Protein kinases form the basis of the phosphorylation cascades that transmit cellular signals by facilitating phosphotransfer from ATP to serine, threonine, tyrosine, and histidine residues in a substrate protein or peptide (Kwon et al., 2019). The highly conserved eukaryotic kinase domain contains three catalytic residues essential for binding and positioning ATP in conjunction with metal ion cofactors such as Mn²⁺ and Mg²⁺ to facilitate phosphotransfer (Adams, 2001). Approximately 5% of all plant genes encode proteins containing a kinase domain, and at least one-third of plant proteins are phosphorylated (Kwon et al., 2019). Based on bioinformatic analyses, eukaryotic kinases have been divided into canonical (catalytically active) kinases and pseudokinases, which are usually catalytically inactive due to a non-canonical catalytic residue (Kwon et al., 2019). The catalytic residues include aspartate in the DFG subdomain that binds divalent cations to co-ordinate β and γ phosphates of ATP, the lysine in the VAIK subdomain that anchors and orients ATP for substrate transfer and the aspartate in the HRD domain of the catalytic loop that is a base acceptor to achieve proton transfer (Adams, 2001). Over 17% of predicted protein kinases encoded in the Arabidopsis genome are classified as peudokinases and recent research on plant pseudokinases confirm that they perform distinct biological functions in plant defense against microbial pathogens through establishing protein-protein interactions, as opposed to phosphotransfer (Kwon et al., 2019). For example, several kinase catalytic subdomains of the ZED1 pseudokinase influences its binding to PBL proteins and the ZAR1 nucleotide-binding leucine-rich repeat (NLR) protein during activation of effector-triggered immunity (ETI) (Bastedo et al., 2019). The BIR2 pseudokinase regulates microbial-associated molecular pattern (MAMP)-induced signaling by binding the BAK1 co-receptor to prevent its untimely interaction with the FLS2 receptor which is involved in perception of the flagellin-derived flg22 (Halter et al., 2014). The GHR1 receptor pseudokinase is involved in inducing stomatal closure following MAMP treatment through its interactions with the calcium-responsive CDPK3 (Hua et al., 2012; Sierla et al., 2018).

While a majority of pseudokinases lack phospho-transfer activity, a small subset maintain atypical kinase activity through catalytic residue substitutions with similar amino acids elsewhere in the catalytic site (Kwon et al., 2019). For example, some pseudokinases including CASK do not require a cofactor to initiate phosphotransfer indicating the divergence of their mechanisms from typical protein kinases (Mukherjee et al., 2010; Lewis et al., 2014; Kwon et al., 2019). The integrin-liked kinase 1 (ILK1) from Medicago and Arabidopsis are the only plant pseudokinases that are active kinases in vitro and in vivo with demonstrated auto- and substrate-phosphorylation activity (Chinchilla et al., 2008; Nemoto et al., 2011; Brauer et al., 2016). ILK1 belongs to the ILK family of Raf-like MAP3Ks found in both metazoans and plants though the nature of ILK catalytic residue degeneration is distinct between kingdoms (Ichimura et al., 2002; Popescu et al., 2017). In animals, the ILKs lack the catalytic aspartate in the HRD domain and have variable changes in the DFG domain (Popescu et al., 2017). Several studies indicate that the metazoan ILKs are inactive, though some evidence suggests that they maintain kinase activity in vitro (Dagnino, 2011; Hannigan et al., 2011). In the plant kingdom, many species encode both pseudokinase ILKs and ILKs with canonical kinase domains, though more ancient lineages including the green algae only encode non-pseudokinase ILKs (Popescu et al., 2017). For example, in Arabidopsis, ILK1, ILK2, and ILK3 encode pseudokinases while ILK4, ILK5, and ILK6 encode kinases. The plant pseudokinase ILKs have a common GFG motif in place of the DFG motif and some species encode ILKs that lack the catalytic aspartate in the HRD domain (Popescu et al., 2017). This suggests that degradation of the ILK kinase domains occurred independently in plants and animals and that the GFG motif is maintained across evolutionary time in plants, potentially due to a link with ILK function. However, it remains unclear how frequently these atypical catalytic residues occur in the closely related Raf-like MAP3K family members.

Previous work on pseudokinase ILKs from Arabidopsis (ILK1) and Medicago (MsILK1) indicate that they are able to phosphorylate substrates in vitro with an unusual preference for the Mn²⁺ cofactor compared to Mg²⁺ in keeping with their modified cofactor-binding domain (DFG) (Chinchilla et al., 2008; Brauer et al., 2016). While mutation of the aspartate residue in the HRD catalytic loop to asparagine eliminates kinase activity in most protein kinases, this mutation increased ILK1 activity further demonstrating the uniqueness of its enzymatic mechanism (Strong et al., 2011; Brauer et al., 2016). Ectopic expression of the Medicago ILK1 containing the same HRD- > HRN mutation in Arabidopsis reduced lateral root formation relative to the native Medicago ILK1, suggesting a role for MsILK1 in root development (Chinchilla et al., 2008). Complementation of the ilk1 knock out mutant in Arabidopsis with kinase-dead or kinase-active ILK1 revealed that ILK1 activity is essential for pattern-triggered root growth inhibition in response to flg22, pep1, and elf18 (Brauer et al., 2016; Popescu et al., 2017). Kinase function of the ILK1 pseudokinase as also required for resistance to the bacterial pathogen Pseudomonas syringae pv. tomato DC3000 in the leaf (Brauer et al., 2016; Popescu et al., 2017). Our previous proteomic screens using functional protein microarrays (FPM) and a kinase client assay (KiC) revealed seven transporters that interacted with ILK1, including the HAK5 potassium transporter that mediates potassium uptake under limiting conditions (Qi et al., 2008). Further confirmation of the ILK1-HAK5 interaction in vivo and functional analysis of the hak5-3 knock out mutant indicate that the role of ILK1 in defense and MAMP response seems to be partly dependent on its maintenance of HAK5 potassium transporter stability in conjunction with the calmodulin-like protein CML9 (Brauer et al., 2016). Indeed, MAMP treatment triggers bulk loss of K⁺, Mn²⁺ and accumulation of P^3–, S^2–, Zn²⁺, and Fe³⁺ in Arabidopsis seedlings, where K⁺ and Mn²⁺ are influenced by ILK1 expression (Brauer et al., 2016). Recent work indicates that ILK1 also promotes transcriptional changes of cell wall modifying enzymes in response to MAMPs as well as lignification of the root xylem which has previously been linked to maintenance of ion transport throughout the plant (Boyce et al., 2004; Dimlioglu et al., 2022). Kinase-independent functions for ILK1 include the promotion of germination during salt and osmotic stresses suggesting that much like the metazoan ILKs that scaffold multiple proteins to promote cell adhesion, ILK1 may also interact with more than one protein in vivo (Brauer et al., 2016; Chastney et al., 2021). Together, this suggests a role for ILK1 in maintaining ion homeostasis and that this regulation is essential for plant adaptive responses to biotic and abiotic stress.

Two non-pseudokinases ILK4 and ILK5 have overlapping functions with ILK1 in regulating root lignification, secondary root formation and root growth during osmotic stress or following MAMP treatment in Arabidopsis (Dimlioglu et al., 2022). It is unclear if this functional overlap is due to similar phosphorylation targets or protein interactors between the ILKs. Recent advances place ILK5 as a true MAP3K functioning downstream of the extracellular ATP receptor P2K1 to promote pattern-triggered responses (Kim D. et al., 2022). Ligand-activated P2K1 phosphorylates ILK5 which in turn initiates the MKK5-MPK3/MPK6 phosphorylation cascade to induce stomatal closure and resistance to P. syringae DC3000 (Kim D. et al., 2022). Thus, the ILKs may contribute to plant responses to multiple patterns through the endogenous ATP signaling pathway which is activated in response to wounding and abiotic stress (Matthius et al., 2019). Externally applied ATP inhibits root elongation and increases cell wall lignification potentially through elevation of auxin (Matthius et al., 2019). A similar mechanism regulates root growth inhibition in response to the pep1 endogenous pattern, where pep1 triggers altered distribution of auxin efflux carriers and auxin accumulation which induces premature cell differentiation and inhibited root growth (Jing et al., 2019). Whether ATP plays a role in pep1 or microbial pattern-induced root growth inhibition remains unclear.

Understanding the plant ILK interactions and phosphorylation patterns would expand our comprehension of the active pseudokinases and facilitate future work on modeling alternative active sites. While our previous work focused on transporter interactors of ILK1, we expand this approach in this work by using two unbiased in vitro proteomic screens to identify putative ILK1 interacting proteins. The importance of the Mn²⁺ or Mg²⁺ cofactors on ILK1 phosphorylation of target peptides is also evaluated to determine the relationship between substrate selectivity and cofactor binding for these atypical kinases.

Here, we describe the putative protein-protein interactions and phosphorylation activity of the atypical protein kinase ILK1. Members of the ILK family were the only pseudokinases identified within the Raf-like MAP3Ks and the ILK1 is one of the only pseudokinases in plants with demonstrated phosphotransfer activity demonstrated in vitro and in vivo (Brauer et al., 2016). In line with its unusual metal ion cofactor-binding domain, we demonstrate that ILK1 phosphorylates a broader range of peptides in the presence of Mn²⁺ compared to Mg²⁺ but that it primarily phosphorylates serine and threonine residues. Our survey of ILK1 protein interactors revealed interactions with diverse proteins, including transporters, protein kinases, cell wall modifying enzymes, and hormone receptors. Together, this indicates that ILK1 is an active kinase with a potentially broad range of interactions with diverse substrates. While its kinase-dependent role in MAMP-induced growth and defense responses seems to align with its regulation of the HAK5 potassium transporter, we cannot exclude the possibility that additional phosphorylation targets of ILK1 may exist in the cell. Indeed, it is increasingly apparent that protein kinases interact with a wide range of proteins in vitro and in vivo, suggesting that the functional specificity of kinases is related to their localization, protein complexing, and posttranslational modifications as opposed to interaction specificities (Popescu et al., 2009; Brauer et al., 2018).

Root growth is influenced by both potassium and MAMPs through ILK1 and HAK5 indicating shared components of potassium and immune signaling pathways (Qi et al., 2008; Brauer et al., 2016). Blocking K⁺ channels abolishes flg22-induced root growth inhibition which is counterintuitive if the function of K⁺ in this process is strictly to promote nutrition (Brauer et al., 2016; Shabala, 2017). The importance of potassium in defense is further underlined by the fact that the rice blast pathogen Magnaporthe oryzae targets the AKT1 potassium channel using the AvrPiz-t effector to promote virulence during infection (Shi et al., 2018). However, it remains unclear if the role of K⁺ in plant defense is due to a direct influence on signaling or if these observations reflect a broader scope of changes in ion homeostasis and their impact on immunity as K⁺ influences transport of diverse nutrients and influences membrane charge and fluidity (Nieves-Cordones et al., 2019). MAMP exposure induces K⁺ and Mn²⁺ loss and accumulation of P^3– S^2–, Zn²⁺, and Fe³⁺ in Arabidopsis seedlings within minutes of flg22 application though this shift is transient since Fe³⁺ accumulation returns to control levels by 6 h (Brauer et al., 2016; Nobori et al., 2018). Indeed, suppression of phosphate transport through the PHT1;4 uptake transporter in roots is regulated downstream of MAMP recognition by BIK1 and PBL1 receptor-like cytosolic kinases and influences infection by Ralstonia solanacearum (Dindas et al., 2022). On the pathogen side, P. syringae DC3000 sulfur starvation responses are induced when plants are pre-treated with flg22 indicating that bacteria may sense the nutritional shift in plants during immune activation (Lovelace et al., 2018). The pH of the plant rhizosphere is manipulated by bacterial release of gluconic acid and influenced by bacterial amino acid synthesis, which has a direct impact on root defense responses and microbiome profiles (Yu et al., 2019; Liu et al., 2021). This indicates the dynamic interplay between plant nutrition status and plant-microbe interactions.

At the cellular level, MAMP recognition induces transient influx of Ca²⁺, H⁺, and efflux of K⁺, Cl^–, and NO3^– (Jeworutzki et al., 2010). Cation transport contributes to the rate but not the extent of membrane depolarization which is primarily mediated by anions (Jabs et al., 1997; Jeworutzki et al., 2010; Brauer et al., 2016). Cation fluxes through La³⁺-sensitive channels are essential to trigger MAMP-triggered plasma membrane depolarization and defense compound production (Jabs et al., 1997; Jeworutzki et al., 2010). While these effects are attributed to Ca²⁺-depending signaling, La³⁺ blocks both Ca²⁺ and K⁺ channels in plants and thus a more nuanced understanding of the relationship between cation fluxes and immune signaling is warranted (Terry et al., 1992). For example, K⁺ levels influence the sensitivity of the plasma membrane to flg22, potentially through its influence on membrane polarity (Chi et al., 2021). Mn²⁺ homeostasis also changes in response to flg22 which could impact cofactor availability for ILK1 as well as tyrosine phosphorylation of pattern recognition receptors such as the lipid MAMP-binding LORE receptor and the elf18-binding EFR receptor (Macho et al., 2015; Luo et al., 2020). In mammals, cation leakage from pore-forming toxins or atypical cation pore formation is detected by the cell to induce MAPK signaling (Kloft et al., 2009; Olabisi et al., 2016). In plants, recent work indicates that the NB-LRRs form cation-conducting channels in the plasma membrane upon avirulence factor recognition to activate ETI (Kim N.H. et al., 2022). These resistosome channels can conduct calcium, but may also conduct other cations (Kim N.H. et al., 2022). The signaling mechanisms connecting cation efflux with downstream activation of ETI signaling remains unclear but could involve similar mechanisms as PTI. For example, cation-induced changes in plasma membrane polarity influences the voltage-regulated vacuolar two-pore channels and potassium channels including GORK and SKOR (Kurusu et al., 2005; Wu et al., 2018). Membrane polarity also influences membrane fluidity and the microdomains in which immune signaling proteins are recruited and thus disruption of these microdomains or protein interactions therein could trigger signaling pathways leading to defense induction (Malinsky et al., 2016). Alternatively, cations movement may be sensed directly as has been recently demonstrated during salt stress, where GIPC sphingolipids bind Na⁺ ions to activate Ca²⁺ influx channels and downstream signaling pathways regulating root growth (Jiang et al., 2019). Our understanding of the influence of cation fluxes on inducible immunity remains unclear, but the recent breakthroughs in our understanding of resistosomes will no doubt advance our progress in this area.

This work improves our understanding of the atypical protein kinases and reinforces the notion that functionally relevant alternative protein kinase mechanisms exist. Further characterization of the ILK family and their role as either scaffolding proteins or protein kinases will help to illuminate the function of these pseudokinases in plant defense and stress responses.

The original contributions presented in this study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

EB and SP conceived of the study and designed the experiments. EB and NA conducted the proteomic experiments. EB, NA, and GP analyzed the data. EB wrote the manuscript. All authors interpreted the data and contributed to writing the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.