Brachypodium distachyon ERECTA-like1 protein kinase is a functional guanylyl cyclase

1. Abstract 2. Introduction 3. Materials and methods 3.1 Expression, purification and electrophoretic analysis of the GST-BdERL1 fusion protein 3.2 Site-directed mutagenesis of the BdERL1 protein 3.3 Determination of BdERL1 WT and mutant guanylyl cyclase activity 3.4 Mass spectrometry analysis of cyclic nucleotides (cGMP and cAMP) 3.5 Kinase activity assay 3.6 Computational modeling 4. Results and discussion 4.1 Identification of a GC catalytic motif in BdERL1 4.2 Computational assessment of the BdERL1 GC domain 4.3 In vitro analysis of the GC activity of wild-type BdERL1 4.4 In vitro analysis of the GC activity of BdERL1 mutants 4.5 In vitro analysis of the kinase activity of wild-type BdERL1 5. Conclusions 6. Author contributions 7. Ethics approval and consent to participate 8. Acknowledgment 9. Funding 10. Conflict of interest 11. References


Abstract
The plant proteins called ERECTA family play important role in inflorescence architecture, stomatal patterning and phloem-xylem organization. ERECTA proteins belong to the moonlighting proteins family containing the guanylyl cyclase (GC) catalytic center embedded within the intracellular kinase domain. This characteristic architecture of ERECTA proteins prompted us to experimentally confirm of enzymatic activity of one of these, BdERL1 (ERECTA-like1 from Brachypodium distachyon). We have shown that BdERL1 is dual-function protein with both kinase and GC activity. Moreover, our mutagenesis studies also revealed the catalytic roles of key conserved amino acid residues at the GC center and importantly, probing of the kinase and GC with Ca 2+ and/or cGMP, shed light on the intramolecular regulations of BdERL1. responses to biotic and abiotic environmental stress factors [1]. The interaction between the ERECTA family and plant hormone signaling (such as auxin, gibberellins or ethylene) has also been confirmed [2][3][4][5]. Interestingly, a few years ago, three receptor kinases belonging to the ERECTA family from A. thaliana, At2g26330 ERECTA (ER), At5g62230 ERECTA-like1 (ERL1), and At5g07180 (ERL2), were bioinformatically selected as potential guanylyl cyclases (GCs) [6]. In silico identification of a 14 amino acid (aa) long catalytic center, [KS]X[SCG]X (10)[KR]X(0, 3)[DHSE], within these proteins suggests that they are able to convert GTP into cGMP. However, biochemical proof of the GC activity of these kinases still needs to be discovered. The amino acid sequence of ERECTA proteins indicates that they belong to the moonlighting family of proteins containing a GC catalytic center embedded within the intracellular kinase domain. Many of these moonlighting proteins are receptor kinases that show a cytosolic dominant function, kinase functionality, an additional cryptic function, cyclase activity, and more importantly, regulatory roles between and within the proteins localized to their cellular microenvironments, which are poorly understood [7,8]. As such, many of the biological processes and responses affected by this class of second messengers have not been fully elucidated.
In this report, we describe the biochemical characterization of BdERL1 from Brachypodium distachyon (NCBI accession no. XP_003561134.3). The purpose of our study was to establish the moonlighting function of the GC in BdERL1 and to determine the catalytic roles of the two functionally assigned and highly conserved amino acids at positions 3 and 14 within the GC catalytic center of BdERL1, as well as to investigate how a functional GC in BdERL1 might affect its kinase activity.

Expression, purification and electrophoretic analysis of the GST-BdERL1 fusion protein
The studied fragment of BdERL1 cDNA (NCBI no. XM_003561086.4) was amplified by PCR using the following primers: 5 ′ -GGGGCCCCTGggatccATGAGGCTGACAGAGAATTTG AGC-3 ′ (forward) and 5 ′ -AGTCACGATgcggccgcTCACA TTGTGTGCTTGGATATGACC-3 ′ (reverse) (lower case letters indicate the BamHI and NotI restriction sites, respectively). The sequence of the studied cDNA fragment encodes the intracellular kinase domain with a conserved GC motif (648-982 aa).
The PCR product was introduced into the pGEX-6P-1 expression vector linearized with BamHI and NotI restriction enzymes using In-Fusion cloning technology (In-Fusion HD Cloning Kit, Clontech). The correctness of the cloned cDNA sequence was verified by sequencing. The E. coli BL21 strain, transformed with the resulting plas-mids, was used to produce GST-tagged proteins. Cells were grown in LB medium supplemented with 2% glucose at 37 • C under vigorous aeration. Fusion protein production was induced by adding isopropyl β-D-thiogalactoside (IPTG) to a final concentration of 1 mM and incubated at 22 • C for 4 h in glass vessels connected to a BioFlo 120 bioprocess control station (Eppendorf). The pH was controlled at 6.5 (± 0.2), the DO (dissolved oxygen) parameter was set to 30%, and the agitation speed was 300 rpm. Bacteria were collected by centrifugation, and the pellet was suspended in lysis buffer (50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 5 mM EDTA, 5 mM EGTA, 1 mM PMSF, protease inhibitor cocktail and 0.2 mg/mL lysozyme) and disrupted by 5 × 15 s of sonication. The soluble fraction was separated by centrifugation at 12 000 × g for 35 min at 4 • C, and the GSTtagged proteins were adsorbed onto glutathione-Sepharose 4B beads. After the column was washed with buffer containing 50 mM Tris-HCl (pH 8.0) and 150 mM NaCl, the GST-BdERL1 complex was eluted with 10 mM glutathione in 50 mM Tris-HCl (pH 8.0). For a control expression vector, pGEX-6-P1 was used, or GST alone was purified as described above.
The homogeneity and purity of the protein fractions were analyzed by 10% (v/v) SDS-PAGE, and the gels were stained with Coomassie blue.

Site-directed mutagenesis of the BdERL1 protein
It was previously described that functionally assigned residues of the GC motif can be replaced with another amino acid to estimate the importance or relevance of these residues to maintain the functional configuration of the catalytic center [7]. The G residue localized at position 3 within the GC catalytic motif has been predicted to determine substrate specificity for GTP. We mutated the G residue (842) in the catalytic center to D (mutant G 842 D). Additionally, the K residue located at position 14 in the GC center is responsible for stabilizing the transition state from GTP to cGMP. In our experiment, we mutated the K residue (853) to A (mutant K 853 A).
Site-directed mutagenesis was performed using a QuikChange II XL site-directed mutagenesis kit (Agilent).
A nonmethylated double strand was synthesized using 125 ng of forward (5'-gagcaactctagaagaacgatatcaaagctgtatacatcagattt-3') and 125 ng of reverse (5'aaatctgatgtatacagctttgatatcgttcttctagagttgctc-3') primers for mutant G 842 D and 125 ng of forward (5'gaatcattgtcaacagccgccttccccgtgagcaactc-3') and 125 ng of reverse 5 ′ -gagttgctcacggggaaggcggctgttgacaatgattc-3') primers for mutant K 853 A from a plasmid DNA containing the BdERL1 gene. The sites where mutagenesis occurred are underlined. PfuUltra High Fidelity DNA Polymerase (Agilent) was used in accordance with the manufacturer's instructions to amplify the plasmids. The original methylated template plasmids were digested with the DpnI restriction enzyme, leaving the amplified plasmid, which was transformed into E. coli XL10-Gold ultracompetent cells (Agilent). Single colonies were selected, and the clones were analyzed by DNA sequencing.

Determination of BdERL1 WT and mutant guanylyl cyclase activity
The guanylyl cyclase activity of the studied proteins was determined by estimating the rate of cGMP formation. For the BdERL1 WT and mutant assay, the reaction mixture contained 50 mM Tris-HCl buffer (pH 7.5), 5 mM MgCl 2 , 5 mM MnCl 2 ,GTP as a substrate at concentrations between 0.25-2 mM and 5 µg of the recombinant fusion enzyme (in some variants supplied with 1 µM Ca 2+ ions and/or calmodulin (AtCaM1; 0.5-10 µg)) in a final volume of 100 µL. After incubation at 30 • C for 15 min, the reaction was stopped by boiling at 100 • C for 10 min, and then the samples were centrifuged at 16 000 × g for 10 min.
Analogically, the cross reactivity of the studied proteins with ATP as a substrate at concentrations between 0.25-2 mM was analyzed.
A negative control with the pure GST protein was performed to exclude its cyclase activity.

Mass spectrometry analysis of cyclic nucleotides (cGMP and cAMP)
Liquid chromatography-tandem mass spectrometry (LC-MS/MS) was used as the method for the GC/AC activity assay. The enzymatic activity was defined as the amount of cGMP/cAMP produced by 1 mg of protein per minute.
LC-MS/MS experiments were performed using the Nexera UHPLC and LCMS-8045 integrated system (Shimadzu Corporation). The ionization source parameters were optimized in positive ESI mode using pure cGMP or cAMP dissolved in HPLC-grade water (Sigma). Samples were separated using a reversed-phase C18 column (150 × 2.1 mm, 2.6 µm, Kinetex) in 10% methanol with 0.1% (v/v) formic acid) (solvent A (water with 0.1% (v/v) formic acid), solvent B methanol with 0.1% (v/v) formic acid) with a flow rate of 0.3 mL/min. The interface voltage was set at 4.0 kV for positive electrospray (ES+). Data acquisition and analysis were performed with the LabSolutions workstation for LCMS-8045.
The calculation of cyclic nucleotide content was based on standard curves for cGMP and cAMP ranging from 0 to 200 pg (0-0.7 pmol).

Kinase activity assay
For quantification of the BdERL1 protein kinase activity, a Kinase-Glo plus luminescent kinase assay (V3771, Promega) was used according to the manufacturer's protocol. Briefly, purified intracellular domain of BdERL1 (1 µg) was added to a 50 µL reaction mixture (25 mM Tris-HCl, pH 7.5, 5 mM MgCl 2 , 1 mM DTT, 1 µg/µL Histone Type III-S, and 10 µM ATP supplied with 0.1 or 1 µM cGMP or 1 µM Ca 2+ ions and/or calmodulin (At-CaM1; 0.5-10 µg)). Reactions were performed at 30 • C for 15 min and stopped by adding Kinase-Glo reagent. After equilibrating the mixture at room temperature for 10 min, the luminescence was monitored using a Synergy HT Multi-Mode Microplate Reader. The results are expressed as relative luminesce units, and experiments were performed in triplicate.

Identification of a GC catalytic motif in BdERL1
A highly conserved and stringent GC search motif for higher plants was constructed, and its continuous refinement in recent years has allowed for the discovery of a number of functional candidate GCs in the plant kingdom [13][14][15][16][17][18][19][20]. Based on bioinformatics prediction of the GCPred tool (http://gcpred.com), the BdERL1 protein was identified as a potential new guanylyl cyclase harboring a conserved GC motif that is localized within the cytosolic C-terminus of the studied protein (BdERL1 GC center 840−853 ) [21]. It is a common feature for all identified transmembrane plant guanylyl cyclases that a conserved GC center is embedded at a moonlighting site within an intracellular kinase domain. The GC motif present within BdERL1 [840-SFGIVLLELLTGKK-853] possesses all three highly conserved functional amino acids localized at positions 1, 3 and 14 (Fig. 1A), where serine at position 1 forms a hydrogen bond with guanine, glycine at position 3 confers substrate specificity for GTP and lysine at position 14 stabilizes the transition state from GTP to cGMP. Additionally, two amino acids away from the C-terminal end of the GC sequence is aspartic acid [D], which is responsible for interacting with Mg 2+ /Mn 2+ ions [7,8,19,20] (Fig. 1B). Moreover, bioinformatics analysis of the amino acid sequence of BdERL1 revealed that the calmodulin-binding site is localized within the kinase domain of the protein and did not show the presence of the EF hand motif (http: //calcium.uhnres.utoronto.ca/ctdb/ctdb/sequence.html; http s://www.uniprot.org/uniprot/I1H187) (Fig. 1B).
Alignment of the BdERL1 aa sequence with the three most experimentally studied ERECTA proteins from A. thaliana (AtER At2g26330; AtERL1 At5g62230; AtERL2 At5g07180) revealed the highest (72%) identity with the AtERL2 protein. All of these proteins possess a 14 aa long GC motif with highly conserved functional residues and architecture characteristics for moonlighting kinases. The significant level of similarity suggests common (moonlighting) functions for these proteins in both plant species. It was revealed that AtERL2 is involved in the determination of epidermal cell division, i.e., whether the cell should divide proliferatively to produce pavement cells or divide asymmetrically to generate stomatal complexes [22]. The GC activity has not been studied in any ERECTA protein from A. thaliana thus far; therefore, BdERL1 is the first experimentally confirmed guanylyl cyclase belonging to the ERECTA family.

Computational assessment of the BdERL1 GC domain
We performed computational assessments using structural and molecular docking strategies to determine the catalytic feasibility of the BdERL1 GC domain. Since the BdERL1 kinase domain has 44% amino acid identity to AtBRI1 and an even higher identity at the corresponding GC center (78%), we modeled BdERL1 against the At-BRI1 crystal structure (PDB ID: 5LPV) [9]. We show in this model that the GC catalytic center is solvent exposed, allowing for substrate interactions and also presumably for catalysis ( Fig. 2A). Furthermore, the GC center of BdERL1 contains a characteristic alpha-helix barrel followed by a loop secondary fold that is an innate feature of experimentally validated plant GCs and ACs identified by this motifbased approach [20]. Further probing of the GC center by molecular docking of GTP suggested that GTP can dock at the GC center with a good free energy value (-5.6 kcal/mol) and a favorable binding pose where the negatively charged phosphate end of GTP points toward K853 (position 14 of the motif) while the guanine end is surrounded by negatively charged residues including S840 (position 1 of the motif) (Fig. 2), similar to the structurally resolved GC centers [23]. We also performed the same docking simulation with ATP and found that the mean binding affinity was lower than that obtained with GTP (data not shown).

In vitro analysis of the GC activity of wild-type BdERL1
To test whether BdERL1 is able to generate cGMP in vitro, the fragment (BdERL1 648−982 ) containing the predicted GC center was first expressed in E. coli and affinity purified. The molecular mass of recombinant GST-BdERL1 protein predicted in silico was 64.3 kDa (https: //web.expasy.org/cgi-bin/compute_pi). The prediction was consistent with the molecular weight of the GST-BdERL1 protein band observed on the polyacrylamide gel (Fig. 3).
Experimental analysis confirmed the bioinformatics prediction of the webtool GCPred [21] as well as the subsequent computational assessments involving both structural assessments and molecular docking simulations, thus suggesting that BdERL1 can act as a functional guanylyl cyclase. The highest GC activity of the BdERL1 protein was observed in the presence of 1.5 mM GTP, magnesium, and manganese ions, where the cGMP concentration reached its maximum of~70 pmol mg protein −1 ·min −1 at 30 • C. The enzyme exhibited a K m of 0.294 mM for GTP, and the V max was 72 fmol cGMP µg protein −1 ·min −1 (Fig. 4A,  Fig. 5). The BdERL1 activity was similar or higher compared to that of the results previously determined for other plant so-called moonlight proteins with guanylyl cyclase  activity, BdPepR2 [24], HpPepR1 [18], AtPepR1 [25], At-BRI1 [26]. However, there have been some concerns about the biological significance of the low amounts of cGMP and reduce GC activity compared to their animal counterparts. The role of guanylyl cyclase activity in moonlighting proteins has been discussed and questioned as other studies using less sensitive methods have not detected cGMP in case of AtBRI1 [9,27].
Moreover, the BdERL1 protein was also able to convert ATP to cyclic AMP, but the catalytic activity was approximately two times lower than that with GTP as the substrate (Fig. 4B). This suggests a significant reduction in affinity for ATP at the catalytic center, which is consistent with the data obtained from the molecular docking simulations, which also predicted a lower binding affinity for ATP compared to GTP at the catalytic center. Crossreactivity is not uncommon for plant GCs and ACs since substrate discrimination is normally poor for moonlighting centers, and this innate characteristic has been described in previous reports [13,16,17]. Notably, the addition of an equivalent amount of ATP and GTP (1:1 ratio) into the reaction mixture generated cGMP and cAMP in amounts that were both lower than those achieved by their respective substrates alone, although the concentration of the cAMP product remained two times lower than that of cGMP, thus clearly showing a preference for GTP over ATP as the substrate (Fig. 4B). Since nucleotide triphosphates share very Previous reports have shown that the activity of plant transmembrane GCs called moonlight proteins can be modulated by calcium ions [28][29][30]. In vitro experiments have shown that the GC activity of the phytosulfokine receptor AtPSKR1 is significantly enhanced by calcium at physiological levels (0.1-10 µM), and more impor-tantly, the same concentration of Ca 2+ that activates GC activity also exerts an inhibitory effect on kinase activity [31]. The authors suggested that Ca 2+ can play the role of a molecular switch for the guanylyl cyclase/kinase activity of select plant proteins. Based on previous studies, we therefore examined the effects of calcium (0-10 µM) on the GC activity of BdERL1. The results showed a subtle but statistically significant impact from Ca 2+ (1 µM)  on the cGMP level produced by BdERL1 (Fig. 4C). Since no known direct calcium ion binding sites have been detected for plant GCs, we decided to examine the possibility of BdERL1 regulation via the Ca 2+ /CaM calmodulin pathway [28,31]. Bioinformatics analysis of the amino acid sequence of BdERL1 revealed that the calmodulinbinding site is localized within the kinase domain of the protein (http://calcium.uhnres.utoronto.ca/ctdb/ctdb/sequence. html) (Fig. 1B). Therefore, we experimentally tested the effect of calmodulin isoform 1 (AtCaM1) from A. thaliana on the GC activity of BdERL1. However, the GC activity of BdERL1 was not affected by calmodulin (Fig. 4C). Additionally, the amount of AtCaM1 (0.5-10 µg) also had no effect, the activity remained the same level (data not shown). The 3D model of BdERL1 also revealed the absence of a notable cleft or cavity that may accommodate CaM. It is therefore likely that Ca 2+ modulation of moonlighting plant GCs is not achieved through the Ca 2+ /CaM calmodulin pathway (Fig. 2C).

In vitro analysis of the GC activity of BdERL1 mutants
The wild-type (WT) BdERL1 protein (without amino acid sequence changes) and two BdERL1 proteins with mutations to the GC catalytic center were generated. It was previously described in an animal retinal guanylyl cyclase that substitution of the amino acid that confers substrate specificity can convert a GC into an AC [32]. In the case of plant GCs, the amino acid at position 3 of the motif was hypothesized to confer substrate specificity [20,33], and therefore, we mutated this amino acid (G 842 ) in BdERL1 to the negatively charged aspartic acid residue (D) to investigate whether this mutation affects the GC and AC catalytic activities. As previously described, sitedirected mutagenesis by substituting amino acids at the positions responsible for substrate specificity can turn a GC into an AC and vice versa [8,32,34]. However, our results showed that mutation at position 3 of BdERL1 (G 842 D) did not affect GC and AC activities, as the cGMP and cAMP levels produced by the mutant protein were comparable to those produced by WT BdERL1 (Fig. 6). This result implies that the catalytic activity of BdERL1 is regulated by other means, such as localized substrate concentrations, which could in turn behave transiently to precisely modulate the GC/AC activities of BdERL1 in response to upstream signals, much like in human interleukin-1 receptor associated kinase 3 (IRAK3) GC, which was also identified by this motif-based approach [35]. Interestingly, a similar result was observed with the ortholog AtBRI1, in which BdERL1 shares high identity with both the kinase and GC regions. In AtBRI1, mutating glycine at position 3 of the catalytic center to a positively charged lysine did not yield any change in GC activity compared to the wild-type en-zyme [23], although this mutation in AtPSKR1 GC resulted in reduced GC activity [14], an effect that was confirmed in a separate subsequent study by [29]. Notably, another study on the GC in AtPNP-R1 (a natriuretic peptide receptor from A. thaliana) revealed that threonine at position 3 of the motif, when substituted with the negatively charged glutamic acid, resulted in a nearly 4-fold reduction in the cGMP level but had no effect on AC activity; however, mutating threonine to aspartic acid did increase the cAMP generation of this AtPNP-R1 GC by as much as threefold [16]. Clearly, mutation to highly conserved amino acids at the GC catalytic center can alter the tertiary structure and surface charges or the hydrophobicity of the cavity responsible for substrate binding to different extents, and as such, candidate GC or AC proteins must be evaluated both structurally and experimentally for substrate preference and substrate binding affinities [8].
In a second mutated BdERL1 protein, the amino acid lysine (K) at position 14 was mutated to alanine (A) (K 853 A). Based on the 3D model of BdERL1, it can be assumed that lysine at position 14 plays a very important role in the interaction with the negatively charged phosphate of GTP. Computational simulations of GTP docking with the GC catalytic centers of four previously identified plant guanylyl cyclases (AtPSKR1, AtPEPR1, At-BRI1, AtWAKL10) all revealed that mutation of the amino acid at position 14 to leucine resulted in reduced affinity for GTP [8]. Indeed, in vitro analysis of the enzymatic activity of the BdERL1(K 853 A) mutated protein showed that both the cGMP (~30 pmol mg protein −1 ·min −1 ) and cAMP (~15 pmol mg protein −1 ·min −1 ) levels generated by the BdERL1(K 853 A) mutant protein were two times lower than those generated by WT BdERL1 (Fig. 6). Consistently, previous experimental findings have shown that mutating these highly conserved positively charged residues to AA in At-BRI1 resulted in a marked decrease in GC activity [23]. Therefore, both in silico and in vitro analyses suggest that a positively charged amino acid [K or R] at position 14 of the GC catalytic center is crucial for optimal GTP/ATP catalysis.

In vitro analysis of the kinase activity of wild-type BdERL1
The architecture of previously characterized LRR-RLKs reveals that they are moonlighting proteins with GCs embedded within the primary kinase domain [36,37], and sequence analysis has also disclosed the same domain organization for BdERL1. Here, we showed that BdERL1 is also a functional kinase, and more importantly, its activity is inhibited by cGMP (Fig. 7A). The addition of cGMP at concentrations of 0.1 or 1 µM to the reaction mixture resulted in a twofold reduction in BdERL1 kinase activity. This result affirmed our hypothesis that cGMP, as a product of the moonlighting GC domain, can directly modulate kinase activity, an intramolecular regulatory feature that is consistent with moonlighting domains in complex proteins [17,35,38,39]. This regulatory effect has also been observed in previously characterized GC kinases AtPSKR1 and AtBRI1 [14,23].
Calcium ions were previously proposed to be a bimodal switch between GC and the kinase activity of AtP-SKR1 [29]. The kinase activity of AtPSKR1 was directly inhibited by free calcium ions, and its guanylyl cyclase activity was enhanced at similar calcium concentrations. Consistent with the findings of AtPSKR1, the molecular switch function of calcium also operates in BdERL1, where Ca 2+ reduces the kinase activity while acting as an activator of the moonlighting GC domain, which will in turn generate more cGMP to reduce its kinase activity even further (Fig. 7A). In plants, calcium may directly or indirectly regulate protein kinases [30,40], while bioinformatics and in vitro analyses have also revealed the presence of a calmodulin-binding motif within the kinase domains of AtPSKR1 and AtBRI1; thus, it is conceivable that calcium may modulate the cyclase/kinase activities indirectly via CaM isoforms [23,31]. However, although Ca 2+ significantly inhibited the kinase activity of BdERL1, there was no significant difference in its activity with the addition of calcium alone or with the addition of both calcium and calmodulin, as both of these conditions inhibited the kinase to a similar extent (Fig. 7B). Moreover, the amount of AtCaM1 (0.5-10 µg) also had no effect, the activity remained the same level (data not shown). Although bioinformatics analysis has suggested the presence of a calmodulin-binding site in the kinase domain, the structure lacks a clear cavity for the accommodation of CaM, as revealed by our structural analysis with the BdERL1 model (Fig. 2C).

Conclusions
In conclusion, we employed a combination of computational and experimental approaches to identify a functional GC that is embedded within the kinase of BdERL1. Moreover, our mutagenesis studies revealed the catalytic roles of key conserved amino acid residues at the GC center. Importantly, probing the kinase and GC with Ca 2+ and/or cGMP shed light on the intramolecular regulation of BdERL1, thus confirming the dual activity of its cytoplasmic region. Given the involvement of ERL1 in inflorescence architecture, stomatal patterning and phloemxylem organization [41][42][43], our results not only reveal an important signaling component in the LRR-RLK family of proteins but can also guide research that focuses on ERECTA-dependent receptor-ligand signaling during plant development and responses to biotic and abiotic stresses [42,44].

Author contributions
All authors have read and approved the final manuscript.

Ethics approval and consent to participate
Not applicable.

Acknowledgment
Not applicable.

Conflict of interest
The authors declare no conflict of interest.