Updating the Mechanism of Bicarbonate (HCO₃⁻) Activation of Soluble Adenylyl Cyclase (sAC)

Abstract

Soluble adenylyl cyclase (sAC) is molecularly and biochemically distinct from other mammalian nucleotidyl cyclases. It is uniquely regulated directly by bicarbonate (HCO₃⁻) and calcium (Ca²⁺) ions and is responsive to physiologic fluctuations in levels of its substrate, adenosine triphosphate (ATP). Our initial in vitro biochemical studies suggested two mechanisms for HCO₃⁻-dependent elevation of sAC activity: increasing catalytic rate and relieving inhibition observed in the presence of supraphysiological levels of substrate, ATP. Structural and mutational studies revealed that HCO₃⁻ increases catalytic rate via the disruption of a salt bridge that facilitates productive interactions with the substrate. Here, we demonstrate that the HCO₃⁻ stimulation observed under supraphysiological ATP concentrations is due to the mitigation of ATP-dependent acidification. Therefore, we conclude that the sole physiologically relevant mechanism of HCO₃⁻ regulation of sAC is through its pH-independent effect facilitating productive substrate binding to the catalytic site.

1. Introduction

Many intracellular signal transduction pathways are mediated by the ubiquitous second messenger cyclic adenosine monophosphate (cAMP). Adenylyl cyclases (ACs) produce cAMP by cyclizing adenosine triphosphate (ATP). In mammalian cells, there are 10 known AC isoforms encoded by the genes ADCY1-10, which can be divided into two classes: transmembrane adenylyl cyclases (tmAC) and soluble adenylyl cyclases (sAC). The nine tmACs (ACDY1-9) are regulated by heterotrimeric G proteins and mediate cAMP responses downstream from hormone and neurotransmitter signaling via G protein-coupled receptors [1,2]. Conversely, sAC (ACDY10) is not bound to the cell membrane and instead is localized to various intracellular compartments including the cytosol, mitochondria, and nucleus [3,4,5,6,7,8,9]. As with other Class III nucleotidyl cyclases, sAC activity requires the intramolecular dimerization of two related catalytic domains (C1 and C2) [10]. Regulation of sAC occurs directly through HCO₃⁻ and Ca²⁺ ions [11,12,13], and its activity is sensitive to physiologically relevant ATP fluctuations [11,14].

While Ca²⁺ ions stimulate sAC activity by increasing the apparent affinity for the substrate ATP [11], HCO₃⁻ stimulates both Mg²⁺-dependent and Mg²⁺/Ca²⁺-dependent cyclase activity with an EC₅₀ within the physiologic intracellular range of HCO₃⁻ in cells (i.e., 10–25 mM) [11,13]. Initial in vitro cyclase assays on purified sAC protein suggested that HCO₃⁻ activation occurred via direct binding and was independent of changes in pH [13]. Subsequent kinetic studies predicted two mechanisms for HCO₃⁻ activation of sAC activity, increasing max velocity (V_max) and relieving a putative substrate inhibition observed under conditions with supraphysiological levels of ATP [11]. Structural, mutational, and biochemical studies of human sAC identified the bicarbonate binding site (BBS) and revealed the molecular mechanism for the HCO₃⁻-dependent elevation of the enzyme’s k_cat [10,15,16]. The BBS is adjacent to the ATP-binding catalytic site, and both sites include residues from C1 and C2. Two important residues in the BBS are Lys95 and Arg176. In apo human sAC, a salt bridge is formed between Arg176 and a residue essential for ATP conversion, Asp99. The Arg176–Asp99 salt bridge creates an inhibitory interaction. HCO₃⁻ binding engages Arg176, breaking the salt bridge and releasing Asp99, which allows it to flip towards the active site, facilitating ion site formation and ATP binding and turnover. Arg176 acts as a “switch” connecting the active site and the BBS, meaning its movement dictates whether the enzyme is in the active or inactive conformation. Thus, direct binding of HCO₃⁻ to the BBS favors productive ATP binding elevating the enzyme’s V_max [10,15].

In contrast, structural and kinetic studies provided no insight into the molecular basis for the substrate inhibition observed at supraphysiological ATP levels (>7.5 mM) and the hypothesized second HCO₃⁻-dependent activation mechanism via mitigation of this inhibitory effect. Here, we investigate this second mechanism of HCO₃⁻-dependent stimulation of sAC in vitro activity.

2. Results

Our previous studies exploring sAC sensitivity to pH focused exclusively on the effects of basic pH and its possible role in HCO₃⁻-induced sAC activation [13]. Here, we expanded these studies to ask how sAC in vitro activity is affected by acidic pH. When assayed at a physiologic or slightly basic pH, the cyclase activity of sAC is unchanged. However, as the pH becomes more acidic, cyclase activity decreases until there is almost no measurable cyclase activity at a pH of 5.5 (Figure 1). This acid sensitivity is consistent with the catalytic roles of several acidic residues [10,15]. With the appreciation that the cyclase activity of sAC is highly acid-sensitive, we revisited whether the previously used assay conditions supplied sufficient buffering capacity to neutralize the pH of solutions containing supraphysiological levels of ATP, whose disodium salt (the form used in all our assays) is a weak polyacid. Our previous kinetic experiments were performed in the presence of 50 mM of Tris at pH 7.5 [11]. The pH of 10 mM of disodium ATP dissolved in water in the presence of 50 mM of Tris proved to be acidic, at ~6.1. In contrast, 100 mM of Tris (pH 7.5) was adequate to buffer 10 mM of disodium ATP to remain at ~7.2 (Figure 2). Interestingly, an addition of 40 mM of HCO₃⁻, a concentration which had successfully mitigated the previously observed substrate inhibition [11], neutralized the pH of 10 mM of ATP solution in 50 mM of Tris to ~6.8 (Figure 2). These data suggested that the substrate inhibition observed at high levels of ATP, which was mitigated by HCO₃⁻ in previous experiments, may be due to inadequate buffering by 50 mM of Tris.

We tested this hypothesis by measuring cyclase activity of sAC in the presence of two concentrations of ATP (2 and 10 mM), two concentrations of Tris (50 and 100 mM), and in the presence and absence of HCO₃⁻. At physiological ATP levels (2 mM), there was no significant difference between the activity observed in the presence of 50 mM of Tris relative to 100 mM of Tris, and at both pHs, HCO₃⁻ stimulated sAC activity to a similar degree (Figure 3A). In contrast, when measured in the presence of supraphysiological levels of ATP (10 mM), there was a stark difference between the assays buffered with 50 mM of Tris, where the pH decreased to ~6.1, compared with the assays buffered with 100 mM of Tris, which remained at pH ~7.5. In the assay with 50 mM of Tris, the acidity caused by high levels of ATP reduced sAC activity considerably (Figure 3B). However, when sAC activity was measured in the presence of 10 mM of ATP and 40 mM of HCO₃⁻, which mitigated the acidifying effects of 10 mM of ATP in 50 mM of Tris (Figure 2), the activity was indistinguishable between 50 mM and 100 mM of Tris (Figure 3B).

Finally, we directly compared ATP kinetics in 50 mM of Tris (pH 7.5) with that in 150 mM of Tris (pH 7.5). These experiments were performed at 150 mM of Tris pH 7.5 to ensure adequate buffering for the highest levels of ATP tested (15 mM). Since Ca²⁺ stimulates sAC by decreasing the enzyme’s K_M for substrate ATP, Ca²⁺ and HCO₃⁻ work synergistically [11]. In the absence of HCO₃⁻ (i.e., Mg²⁺/Ca²⁺ ATP alone), we observed inhibition at high ATP levels (>5 mM) only in the reactions buffered by 50 mM of Tris; in the reactions buffered by 150 mM of Tris, the K_M for ATP was ~1 mM, as previously reported, and importantly, there was no evidence of substrate inhibition even at the highest ATP concentrations tested (Figure 4A). In contrast, in the presence of 40 mM of HCO₃⁻, there was no discernable difference in the kinetics measured in the presence of 50 mM of Tris compared with 150 mM of Tris. In both cases, the measured K_M was between 1 and 2 mM (Figure 4B), consistent with previous reports [11]. These observations reveal that the substrate inhibition observed in 50 mM of Tris in the absence of HCO₃⁻ is due to the acid sensitivity of the enzyme, and the previously hypothesized HCO₃⁻-dependent stimulation of activity at supraphysiological levels of ATP was merely due to its ability to mitigate the acidification due to the insufficient buffering capacity of 50 mM of Tris.

In conclusion, the data presented in this paper reveal that the buffering capacity used previously was insufficient at very high levels of ATP and updates the mechanism for HCO₃⁻ activation of sAC. The previous hypothesized ATP inhibition was in fact due to a decrease in cyclase activity at pH values less than 6.5. Therefore, the only known mechanism for HCO₃⁻ activation of sAC is through direct binding to the enzyme, which breaks an inhibitory salt bridge (Figure 5) [10,15]. Moreover, moving forward, the optimal conditions for ATP kinetics of sAC to properly control pH are higher concentrations of Tris or ATP stock solutions with pre-adjusted pH to avoid pH shifts in the assay solution.

3. Materials and Methods

3.1. Materials

Magnesium chloride (MgCl₂), calcium chloride (CaCl₂), manganese (II) chloride (MnCl₂), sodium bicarbonate (NaHCO₃), dithiothreitol (DTT), bovine serum albumin (BSA), disodium ATP (non-radioactive), cAMP (non-radioactive), Tris-HCl, MES, and MOPS were all purchased from Sigma-Aldrich (St. Louis, MO, USA). [α-³²P]-ATP and [2,8-³H]-cAMP were purchased from PerkinElmer (Waltham, MA, USA).

Heterologously expressed and purified human truncated sAC (sAC_t) protein was used for all biochemical assays. The sAC_t protein was N-terminal tagged with GST and purified from Sf9 cells via baculovirus expression [11].

3.2. In Vitro Adenylyl Cyclase Activity Assay

All in vitro adenylyl cyclase activity assays utilized the classical “two-column” method, developed by Salomon [17]. In this assay, conversion of [α-³²P]-ATP to [³²P]-cAMP is quantified by sequential Dowex and Alumina chromatography resins to purify the generated [³²P]-cAMP. Cyclase assays were performed in 100 μL of total reaction volume using 25–600 ng of sAC_t, 50–150 mM of Tris-HCl (pH 7.5), 3 mM of DTT, BSA 0.03%, substrate [α-³²P]-ATP, and either MnCl₂, or MgCl₂, and/or CaCl₂, and/or NaHCO₃ concentrations as indicated. Reactions were initiated by the addition of 10,000,000–1,000,000 counts/minute of [α-³²P]-ATP. Reactions were incubated for 30 min at 30 °C, quenched by adding 200 μL of SDS 2%, and cAMP was quantified following sequential Dowex and Alumina chromatography. An internal standard of ~10,000 counts/minute of [³H]-cAMP was added to each reaction.

For Figure 1, MES or MOPS buffer was used instead of Tris to maintain the pH of the reaction mixtures at pH levels other than 7.5. For ATP kinetics curves (Figure 4), [α-³²P]-ATP was diluted in 2/3 steps, starting at 15 mM, thus keeping the specific activity equivalent in every reaction.

3.3. Statistical Analysis

Plotting data, curve fitting, and statistical analyses were performed using the GraphPad Prism software (GraphPad, version 10.2.3, San Diego, CA, USA). All data are shown as the mean ± SD or SEM.