Introduction

Linker histone (LH) protein or H1 is one of the five main histone protein families that are components of chromatin in eukaryotic cells (Luger et al. 1997; Davey et al. 2002; Alberts et al. 2002; Albert et al. 2007). Other protein families are H2A, H2B, H3 and H4, which exist in pairs (octamer) together with DNA wound around them, forming a nucleosome (Alberts et al. 2002; Albert 2007; Dekker 2008; Naumova et al. 2013). This nucleosome component or nucleosome ‘bead’ is actually the basic unit of chromatin or simply the first level of chromatin compaction (Alberts et al. 2002; Albert et al. 2007). Unlike the other histone proteins, LH does not really make up the nucleosome ‘bead’; instead, it positions itself on top of the structure, keeping in place the DNA that is wrapped around the nucleosome. In terms of protein composition in the structure, LH constitutes half the amount of the other individual four histone proteins, which principally contribute two molecules each to the nucleosome ‘bead’. In addition to binding to the nucleosome, LH binds to the ‘linker DNA’ (approximately 20–80 nucleotides in length) region between nucleosomes, helping stabilize the zig-zagged 30-nm chromatin fibre (Jeon and Berezney 1995).

While most of the LH protein in the nucleus is bound to chromatin, LH molecules shuttle between chromatin regions at a fairly high rate (Misteli et al. 2000; Chen et al. 2006). It is absolutely difficult to understand how such a dynamic protein could be a structural component of chromatin. Apparently, it has been suggested that the steady-state equilibrium within the nucleus still strongly favours association between LH and chromatin. This then means that despite its dynamics, the vast majority of this protein, at any given time point, is chromatin-bound (Bustin et al. 2005). LH compacts and stabilizes DNA under force and during chromatin assembly. This thus suggests that the dynamic binding of this protein may provide protection for DNA in situations where nucleosomes need to be removed (Xiao et al. 2012). Another important feature of the LH protein family is its heterogeneity. Multiple LH sub-types exist and are expressed in organisms as diverse as plants and humans, with eleven of them, namely, H1.0, H1.1, H1.2, H1.3, H1.4, H1.5, H1t, H1T2, H100, H1LS1 and H1X, found in mammals (Parseghian et al. 1994).

Like the other histone proteins, the LH family is extensively post-translationally modified. Such post-translational modifications (PTMs) include serine and threonine phosphorylation, lysine acetylation, lysine methylation, ADP ribosylation, ubiquitination, formylation, PARylation and O-glycosylation (Poirier et al. 1982; Garcia et al. 2004; Villar-Garea and Imhof 2006; Jiang et al. 2007; Wiśniewski et al. 2007, 2008; Deterding et al. 2008; Snijders et al. 2008; Lu et al. 2009; Bonet-Costa et al. 2012; Kim et al. 2015; Sarg et al. 2015). These PTMs are involved in the coordination of a variety of processes in the cell, although such involvement is somewhat less studied in LH compared to the other histone proteins. Nonetheless, LH phosphorylation and O-glycosylation have been shown to play a key role in chromatin compaction and remodelling; heterochromatin formation; DNA replication, transcription and repair; apoptosis; microtubule organization; and protein expression and PTM (Thoma and Koller 1977; Thoma et al. 1979; van Holde and Zlatanova 1996; Calikowski et al. 2000; Kim et al. 2012; Harshman et al. 2013). In addition, LH is also involved in other key cellular and biological processes (Hergeth and Schneider 2015) such as embryogenesis, where it controls the expression of pluripotency genes (Tanaka et al. 2003); reproduction, where it controls the differentiation of sperm cells (Martianov et al. 2005); and disease resistance, where it regulates the innate immune and stress response genes (Studencka et al. 2011).

However, despite the fact that all these processes listed above are typically mediated by the second messenger molecule 3ʹ,5ʹ-cyclic adenosine monophosphate (cAMP) generated by adenylyl cyclases (ACs) (Ito et al. 2014; Kasahara et al. 2016; Vaz Diaz et al. 2019), no histone protein to date, including LH, has been shown to possess AC activity. The only information available so far is the existence of a LH-like or HON4 protein (encoded by the At3g18035 gene) in Arabidopsis thaliana (AtLHL), annotated to be an AC (Gehring 2010). Therefore, focusing on these premises, we targeted AtLHL in this study to determine if it is an AC and perhaps to be able to at least better contextualize its currently known functions.

Materials and Methods

At3g18035 Gene Sequence and AtLHL Protein Sequence Analysis

Complete copy DNA (cDNA) and amino acid sequences of At3g18035 and AtLHL respectively, were retrieved from The Arabidopsis Information Resource (TAIR) (https://www.arabidopsis.org/), followed by analysis of the AtLHL sequence for presence of the AC catalytic centre (Gehring 2010) using the PROSITE database located within the Expert Protein Analysis System (ExPASy) proteomics server (https://www.expasy.org/). In addition, both the presence and location of the AC centre in AtLHL were further confirmed by ACPred, available at http://gcpred.com/acpred/ (Xu et al. 2018).

Computational Analysis of the AtLHL Protein

A 3-dimensional (3D) model of the AtLHL protein was constructed by artificial intelligence using its AlphaFOLD beta version with low predicted error and very high confidence (pLDDT > 90) (Varadi et al. 2022). This software uses a neural network-based model of artificial intelligence to predict protein structures from their amino acid sequences at an atomic level of accuracy. It first aligns the amino acid sequence input with sequences of known structures for pair-wise representation. The representation is then used to produce atomic coordinates for each residue, thus predicting the necessary rotation and then assembling a structured chain of amino acid residues. Its developers freely provide the source code for access to trained modellers and a script for predicting structures of novel input sequences (Varadi et al. 2022). In our case, the full-length amino acid sequence of AtLHL was submitted to the AlphaFOLD database followed by downloading of the model with the highest quality (based on C-scores). The downloaded model was then visualized and analysed using UCSF ChimeraX next-generation molecular visualization program (v.1.10.1.) (Pettersen et al. 2021). SeeSAR 3D (v.12.0.1) desktop modelling platform was next used to perform docking of ATP (PubChem ID: 5957) to the AC centre of the selected AtLHL model via FlexX docking functionality (Gastreich et al. 2006; Trott and Olson 2010). A structural alignment was conducted by fragment assembly simulations based on iterative templates using the iterative threading assembly refinement (I-TASSER) server to match AtLHL to an experimentally confirmed structure in the PDB library (Zhang 2008). The model with the highest C-score was analysed using PyMOL (v.1.7.4.) (Schrödinger LLC, New York, USA) and then adopted in the study.

Cloning of the At3g18035 Gene

Total RNA was extracted from 6-week-old A. thaliana ecotype Columbia-0 (Col-0) seedlings using the RNeasy plant mini kit, in combination with DNase 1 treatment, as instructed by the manufacturer (Qiagen, Crawley, UK). At3g18035 cDNA synthesis from the total RNA and subsequent amplification of the AtLHL301−480 gene fragment from the cDNA, were simultaneously performed in the presence of two sequence-specific primers (forward: 5ʹ-GGAAGGCCTAGGAGAGTTGTTGACCCTAGC-3ʹ and reverse: 5ʹ-GAACAGAGCTTCTTGCATTGCCTCTGCTTC-3ʹ), using a Verso 1-Step RT-PCR kit and in accordance with the manufacturer’s instructions (Thermo Scientific, Rockford, USA). The PCR product was then cloned into a pTrcHis2-TOPO expression vector via the TA cloning system (Invitrogen Corp., Carlsbad, USA) to make a pTrcHis2-TOPO:AtLHL301−480 fusion expression construct with a C-terminus His purification tag.

Expression of the AtLHL301−480 Protein

For expression of the recombinant AtLHL301−480 protein, competent Escherichia coli BL21 Star pLysS cells (Invitrogen, Carlsbad, USA) were transformed (through heat shock at 42 °C for 2 min) with the pCRT7/NT-TOPO:AtLHL301−480 fusion construct and grown in double strength yeast-tryptone (2YT) media (16 g/L tryptone, 10 g/L yeast extract, 5 g/L NaCl and 4 g/L glucose; pH 7.0) containing 100 µg/ml ampicillin and 34 µg/ml chloramphenicol, on an orbital shaker (250 rpm) at 37 °C. Protein expression was induced by the addition of isopropyl-β-D-thiogalactopyranoside (IPTG, Sigma-Aldrich Corp., MO, USA) at a final concentration of 1 mM and when the optical density (OD600) of the cell culture had reached 0.5 (approximately 3 h). The culture was then left to grow for a further 3 h at 37 °C.

Purification of the AtLHL301−480 Protein

The resultant expressed recombinant AtLHL301−480 protein was purified by preparing a cleared cell lysate of the induced E. coli cells under non-native denaturing conditions, whereby the harvested cells were resuspended in lysis buffer (8 M urea, 100 mM NaH2PO4, 10 mM Tris-Cl; pH 8.0, 500 mM NaCl, 20 mM β-mercaptoethanol, 7.5% (v/v) glycerol) at a ratio of 1 g pellet weight to every 10 ml buffer volume, mixed thoroughly using a mechanical stirrer at 24 °C for 1 h and then centrifuged at 2,500 × g for 15 min. Supernatant was collected as the cleared lysate and transferred to 2 ml of 50% (w/v) nickel-nitriloacetic acid (Ni-NTA) slurry (Sigma-Aldrich Corp., MO, USA) that had been pre-equilibrated with 10 ml of lysis buffer and the lysate/slurry mixture then gently swirled on a rotary mixer for 1 h at 24 °C. This step allowed for binding of the AtLHL301−480 protein onto the Ni-NTA resin. The lysate–resin mixture was loaded into an empty XK16 column (Bio-Rad Laboratories Inc., CA, USA) and allowed to settle and flow-through discarded. The protein-bound resin was then washed three times with 30 ml of wash buffer (8 M urea, 100 mM NaH2PO4, 10 mM Tris-Cl; pH 8.0, 500 mM NaCl, 20 mM β-mercaptoethanol, 7.5% (v/v) glycerol and 40 mM imidazole) to remove unbound proteins.

Refolding of the AtLHL301−480 Protein

The washed protein-bound resin was equilibrated with 2 ml of gradient buffer (8 M urea, 200 mM NaCl, 50 mM Tris-Cl; pH 8.0 and 20 mM β-mercaptoethanol) before the column was connected to a Bio-Logic F40 Duo-Flow chromatography system (Bio-Rad Laboratories Inc., CA, USA) programmed to run a linear refolding gradient. The refolding gradient for the denatured recombinant AtLHL301−480 was then performed by linearly diluting the 8 M gradient buffer to 0 M urea concentration with a refolding buffer (200 mM NaCl, 50 mM Tris-Cl; pH 8.0, 500 mM glucose, 0.05% (w/v) polyethyl glycol, 4 mM reduced glutathione, 0.04 mM oxidized glutathione, 100 mM non-detergent sulfobetaine and 0.5 mM phenylmethanesulfonyl fluoride (PMSF)) over 10 h at a flow rate of 0.5 ml/min. After refolding, the renatured recombinant AtLHL301−480 was eluted in 2 ml of elution buffer (200 mM NaCl, 50 mM Tris-Cl; pH 8.0, 250 mM imidazole, 20% (v/v) glycerol and 0.5 mM PMSF). The eluted native protein fraction was then de-salted and concentrated using a Spin-XUF filtration/concentration device with a molecular weight cut-off (MWCO) point of 3000 Da and in accordance with the manufacturer’s instructions (Corning Corp., NY, USA). Protein concentration was determined by the Bradford method (Bradford 1976) and ND2000 nanodrop spectrophotometer (Thermo Scientific Inc., MA, USA) before the recombinant protein was stored at −20 °C.

Testing for the In Vitro AC Activity of AtLHL301−480

The probable in vitro AC activity of the purified recombinant AtLHL301−480 was tested by incubating 5 µg of the protein in 50 mM Tris-Cl (pH 8.0) containing 5 mM Mg2+ or Mn2+ and 1 mM ATP, with or without 250 µM Ca2+ or 50 mM HCO3−, in a final volume of 200 µl, followed by measurement of the generated cAMP. Background cAMP levels in control reactions were measured in tubes containing all the other components but no protein or Ca2+ or HCO3−. All incubations were performed at room temperature (24 °C) for 20 min and terminated by the addition of 10 mM EDTA followed by boiling for 3 min and cooling on ice for 2 min before centrifugation at 2,500 × g for 3 min. The resulting supernatants were assayed for cAMP content using the cAMP-linked enzyme linked immunosorbent assay (ELISA) kit, following its acetylation protocol and as is described by the supplier’s manual (Sigma-Aldrich Corp., MO, USA; code: CA201). The anti-cAMP antibody in this assaying system is highly specific for cAMP and has approximately a 106 times lower affinity for 3′,5′-cyclic guanosine monophosphate (cGMP). In all cases, each experiment was performed in triplicate (n = 3) using three different protein extracts that had been independently prepared.

Detection of cAMP by Mass Spectrometry

Acetylated cAMP samples from the in vitro AC activity assays were also assayed by tandem liquid chromatography mass spectrometry (LC-MS/MS). In this method, samples were introduced into a Waters API Q-TOF Ultima mass spectrometer (Waters Microsep, Johannesburg, RSA) with a Waters Acquity UPLC at a flow rate of 180 ml/min. Separation was achieved in a Phenomenex Synergi (Torrance, CA) 4 µm Fusion-RP (250 × 2.0 mm) column when a gradient of solvent ‘A’ (0.1% formic acid) and solvent ‘B’ (100% acetonitrile) was applied over 18 min. During the first 7 min, the solvent composition was kept at 100% ‘A’ followed by a linear gradient of up to 80% ‘B’ for 3 min and then a re-equilibration to the initial conditions. An electrospray ionization in the negative (W-) mode was used at a cone voltage of 35 V, to detect molecules and generate chromatograms.

Testing for the Ability of AtLHL301−480 to Complement cyaA Mutation in E. coli

The E. coli mutant strain, SP850 (lam-, el4-, relA1, spoT1, cyaA1400 (:kan), thi-1) (Shah and Peterkofsky 1991; Ullmann and Danchin 1983), deficient in the AC gene (cyaA), was obtained from the E. coli Genetic Stock Centre (Yale University, New Haven, USA; accession No. 7200). The strain was prepared to be chemically competent followed by its transformation with the pTrcHis2-TOPO:AtLHL301−480 fusion construct (through heat shock at 42 °C for 2 min). The transformed bacteria together with the non-transformed cells were then grown at 37 °C in Luria-Bertani (LB) media containing kanamycin (15 µg/ml) up until their cell culture had reached an optical density (OD600) of 0.5. Both groups of cells were streaked on MacConkey agar supplemented with 15 µg/ml kanamycin and 0.5 mM IPTG (Sigma-Aldrich Corp., Missouri, USA) (for transgene induction) before the streaked media was incubated for 40 h at 37 °C, for visual evaluation. After incubation, an ability of the induced transformed mutant cells to now ferment lactose would then be considered as an indication of the expressed recombinant AtLHL301−480’s ability to generate cAMP from ATP, as a functional AC. As a result, the induced transformed cells would turn deep red or purple (just like wild-type cells), while the mutant control cells would remain yellowish or colourless (Shah and Peterkofsky 1991; Ullmann and Danchin 1983).

Electrochemical Evaluation of the In Vitro AC Activity of AtLHL301−480

Various electrodes were prepared in a 20 ml cell system, using GCE bio-electrodes (BioAnalytical Systems, West Lafayette, IN, USA) (polished with 1.0, 0.3 and 0.05 µm alumina (Buehler, IL, USA)) and washed with distilled water before ultrasonication for 5 min in distilled water and 5 min in ethanol, and drying in a stream of N2 for 10 s before drop coating), whereby the control electrode (a 0.071 cm2 glassy carbon (GCE)) was left uncoated, whereas the basal electrode (a Ag/AgCl platform with a 3 M NaCl salt bridge (GCE)) was coated with the AtLHL301−480 protein (5 µg) while test electrodes (auxiliary platinum wires (GCE)) were each coated with the AtLHL301−480 protein (5 µg) pre-incubated with each of the selected and tested AC co-factors (5 mM Mg2+ or Mn2+) or modulators (250 µM Ca2+ or 50 mM HCO3−). These electrodes were then connected to a BAS Epsilon electrochemical workstation (BioAnalytical Systems, West Lafayette, IN, USA), followed by recording of the resultant square wave voltammetries with a computer interface linked to the workstation at a potential scan rate of 2 mVs−1 from the initial potential to the Ei = + 300 mV switch potential, and the Eλ = 350–1000 mV experimental potential. Wherever there was binding of a co-factor or modulator by the protein, the reading of such a test coated electrode would be expected to be higher than that of the uncoated control electrode and that of the basal electrode coated with the AtLHL301−480 alone. All experiment recordings were carried out at 25 °C at a constant amplitude of 25 mV and a fixed frequency of 15 mV. The method is further detailed elsewhere (Mulaudzi et al. 2011).

Searching for Calcium-Binding Sites in AtLHL

Three different methods were used for this task. In the first method, AtLHL sequence was manually searched for presence and frequency of the 9-residue calcium-binding site commonly known as the RTX (repeat-in-toxin) motif, GGXGXDXHX, where X is any amino acid and H is any hydrophobic residue (Grzybowska 2018). In the second method, AtLHL sequence was aligned with five selected Arabidopsis proteins (At1g05990, At3g43810, At2g17290, At5g23580 and At4g23650) containing the EF-hand motif (Grzybowska 2018) to which calcium binds, followed by homology search using MAFFT (https://mafft.cbrc.jp/alignment/server/) (Katoh et al. 2019). In the last method, the prediction program of the Calmodulin Target Database (http://calcium.uhnres.utoronto.ca/) was used to search for presence and frequency of the 17-residue calmodulin-binding site; KLWKKLLKLLKKLLKLG (Kauer et al. 1986) in the AtLHL sequence. The program predicts the calmodulin-binding site sequences based on criteria such as hydropathy, α-helical propensity, residue weight, residue charge, hydrophobic residue content and helical and occurrence of particular residues. In this regard, sequences are scored from 0 to 9, with the most likely binding site assigned a series of 9 s (Yap et al. 2000).

Searching for the AC Centre in Other LH Family Proteins

Amino acid sequences of AtH2A, AtH2B, AtH3 and AtH4 were retrieved from TAIR (https://www.arabidopsis.org/). The retrieved sequences were uploaded to the MAFFT online alignment tool (https://mafft.cbrc.jp/alignment/server/) and then alignment ran with default settings to search for sequence similarities at the AC centre.

Reaction Kinetics of the Recombinant AtLHL301−480 Protein

Individual reaction settings of the increasing concentration of ATP (0.0; 0.5; 1.0; 1.5 and 2.0 mM) were prepared in 200 µl of 50 mM Tris-Cl (pH 8.0) containing 5 µg AtLHL and 5 mM Mn2+, followed by incubation at room temperature (24 °C) for 20 min. After incubation, each reaction was terminated by the addition of 10 mM EDTA followed by boiling for 3 min and cooling on ice for 2 min before centrifugation at 2500 × g for 3 min. The resulting supernatants were then assayed for cAMP content using the cAMP-linked ELISA kit, following its acetylation protocol as is instructed by the supplier (Sigma-Aldrich Corp., MO, USA; code: CA201). The initial velocity for each of the used ATP concentrations was then used to sketch a Hanes plot, followed by determination of the reaction kinetic constants (Km and Vmax) of AtLHL301−480 from the same plot. Km was determined as the negative value of the x-intercept (x = −Km, when y = 0) of a linear fit of the data, while Vmax was calculated from the y-intercept (y = Km/Vmax, when x = 0) of the same linear fit (Irving et al. 2012).

Statistical Analysis

All generated enzyme immunoassaying data was subjected in triplicate sets (n = 3) to one-way analysis of variance (ANOVA) (Super-Anova, Statsgraphics version 7, 1993; Statsgraphics Corp., The Plains, VI, USA). Wherever ANOVA revealed significant differences between treatments, means were separated by post hoc Student–Newman–Keuls (SNK) multiple range test (p < 0.05).

Results

Structure and Computational Analysis of AtLHL and its AC Centre

The identification of ACs in plants has mostly involved querying protein sequences with an AC motif (Fig. 1A) derived from a guanylyl cyclase (GC) search motif (Ludidi and Gehring 2003) through modification at position 3, changing [CTGH] to [DE] (Gehring 2010). This modification is primarily based on previous findings, which indicate that the conversion of GCs into ACs and vice versa could be easily achieved through a single mutation in the amino acid that confers substrate specificity (Tucker et al. 1998; Roelofs et al. 2001). In this study, when the amino acid sequence of AtLHL was queried by the AC motif, a matching hit was detected towards its C-terminal end (amino acids 382–399) (Fig. 1B). We also submitted the whole AtLHL sequence to ACPred, which is available at http://gcpred.com/acpred/ (Xu et al. 2018), for prediction of the AC centre in the AtLHL protein. The software then identified the region K382 to D399 (Fig. 1B) that fits the [RKS]X[DE]X{9,11}[KR]X{1,3}[DE] AC motif (Fig. 1A) and its search hit (Fig. 1B). Next, computational analysis of the AtLHL protein was undertaken to assess and determine the ability of its AC centre to bind ATP and catalyse its subsequent conversion into cAMP. The full-length model of AtLHL was prepared by artificial intelligence followed by docking simulations, which then showed that in this model, the AC centre is solvent-exposed, thus allowing for unimpeded substrate interactions and ultimately catalysis (Zhou et al. 2021) (Fig. 1C–F).

Fig. 1
figure 1

Structural features and computational analysis of AtLHL. A The 14 amino acid AC search motifs derived from annotated and experimentally tested GC and AC catalytic centres (Ludidi and Gehring 2003; Gehring 2010). The residue forming hydrogen bonding with purine at position 1 is highlighted in red; the residue conferring substrate specificity in position 3 is highlighted in blue; while the amino acid in position 14, stabilizing the transition state from ATP to cAMP, is highlighted in red. The amino acid [DE] at 1–3 residue downstream from position 14 participates in Mg2+/Mn2+-binding and is coloured green (Ludidi and Gehring 2003; Gehring 2010). B The complete amino acid sequence of AtLHL with the AC catalytic centre towards its C-terminus (amino acids 382–399) highlighted in bold and underline, and the 180 amino acid sequence fragment expressed and tested for AC activity indicated within the inverted red triangles. C 3D rainbow depiction of the AtLHL ribbon model as affirmed by PyMOL, where the N → C orientation is shown as blue → red. D Surface model (crystal structure) of AtLHL, highlighting the solvent-exposed AC centre (white), wherein residues of the centre are labelled with single letter codes in black. Docking of ATP at the AC centre and interaction of ATP with key residues in the catalytic centre of AtLHL shown as stick (tail) and dots (head) in the ribbon model E, and club (phosphate) and ball (purine) in the surface model F. AtLHL was modelled using AlphaFOLD (Varadi et al. 2022), while ATP docking simulation was performed using the FlexX functionality of SeeSAR (v12.0.1) (Gastreich et al. 2006)

Full size image

Testing and Confirmation of the AC Activity of AtLHL301−480

To assess the AC activity of AtLHL, a fragment sequence of the At3g18035 gene (amino acids 301–480), harbouring the AC motif (Fig. 1B), was cloned into a prokaryotic system and expressed into a 23.300 kDa AtLHL301−480 His-tagged recombinant protein (Fig. 2A). To check if the AC centre of AtLHL can generate cAMP in vitro, the expressed recombinant AtLHL301−480 was affinity purified (Fig. 2B, left inset) and tested in a reaction mixture containing ATP as substrate, Mn2+ or Mg2+ as co-factor and Ca2+ or HCO3− as modulator, followed by measurement of cAMP by enzyme immunoassay. Maximum activity was reached after 20 min of the reaction system (Fig. 2B, right inset), generating about 103.25 fmols/µg protein of cAMP in the presence of Mn2+ and approximately 21.50 fmols/µg protein of cAMP in the presence of Mg2+ compared to only about 15.15 fmols/µg protein of cAMP of the control reaction (Fig. 2B). Besides being strictly Mn2+-dependent, the catalytic activity of AtLHL301−480 was also significantly enhanced by both Ca2+ and HCO3−, reaching activity levels of around 150.00 and 126.50 fmols/µg protein of cAMP respectively when Mn2+ is the co-factor (Fig. 2B). cAMP was also measured by LC-MS/MS, and in a reaction mixture containing AtLHL301−480, ATP and Mn2+, the product was detected (Fig. 2C) at almost 90% level of cAMP (Fig. 2D). This additional method, therefore, confirmed presence of cAMP in the reaction mixture, thus validating the ELISA technique and, at the same time, confirming AC activity for the recombinant AtLHL301−480 protein.

Fig. 2
figure 2

Testing and confirmation of the AC activity of AtLHL301−480. A Sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS-PAGE) of protein fractions (stained with Coomassie brilliant blue) from the induced (IPTG) and un‐induced (Cont.) cell cultures, where (M) is the molecular weight marker and the arrow marking the expressed recombinant AtLHL301−480 protein. B cAMP generated by 5-µg recombinant AtLHL301−480 in the presence of 1 mM ATP and 5 mM Mg2+ or Mn2+ or 1 mM ATP and 250 µM Ca2+ or 50 mM HCO3− when 5 mM Mn2+ ion is the co-factor. Control reaction contained all other components except the protein, Ca2+, and HCO3−. Left insert: a Coomassie brilliant blue-stained gel after resolution of the affinity purified His-tagged recombinant AtLHL301−480 (arrow) by SDS-PAGE. Right insert: time course calibration curve. Data are mean values (n = 3) and error bars show standard error (SE) of the mean. Asterisks denote values significantly different from those of control (p < 0.05) determined by the analysis of variance (ANOVA) and post hoc Student–Newman–Keuls (SNK) multiple range tests. C An extracted mass chromatogram of the m/z 328 [M-1]−1 ion of cAMP generated by 5 µg of AtLHL301−480 in a reaction system containing 50 mM Tris-Cl; pH 8.0, 1 mM ATP, and 5 mM Mn2+ after 20 min. D Mass of the resultant cAMP peak in the chromatogram. E The AC centre of AtLHL301−480 complemented the cyaA mutant E. coli (SP850) to ferment lactose. AtLHL301−480-expressing SP850 E. coli cells showed a strong reddish colour as if they were wild-type cells (Shah and Peterkofsky 1991; Ullmann and Danchin 1983), while cyaA mutant cells yielded yellowish colonies

Full size image

To check if the AC centre of AtLHL can rescue AC-deficiency in E. coli, the AtLHL301−480 protein was expressed in the E. coli host strain, SP850, lacking the AC gene (cyaA), essential for lactose fermentation (Shah and Peterkofsky 1991; Ullmann and Danchin 1983). As a result of this mutation, the SP850 mutant cells remained yellowish in colour when grown on MacConkey agar, while in contrast, the AtLHL301−480-expressing SP850 cells formed deep reddish colonies (Fig. 2E) much like what wild-type E. coli does (Shah and Peterkofsky 1991; Ullmann and Danchin 1983). This thus, indicated a functional AC centre in AtLHL.

Characterization of the AC Activity of AtLHL301−480

To check if AtLHL physically interacts with its catalytic co-factors or modulators during catalysis and/or activation, AtLHL301−480 alone or AtLHL301−480 pre-incubated with individual co-factors or modulators was used to coat electrodes, followed by recording of square wave voltammetries using a computer interface linked to a workstation at a potential scan rate of 2 mVs−1 from the initial potential to the Ei = +300 mV switch potential and the Eλ = 350–1000 mV experimental potential. Peak detection occurred at around 0.23 V, whereby the control bare electrode showed a peak of ~1.27 A, while AtLHL301−480 alone, AtLHL301−480 plus Mg2+, AtLHL301−480 plus Mn2+, AtLHL301−480 plus HCO3− and AtLHL301−480 plus Ca2+ showed peaks of ~1.73, ~1.99, ~2.25, ~3.26 and ~5.00 A, respectively (Fig. 3A), thus signifying physical interaction between AtLHL and its catalytic co-factors or modulators. Since the co-factor binding site of AtLHL was already known to exist within its AC centre (Gehring 2010), we then sought to search for the possible existence of the calcium binding site or sites within this protein so that this could better explain the observed physical interaction between AtLHL301−480 and its catalytic modulators. Using the Calmodulin Target Database program (http://calcium.uhnres.utoronto.ca/) that predicts calmodulin-binding site sequences (Yap et al. 2000) based on various criteria (Kauer et al. 1986) to search the AtLHL amino acid sequence, a single 16-residue calmodulin-binding sequence (IVQASVMAGIMKRRGR) was identified as the most likely binding site for Ca2+ in AtHLH (Fig. 3B). Finally, the reaction kinetics of AtLHL as functional AC enzyme were then calculated, and a Km constant of around 0.7 mM and Vmax constant of approximately 9.2 fmol/min/μg protein were obtained (Fig. 3C).

Fig. 3
figure 3

Characterization of the AC activity of AtLHL301−480. A Square wave voltammograms showing the response of AtLHL301−480 in 5 ml of 50 mM Tris-HCl buffer (pH 8.0) when it is on its own (lime green) or after its pre-incubation with either 5 mM Mg2+ (black) or 5 mM Mn2+ (pink) or 50 mM HCO3− (green) or 250 µM Ca2+ (blue). The square wave voltammogram of the uncoated/bare control electrode is shown in red. The used potential window was −0.35 to 1.0 V, while the scan rate was 2 mVs−1. The scan shows peaks at ~0.23 V. B Prediction of Ca2+ binding sites in AtLHL using the Calmodulin Target Database program, showing a single 16-residue calmodulin-binding sequence (IVQASVMAGIMKRRGR) as the only most likely binding site (blue). C Hanes plot for the determination of the reaction kinetics of AtLHL, whereby Km was determined as the negative value of the x-intercept (x = −Km, when y = 0) of the linear fit and Vmax calculated from the y-intercept (y = Km/Vmax, when x = 0) of the same linear fit

Full size image

Discussion

Adenylyl cyclases (ACs) are enzymes capable of catalysing the conversion of adenosine 5′-triphosphate (ATP) to the second messenger, 3',5'-cyclic adenosine monophosphate (cAMP) (Robison et al. 1968; Goodman et al. 1970; Gerisch et al. 1975). In plants, cAMP in turn controls various downstream processes such as the cell cycle (Ehsan et al. 1998), growth of pollen tubes (Tezuka et al. 1993; Malho et al. 2000; Moutinho et al. 2001; Vaz Dias et al. 2019) and response to stress (Jin and Wu 1999; Ma et al. 2009; Sabetta et al. 2019; Blanco et al. 2020). Across the different species of land plants on earth (ranging from mosses and herbs to woods), a total of twenty-nine ACs have so far been identified, using various methods (Yuan et al. 2022; Liu et al. 2023). Apparently, while methods such as omics analysis and homologous cloning have been successfully used to identify ACs in other plants, the discovery of ACs in Arabidopsis thaliana, Glycine max and Physcomitrella patens has mostly been through a systematic approach, which involved identification of key amino acid residues in the catalytic centre of known and experimentally tested nucleotide cyclases (NCs) (Ludidi and Gehring 2003). In that approach, a GC search motif (Ludidi and Gehring 2003) at position 3 was changed from [CTGH] to [DE] to generate a rationally designed search motif specific for ACs (Fig. 1A) (Gehring 2010). This substitution was based on previous findings, which indicated that the conversion of GCs into ACs and vice versa could be achieved by a single mutation in the amino acid residue that confers substrate specificity (Tucker et al. 1998; Roelofs et al. 2001).

Using this systematic approach, a total of sixteen ACs have been discovered, i.e. eleven in A. thaliana, four in P. patens and only one in G. max. The Arabidopsis ACs are AtPPR-AC (Ruzvidzo et al. 2013), AtKUP7 (Al-Younis et al. 2015), AtClAP (Chatukuta et al. 2018), AtKUP5 (Al-Younis et al. 2018), AtLRRAC1 (Bianchet et al. 2019; Ruzvidzo et al. 2019), AtNCED3 (Al-Younis et al. 2021), AtAC (Sehlabane et al. 2022), AtMEE (Kawadza et al. 2022) and AtTIR1, AtAFB1 and AtAFB5 (Qi et al. 2022), while the moss ACs are PpAFB1, PpAFB2, PpAFB3 and PpAFB4 (Qi et al. 2022), and the soybean AC is GmAC (Bobo et al. 2022). AtPPR-AC is annotated to play a role in chloroplast biogenesis and the restoration of cytoplasmic male sterility (CMS) (Ruzvidzo et al. 2013), while AtClAP is predicted to have a role in endocytosis and plant defence (Chatukuta et al. 2018). The two AtKUPs are responsible for K+ ion flux (Al-Younis et al. 2015, 2018). AtLRRAC1 has a role in pathogen defence (Bianchet et al. 2019; Ruzvidzo et al. 2019), while AtNCED3 is involved in the biosynthesis of the stress hormone abscisic acid (ABA) (Al-Younis et al. 2021). AtAC is known to be transcriptionally upregulated in response to biotic stress (Sehlabane et al. 2022), while AtMEE is involved in embryogenesis and response to abiotic stress (Kawadza et al. 2022). TIR1 and all the AFB ACs from A. thaliana and P. patens are auxin receptors involved in the regulation of root growth mediated by auxin (Qi et al. 2022), while GmAC has a role in early plant development and stress response (Bobo et al. 2022).

Apparently, there are also four other plant ACs, discovered through either the omics analysis or homologous cloning methods, that harbour the same rationally designed AC search motif, i.e. two from Zea mays, one from Nicotiana benthamiana and one from A. thaliana. The maize ACs are ZmPSiP (Moutinho et al. 2001) and ZmRPP13-LK3 (Yang et al. 2021), while the tobacco AC is NbAC (Ito et al. 2014), and the Arabidopsis AC is AtDK4 (Vaz Dias et al. 2019). ZmPSiP is responsible for the polarized growth and re-orientation of pollen tubes (Moutinho et al. 2001), while ZmRPP13-LK3 participates in ABA-mediated resistance to heat stress (Yang et al. 2021). NbAC has a role in tabtoxinine-β-lactam-induced cell death during the development of wildfire disease (Ito et al. 2014) while AtDK4 plays a role in nitric oxide (NO)-dependent pollen tube guidance and fertilization (Vaz Dias et al. 2019).

Notably, besides these twenty ACs harbouring the rationally designed AC search motif across all plants, nine other plant ACs, all lacking the rationally designed AC search motif, are known. These include one from Hippeastrum hybridum, which is HpAC1 (Świeżawska et al. 2014); one from Marchantia polymorpha that is MpAC (Kasahara et al. 2016); two from Brachypodium distachyon, namely, BdTTM3 and BdGUCD1 (Świeżawska et al. 2020; Duszyn et al. 2022); two from Malus domestica, namely, MdTTM1 and MdTTM2 (Yuan et al. 2022); and three from Ziziphus jujuba, which are ZjAC1, ZjAC2 and ZjAC3 (Liu et al. 2023). HpAC1 is involved in responses to infection by the plant fungal pathogen, Phoma narcissi, and also injuries through mechanical damage (Świeżawska et al. 2014), while MpAC has a role in male organ and cell development (Kasahara et al. 2016). BdTTM3 is responsible for responses to mechanical wounding (Świeżawska et al. 2020), while BdGUCD1 is involved in jasmonic acid (JA)-mediated responses to Fusarium pseudograminearum infection (Duszyn et al. 2022). MdTTM1 and MdTTM2 currently do not have any known function(s) (Yuan et al. 2022), while ZjAC1, ZjAC2 and ZjAC3 are involved in the significant acceleration of seed germination, root growth and flowering, respectively (Liu et al. 2023).

Interestingly, in A. thaliana, there exists another additional protein annotated to be an AC at NCBI (https://www.ncbi.nlm.nih.gov/protein/51968402). This protein, termed linker histone-like (AtLHL) or HON4 protein, coded for by the At3g18035 gene, has the rationally designed AC search motif (Fig. 1B), however, has never been functionally confirmed as an AC. This is also despite the fact that the protein is known to be primarily involved in a number of key cellular processes essentially mediated by cAMP. Such processes include (i) chromatin formation, where AtLHL binds to the nucleosome and linker DNA (Jeon and Berezney 1995); (ii) embryogenesis, where the protein controls the expression of pluripotency genes (Martianov et al. 2005); (iii) reproduction, where AtLHL controls the differentiation of sperm cells (Tanaka et al. 2003); and (iv) disease resistance, where the protein regulates the innate immune and stress response genes (Studencka et al. 2011) and drought stress response (Ascenzi and Gantt 1997; Scippa et al. 2000; Wu et al. 2022). Therefore, based on this basis, we then sought to assess and establish if this additional Arabidopsis protein candidate could also be a functional AC.

We then started the assessment by using ACPred, which is a prediction tool designed to identify motif-based AC centres in proteins with multiple domains (Xu et al. 2018), to predict the presence and location of the AC centre in AtLHL. Supported with the 3D model of AtLHL generated by AlphaFOLD using artificial intelligence (Varadi et al. 2022), the identified centre was found to encompass amino acid residues K382 to D399 (Fig. 1C and D), which furthermore, concurred with the information and data from NCBI (https://www.ncbi.nlm.nih.gov/protein/51968402). The FlexX functionality of SeeSAR (v12.0.1) (Gastreich et al. 2006) was then used to carry out docking simulations of the substrate ATP at this AC centre, and the software predicted good affinity for the ATP at the centre (Fig. 1E and F). Additionally, both the artificial intelligence and SeeSAR showed that in the AtLHL model, the AC centre is solvent-exposed, thus indicating its unimpeded access to the centre and ultimately catalysis (Zhou et al. 2021). Notably, this same outcome is so much consistent with what was also obtained previously for AtKUP7 (Al-Younis et al. 2015), AtClAP (Chatukuta et al. 2018), AtKUP5 (Al-Younis et al. 2018), AtLRRAC1 (Bianchet et al. 2019; Ruzvidzo et al. 2019), AtNCED3 (Al-Younis et al. 2021), AtAC (Sehlabane et al. 2022) and MdTTM1 and MdTTM2 (Yuan et al. 2022).

We then cloned a fragment of the At3g18035 gene and expressed a truncated version of the AtLHL protein (AtLHL301−480) harbouring the AC search motif (Fig. 1B) as a His-tagged fusion recombinant product of approximately 23.300 kDa (Fig. 2A). When purified (Fig. 2B, left inset) and tested for in vitro AC activity, using ELISA, after a 20-min reaction time (Fig. 2B, right inset), the recombinant AtLHL301−480 showed a Mn2+-dependent activity that is positively enhanced by Ca2+ and HCO3− ions (Fig. 2B). This very same result was also obtained via LC-MS/MS (Fig. 2C and D), another analytical method capable of specifically and sensitively detecting cAMP levels at femtomolar concentrations, thus validating the ELISA technique and also confirming the AC activity of the AtLHL301−480 recombinant protein.

In order to validate the AC activity of AtLHL (detected by ELISA and confirmed by LM-MS/MS), recombinant AtLHL301−480 was expressed in a cyaA Escherichia coli mutant strain (SP850) to see if it could rescue the mutant. This mutant strain has a catabolic defect of lacking the only AC system available in E. coli, necessary for lactose fermentation; therefore, its rescue by any foreign protein to metabolize lactose signifies AC function for such a protein (Shah and Peterkofsky 1991; Ullmann and Danchin 1983). In our case, AtLHL301−480 rescued the SP850 strain (Fig. 2E), thereby validating the AC function of AtLHL. Notably, this outcome is also very much consistent with what was previously obtained for ZmPSiP (Moutinho et al. 2001), AtPPR-AC (Ruzvidzo et al. 2013), HpAC1 (Świeżawska et al. 2014), AtKUP7 (Al-Younis et al. 2015), MpAC (Kasahara et al. 2016), AtClAP (Chatukuta et al. 2018), AtKUP5 (Al-Younis et al. 2018), AtLRRAC1 (Bianchet et al. 2019; Ruzvidzo et al. 2019), ZmRPP13-LK3 (Yang et al. 2021), BdTTM3 (Świeżawska et al. 2020), AtNCED3 (Al-Younis et al. 2021), AtAC (Sehlabane et al. 2022), MdTTM1 and MdTTM2 (Yuan et al. 2022), AtAFB1 and AtAFB5 (Qi et al. 2022), GmAC (Bobo et al. 2022), AtMEE (Kawadza et al. 2022) and ZjAC1, ZjAC2 and ZjAC3 (Liu et al. 2023).

Apparently, the observed outcome whereby AtLHL301−480 exhibited a relatively higher (~4.6-fold) in vitro AC activity with Mn2+ as opposed to Mg2+ (Fig. 2B), proposes that AtLHL could be a soluble AC (sAC) because all sACs prefer Mn2+ to Mg2+ ion as a co-factor of activity and are intracellularly localized (Braun and Dods 1975; Steer and Levitzki 1975; Sehlabane et al. 2022). In A. thaliana, AtLHL is localized in the nucleus (https://www.arabidopsis.org/) just like all other known sACs (Sehlabane et al. 2022). Moreover, the subsequent activation of AtLHL301−480 by both calcium (~1.5-fold) and hydrogen carbonate (~1.2-fold) (Fig. 2B), further points to our earlier assumption that AtLHL could be a sAC because only sACs and not transmembrane ACs (tmACs) are functionally activated by the Ca2+ and HCO3− ions (Chen et al. 2000; Kamenetsky et al. 2006; Sehlabane et al. 2022); whereby activation by Ca2+ is through physical interaction via a Ca2+-binding protein (calmodulin) (Kamenetsky et al. 2006), whereas that by HCO3− is thought to be through the alteration of pH (Chen et al. 2000). Interestingly, the molecular interaction of AtLHL301−480 with either its co-factors (Mg2+ and Mn2+) or modulators (Ca2+ and HCO3−) during its catalysis or activation, respectively, was actually found to be physical (Mulaudzi et al. 2011) (Fig. 3A), proposing for the possible presence of Mg2+, Mn2+, Ca2+ and/or HCO3− binding sites in this protein.

Arguably, the probable existence of Ca2+ binding sites in AtLHL and its associated activation by this very same metal ion, strongly prompted our interest to search for such sites in the protein. The search criterion used targeted for the presence and frequency of any or all of the calcium-binding targets, i.e. (i) the EF-hand domain (Grzybowska 2018), (ii) the 9-residue RTX motif (GGXGXDXHX) (Grzybowska 2018) and (iii) the 17-residue calmodulin-binding sequence (e.g. KLWKKLLKLLKKLLKLG) (Kauer et al. 1986). As is shown in Fig. 3B, a single 16-residue calmodulin-binding sequence (IVQASVMAGIMKRRGR) was picked up between amino acids 251 and 268 and predicted to be the most likely binding site for Ca2+ in AtLHL (Yap et al. 2000). This outcome, besides strongly supporting our findings from the electrochemical analysis, also consistently and firmly corresponded with both the established binding properties and known key functions of AtLHL as an H1 family protein, whereby it is specifically responsible for binding to the nucleosome and linker DNA of the chromatin structure (Jeon and Berezney 1995).

Finally, after unequivocally establishing that AtLHL is a bona fide AC protein, we then went on to assess, determine and evaluate its reaction kinetics as an enzyme. As is seen in Fig. 3C, AtLHL was found to have a Km constant of around 0.7 mM and a Vmax constant of approximately 9.2 fmol/min/μg protein, comparable to other previously confirmed plant ACs. AtTIR1, AtAFB1 and AtAFB5 have Km constants of 0.644, 0.602 and 0.675 mM and Vmax constants of 7.462, 8.615 and 10.45 fmol/min/μg protein, respectively (Qi et al. 2022), while AtKUP7, AtClAP and MdTTM2 have Vmax constants of 2.2, 7.3 and 8.3 fmol/min/μg protein, respectively (Al-Younis et al. 2015; Chatukuta et al. 2018; Yuan et al. 2022). Apparently, all plant ACs seem to have very low kinetics levels (high Km and low Vmax constants), suggesting that AC activity is not their main function but rather a secondary function as multi-functional or moonlighting proteins. The moonlight nature is typically exemplified in proteins such as AtKUP7, MpCAPE, AtNCED3, MdTTM1, AtDGK4 and AtTIR1, where AC activity is respectively combined with the permease, phosphodiesterase, dioxygenase, hydrolase, nitric oxide-binding and auxin perception activities (Al-Younis et al. 2015, 2021; Kasahara et al. 2016; Vaz Dias et al. 2019; Yuan et al. 2022; Qi et al. 2022). In AtTIR1, which is an auxin receptor, AC activity tightly controls auxin perception and ultimately the protein’s main biological function of regulating root growth (Qi et al. 2022). This could be true for AtLHL, whereby the identified AC activity tightly controls its core functions, particularly those that it performs together with the other histone family proteins. This is not surprising because when we searched for presence of the AC centre in AtH2A, AtH2B, AtH3 and AtH4, none of these histone family proteins was found to possess the centre (results not shown).

Conclusion

In this study, we provide practical evidence that a linker histone-like or HNO4 protein from Arabidopsis thaliana (AtLHL) is a bona fide adenylyl cyclase (AC) capable of generating the second messenger, cAMP from ATP. AtLHL thus becomes the thirteenth AC to be identified in Arabidopsis and also the thirtieth AC identified in plants in general. Thus, considering that AtLHL appears to be the only histone family protein with AC activity and that it is typically involved in a number of key cellular processes such as chromatin formation (Jeon and Berezney 1995), embryogenesis (Martianov et al. 2005), reproduction (Tanaka et al. 2003), drought resistance (Ascenzi and Gantt 1997; Wu et al. 2022; Scippa et al. 2000) and disease resistance (Studencka et al. 2011), it is pertinent that more work is undertaken to perhaps shed more light onto its probable modes of action and at least functional significance in plants, particularly crops. Work involving the inactivation of its AC activity through site-directed mutations together with some in planta (in vivo) studies would be most ideal.