Cyclic guanosine monophosphate (cGMP) hydrolyzing phosphodiesterase 5 (PDE5) is multidomain protein, in which cGMP binding to the regulatory GAFa domain allosterically increases the cGMP hydrolytic activity of the catalytic domain. In addition to cyclic nucleotides, GAF domains in phosphodiesterases have been proposed to be regulated by sodium (Na+) ions. The current study was initiated to investigate the effect of Na+ ions on the structure of the GAFa domain of PDE5. Bioluminescence Resonance Energy Transfer (BRET) experiments with various sensor constructs (the GFP2-GAFa-Rluc, GFP2-PDE5-Rluc containing the full-length PDE5, and a control, GFP2-Rluc, construct that does not contain any domains from PDE5) showed a general salt (NaCl and KCl)-induced reduction in the BRET efficiency. Independent determination of Rluc and GFP2 emission of the GFP2-Rluc construct revealed that increased salt concentration enhances Rluc activity without concomitantly increasing GFP2 fluorescence, thus providing a basis for the decrease in the BRET efficiency. The isolated GAFa domain sensor showed similar changes in BRET efficiency with varying concentrations of cGMP at both low (10 mM) and physiologically relevant (100 mM) NaCl concentrations. However, the full-length PDE5 sensor failed to respond to both cGMP and sildenafil at low (10 mM) NaCl concentration. These results suggest a role for NaCl in regulating the structural changes in PDE5, and thus NaCl should be included at 100 mM in BRET assays to detect ligand-induced conformational changes. Additionally, the increase in Rluc activity at higher salt concentrations reported here could be utilized to detect changes in ionic strengths in live cells.

Phosphodiesterase 5 is a key regulator of cGMP signaling in a variety of cell types including vascular smooth muscle cells, and in which it functions to control their contractility^[1]. It is a multidomain protein with two N-terminally located regulatory GAF domains (GAFa and GAFb) in tandem, and C-terminally located catalytic domain. While the GAFb domain is not known to bind any ligand, the GAFa domain binds cGMP with high specificity^[2], and cGMP binding to the GAFa domain allosterically regulates the activity of the catalytic domain^[3][4]. The structural change induced upon cGMP binding to the GAFa domain has been utilized in developing highly specific, live cell cGMP sensors^[2][5][6]. The catalytic domain of PDE5 specifically hydrolyzes cGMP^[7][8], and has been targeted with a number of pharmaceutical inhibitors, including sildenafil, for treatment of diseases such as penile erectile dysfunction and pulmonary hypertension^[9][10]. Structural studies have revealed that the binding of these inhibitors is associated with specific structural changes in the catalytic domain^[11]. Ligand binding-induced structural changes in both the GAFa and the catalytic domains in the full-length PDE5 have been extensively explored by developing a Bioluminescence Resonance Energy Transfer (BRET)-based intramolecular conformational sensor^[12].

BRET is being increasingly utilized for monitoring activation of proteins and signaling pathways in live cells^[13][14][15]. The underlying mechanism behind BRET is the non-radiative transfer of energy from a bioluminescent donor such as Renilla luciferase (Rluc) to a fluorescent acceptor such as GFP2 (a variant of GFP optimized for excitation by Rluc emission)^[16]. Key parameters that affect the resonance energy transfer efficiency include the distance between the donor and acceptor, their relative orientation, the quantum yield of the donor and the refractive index of the medium^[17]. A typical BRET assay is performed by incubating live cells or cell lysates with the Rluc substrate, and measuring the GFP2 fluorescence and Rluc emission. BRET efficiency is determined as a ratio of GFP2 and Rluc emissions.

To determine the effect of buffer NaCl concentration on ligand-induced changes in BRET efficiency of PDE5 constructs.

Cyclic nucleotide binding to the tandem GAF domains allosterically regulates the activity of the catalytic domain of CyaB1 and CyaB2, adenylyl cyclases from Anabaena^[18][19]. Additionally, Na+ ions have been shown to inhibit the GAF domain-mediated activation of the catalytic domain in these proteins^[20]. This inhibitory effect was found to be specific to Na+ ions, and other monovalent cations did not induce such effects on the GAF domain^[20]. Importantly, the inhibitory effect of Na+ ion was observed even when the GAF domains from the Anabaena adenylyl cyclase was exchanged for the GAF domains of rat PDE2 indicating the conservation of this mode of regulation across GAF domains^[20]. Specific interaction aside, altering the salt concentration will result in an alteration in the ionic strength of the buffer, unless controlled for. An alteration in the ionic strength could affect interactions between amino acids residues in a protein by electrostatic screening^[21][22] leading to changes in the structural properties of the protein^[23]. Therefore, an attempt was made to determine the role of Na+ ions on the structure of the cGMP-binding GAFa domain of PDE5. A BRET-based GAFa conformational sensor construct was utilized for this^[2]. Lysate prepared from HEK 293T cells transfected with the GFP2-GAFa-Rluc sensor construct was incubated with varying concentrations of NaCl, and intramolecular BRET efficiency (measured as a ratio of GFP2 emission and Rluc emission) was monitored^[2][12]. Increasing NaCl concentration in the buffer resulted in a concentration-dependent reduction in the BRET of the GAFa construct with an effective concentration value for 50% reduction in BRET efficiency (EC50) of 70 ± 43 mM (Fig. 1A). Since BRET efficiency is dependent upon the structure of the protein^[12][24], this result suggests the possibility of a structural regulation of the GAFa domain by Na+ ions. This was then followed up with experiments with the full-length PDE5 (A2 isoform) domain that contains two GAF domains in tandem^[1]. Similar to the isolated GAFa domain, the full-length PDE5 also showed a concentration-dependent reduction in the BRET efficiency with increasing NaCl concentrations with an EC50 value of 218 ± 140 mM (Fig. 1B). However, contrary to the specific inhibitory effect of Na+ seen with the Anabaena proteins^[20], increasing concentrations of KCl also reduced the BRET efficiency (EC50 value of 268 ± 80 mM) suggesting that it could be a general effect of the increase in the salt concentration in the buffer. To confirm this, a construct containing only the GFP2 and Rluc domains, but no domains from PDE5, was used. As seen with the GAFa and full-length PDE5 sensors, the GFP2-Rluc construct also showed a dose-dependent reduction in the BRET efficiency (EC50 values of 95 ± 55 mM and 93 ± 10 mM for NaCl and KCl, respectively) (Fig. 1C). These results suggest that the higher salt concentrations in the buffer negatively impact the energy transfer between Rluc and GFP2.

The reduction in the BRET efficiency at higher salt concentrations could occur due to either a change in the structure of the proteins resulting in a reduction in the energy transfer or a decrease in the quantum yield of GFP2 protein such that the fluorescence output of GFP2 is reduced, even though a similar amount of non-radiative energy is transferred^[17]. To delineate the mechanism behind this, the activity of the Rluc and the fluorescence of the GFP2 proteins in the GFP2-Rluc construct were independently monitored. Increasing the salt concentration of the buffer resulted in a dose-dependent increase in the emission of the Rluc protein with EC50 values of 183 ± 63 mM and 198 ± 62 mM for NaCl and KCl, respectively (Fig. 1D), without significantly impacting the GFP2 emission of the fusion protein (Fig. 1E). Similar increase in the emission was also observed for a construct of the Rluc protein only (EC50 of 154 ± 5 mM) (Fig. 1F). These results suggests that the decrease in the BRET efficiency of the PDE5 constructs is due to an increase in Rluc emission without a concomitant increase in the fluorescence of the GFP2 protein. The increase in the Rluc emission at higher salt concentrations is contrary to the sensitivity of the Firefly luciferase, which is an ATP-consuming enzyme, to higher ionic strengths^[25].

While the general increase in the Rluc activity observed in the presence of higher salt concentrations is interesting in itself, it is not obvious as to how alterations in the salt concentration would impact the ligand-binding induced conformational change in a sensor protein. Therefore, experiments were performed to understand the impact of NaCl in the ability of BRET-based, PDE5 conformational sensors^[2][12] to detect ligand-induced conformational changes. The salt concentration of 10 mM and 100 mM were chosen based on the EC50 value of BRET efficiency reduction (Fig. 1A). Incubation of lysates prepared from cells expressing the GAFa sensor construct with varying concentrations of cGMP in the presence of low (10 mM) or physiologically relevant (100 mM) NaCl revealed similar increases in the BRET efficiency (EC50 values of 38 ± 3 vs. 22 ± 5 nM at 10 and 100 mM NaCl, respectively). On the other hand, incubation of lysates prepared from cells expressing the full-length PDE5 sensor construct with varying concentrations of cGMP or sildenafil at low (10 mM) NaCl concentration revealed a lack of change in the BRET efficiency of the sensor (Fig. 1H & I), which was seen at 100 mM NaCl concentration (EC50 values of 104 ± 50 and 100 ± 52 nM for cGMP and sildenafil, respectively) (Fig. 1H & I). 

The lack of change in the BRET efficiency of the full-length PDE5 sensor with cGMP in the presence of low NaCl concentration could be a result of enzymatic hydrolysis of cGMP by the catalytic domain of PDE5. The use of non-hydrolyzable cGMP analogues may thus confirm the effect of salt on PDE5 conformational change. However, experiments described here were performed in the presence of ethylenediaminetetraacetic acid (EDTA), which would chelate metal ions in the buffer, and thus render the catalytic domain of PDE5 inactive. Therefore, the absence of change in BRET efficiency is not a consequence of hydrolysis of cGMP added during the assay. It is also important to note that no change in BRET was observed in the presence of sildenafil citrate at low NaCl concentration. Therefore, the requirement for 100 mM NaCl for the detection of ligand-induced conformational changes in PDE5 suggests a role for Na+ ions by either specific binding to PDE5, or buffer ionic strength in regulating the structure of PDE5. Since the GAFa sensor responded equally well to cGMP in the two NaCl concentrations, it is possible that the GAFb domain in PDE5 is responsible for the sensitivity to NaCl concentration.

The present study is directed towards understanding the effect of NaCl concentration on the structure as well as ligand-induced conformational changes in PDE5-based constructs using the BRET technology. The enzymatic activity of Rluc protein used as a donor in the BRET technology is increased upon an increase in the buffer salt (NaCl as well as KCl) concentration resulting in a net decrease in the BRET efficiency of all GFP2 and Rluc containing sensor constructs tested here. The GAFa domain-based cGMP sensor responds to cGMP at both low as well as physiologically relevant NaCl concentrations. However, the response of the full-length PDE5-based sensor is sensitive to NaCl concentrations, and fails to show changes in BRET efficiency at low NaCl concentrations. Observations reported here will be useful in both designing in vitro conformational sensor assays^{[2][12][24][26]} as well as live cell assays involving changes in intracellular ionic strength^[27][28].

One of the limitations of the present study is the use of crude lysates for determining both the BRET efficiency as well as the luciferase activity. While lysates prepared from transfected cells is experimentally straightforward, the presence of other cellular components could potentially impact the measurements. To alleviate this issue, assays should be performed with purified proteins. Additionally, the effect of increased salt concentrations on the structures of the proteins should be assessed independently using assays such as circular dichroism^[20] or hydrogen-deuterium exchange^[26].

The increase in luciferase activities of various sensor constructs reported in the current study is determined at a fixed wavelength. It is possible that the increase in the salinity of the buffer alters the emission spectra of Rluc in a manner similar to the red shift seen in the emission peak of Rluc upon mutation of specific residues^[29] or of firefly luciferase in vivo^[30]. A full spectral characterization of the Rluc protein at various salt concentrations should be undertaken to determine any shift in the emission spectra. Additionally, while it is assumed that the increase in Rluc emission is due to generic effect of the salt concentration on the structural properties of the protein, it is possible that the increased salt concentration alters certain a specific charge-charge interaction in the protein. Therefore, a structure-guided effort should be undertaken to delineate the mechanism of the increase in Rluc activity at high salt concentrations.

Intracellular ionic strength is tightly controlled in a living cell and an alteration in the intracellular ionic strength could activate specific signaling pathways in the cell such as activation of ion channels or transporters as a counter measure^[27][31]. Indeed, a specific class of proteins have been evolved for the purpose of detecting changes in the intracellular ionic strengths^[28]. However, there is a lack of genetically encoded, synthetic reporter for intracellular ionic strength. The sensitive activation of Rluc upon increase in the salt concentration reported here points towards its utilization as a luminescence-based, live cell indicator of changes in the intracellular ionic strength.

Plasmid constructs

Previously reported GAFa (GFP2-GAFa-Rluc)^[2] and the full-length PDE5 (GFP2-PDE5-Rluc)^[12] sensor plasmid constructs have been used here. The GFP2-Rluc and Rluc plasmids were purchased from Perkin-Elmer and used as such.

Cell culture and transfection

Human embryonic kidney (HEK) 293T cells were maintained in Dulbecco’s modified Eagle's media (DMEM) with 10% fetal calf serum, 120 mg/L penicillin and 270 mg/L streptomycin at 37°C in a 5% CO2 humidified incubator. Transfections of the cells with specific plasmid DNA were performed with polyethyleneimine lipid according to manufacturers’ protocols.

BRET assays

All BRET assays were performed in vitro using the BRET2 assay components i.e. acceptor- GFP2, donor- Rluc and Rluc substrate- Coelenterazine 400a (Molecular Imaging Products)^[2]. HEK 293T cells transfected with appropriate plasmids were lysed in a buffer of 50 mM HEPES (pH 7.5), containing 2 mM EDTA, 1 mM dithiothreitol, 100 mM NaCl, 10 mM sodium pyrophosphate, 80 µM β-glycerophosphate, 1 mM benzamidine, 1 µg/mL aprotinin, 1 µg/mL leupeptin, 5 µg/mL soybean trypsin inhibitor, 100 µM sodium orthovanadate and 10% glycerol. Cells were lysed by a brief sonication and lysates were centrifuged at 13,000 g for 1 h at 4°C, and the cytosol was collected. Aliquots of the cytosol were incubated with varying concentrations of salts diluted in a buffer containing 50 mM HEPES, pH 7.5 in a total volume of 40 µL at 37°C for 10 min. In experiments to determine the change in the BRET efficiency upon cGMP and sildenafil binding to the GAFa and the PDE5 catalytic domain, lysates were incubated with or without 1 mM cGMP and 100 µM sildenafil citrate under the same conditions. The Rluc substrate coelenterazine 400a (Molecular Imaging Products) was added to a final concentration of 5 µM, and emissions were collected for 0.8 s in a Victor3 microplate reader (Perkin Elmer). Emission filters used for Rluc and GFP2 emission were 410 nm (bandpass 80 nm) and 515 nm (bandpass 30 nm), respectively. BRET was calculated as the ratio of GFP emission per second to Rluc emission per second, and the average of three such measurements is reported.

Statistical analysis

All experimental data were analyzed using GraphPad Prism 6, and represent the mean ± S.E.M.

This work was supported by the Department of Biotechnology, Government of India, and a fellowship to KHB from the Council for Scientific and Industrial Research, Government of India.

We acknowledge members of the laboratory for useful discussions. KHB acknowledges the Senior Research Fellowship from the Mechanobiology Institute, National University of Singapore, Singapore.

The authors declare no conflicts of interest.

Not applicable.

No fraudulence is committed in performing these experiments or during processing of the data. We understand that in the case of fraudulence, the study can be retracted by ScienceMatters.