Bac-to-Bac baculovirus expression system and Sf9 cells adapted to Sf-900 II SFM media (Cat #, 11496015, Cat # 10902104) used in this project were from ThermoFisher Scientific. Sf9 cell expansion, culture, and subsequent baculovirus experiments were performed following the company’s manual with slight modification (O'Shaughnessy and Doyle, 2011; Date et al., 2017). Three constructs were used in this project: human ABCC4 (hABCC4), Lamprey-CFTR (Lp-CFTR), and human CFTR (hCFTR), which were subcloned into the pFastBac Vector that contains a 6xhis tag by the Emory Integrated Genomics Core at Emory University. The three constructs were transformed into DH10Bac™ E. coli (ThermoFisher Scientific, Cat #, 10361012) and the bacmid DNA of the three constructs were purified using a ZR BAC DNA miniprep Kit (ZYMO RESEARCH, D4048). The bacmid DNAs were verified by PCR and full sequencing of the inserts. Sf9 cells were then chemically transfected with each individual BAC DNA using Cellfectin II reagent (ThermoFisher Scientific, Cat #, 10362100). The P0 viruses of the three BAC DNAs were collected and further amplified using Sf9 cells. P1 viruses of the three constructs were then collected, aliquoted, and stored at −80°C. P1 viruses of hABCC4, Lp-CFTR, and hCFTR were further used to infect Sf9 cells for protein expression.

Non-infected Sf9 cells and infected Sf9 cells were harvested by centrifugation at 1000xg for 10 min, washed with 5 mL of 1 X PBS, and pelleted at 1000xg for 10 min. The pellets were promptly frozen with liquid nitrogen and stored at −80°C. The cell pellets were resuspended in Buffer 1 (50 mM Tris-HCl, 250 mM sucrose, 0.25 mM CaCl₂, pH 7.4) mixed with 1X protease inhibitors (Sigma P8340-1 mL, 1: 100 dilutions for final use) right before experiments (Hardy et al., 2019a; Hardy et al., 2019b). Cells were lysed using an ice-cold Dounce Homogenizer, 30 strokes with the loose pestle and then 30 strokes with the tight pestle on ice. The cell lysates were centrifuged at 750 g for 10 min at 4°C to remove cell debris; the supernatant was then ultracentrifuged at 37,500 rpm for 1 h at 4°C (Optima XPN-100 Ultracentrifuge, BECKMAN COULTER). The membrane pellets were resuspended in ice cold Buffer 2 (50 mM Tris-HCl, pH 7.4, 250 mM sucrose), aliquoted, and stored at −80°C. Proteins of hABCC4, Lp-CFTR, and hCFTR were evaluated with Western blotting and verified with anti-His Rabbit polyclone antibody (Cell Signaling, Cat #, 2365).

The total protein concentrations of membranes, including from controls (untransfected Sf9 cells), and Sf9 cells expressing hABCC4, Lp-CFTR, and hCFTR, were measured using a bicinchoninic acid (BCA) assay kit (ThermoFisher Scientific, Cat #23225). Specified amounts of individual Sf9 membranes containing target proteins and different amounts of the purified rat His-tag protein (PRHP) (Rat-Tie-2-His-tag protein, R&D, Cat # 10458T2050) were prepared in dithiothreitol-containing Laemmli sample buffer to a total of 42 µL each and incubated at room temperature for 15 min before loading in a gel (4%–15% Mini-PROTEAN TGXTM Precast Protein Gels, Bio-Rad, Cat # 4561084). The loaded samples were electrophoresed in Tris-glycine running buffer (25 mM tris base, 250 mM glycine, and 0.1% SDS) at 100 V for 60–70 min until the dye front reached the end of the gel. Proteins from the gel were transferred to a polyvinylidene fluoride (PVDF) membrane overnight at 4°C. The membrane was blocked with Intercept Blocking Buffer (LI-COR, Cat # 927-70001) for 1 h at room temperature. Blots were probed with a rabbit anti-His poly-clone antiserum at a dilution of 1:500 (Cell signaling, Cat# 2365), followed by IRDye 800CW Goat anti-Rabbit IgG secondary antibody at a dilution of 1:10,000 (LI-COR, Cat # 926-32211). The blots were visualized using Odyssey imager and software from LI-COR.

The western blotting images were further analyzed using ImageJ (Janes, 2015; McDonough et al., 2015; Kendrick et al., 2020). Band density ratios were calculated relative to the highest PRHP load. The absorbance values for each image were background subtracted, plotted as band density versus load and fit with a hyperbolic equation (Single Rectangular, 2 parameter) in Sigmaplot 12.3 (Palo Alto, CA). The regression line has an adjusted correlation R ² value of 0.97–0.999 (Tang et al., 2023). The protein concentrations of hABCC4, Lp-CFTR, and hCFTR were calculated following the fit result of the standard PRHP load in the same image. Absorbance values were always within the linear range of the standard curves for PRHP abundance.

We adopted the Antimony-phosphomolybdate method developed to detect small amounts of inorganic phosphate (P_i) to evaluate the rate of hydrolysis of ATP by phosphatases/ATPase (Bartolommei et al., 2013; Bartolommei and Tadini-Buoninsegni, 2016). The experimental procedures were carried out at room temperature unless otherwise stated. Each procedure consists of the preparation of a coloring solution, the determination of a calibration curve with a series of standard solutions of P_i, and the spectrophotometric quantification of P_i released by the ATPase protein at subsequent time intervals. Coloring solution is made fresh for every experiment from stock solutions to reach final concentrations of 125 mM H₂SO₄, 0.5 mM ammonium molybdate, 10 mM ascorbic acid, 40 µM tartrate, and double distilled H₂O (ddH₂O) at room temperature. The stock solutions are made and stored individually, including 2.5 M H₂SO₄ (Sigma, 339741), 24 mM Ammonium molybdate solution (Sigma, 1.01180), 0.3 M ascorbic acid (Fisher Scientific, BP351-500), and 4 mM potassium antimony (III) oxide tartrate trihydrate (Sigma, 1.08092). Standard phosphate solutions (KH₂PO₄, Sigma, P0662) are made by serial dilution of 1 mM phosphate stock in ddH₂O to get final solution of P_i from 0 to 100 nM. Thirteen standard solutions were prepared at 100 µL each to which 900 µL of coloring solution was added (final volume of 1,000 µL per tube). Color development starts immediately, stops when it plateaus and remains stable for hours. 200 μL of 1,000 µL of every standard solution was added to a 96 well transparent plate (BRANDplates, Cat# 781601) and the color was detected with absorbance at 850 nm (SpectraMax iD3 driven with SoftMax Pro7, Molecular Devices). The calibration curve was obtained by plotting the absorbance versus P_i concentration and fitted with the Polynomial, Linear equation in SigmaPlot 12.3.

The ATPase activity measurement was performed using the crude membranes purified from control Sf9 cells and Sf9 cells expressing hABCC4, Lp-CFTR, or hCFTR based on the study by Bartolommei and colleagues with modification (Bartolommei et al., 2013; Bartolommei and Tadini-Buoninsegni, 2016). The following solutions were mixed sequentially: Buffer 2, protein kinase A (2,500 units/µL) (PKA catalytic subunit, New England Biolabs, Cat #P6000), MgCl₂, various concentrations of Na₂ATP (Sigma, Cat# A6419), and finally the membrane samples to the total volume of 100 µL for every measurement in room temperature. No PKA was added for ATPase activity measurement of hABCC4. The volume of Buffer 2 was adjusted to retain final volume. The final concentration of MgCl₂ is 5 mM for all the measurement (Sigma, M2670). For Sf9 membranes bearing Lp-CFTR and hCFTR, the tubes were incubated at 30°C for 1 h, then 900 µL fresh color solution was added to each tube to stop the reaction and incubated at room temperature for 2 h for color development. For control only and membranes bearing hABCC4, solutions were mixed sequentially to a final volume of 100 µL for each tube at room temperature and then were incubated at 37°C for 20 min instead. Then 900 µL of fresh color media was added to each tube and these were then maintained at room temperature for 2 h for hABCC4. A final volume of 200 µL of every sample was added to the same 96 well transparent plate with their standard solutions and the plate was read with absorbance at 850 nm. This test condition for hABCC4 (20 min incubation at 37°C) was widely used by previous publication (Sauna et al., 2004; Huang et al., 2023). The test conditions for hCFTR/Lp-CFTR (1 h at 30°C) were selected to enable the best efficacy of PKA (NEB, Cat #P6000) (Zhang and Chen, 2016; Liu et al., 2017). We used these experimental conditions in most of this project with the exception of the final figure for direct comparison of hABCC4 and hCFTR (1 h, 30°C incubation), as listed in the results.

Results are presented as mean ± SEM. Data was analyzed using Sigmaplot 12.3. Student’s t-test were performed for single comparisons. A p-value <0.05 was accepted to be a statistically significant difference between groups.

The method developed by Bartolommei and colleagues for measurement of ATPase activity has been used to measure the ATPase activity of Na/K-ATPase as well as ABC transporter ABCG2 expressed on the purified crude membrane (Bartolommei et al., 2013; Khunweeraphong et al., 2017). We have adopted this technique for our project due to its remarkable efficiency, sensitivity, and integrity.

We freshly prepared 13 standard solutions of P_i in 100 µL final volume (including 0, 2, 5, 10, 20, 30, 40, 50, 60, 70, 80, 90,100 nmol) to obtain the calibration curve. Then, we added 900 µL of freshly mixed color solution and incubated the tubes at room temperature. Color started to develop immediately. To evaluate the color development, we observed the process at various intervals. As shown in Figure 1, we first detected the color absorbance at 10 min and proceeded with further detection until 240 min. The absorbance readings exhibited a continued increase until 90 min incubation, and then exhibited little change after 120 min. We fitted the data for these time courses with the Polynomial, Linear equation in SigmaPlot 12.3. The adjusted R ² shifted from 0.9413 for 10 min duration to 0.9992 for 240 min duration. We have repeated the process for three different experiments in three individual days and the results are consistent. We then chose 120 min as the standard color development duration for this project.

Our lab has been using MgATP (Sigma, A9187) and PKA (cAMP-dependent protein kinase, Catalytic subunit, Promega, Cat #V516A) to activate CFTR channels to study structure-function relationship in excised inside-out patches for decades. However, different groups studying CFTR have used diverse sources of ATP and PKA for their ATPase activity measurements. To help identify the experimental conditions leading to the most reproducible results for this study, we evaluated multiple different experiment combinations for their possible contributions of contaminants in the ATPase activity measurement. To identify conditions leading to the lowest apparent generation of P_i in the absence of ABC transporter protein, we tested ten different conditions including changes in concentration of MgCl₂, MgATP, Na₂ATP, and PKA. A representative calibration curve and summarized data are shown in Figures 2A,B, respectively. The data showed that the MgATP stock solution made with ddH₂O contained phosphate. In addition, the storage buffer for PKA from Promega is 350 mM potassium phosphate, which matched our data showing that the amount of P_i in PKA from Promega is the highest among all the groups (Promega, V516A). Furthermore, buffer 2, the MgCl₂ solution, the Na₂ATP, as well as PKA from NEB contained minimal P_i and thus were used for the remainder of the study.

A large population of proteins that carry ATPase activity are universally expressed in cells including Sf9 cells, such as ATP binding cassette family members, Na⁺/K⁺ ATPase, Ca²⁺-ATPase and more (Nitsche et al., 2018; Forzano et al., 2022). The endogenous ATPase activity in Sf9 cells could cause inaccuracy in measurements of ATPase activities of hABCC4, hCFTR, and Lp-CFTR. To estimate the possible effects of endogenous ATPase, we evaluated the endogenous ATPase activity of the crude membranes from untransfected Sf9 cells. A representative standard curve is shown in Figure 3A, and summary data are shown in Figures 3B,C, respectively. In the absence of Na₂ATP (0 ATP), the crude membranes of untransfected Sf9 cells exhibited minimal ATPase activity with little P_i production regardless of the presence of PKA (Figure 3B). Not surprisingly, the membrane proteins produced significant amounts of P_i in the presence of 5 mM Na₂ATP without PKA. Furthermore, the production of P_i was significantly higher in the presence of both ATP and PKA (Figure 3C). The data suggest that: 1) there is minimal contaminating ATP in the sample. 2) Some endogenous ATPase proteins were active in a PKA-independent manner. 3) PKA phosphorylated and activated endogenous proteins with ATPase activity and thus induced higher P_i production compared to ATP condition alone. In summary, the cell membranes from untransfected Sf9 cells contain endogenous proteins with ATPase activity, including proteins which exhibited PKA-dependent activity.

The results of a BCA assay only offer the total amount of protein in a sample. To evaluate the expression level of the three targeted protein, we adopted a semi-quantitative method utilizing immunoblotting to compare the ABC protein samples with a known loading protein. Figure 4 shows a representative western blot experiment with varying amounts of purified His-tag rat protein and three samples assessed with anti-His antibody. The band intensity was measured using ImageJ. The representative standard curve was fit with a hyperbolic, Single Rectangular, 2 Parameter equation in Sigmaplot 12.3 (Figure 4B). The concentration of each target protein was calculated based on their band intensities using the fit results from the standard curve. Summary data for the three target proteins are shown in Figure 4C where each symbol reflects analysis of a separate aliquot from the same membrane prep. Although the expression levels of each protein calculated by the method varied, no significant difference was observed among them. The results are not surprising because the overexpression of heterologous proteins could lead to cell death in the baculovirus/Sf9 cell system as well as in other systems. We chose to use the average protein concentration for each protein for the following experiments.

Since activation of the hABCC4 protein does not require phosphorylation by PKA but only requires ATP, we measured the intrinsic ATPase activity of hABCC4 (i.e., in absence of substrate) in multiple ATP concentration conditions without PKA. We used membranes from untransfected Sf9 cells as control and loaded the same amount of total protein in every measurement. The ATPase activity of hABCC4 was ATP-concentration dependent (Figure 5A). We calculated total P_i released by hABCC4 by subtracting the amount of P_i released by the control group from that released by the hABCC4 group and divided by the estimated amount of hABCC4 protein (using the average concentration of hABCC4 calculated from Figure 4). We calculated the dissociation constant (K_d) of ATP by fitting the data with Michaelis-Menten fits in Sigmaplot 12.3. K_ds for the control group and background-subtracted hABCC4 are summarized in Figure 5B. K_d of hABCC4 for ATP was significantly higher than that of control group. We presented the estimated maximum ATP binding affinity (B_max) for the control group and background-subtracted hABCC4 in Figure 5C. In addition, the B_max for hABCC4 was 5.53 ± 1.35 mM (n = 8), which was significantly higher than for the control group. The maximum ATPase activity (V_max) for hABCC4 also was significantly higher than for the control group (Figure 5D).

To make a comparison between hCFTR and Lp-CFTR, we measured their ATPase activity, subtracting activity of the endogenous proteins from untransfected Sf9 cell membranes as controls, in the same experimental conditions as shown in Figure 6. Both hCFTR and Lp-CFTR exhibited a PKA-concentration dependent ATPase activity (Figures 6A,B). Under the same experimental conditions, the ATPase activity of hCFTR was significantly higher compared to that of Lp-CFTR (Figure 6C). These data supported our previous findings that channel activity of hCFTR is driven by PKA-mediated phosphorylation and ATP hydrolysis, and is significantly stronger than that of Lp-CFTR (Cui et al., 2019).

It is known that both hABCC4 and hCFTR contain two ATP binding sites, but only one site exhibits ATPase activity. As we have shown above, hABCC4 and hCFTR in Sf9 cell crude membranes exhibited similar protein expression levels (Figure 4). We compared the ATPase activity of hABCC4 and hCFTR evaluated with the same conditions of 1 mM Na₂ATP and 5 mM MgCl₂ at 30°C for 1 h incubation except with addition of PKA in the case of hCFTR (4 µL PKA from NEB = 10,000 units from NEB). The presence of PKA is a prerequisite for activating hCFTR (Liu et al., 2017; Zhang et al., 2017). The ATPase activity of hABCC4 is about twice that of hCFTR (Figure 7).