Regulation of prokaryotic adenylyl cyclases by CO₂

Abstract

The Slr1991 adenylyl cyclase of the model prokaroyte Synechocystis PCC 6803 was stimulated 2-fold at 20 mM total C_i (inorganic carbon) at pH 7.5 through an increase in k_cat. A dose response demonstrated an EC₅₀ of 52.7 mM total C_i at pH 6.5. Slr1991 adenylyl cyclase was activated by CO₂, but not by HCO₃⁻. CO₂ regulation of adenylyl cyclase was conserved in the CyaB1 adenylyl cyclase of Anabaena PCC 7120. These adenylyl cyclases represent the only identified signalling enzymes directly activated by CO₂. The findings prompt an urgent reassessment of the activating carbon species for proposed HCO₃⁻-activated adenylyl cyclases.

INTRODUCTION

Inorganic carbon (C_i) is fundamental to the physiology of all organisms. CO₂ and HCO₃⁻ exist in a pH-dependent equilibrium and are the major biologically active forms of C_i. CO₂ and HCO₃⁻ are vital to such diverse physiological processes as photosynthetic carbon fixation [1], pH homoeostasis [2] and carbon metabolism [3]. Study of C_i biology is essential to understand these vital physiological processes. Relatively little is known of the signalling mechanisms through which prokaryotic and eukaryotic cells directly detect CO₂/HCO₃⁻ fluctuations [4]. The identification of C_i-activated signalling molecules and their role in physiology is fundamental to understanding the diverse roles of C_i in biology. Currently, no signalling enzymes directly activated by CO₂ are known.

The mammalian sAC (soluble adenylyl cyclase) synthesizes the second messenger 3′,5′-cAMP and is stimulated by HCO₃⁻ [5,6]. It was observed that HCO₃⁻ regulation of AC (adenylyl cyclase) was conserved in a cyanobacterial AC, CyaC of Spirulina (Arthrospira) platensis, which had significant sequence similarity in the AC domain to sAC [6]. More recently, an active-site Asp→Thr polymorphism in the Class III AC family has been proposed as a marker for HCO₃⁻-responsiveness [7]. On this basis, proposed HCO₃⁻-responsive ACs are predicted to be widespread among the genomes of prokaryotes and eukaryotes [8]. To date, C_i regulation of AC has been confirmed in prokaryotes as diverse as Anabaena PCC 7120, Mycobacterium tuberculosis, Stigmatella aurantiaca and Chloroflexus aurantiacus [7,9]. An implicit assumption is made in the literature that the activating C_i ligand for AC is HCO₃⁻, on the basis that the ionic form is more likely to bind in the active site than CO₂. Identification of the activating carbon ligand for AC is essential to validate or question the relevance of significant recent literature in the field.

The photosynthetic cyanobacteria are an excellent model for investigating C_i signalling through AC, since hypothesized HCO₃⁻-responsive ACs are widespread in these organisms and C_i has clearly defined roles in their physiology. Here we demonstrate that the single Class III AC, Slr1991 (Cya1), of the unicellular cyanobacterium Synechocystis PCC 6803, is activated by C_i. Furthermore, we demonstrate, surprisingly, that the activating ligand for this enzyme is CO₂ and not HCO₃⁻. A previously characterized proposed HCO₃⁻ regulated AC, CyaB1 of Anabaena PCC 7120, is also shown to respond to CO₂ rather than HCO₃⁻. The present work provides the first evidence for AC as a CO₂-activated signalling molecule. This original finding prompts an immediate reassessment of the true activating carbon species in reported HCO₃⁻-responsive ACs.

MATERIALS AND METHODS

Recombinant proteins

DNA corresponding to amino acids 120–337 of slr1991 was isolated by PCR from the genomic DNA of Synechocystis PCC 6803, subcloned into pQE30, and fitted with an N-terminal MRGSH₆GS dodecapeptide affinity tag. Constructs were confirmed by double-stranded sequencing. Slr1991_120–337 protein was expressed in Escherichia coli M15 [pREP4] cells at 25 °C, for 3 h with 300 μM isopropyl β-D-thiogalactoside. Pelleted cells were washed with 50 mM Tris/HCl (pH 8.5)/1 mM EDTA, resuspended in 50 mM Tris/HCl (pH 8.5)/250 mM NaCl/10 mM 1-thioglycerol, lysed by sonication (1×150 s) and protein was purified from the supernatant with Ni²⁺-nitrilotriacetic acid (Qiagen) as previously described [10]. CyaB1_595–859 protein was generated as previously described [10]. Primer sequences are available on request from M. J. C.

AC assays

AC assays were performed at 40 °C in a final volume of 100 μl and typically contained 50 mM buffer, 2 mM MnCl₂, 2 mM [2,8-³H]cAMP (150 Bq) and [α-³²P]ATP (25 kBq) as substrate, if not stated otherwise [11]. Protein concentrations were adjusted to maintain substrate utilization at less than 10%. Kinetic constants were determined over a concentration range of substrate (Mn²⁺-ATP) of 1–100 μM. The following buffers were used: pH 6.5, Mes; pH 7.0–7.5, Mops; pH 8.0–8.5, Tris/HCl; and pH 9.0, Ches [2-(N-cyclohexylamino)ethanesulphonic acid]. Enzyme, buffer and substrate were all prepared at the appropriate pH for the required assay. CO₂ was quantified by titration against NaOH. The assay pH was stable over a period of at least 40 min. All errors correspond to the S.E.M. If absent, errors are smaller than the symbol used to depict the data point.

RESULTS AND DISCUSSION

The cya1 (slr1991; http://www.kazusa.or.jp/cyano/Synechocystis) gene of the unicellular cyanobacterium Synechocystis PCC 6803 encodes an enzyme consisting of a single FHA (forkhead associated) domain and a Class III AC domain that contains an Asp→Thr polymorphism associated with a putative HCO₃⁻ responsiveness [7,12]. We expressed the AC domain of Slr1991 as a purified recombinant protein (Figure 1a). The purified wild-type protein had a significant AC specific activity in the presence of both Mg²⁺-ATP (154±2.0 pmol of cAMP·min⁻¹·mg⁻¹, n=8) and Mn²⁺-ATP (5816±87 pmol of cAMP·mg⁻¹·min⁻¹, n=8) under optimal conditions (pH 9.5, 40 °C, 0.3 mM ATP and 8 μM protein).

Figure 1. AC activity of purified recombinant Slr1991_120–337.

(a) Purification of recombinant Slr1991_120–337 (SDS/PAGE analysis and Coomassie Blue staining). A 1.5 μg portion of protein was applied and molecular-mass standards (in kDa) are indicated. (b) Slr1991_120–337 specific activity (n=8) in the presence of 20 mM total C_i/salt (0.6 μM protein and 20 μM Mn²⁺-ATP, pH 7.5).

The Slr1991_120–337 protein had a pH optimum of 9.5 and a temperature optimum of 40 °C. The enthalpy of activation (E_A) derived from the linear arm of an Arrhenius plot using Mn²⁺-ATP was 33.5±1.4 kJ·mol⁻¹ (n=6). We investigated whether Slr1991_120–337 was regulated by C_i with a view to determining the identity of the activating species, CO₂ or HCO₃⁻. Slr1991_120–337 specific activity was stimulated 2-fold by 20 mM total C_i (1.2 mM CO₂/18.8 mM HCO₃⁻) at pH 7.5 compared with Cl⁻. Stimulation was independent of cation and robust to 95% confidence intervals (Figure 1b). A previous report by Masuda and Ono [13] had not observed stimulation of Slr1991 by C_i at pH 7.5. We noted that an extended assay period (40 min) was required to observe robust C_i activation of Slr1991 at pH 7.5. Although Masuda and Ono [13] did not report the assay time, this is the most likely cause of the discrepancy.

We determined the kinetics of activation of Slr1991_120–337 by C_i (Table 1). Slr1991_120–337 showed Michaelis–Menten kinetics in the presence of both Cl⁻ and Ci. The K_m value for Mn²⁺-ATP was greater in the presence of C_i than Cl⁻, but V_max values were proportionately greater for C_i than Cl⁻. The overall result was that C_i increased turnover rate (k_cat). A dose–response curve with increasing C_i was performed at a reduced pH (6.5) to eliminate problems with enzyme inhibition at >20 mM total C_i at pH 7.5 in the presence of Mn²⁺-ATP (Figure 2). The experiment revealed a maximum 8-fold stimulation with an apparent EC₅₀ for C_i of 52.7±1.0 mM (n=6) (20.4 mM CO₂/32.3 mM HCO₃⁻).

Table 1. Kinetic parameters for Slr1991_120–337.

Protein at 0.6 μM was assayed at pH 7.5 in the presence of 20 mM salt (n=6).

		Value
Parameter	Addition…	Cl−	HCO3−
Vmax (nmol of cAMP·min−1·mg−1)	0.74±0.01	1.13±0.03
Km,ATP (μM)	11.4±0.7	16.2±1.3
Hill slope	1.01±0.01	1.03±0.03
kcat (min−1)	0.018	0.027

Figure 2. Response of wild-type Slr1991_120–337 to C_i.

Slr1991_120–337 specific activity (n=6) was plotted against increasing total C_i (‘inorganic carbon’). The assay mixture contained 1.5 μM protein and 20 μM Mn²⁺-ATP, pH 6.5, with Na⁺ as cation. The total salt concentration was adjusted to 200 mM with NaCl.

We investigated the response of Slr1991_120–337 to total C_i at various pH values to gain an insight into whether the enzyme is responsive to CO₂ and/or HCO₃⁻. The experiment was performed using Mg²⁺-ATP as substrate since Mg²⁺ cofactor is more soluble than Mn²⁺ in the presence of C_i at alkaline pH. Intriguingly, relative stimulation (C_i/NaCl) varied from 1.1 at pH 8.5 (0.3 mM CO₂/39.1 mM HCO₃⁻/0.6 mM CO₃²⁻) to 2.4 at pH 6.5 (15.5 mM CO₂/24.5 mM HCO₃⁻) (Figure 3a). This is consistent with a role for CO₂ as opposed to HCO₃⁻ as the activating carbon species, but may also be due to the altered protonation status of the enzyme limiting the ability of Slr1991 to respond to HCO₃⁻ at elevated pH. We therefore sought direct evidence for regulation of Slr1991_120–337 by CO₂ and/or HCO₃⁻ by analysis under conditions of C_i disequilibrium when a single predominant carbon species, CO₂ or HCO₃⁻, is present at a defined pH. We exploited the fact that acquisition of the equilibrium between CO₂ and HCO₃⁻ is significantly lowered at reduced temperature in the absence of carbonic anhydrase and is a well established method for identifying the C_i substrate for CO₂/HCO₃⁻-fixing enzymes [14]. We followed the acquisition of the CO₂/HCO₃⁻ equilibrium by measuring the pH of a weakly buffered (5 mM) Mes solution on addition of 20 mM CO₂ or 20 mM NaHCO₃ in the presence or absence of carbonic anhydrase at 0 °C (Figure 3b). On the basis of these results we defined conditions for assaying AC under conditions of disequilibrium using 20 mM CO₂ or HCO₃⁻ as a 10 s assay period at 0 °C after addition of C_i. Under these conditions, C_i is predominantly in the form added to the assay (CO₂ or HCO₃⁻) and has not significantly advanced toward the equilibrium determined by assay pH (held with 100 mM Mes). Control experiments demonstrated that, under the conditions used for the assay, the final pH was equivalent when CO₂, HCO₃⁻ or Cl⁻ were added, demonstrating that any observed stimulation was due to addition of C_i and not a change in assay pH (results not shown).

Figure 3. Activation of AC by CO₂.

(a) Ratio of the specific activities of Slr1991_120–337 when assayed in the presence of 40 mM total C_i or NaCl at various pH values (8 μM protein, 1 mM Mg²⁺-ATP and 20 mM Mg²⁺). The inset shows the percentage of total C_i made up by CO₂ and HCO₃⁻ over the pH range tested. (b) Change in pH of a 5 mM Mes solution (starting pH 6.4) on addition of 20 mM NaHCO₃ (↑) in the presence (□) or absence (△) of 132 units of carbonic anhydrase at 0 °C. (c) cAMP produced by Slr1991_120–337 under conditions of C_i disequilibrium (50 μM Slr1991_120–337 protein, 0 °C, 10 s, 20 mM CO₂/NaHCO₃/NaCl, 100 mM Mes, pH 6.5, 150 μM Mn²⁺-ATP). (d), cAMP produced by CyaB1_595–859 under conditions of C_i disequilibrium (38 μM CyaB1_595–859 protein, 0 °C, 10 s, 20 mM CO₂/NaHCO₃/NaCl, 100 mM Mes, pH 6.5, 150 μM Mn²⁺-ATP, n=6).

We assayed Slr1991_120–337 under conditions of C_i disequilibrium and observed, surprisingly, that CO₂, but not HCO₃⁻, stimulated the enzyme (Figure 3c). We investigated whether this highly significant result was unique to Slr1991 or of more general significance. The CyaB1_595–859 protein of Anabaena PCC 7120 was previously shown to respond to HCO₃⁻/CO₂, but the activating species was not demonstrated [7]. Consistent with the findings for Slr1991_120–337, CyaB1_595–859 was also stimulated by CO₂, but not HCO₃⁻, under conditions of C_i disequilibrium (Figure 3d).

These results demonstrate that, for at least two randomly selected prokaryotic ACs, Slr1991 and CyaB1, the activating carbon species is dissolved CO₂ and not the more ably binding HCO₃⁻ species. These enzymes therefore represent the first identified signalling molecules demonstrated to respond directly to CO₂. HCO₃⁻ regulation of AC has been proposed, but not proven, for enzymes from species as diverse as the cyanobacterium Spirulina platensis, the encapsulated yeast-like fungus Cryptococcus neoformans, the yeast Candida albicans, the photosynthetic bacterium Chloroflexus aurantiacus and mammals [6,9,15,16]. An urgent examination of these systems is required to prove whether the ACs defined from these species respond to HCO₃⁻ or to CO₂ as described here.