐interacting proteins specifically bind to GTP‐loaded RHO and ROP proteins, and are therefore often used as RHO activity sensors. To test the specific activation of HvRACB in planta, we used HvCRIB46, which represents a fragment of HvRIC171 containing the CRIB domain and was shown before to interact with CA HvRACB‐G15V but not DN HvRACB‐T20N (Schultheiss et al., 2008). Based on this, we established a FRET‐based activity sensor probe containing N‐terminally meGFP‐tagged HvRACB and C‐terminally mCherry‐tagged HvCRIB46 on individual plasmids. To reduce interference of endogenous signalling components in barley, the measurements were performed in Nicotiana benthamiana. meGFP‐HvRACB WT did not interact with the negative control GST‐mCherry (average meGFP lifetime of 2.62 ns), but the meGFP‐HvRACB WT lifetime was significantly reduced when meGFP‐HvRACB WT was co‐expressed with HvCRIB46‐mCherry (2.57 ns on average). This probably reflects the ability of WT HvRACB to switch between GDP‐ and GTP‐loaded forms also in N. benthamiana. As a positive control, the interaction of meGFP‐CA HvRACB‐G15V with HvCRIB46‐mCherry was measured at 2.35 ns on average (Figure 5d). When additionally co‐expressing HA‐HvGEF14 with meGFP‐HvRACB WT and HvCRIB46‐mCherry, we measured a significant decrease in meGFP fluorescence lifetime, suggesting enhanced abundance of activated CRIB46‐binding HvRACB‐GTP. HA‐HvGEF14 protein stability was verified via western blot after FRET‐FLIM measurements (Figure S8). Together, this supports that HvGEF14 can facilitate HvRACB to switch into the GTP‐bound signalling‐activated conformation in planta.

supports barley susceptibility towards penetration by Bgh

Because HvGEF14 is expressed in barley leaf epidermal cells and HvGEF14 interacts directly with HvRACB, we tested the potential involvement of this exchange factor in the interaction of barley with Bgh. After transient single cell overexpression or RNAi‐mediated silencing of HvGEF14 in barley epidermis cells and subsequent inoculation with fungal spores, we scored the penetration efficiency for each transformed and attacked cell in at least five independent experiments. Every experiment represents a mean score of at least 50 observed plant–fungus interactions. On average, we observed a significant increase of successful fungal penetration from 34.6% to 46.5% (relative increase of 34%) in HvGEF14 overexpressing cells when compared to an empty vector control (Figure 6). We assessed the efficacy of RNAi‐mediated gene silencing to 53% by measuring the reduction of GFP fluorescence intensities of single GFP‐HvGEF14 expressing barley cells co‐transformed with empty vector or the RNAi silencing construct (Table S4). The knockdown of HvGEF14 by RNAi then led to the opposite effect of overexpression: a decrease in penetration rate from 26.4% in the controls to only 16.5% in cells in which HvGEF14 was silenced by RNAi (relative decrease of 38%). Absolute penetration rates are lower in RNAi experiments due to the longer incubation time of used leaf segments after transformation. In addition, the results considerably varied from experiment to experiment, even in the controls. Accordingly, a p value of 0.07 was computed during statistical analysis of the means and supports a trend towards higher resistance after silencing HvGEF14 (Figure 6). Genetic evidence thus suggests that HvGEF14 supports the susceptibility of barley epidermal cells to penetration by Bgh.

FIGURE 6

HvGEF14 affects Blumeria graminis f. sp. hordei (Bgh) penetration efficiency. Bgh penetration efficiency in barley epidermis cells overexpressing (OE) HvGEF14 full length compared to mean penetration efficiency in cells with pGY1 empty vector (control OE) or after knockdown (KD) of HvGEF14 via RNAi compared to mean penetration efficiency of pIPKTA30N empty vector (control KD). Data points each show penetration efficiency of a minimum of 50 plant–fungus interactions. The mean values of all experiments are indicated with bars. Statistical significance of differences of the mean calculated with the t test in Rstudio (v. 1.2.5033) after assessing for normal distribution with the Shapiro–Wilk test.

DISCUSSION

The barley susceptibility factor HvRACB has been studied to understand molecular mechanisms of its role in supporting fungal entry into barley epidermal cells (Engelhardt et al., 2020). The transition from a GDP‐bound towards the GTP‐loaded signalling activated state of HvRACB is probably important in this context. As shown in the model species A. thaliana and O. sativa, PRONE‐GEFs can facilitate the activation of ROPs. In this work, we therefore investigated the role of a barley PRONE‐GEF candidate. We show the role of epidermis‐expressed and transcriptionally Bgh‐regulated HvGEF14 in ROP activation and susceptibility to fungal penetration (Figure 7).

FIGURE 7

Scheme of hypothetical HvRACB‐dependent signalling. HvGEF14 is a novel PRONE‐GEF facilitating the exchange of GDP to GTP bound to the susceptibility factor and small GTPase HvRACB. The membrane‐associated HvRACB‐GTP interacts with downstream executors, such as HvRIC171, to initiate changes in cellular organization that lead to fungal haustorium accommodation. GAPs (GTPase activating proteins) support GTP hydrolysis to GDP and inorganic phosphate (Pi), thereby switching off HvRACB. Possible interactions between unknown receptor‐like kinases (RLKs) and HvGEF14 are indicated with a dashed arrow. Fungal virulence effectors and guanine nucleotide dissociation inhibitors (GDIs) are not depicted. Adapted from Engelhardt et al. (2020).

According to our phylogenetic analysis, we suggest that PRONE‐GEF14 has a unique position in the evolution of PRONE‐GEFs. A BLAST search with the primary sequence of AtGEF14 reveals that the closest homologue to PRONE‐GEF14 in the moss Physcomitrium patens is PpGEF1 but there is no PRONE‐GEF14 found in the moss (Eklund et al., 2010). In addition, in the liverwort Marchantia polymorpha, only one PRONE‐GEF can be found in the genome. The primary sequence of this PRONE‐GEF, KARAPPO, is also most similar to AtGEF1 (Figure S6 and Hiwatashi et al., 2019). However, the ancient angiosperm species Amborella seems to encode a GEF14 orthologue (protein accession XP_006878646). Hence, PRONE‐GEF14 proteins might have evolved early in angiosperms before separation of monocots and dicots. A unique position of GEF14 proteins in the phylogeny of PRONE‐GEFs is further supported by its comparatively low sequence conservation of the PRONE domain when compared to all other PRONE‐GEFs (Figure S2). Due to high confidence bootstrap analysis and comprehensive annotation of O. sativa and H. vulgare PRONE‐GEFs, we propose to base future comparisons of angiosperm PRONE‐GEFs on the presented phylogeny (Figure 1).

The HvGEF14 transcript level was higher in epidermal peels when compared to whole leaves, suggesting a specific function in epidermal cells. Possible candidate ROPs to interact with HvGEF14 are the epidermal cell‐expressed small GTPases HvRACB, HvRACD, HvRAC1, HvRAC3, and HvROP6. When barley leaves were challenged with Bgh, several susceptibility‐related barley ROPs had slightly lower transcript levels compared to noninoculated controls (Schultheiss et al., 2003). This gene expression profile in the epidermis is reminiscent of the HvGEF14 expression in barley epidermis during Bgh attack (Figure 3). The powdery mildew effector Bgh‐ROP‐interactive peptide 1 can interact with HvRACB and supports fungal virulence (Nottensteiner et al., 2018). We speculate that downregulation of HvGEF14 transcripts and some barley ROPs could reflect a plant response to fungal interference with host ROP signalling, to which the plant reacts by countermeasures and downregulation of the susceptibility pathway.

GEFs are supposed to interact with signalling‐inactive GDP‐loaded ROP versions. This study, however, did not show direct protein–protein interaction of HvGEF14 with GDP‐bound dominant‐negative DN HvRACB‐T20N in vitro or in vivo, but with WT HvRACB and CA HvRACB‐G15V (Figure 4). The interaction of PRONE‐GEFs with CA ROPs is not unheard of, however, as previous studies have shown. For instance, OsGEF1 interaction with OsRAC1 mutants was shown via split‐Venus fluorescence complementation assays in protoplasts. Both the constitutively activated OsRAC1‐G19V as well as the dominant negative OsRAC1‐T24N variants were able to reconstitute Venus fluorescence at the plasma membrane (Akamatsu et al., 2013). Additionally, similar to studies in A. thaliana, which have previously shown an interaction of PRONE‐GEFs with D121N‐like low nucleotide affinity mutants of AtROPs (Akamatsu et al., 2013; Berken et al., 2005; Denninger et al., 2019; Gu et al., 2006), we observed an interaction of HvRACB‐D121N with HvGEF14 in yeast but not consistently in planta. The initial discovery of the A. thaliana PRONE‐GEFs was made in a Y2H screen using AtROP4‐D121N (Berken et al., 2005) and a similar strategy was used to find PRONE‐GEFs as activators of OsRAC1 (Akamatsu et al., 2013). Furthermore, a global investigation into AtGEF–AtROP1 interactions showed that AtGEF14 preferably interacts with a D121A/C188S mutant of AtROP1 in vitro (Gu et al., 2006). The D121N mutation might lead to a similar protein conformation as the nucleotide‐free ROP, which has been crystallized in complex with the PRONE domain of AtGEF8 (Thomas et al., 2007). Together, a picture emerges supporting that the GEF–ROP interaction is most stable in an intermediate, non‐nucleotide‐bound state. Consequently, with regard to DN HvRACB‐T20N, its conformation and high affinity to GDP might prevent completely the initial ROP‐GEF interaction phase, which is usually quickly followed by GDP release. By contrast, activated ROPs might stay in contact with GEFs for ROP activity feedback regulation through complexes formed with ROP executers, as discussed before (Wu & Lew, 2013). Because PRONE‐GEFs including HvGEF14 can form dimers, GEF‐GEF‐ROP(GTP) interaction could also recruit further ROP‐GDP for activation and therefore create a positive feedback loop to form nanodomains of ROP activity (Smokvarska et al., 2021).

Subcellular localization of signalling protein complexes is vital to ROP‐mediated processes. On activation ROPs relocate to, or are stabilized in their association with, the plasma membrane. There, they interact with downstream effectors/executers to facilitate, amongst other functions, polar growth processes (Kawano et al., 2014; Poraty‐Gavra et al., 2013; Schultheiss et al., 2003). During A. thaliana pollen tube growth, for example, the localization of activated AtROP1 to the apical plasma membrane regulates tip growth (Gu et al., 2004). This specific localization of activated AtROP1 to the membrane leads to the accumulation of downstream executers like AtRIC4 in the same compartment. In the presence of dominant negative AtROP1, however, the RIC protein localizes to the cytoplasm (Gu et al., 2005). In addition, the co‐expression of a ROP–GAP, which renders AtROP1 inactive, also results in the relocation of AtRIC4 to the cytoplasm. The correct localization of activated AtROP1 as well as AtRIC4 was observed to be crucial for downstream signal transduction and fine‐tuning of the growth process (Hwang et al., 2005). To assess HvRACB signalling activity status in planta, we made use of the previously published recruitment of HvRIC171 by CA HvRACB‐G15V to the cell periphery (Schultheiss et al., 2008). We measured higher mCherry‐HvRIC171 localization at the cell periphery in the presence of transiently overexpressed HvGEF14 (Figure 5). This suggests that the overexpression of HvGEF14 leads to a higher ratio of activated endogenous ROPs, which in turn recruit HvRIC171 towards the cell periphery. Our observation could even indicate that HvGEF14 functions in the specific susceptibility pathway of HvRIC171 in the interaction with Bgh because HvRIC171 can also support fungal invasion into epidermal cells (Schultheiss et al., 2008). In this way, HvGEF14 might activate HvRACB, which then associates with the cell periphery to where it recruits HvRIC171. In the case of a fungal attack, this localization of activated HvRACB and HvRIC171 was observed in context of successful fungal penetration (Engelhardt et al., 2021; Schultheiss et al., 2003, 2008). Interestingly, CA RACB‐G15V likewise recruits MICROTUBULE‐ASSOCIATED ROP GTPASE ACTIVATING PROTEIN1 (MAGAP1) to the cell periphery and MAGAP1 partially accumulates at the cell periphery or haustorial neck when Bgh successfully penetrates (Hoefle et al., 2011). Hence, interaction of activated HvRACB with the negative regulator HvMAGAP1 or other ROP GTPase activating proteins might lead to hydrolysis of RACB‐bound GTP and limit the efficiency of GEF14 in inducing susceptibility towards Bgh (Figure 7).

A more direct way to measure ROP activation in planta is via CRIB‐based ROP activity sensors that have previously been adapted for plants (Kawano et al., 2010; Wang et al., 2018; Wong et al., 2018). The so‐called Ras and interacting chimeric unit (Raichu) sensor includes the fluorophores Venus and CFP as FRET‐pair. ROP activity is monitored by its interaction with the CRIB domain of a downstream executer. In a similar manner, we used the interaction of HvRACB with the CRIB domain of HvRIC171 in FRET‐FLIM assays (Denay et al., 2019). We observed a stronger interaction of HvRACB and HvCRIB46 when we co‐expressed HvGEF14, which strongly suggests that HvGEF14 can activate HvRACB in planta (Figure 5d). Additionally, transient overexpression of HvGEF14 in barley epidermal cells led to a significant increase in Bgh penetration success with 34% higher relative penetration events on average when compared to controls (Figure 6). This is comparable to enhanced fungal penetration during overexpression of CA HvRACB‐G15V (Schultheiss et al., 2003), as well as the overexpression of HvRACB‐downstream executers such as HvRIC171 (Schultheiss et al., 2008) and HvRIPb (McCollum et al., 2020). Transient silencing of HvGEF14, on the other hand, rendered barley epidermis cells on average 38% more resistant to Bgh penetration (Figure 6). Even though this effect could be observed in every repetition when comparing HvGEF14 RNAi with its respective control, the extent of fungal penetration varied amongst repetitions so that a p value of 0.05 was not met after statistical analysis. Considering, however, that inoculation with Bgh naturally decreases the transcription of HvGEF14 (Figure 3), additional ectopic knockdown perhaps cannot be expected to exert a major additional effect. Taken together, these assays point to a role of HvGEF14 in supporting the accommodation of Bgh infection structures in barley epidermal cells, similar to and possibly in cooperation with the susceptibility factor HvRACB and other epidermis‐expressed HvROPs.

In conclusion, HvGEF14 is a bona fide barley PRONE‐GEF that interacts with barley ROPs. The interaction with susceptibility‐related barley ROPs might lead to ROP activation and therefore facilitate ROP functions in susceptibility to invasion by Bgh. Research on A. thaliana PRONE‐GEFs has highlighted the interplay of different PRONE‐GEFs in polar growth processes like root hair formation or pollen tube growth (Denninger et al., 2019; Li et al., 2020). It remains to be studied if HvGEF14 function is choreographed in a similar manner to other barley PRONE‐GEFs. In addition, other PRONE‐GEF interactors, such as potential upstream RLKs, remain to be investigated (Figure 7). In future studies, susceptibility‐ and HvRACB‐related candidate RLKs (Douchkov et al., 2014; Schnepf et al., 2018) will be of interest to link HvRACB signalling to cell surface signal perception. HvGEF14 could hence provide a link between HvRACB and cell surface RLKs, and further help understanding the ROP signalling pathway co‐opted by Bgh in susceptible barley.

EXPERIMENTAL PROCEDURES

Plant and pathogen propagation and maintenance

H. vulgare 'Golden Promise' was grown at 20°C, 50% humidity, and 16 h light, 8 h dark cycles for 7–8 days in standard potting soil. Bgh was propagated for 7–21 days on Golden Promise in a climate chamber at 18°C and 65% humidity, with 16 h light, 8 h dark cycles.

Cloning procedures

Open reading frames (ORFs) of Golden Promise genes were amplified from leaf cDNA with primers (Table S5) designed on the Barley Genome (The International Barley Genome Sequencing Consortium, 2012). Constructs were cloned into Gateway destination vectors (Table S6) using BP and LR clonase (Invitrogen). Plasmids were prepared via column purification (Machery Nagel).

Binary Agrobacterium vectors for transformation in N. benthamiana and subsequent FRET‐FLIM measurements were cloned using a combination of GoldenGate (Engler et al., 2008) and Gateway (Invitrogen) cloning (Table S5). Fusion constructs consisting of a fluorescent protein, a 10× glycine linker, and a protein‐of‐interest were first linked through Esp3I‐mediated GoldenGate cloning and subsequently transferred into Gateway vectors through flanking attBsites. The necessary Esp3I sites and attBsites were introduced via overhang‐PCR. Purified amplicons were assembled through Esp3I‐ and T4 DNA ligase‐mediated restriction‐ligation cloning (Engler et al., 2008). pDONR223 Gateway entry clones were used for transformation of Escherichia coli DH5α. The correct assembly of fusion constructs and integrity of sequences was confirmed via restriction digestion followed by Sanger sequencing. Subsequently, the constructs were shuffled from pDONR223 into the Gateway binary vector pGWB2 (Nakagawa et al., 2007) using Gateway LR reactions. These pGWB2 clones were used for transformation of E. coli DH5α. The integrity of sequences was again confirmed via restriction digestion and Sanger sequencing.

Sequence alignment and phylogenetic tree construction

MUSCLE alignment of 14 A. thaliana, 11 O. sativa (sequences downloaded from TAIR and NCBI on 10.05.2021), and 11 H. vulgare PRONE‐GEFs (MOREX genome v. 3) was performed in SeaView software. A maximum‐likelihood (PhyML) analysis was performed with an LG model, bootstraps with 100 replicate, model‐given amino acid equilibrium frequencies, nearest neighbour interchange tree searching, and five random starts. The resulting TBE tree's design was further adjusted in InkScape.

RNA extraction and cDNA synthesis

Seven‐day‐old H. vulgare 'Golden Promise' was collected in three biological replicates. Whole‐leaf and epidermal peels were cut and frozen in liquid nitrogen. Leaf tissue was ground in a tissue lyser with glass beads. RNA was extracted with TRIzol according to the protocol in Chomczynski and Sachhi (1987) and DNase I digestion was performed. Subsequently, cDNA synthesis was performed from 1 μg RNA with the QuantiTect reverse transcription kit (Qiagen) according to the supplier's protocol. cDNA from 1 μg RNA was diluted 1/10 for further analysis.

RT‐qPCR

Appropriate primers (Table S5) were used in 10 μl reactions with the Maxima 2 × SYBR Green/ROX qPCR Master Mix (Thermo Scientific) and RT‐qPCR was run on Aria Mx3000 (Agilent) with 40 cycles at 60°C for 10 s followed by 72°C for 15 s and a subsequent melting curve (65–95°C). HvUbiquitin was measured as a housekeeping gene (Schnepf et al., 2018) and foldchanges were calculated via the 2−∆∆ method by Livak and Schmittgen (2001).

Y2H assay

Protein–protein interactions were performed as described in the Matchmaker protocol (Clontech). HvRACB (amino acids 1–193) ORF was cloned with a premature stop codon to express a truncated ROP protein lacking its C‐terminal prenylation signal. Saccharomyces cerevisiae AH109 was transformed with pGBKT7 and pGADT7 plasmids (Table S6) containing the specific gene of interest and cultivated for 3–6 days at 30°C on synthetic dropout medium lacking amino acids leucine and tryptophan (SD−L−W). Five millilitres of liquid SD−L−W medium was inoculated with yeast colonies and on overnight cultivation at 30°C dilutions were dropped on SD−L−W and SD−L−W−H (SD lacking amino acids leucine, tryptophan, histidine) or SD−L−W−H−Ade (SD lacking amino acids leucine, tryptophan, histidine, and nucleotide adenosine) plates and incubated at 30°C.

Protein extraction from yeast and N. benthamiana

Transformed yeast was cultured in 4 ml of SD−L−W overnight and centrifuged at 4000 × g for 5 min at 4°C. Cells were washed in 100 μl of 2 M LiAc and subsequently incubated for 5 min at room temperature in 100 μl of 0.4 M NaOH. Pellets were collected and 50 μl of 4× SDS‐sample buffer was added. After vortexing, samples were boiled at 95°C for 5 min and briefly spun down before loading onto an SDS‐polyacrylamide gel (Zhang et al., 2011).

Leaf discs (12 mm tissue punch) were collected from A. tumefaciens‐transformed N. benthamiana leaves 48 h posttransformation and directly frozen in liquid nitrogen. Material was homogenized in a tissue lyser with glass beads and 200 μl of 4× SDS sample buffer was added. After vortexing, samples were boiled at 95°C for 10 min and spun down before loading onto an SDS‐polyacrylamide gel.

SDS‐PAGE and western blot

Extracted proteins were separated by electrophoresis in a 12% polyacrylamide gel and blotted onto polyvinylidene difluoride (PVDF) membrane via semidry western blotting (protein extracted from yeast) or wet western blotting (protein extracted from N. benthamiana). The membrane was blocked with 5% milk in phosphate‐buffered saline and incubated with specific antibodies. Proteins were detected by chemiluminescence with SuperSignal West Dura or FEMTO chemiluminescence substrate (Thermo Scientific).

Transient biolistic transformation of barley epidermal cells

Barley cv. Golden Promise 7‐day‐old detached leaves were transformed by particle bombardment as described previously (McCollum et al., 2020). Two micrograms of plasmid/transformation was used in FRET‐FLIM experiments. For fungal penetration efficiency experiments, 1 μg of plasmid/transformation for the gene of interest and 0.5 μg of plasmid/transformation of transformation marker (pUbi_GUSplus, β‐glucuronidase) were applied.

A. tumefaciens transfection of N. benthamiana leaves

Agrobacterium. tumefaciens GV3101 carrying binary expression vectors pGWB containing meGFP‐HvRACB WT, meGFP‐HvRACB G15V (CA), GST‐mCherry, HvCRIB46‐mCherry, or 3xHA‐HvGEF14 were infiltrated into N. benthamiana leaves according to Yang et al. (2000). Bacterial liquid cultures were grown to OD600 0.5 and mixed in equal amounts including P19 silencing suppressor. Forty‐eight hours after infiltration, FRET‐FLIM measurements were performed and fluo proteins were extracted for western blotting.

FRET‐FLIM

HvGEF14 was N‐terminally tagged with monomeric eGFP as donor and N‐terminal fusions with mCherry of HvRACB variants were used as acceptors. To test HvRACB activation status in planta, binary vectors of meGFP‐HvRACB WT or meGFP‐HvRACB CA were co‐expressed with GST‐mCherry or HvCRIB46‐mCherry with or without co‐expression of 3HA‐HvGEF14 via A. tumefaciens infiltration in N. benthamiana leaves.

Microscopy of transiently transformed H. vulgare epidermis cells and A. tumefaciens‐infiltrated N. benthamiana leaf discs was performed with an Olympus FV 3000 microscope with 488 nm (20 mW) and 561 nm (50 mW) diode lasers. GFP photons were excited with a 485 nm (LDH‐D‐C‐485) pulsed diode laser and time‐correlated single photon counting (TCSPC) was performed with 2× PMA Hybrid 40 photon counting detectors. A minimum of 1000 photon counts was collected and subsequently analysed with the PicoQuant SymPhoTime 64 software. N‐exponential reconvolution and decay curve fitting with daily measured or calculated IRF, for H. vulgare and N. benthamiana, respectively, was applied to gain a fit with χ2 values between 0.9 and 1.2.

Confocal microscopy

Transiently transformed barley epidermis cells were imaged 24 h posttransformation (hpt) with a Leica TCS SP5 confocal microscope with hybrid HyD detectors. mCherry fluorophores were excited with a 561 nm laser and detected at 570–610 nm. GFP fluorescence was excited with a 488 nm argon laser and detected at 500–550 nm.

RNAi efficiency

GFP‐HvGEF14 was transiently overexpressed together with cytosolic mCherry as a transformation marker. In addition, the empty vector of a HvGEF14 RNAi hairpin construct of amino acids 201–401 was cotransformed and the fluorescence intensity of GFP and mCherry was measured in the z‐stack of confocal images taken 48 hpt. RNAi efficiency was determined by the ratio of mean mCherry‐normalized GFP fluorescence in HvGEF14‐silenced cells divided by mCherry‐normalized GFP‐HvGEF14‐expressing control cells.

Fungal penetration efficiency

Barley leaves were fixed on 0.8% water agar and inoculated 24 h after transient transformation with overexpression constructs or 48 h after transient transformation with RNAi constructs with 100 Bgh conidiospores per mm2. Inoculated leaves were incubated for 48 h in a climate chamber at 18–22°C and 16 h light, 8 h dark. Inoculated leaves were stained in 5‐bromo‐4‐chloro‐3‐indolyl β‐d‐glucuronic acid (X‐Gluc) solution 48 h after inoculation and fixed in 80% ethanol. Fungal penetration efficiency was determined with light microscopy as described before (Hückelhoven et al., 2003).

Statistical analysis

Statistical analyses were performed in Rstudio. Global comparison for nonparametric data was assessed with the Kruskal test and pairwise comparisons of nonparametric data were calculated via the Wilcox test with Bonferroni p value adjustment. Outlier tests were performed using the Grubbs test. Figures were prepared with RStudio's ggplot2 package and adjusted in InkScape.

CONFLICT OF INTEREST

The authors declare no conflict of interest.

Figure S1 Arabidopsis thaliana and Oryza sativa PRONE‐GEFs. Maximum‐likelihood analysis and annotation of OsGEFs based on A. thaliana nomenclature in (Berken et al., 2005)

Click here for additional data file.

Figure S2 MUSCLE alignment of Arabidopsis thaliana, Oryza sativa, and Hordeum. vulgare PRONE‐GEFs as the basis for the phylogenetic tree in Figure 1

Click here for additional data file.

Figure S3 HvGEF14 homodimerizes with full length and PRONE domain (amino acids 124–485) in yeast two‐hybrid assay. Protein–protein interaction shown on medium containing amino acid mix without leucine (−L), tryptophan (−W), or histidine (−H). Growth on −L−W medium shows successful plasmid transformation. Representative image of three experiments with the same result. EV, empty vector; +ve control, positive control

Click here for additional data file.

Figure S4 Original images of HvGEF14 full‐length and PRONE domain (amino acids 124–485) interaction with HvRACB variants (wild type, WT; G15V; D121N; T20N) via yeast two‐hybrid asssay. Successful yeast transformation shown on medium containing amino acid mix without leucine (−L) and tryptophan (−W). Medium containing amino acids mix without L, W, and histidine (−H) was used as selection for protein–protein interaction. Yeast colonies were grown in liquid −L−W medium and dropped onto plates in a dilution series. Representative image of three experiments with the same result

Click here for additional data file.

Figure S5 Western blot of yeast proteins on polyvinylidene difluoride (PVDF) membranes. (a) Prey proteins HvGEF14 full‐length (F) and PRONE (P) as well as empty vector (EV) fused to HA‐tag and labelled with anti‐HA‐HRP (3F10; Roche). Detection with DURA chemiluminescence substrate (Thermo Scientific). In brackets: corresponding bait proteins. (b) Bait proteins HvRACB variants fused to Myc‐tag and labelled with c‐Myc (9E10) and m‐IgGκ BP‐HRP (sc‐516,102) antibodies (Santa Cruz Biotechnology). Detection with FEMTO chemiluminescence substrate (Thermo Scientific). In brackets: corresponding prey proteins HvGEF14 full‐length (F), HvGEF14 PRONE (P), and empty vector (EV)

Click here for additional data file.

Figure S6 HvRAC1 wild type (WT), G23V constitutively activated (CA), and T28N dominant negative (DN) variants were tested as bait against prey constructs HvGEF14 full‐length, HvGEF14 124–485 (PRONE domain), or empty vector (EV). Interaction of proteins shown on medium containing amino acid mix without leucine (L), tryptophan (W), histidine (H), and adenine (Ade) (−L−W−H−Ade) in a dilution series. Representative image of three experiments with the same result

Click here for additional data file.

Figure S7 Western blot of yeast proteins on polyvinylidene difluoride (PVDF) membranes. (a) HvGEF14 variants (prey proteins) fused to HA‐tag and labelled with anti‐HA‐horseradish peroxidase (HRP) (3F10; Roche). Detection with DURA chemiluminescence substrate (Thermo Scientific). (b) HvRAC1 (bait proteins) fused to Myc‐tag and labelled with c‐Myc (9E10) and m‐IgGκ BP‐HRP (sc‐516,102) antibodies (Santa Cruz Biotechnology). Detection with FEMTO chemiluminescence substrate (Thermo Scientific)

Click here for additional data file.

Figure S8 Western blot of proteins extracted from Nicotiana benthamiana on polyvinylidene difluoride (PVDF) membranes. (a) 3×HA‐HvGEF14 co‐expressed with meGFP‐HvRACB, HvCRIB46‐mCherry and P19 in lane 1. 3×HA‐CA HvRACB extracted from transgenic barley plants used as positive control in lane 2. Negative control in lane 3 with untagged HvGEF14 co‐expressed with meGFP‐HvRACB, HvCRIB46‐mCherry, and P19. Proteins were labelled with anti‐HA‐horseradish peroxidase (HRP) (3F10; Roche). (b) HvCRIB46‐mCherry co‐expressed with meGFP‐HvRACB, 3xHA‐HvGEF14, and P19 in lane 1. GST‐mCherry co‐expressed with meGFP‐HvRACB and P19 in lane 2. HvCRIB46‐mCherry co‐expressed with meGFP‐HvRACB and P19 in lane 3. HvCRIB46‐mCherry co‐expressed with meGFP‐CA HvRACB and P19 in lane 4. Proteins were labelled with anti‐RFP‐rat mAb (5F8) and anti‐rat (A9542) antibodies (ChromoTek). (c) meGFP‐HvRACB co‐expressed with HvCRIB46‐mCherry, 3xHA‐HvGEF14 and P19 in lane 1. meGFP‐HvRACB co‐expressed with GST‐mCherry and P19 in lane 2. meGFP‐HvRACB co‐expressed with HvCRIB46‐mCherry and P19 in lane 3. meGFP‐CA HvRACB co‐expressed with HvCRIB46‐mCherry and P19 in lane 4. Proteins were labelled with anti‐GFP (B‐2) (sc‐9996) and m‐IgGκ BP‐HRP (sc‐516,102) antibodies (Santa Cruz Biotechnology). Detection with DURA chemiluminescence substrate (Thermo Scientific)

Click here for additional data file.

Table S1 Primary sequence similarity of HvGEFs when compared to consensus sequence shows the highest variability of the HvGEF14 primary sequence

Click here for additional data file.

Table S2 List of barley PRONE‐GEFs with gene/protein identifiers, annotation, PRONE prediction, and protein sequence length

Click here for additional data file.

Table S3 Tissue‐specific HvGEF gene expression based on RNAseq (James Hutton Institute) in fragments per reads per kilobase million (FPKM)

Click here for additional data file.

Table S4 RNAi efficiency of GEF14 hairpin construct in transient transformation of barley epidermal cells

Click here for additional data file.

Table S5 List of primers used in this study

Click here for additional data file.

Table S6 List of plasmids used in this study

Click here for additional data file.