Active Plasma Membrane P-type H+-ATPase Reconstituted into Nanodiscs Is a Monomer*

Background: The plasma membrane H+-ATPase generates electrochemical gradients in plants and fungi. The minimal subunit organization required for activity is not known. Results: We developed a protocol for reconstitution of active H+-ATPase in nanodiscs. Conclusion: The minimal functional unit of the H+-ATPase is a monomer. Significance: The plasma membrane H+-ATPase functions like well characterized cation pumping P-type ATPases. Plasma membrane H+-ATPases form a subfamily of P-type ATPases responsible for pumping protons out of cells and are essential for establishing and maintaining the crucial transmembrane proton gradient in plants and fungi. Here, we report the reconstitution of the Arabidopsis thaliana plasma membrane H+-ATPase isoform 2 into soluble nanoscale lipid bilayers, also termed nanodiscs. Based on native gel analysis and cross-linking studies, the pump inserts into nanodiscs as a functional monomer. Insertion of the H+-ATPase into nanodiscs has the potential to enable structural and functional characterization using techniques normally applicable only for soluble proteins.


Plasma membrane H ؉ -ATPases form a subfamily of P-type
ATPases responsible for pumping protons out of cells and are essential for establishing and maintaining the crucial transmembrane proton gradient in plants and fungi. Here, we report the reconstitution of the Arabidopsis thaliana plasma membrane H ؉ -ATPase isoform 2 into soluble nanoscale lipid bilayers, also termed nanodiscs. Based on native gel analysis and cross-linking studies, the pump inserts into nanodiscs as a functional monomer. Insertion of the H ؉ -ATPase into nanodiscs has the potential to enable structural and functional characterization using techniques normally applicable only for soluble proteins.
The plasma membrane (PM) 3 H ϩ -ATPase is a prominent member of the P-type ATPases, a large superfamily of proteins pumping ions and lipids across cellular membranes. This family of proteins forms a phosphorylated intermediate during the catalytic cycle, hence P-type, and is divided into five subfamilies, P1 to P5, each having different substrate specificities (1,2).
The PM H ϩ -ATPase belongs to the P3 subfamily only found in plant and fungi and is responsible for proton extrusion out of the cell. The resulting proton gradient is used by proton-coupled transporters and secondary active transport of nutrients and metabolites across the PM. In addition to the fundamental role in nutrient uptake, the plant PM H ϩ -ATPase is involved in a number of processes during plant growth and development, such as stomatal movement and cell elongation. Also, responses to both biotic and abiotic stresses often require the activation of the PM H ϩ -ATPase (3)(4)(5).
A large body of information on the structure, function, and regulation of PM H ϩ -ATPases from the P3 subfamily has been compiled during the last decade. Cryo-electron microscopic images of the Neurospora crassa H ϩ -ATPase (6) and the recently solved crystal structure of the Arabidopsis thaliana H ϩ -ATPase isoform 2 (AHA2) (7) revealed the presence of 10 transmembrane helices and three large cytoplasmic domains, including the phosphorylation and ATP-binding sites. The overall structure of AHA2 resembles the one known for the well studied sarco/endoplasmic reticulum Ca 2ϩ -ATPase pumps, which belongs to the P 2 subfamily (1).
Oligomerization is a common feature of members of the P-type ATPase family of pumps. Thus, representative members of the P-ATPase family, such as the sarcoplasmic reticulum Ca 2ϩ pump (8) and the Na ϩ /K ϩ pump (9), self-associate. The monomer of the sarcoplasmic reticulum Ca 2ϩ -ATPase and the ␣-␤ protomers of the Na ϩ /K ϩ -and H ϩ /K ϩ -ATPases are capable of performing all the steps of the reaction cycle (8,10,11). The human PM Ca 2ϩ -ATPase is a P2B-ATPase that resembles PM H ϩ -ATPases by having an extended C-terminal regulatory domain. PM Ca 2ϩ -ATPase isolated from human erythrocytes undergoes reversible, enzyme concentration-dependent oligomerization (12,13). This oligomerization process involves the C-terminal calmodulin-binding domain of the pump (13)(14)(15) and likely results in an activated high affinity state of the pump (16).
An early study concluded that the functional unit of the fungal Neurospora PM H ϩ -ATPase reconstituted with excess lipid in liposomes might be a monomer (17). Subsequent structural studies have revealed the presence of PM H ϩ -ATPase dimers and hexamer complexes, but their functional roles remain to be elucidated (18,19). A distinct characteristic of P-type PM H ϩ -ATPases is the presence of a C-terminal regulatory domain (R-domain) (7), and activation of pump activity occurs by phosphorylation-dependent binding of 14-3-3 regulatory proteins to this domain (20,21). A three-dimensional reconstruction of purified PM H ϩ -ATPase/14-3-3 complex suggested a hexameric arrangement (19) in line with the reported structure from x-ray crystallography on two-dimensional crystals (22,23). Whether the activation of the pump correlates with its oligomerization status is not known. Likewise, it remains uncertain whether the monomeric state is active like it is in other P-type ATPases.
Reconstitution of membrane-embedded proteins in soluble nanoscale lipid bilayers, also termed nanodiscs, is a novel technique in the study of membrane proteins (24). Assembled from membrane scaffold proteins (MSPs), a nanodisc consists of two MSPs encircling a planar lipid bilayer in a double-belt configuration. The amphiphilic helical structure of the MSPs shields the hydrophobic edge of the lipid bilayer and stabilizes discrete disc sizes determined by the length of the MSPs, resulting in diameters ranging from 10 to 17 nm (25,26). Advantages in using this system include water solubility, monodispersity, flexible lipid composition, access to both sides of the bilayer simultaneously, and controlled stoichiometry of disc to target protein. Despite still being a relative novel approach for membrane protein study, several examples of membrane proteins reconstituted into nanodiscs exist, e.g. G-protein-coupled receptors (27)(28)(29)(30)(31) and cytochrome P450s (32)(33)(34)(35). Nanodisc-reconstituted membrane proteins can essentially be handled as soluble proteins in aqueous solution. This facilitates the application of analytical techniques that are normally difficult to use in the study of membrane proteins such as surface plasmon resonance (SPR), nuclear magnetic resonance spectroscopy, as well as electron paramagnetic resonance spectroscopy.
Here, we demonstrate the reconstitution of the A. thaliana PM P-type proton ATPase isoform 2 into the nanodisc. Using native gel analysis, cross-linking, and transmission electron microscopy, we demonstrate that the nanodisc-embedded AHA2 is an active monomer. Reconstitution of the PM H ϩ -ATPase into nanodiscs enables further functional and structural analysis of this family of pumps in complex with regulatory proteins. This is demonstrated by the ability of soluble nanodiscs with reconstituted PM H ϩ -ATPase to be selectively immobilized for SPR analysis.
Expression and Purification of Plasma Membrane H ϩ -ATPase Isoform 2-A plasmid based on the multicopy vector YEp-351 (36) was used for expression in yeast, containing the coding sequence of a 73-amino acid C-terminal truncated A. thaliana PM H ϩ -ATPase isoform 2 (aha2⌬73) (plasmid pMP1280) (37). The coding sequence is under the control of the PMA1 promoter and in fusion with a C-terminal Met-Arg-Gly-Ser-His 6 (MRGSH 6 ) tag. For expression, the Saccharomyces cerevisiae strain RS-72 (MATa ade1-100 his4-519 leu2-3,112) (38) was transformed and cultured essentially as described previously (39). In RS-72, the endogenous yeast PM H ϩ -ATPase PMA1 gene is placed under the control of a genomic galactose-dependent promoter. This strain grows in media containing galactose, whereas growth in glucose-based medium requires the complementation of the yeast PM H ϩ -ATPase by the constitutively expressed A. thaliana PM H ϩ -ATPase. The cells were grown and harvested by centrifugation, and membranes were isolated as described previously (7). All subsequent manipulations were performed at 4°C, and all buffers contained 0.3 mM phenylmethylsulfonyl fluoride and 3 g/ml pepstatin A. Membranes were solubilized using DDM at a 3:1 detergent/protein (w/w) ratio in solubilization buffer (50 mM Mes-KOH, pH 6.5, 20% (v/v) glycerol, 1 mM EDTA, 50 mM KCl, 1 mM dithiothreitol) containing 1.2 mM ATP with gentle agitation for 30 min. Insoluble material was removed by centrifugation for 60 min at 100,000 ϫ g. The supernatant containing solubilized A. thaliana H ϩ -ATPase was diluted 1:1 (v/v) with washing buffer WB500 (50 mM Mes-KOH, pH 6.5, 20% (v/v) glycerol, 0.15% (w/v) DDM, 500 mM KCl, 10 mM imidazole) and incubated for 16 h with nickel-nitrilotriacetic acid resin (1 ml of resin, 30 mg of membrane protein) pre-equilibrated in the same buffer. To minimize unspecific binding, the nickel-nitrilotriacetic acid resin was washed with 10 volumes of washing buffer WB250 (as WB500 with 250 mM KCl) followed by 10 volumes of washing buffer WB50 (as WB500 with 50 mM KCl). Bound proteins were eluted with 2 volumes of elution buffer (50 mM Mes-KOH, pH 6.5, 50 mM KCl, 300 mM imidazole, 20% (v/v) glycerol, 0.5 mM dithiothreitol, and 0.075% (w/v) DDM). Centrifugal concentrators (Vivaspin 20, GE Healthcare) with a 30-kDa molecular mass cutoff were used for buffer exchange of the eluted proteins to solubilization buffer containing 0.075% (w/v) DDM in two washing steps. Purified H ϩ -ATPase was finally concentrated to 5-10 g/liter, frozen in liquid nitrogen, and stored at Ϫ80°C.
Expression and Purification of Membrane Scaffold Proteins-His 7 -tagged membrane scaffold protein MSP1D1 (25) was expressed in Escherichia coli and purified as described previously (24). The His 7 tag on MSP1D1 was removed by treatment with His 6 -tagged tobacco etch virus protease, overnight at room temperature. After cleavage, the sample was loaded on a nickel-Sepharose column (GE Healthcare), and the cleaved membrane scaffolding protein MSP1D1(Ϫ) devoid of the polyhistidine tag was collected in the flow-through, whereas tobacco etch virus protease and any uncleaved protein were retained onto the column.
Reconstitution and Purification of PM H ϩ -ATPase in Nanodiscs-Reconstitution of PM H ϩ -ATPase into nanodiscs was based on the method described previously (24). Lipids in chloroform were mixed in glass tubes to achieve the indicated lipid composition. A lipid mixture of POPC/POPG (3:2, molar ratio) was applied in the reconstitution experiments. Excess chloroform was blown off with nitrogen, and lipids were solubilized at a concentration of 50 mM in MEK buffer (Mes-KOH, 1 mM EDTA, 100 mM KCl, pH 6.5) containing 200 mM octyl glucoside and incubated with purified MSP1D1 or MSP1D1(Ϫ) with gentle agitation overnight at 4°C. Purified aha2⌬73 was added to give a final lipid/MSP/H ϩ -ATPase molar ratio of 64:1: 0.125 and incubated for 1 h at 4°C. Detergent was removed by incubation for 4 h at 4°C with 1 g/ml SM-2 Adsorbent Bio-Beads. Reconstituted nanodiscs were purified and analyzed by size exclusion chromatography (SEC) (Superdex 200 10/300 GL, GE Healthcare) with a flow rate of 0.4 ml/min using MEK buffer as the mobile phase. The column was calibrated according to the manufacturer's instructions using the following known standards: blue dextran (2 MDa); ferritin (400 kDa; 12.2 nm); aldolase (158 kDa; 9.62 nm); conalbumin (75 kDa); ovalbumin (43 kDa; 6.1 nm); carbonic anhydrase (29 kDa); ribonuclease A (13.7 kDa; 3.28 nm), and aprotinin (6.5 kDa) (GE Healthcare). Values for the Stokes diameter of selected standards were plotted at their respective elution volume. Stokes diameter of the reconstituted nanodiscs was estimated by a linear fit of the Stokes diameter of the standard proteins versus the partition coefficient K av (supplemental Fig. S1). A 1-ml nickel-Sepharose column (GE Healthcare) was used to separate nanodiscs containing histidine-tagged aha2⌬73 from empty nanodiscs assembled with MSP1D1(Ϫ). Nanodiscs containing aha2⌬73, from here on denoted ND-AHA2, were eluted using a gradient of 0 -300 mM imidazole over a volume of 5 ml in MEK buffer, then buffer-exchanged to MEK using centrifugal concentrators with a 10-kDa molecular mass cutoff (Vivaspin 6, GE Healthcare), and concentrated to a final volume of about 500 l.
Cross-linking-Cross-linking was performed by incubating samples with 10 mM dimethyl suberimidate (DMS; Pierce) in 50 mM MOPS-KOH, pH 7.9, for 2 h at ambient temperature (ϳ25°C). This cross-linker was specific for amines and had a spacer arm of 11 Å. The reaction was stopped by addition of SDS-PAGE loading buffer.
Blue-native PAGE-Buffer conditions during Blue-native (BN)-PAGE were according to Schagger and von Jagow (42), and a cathode buffer with 0.02% (w/v) Coomassie G-250 was used. Gels were either cast at room temperature using a gravity-based gradient gel maker or bought from Invitrogen (Native-PAGE TM Novex 4 -16% BisTris gels). For cast gels, the final gel dimensions were 20 ϫ 20 cm with a 1.5-mm spacer and gradient gel from 5-15%. The samples were diluted 1:1-5 in 750 mM amino-n-caproic acid, 50 mM BisTris, pH 7.0, and 0.5 mM Na 2 EDTA to prevent protein precipitation and facilitate gel loading. Nativemark TM unstained protein standard (Invitrogen) was used for molecular weight estimation. Values for the Stokes diameter of selected standards were obtained and displayed along with the molecular weight (43)(44)(45). For SDS treatment, samples after centrifugation were preincubated with a final concentration of 0.1% SDS for 10 min at room temperature. To account for the apparent size difference between soluble and membrane proteins in BN-PAGE (46), the running conditions were 0.5 h at 70 V and 5 h at 385 V at 4°C. After running, the gel was destained in 10% ethanol, 10% glacial acetic acid for 16 h to visualize the bands.
Transmission Electron Microscopy-A drop (2 l) of nanodisc solution, typically diluted to an A 280 of about 0.001, was placed on a carbon film-coated, 400 mesh grid and incubated for 30 s. Liquid was wicked away from the grid using a wetted wedge of filter paper, and then negative staining was performed by covering the grid with a drop of aqueous 2.5% uranyl acetate (filtered through a 0.02-m syringe filter) for 2 min. After wicking away excess stain, the grids were air-dried and placed in a Philips CM 100 transmission electron microscope operated at 80 kV. All images were collected at a magnification of ϫ64,000 or 92,000. The diameter of 100 nanodiscs from two to three different grids was measured using ImageJ. In the case of nonspherical nanodiscs, the longest side was measured. Contrast adjustments were carried out to improve clarity of images but did not alter overall appearance. Final image processing, cropping, and mounting of the images were done with Adobe Photoshop CS2 and Illustrator CS2. Empty grids, grids with buffer, and grids with tobacco mosaic virus particles were used as control for the technique.
Surface Plasmon Resonance Analysis-Binding analysis of ND-AHA2 was detected by SPR using a Biacore X100 instrument at 25°C. The penta-His antibody (Qiagen) was covalently immobilized to the CM5 sensor chips (GE Healthcare, Biacore) by amino coupling. The penta-His antibody was buffer-exchanged with Bio-Spin 30 columns (Bio-Rad) to 10 mM sodium acetate, pH 5.0, buffer before injection. The surface was activated by 0.2 M 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide and 0.05 M N-hydroxysuccinimide for 7 min at a flow rate of 10 l/min. The penta-His antibody (0.1 mg/ml) was then injected for 7 min at a flow rate of 5 l/min, and the surface was deactivated by 1 M ethanolamine-HCl for 7 min at a flow rate of 10 l/min. The immobilization of penta-His antibody was done in a standard running buffer (10 mM HEPES, pH 7.4, 150 mM NaCl, 0.05% surfactant P-20, 3 mM EDTA). The penta-His antibody was immobilized only in flow cell 2; flow cell 1 was left unmodified as a reference surface. To capture aha2⌬73 in nanodiscs on the sensor chip, the running buffer was changed to 50 mM MOPS, pH 7.5, 150 mM KCl, 1 mM DTT, 0.5 mM EDTA. Samples were diluted to different concentrations in running buffer. Each sample was injected for 380 s at a flow rate of 5 l/min, and dissociation was observed for 600 s. The sur-face was regenerated by a 30-s injection of 10 mM glycine, pH 2.0, and washed with buffer for 300 s before the next injection. In each case, the bulk effect on the sensorgram was corrected by subtraction of the reference surface sensorgram from the sensorgram of the active surface functionalized with penta-His antibody and subtraction of a blank run.
Other Procedures-Proteins were analyzed by SDS-PAGE using the system of Fling and Gregerson (47). Western blotting was performed using monoclonal anti-penta-His antibodies (Qiagen), polyclonal antibodies raised against the N-terminal region of AHA2 (AHA2 N term ; number 762) (48), polyclonal antibodies raised against the preserved catalytic domains of AHA2 (AHA2 cat ) (49), polyclonal antibodies raised against a 19-amino acid peptide near the C terminus of human ApoA-1 for detection of MSP (ApoA-I Cterm ; Pierce antibodies PA5-21166), and alkaline phosphatase detection. Concentrations of purified proteins were routinely determined by UV spectroscopy (Nanodrop ND-1000 spectrophotometer, Thermo Scientific) using a calculated molar extinction coefficient at 280 nm of 18,450 and 110,350 M Ϫ1 cm Ϫ1 for MSP and aha2⌬73, respectively. For more accurate concentration determinations, samples were subjected to quantitative amino acid analysis (Dept. of Biochemistry and Molecular Biology, University of Southern Denmark). Densitometry analysis of SDS-polyacrylamide gels was performed with a Gel Doc 2000 (Bio-Rad) interfaced with Labworks Image Acquisition and Analysis Software (Ultra-Violet Products, Upland, CA) using known amounts of MSP and aha2⌬73 as standards. ATPase activity was determined essentially as described previously (39). Activity was measured at pH 7, and unless stated otherwise, the buffer contained 3 mM ATP (AppliChem). When solubilized ATPases were used, 0.05% (w/v) sonicated asolectin or POPC/POPG (3:2, molar ratio) was added. The Michaelis-Menten parameters of maximal velocity (V max ) and ATP affinity (K m ) were obtained from plots of the ATPase activity as a function of nucleotide concentration by nonlinear regression of the following equation where v is enzyme activity (nanomoles of P i per min/mg); V max is maximal ATPase activity (nanomoles of P i per min/mg); K m is the Michaelis-Menten constant for ATP (nanomolar), and [S] is substrate concentration (nanomolar). For curve fitting, V max and K m slope calculations, SOLVER in Excel 2007 was used.

Plant PM H ϩ -ATPase Purifies as a Dimer upon Heterologous
Expression in Yeast-The S. cerevisiae strain RS-72 has previously proved to be suitable for the high level expression and purification of A. thaliana PM H ϩ -ATPase AHA2 (50). Here, we used this strain for expression and purification of a C-terminally truncated version of AHA2 (aha2⌬73), which lacks 73 amino acid residues at the C terminus. This deletion removes a regulatory autoinhibitory domain of AHA2 and renders it in a constitutively activated state (39). Furthermore, aha2⌬73 can be expressed in higher yields as compared with AHA2 and is less prone to degradation (data not shown). Under our conditions, a typical purification by Ni 2ϩ affinity batch-binding resulted in ϳ1-2 mg of protein from 1 liter of cell culture. The purified PM H ϩ -ATPase had the expected molecular mass of 97 kDa and was homogeneous as assessed by Coomassie Brilliant Blue-stained SDS-PAGE (Fig. 1A). The protein identity was unequivocally confirmed by Western blot analysis using the anti-His and anti-AHA2 N term antibodies (Fig. 1A). ATPase analysis of the purified aha2⌬73 in detergent-containing buffer in the presence of asolectin or POPC/POPG (3:2 molar ratio) revealed an average ATPase activity of 10 -20 mol of P i /min/mg of protein.
To evaluate the oligomeric state of purified PM H ϩ -ATPase, aha2⌬73 was subjected to BN-PAGE analysis. The protein migrated predominantly as a single band with a size of ϳ200 kDa (Fig. 1B, X1) in the range of dimeric aha2⌬73, with weaker bands visible around ϳ450 kDa (Fig. 1B, X2) and ϳ600 kDa (Fig. 1B, X3). To confirm the oligomeric states of aha2⌬73, cross-linking was performed using the amine-specific crosslinker DMS (Fig. 1C). Following incubation with DMS, PM H ϩ -ATPase migrated as two bands in SDS-PAGE; the PM H ϩ -ATPase monomer at ϳ97 kDa and a band migrating above 170 kDa (indicated by an arrow in Fig. 1C) near the molecular mass correspond to dimeric aha2⌬73 of about 194 kDa. Both bands were recognized in Western blots using an anti-AHA2 cat antibody (data not shown). 2-Fold dilutions of the starting amount of PM H ϩ -ATPase resulted in similar dimer/monomer ratios, excluding concentration effects as a possible explanation for the presence of dimers. Furthermore, some protein retained at the stacking gel, suggesting the formation of higher oligomers or aggregates too large to enter the gel pores. These data suggest that the majority of DDM-solubilized PM H ϩ -ATPase is present as oligomers. Upon reconstitution into asolectin vesicles, aha2⌬73 exhibited ATP-driven proton pumping activity as revealed by ATPdependent fluorescence quenching of ACMA, a dye that accumulates inside the vesicles upon protonation (Fig. 1D). These results demonstrate successful purification of functionally active PM H ϩ -ATPase.
Reconstitution of PM H ϩ -ATPase into Nanodiscs-Purified aha2⌬73 was used for reconstitution into nanodiscs employing the MSP construct MSP1D1 or MSP1D1(Ϫ) devoid of the His 7 tag. An MSP/aha2⌬73 ratio was chosen to give a 4-fold excess of nanodiscs to PM H ϩ -ATPase, facilitating the reconstitution of monomeric PM H ϩ -ATPase. Assembly of stable PM H ϩ -ATPase-nanodisc complexes from MSP, lipids, and detergentsolubilized PM H ϩ -ATPase was verified by size exclusion chromatography. For nanodisc preparation with PM H ϩ -ATPase, the chromatogram displayed a major elution peak with a shoulder on the leading edge of the peak (Fig. 2A). By contrast, nanodisc preparation without PM H ϩ -ATPase showed a large and symmetrical elution peak. SDS-PAGE analysis revealed that the PM H ϩ -ATPase co-elutes with MSP in the shoulder region, although the major peak corresponds to MSP alone representing empty nanodiscs (Fig. 2B). Analysis of the ATPase activity across the elution fractions revealed that the peak of enzymatic activity coincided with ND-AHA2 complexes in the shoulder region (Fig. 2C). Taken together, these results confirm successful reconstitution of functional PM H ϩ -ATPase into nanodiscs.
Separation of Empty and Full Nanodisc Complexes-From initial reconstitution attempts, it was evident that a single size exclusion chromatography step was insufficient to separate ND-AHA2 complexes from empty nanodiscs, i.e. devoid of PM H ϩ -ATPase. To allow for purification of ND-AHA2 complexes, purified His 6 -tagged aha2⌬73 was reconstituted into nanodiscs using the MSP construct MSP1D1(Ϫ) devoid of the polyhistidine tag and subsequently subjected to Ni 2ϩ affinity chromatography (Fig. 3A). Analysis of the proteins eluted from the Ni 2ϩ matrix by SDS-PAGE and Western blotting demonstrated that under this condition, ND-AHA2 bound to the Ni 2ϩ matrix, although the flow-through contained mainly the empty nanodiscs, with residual aha2⌬73 as observed from Western blotting using anti-penta-His antibody (Fig. 3, B and C). Again, analysis of the ATPase activity across the elution profile showed enzymatic activity, indicating that PM H ϩ -ATPase was correctly reconstituted in the nanodiscs (Fig. 3D). Additionally, the ATPase activity confirmed the presence of residual aha2⌬73 in the flow through. Size exclusion chromatography after nickel affinity chromatography on both the flow-through and the eluate confirmed efficient separation of ND-AHA2 from empty nanodiscs; ND-AHA2 eluted in a major peak ahead of the position for empty nanodiscs (Fig. 4A). A small shoulder on the trailing edge of the peak for nickel affinity-purified nanodiscs indicated a small proportion of empty discs, possibly assembled from residual amounts of His 7 -tagged MSP1D1 still present in the reconstitution mixture. SDS-PAGE analysis on fractions from size exclusion on purified ND-AHA2 (Fig. 4B) revealed a complete co-elution pattern of aha2⌬73 and MSP1D1(Ϫ) devoid of the His 7 tag, further confirming the correct assembly of PM H ϩ -ATPase containing nanodiscs.

Analytical Characterization of PM H ϩ -ATPase-containing
Nanodiscs-To determine Stokes diameter of the phospholipid bilayer nanodiscs, the size exclusion chromatography column was calibrated using a standard set of proteins. Based on this calibration, empty nanodiscs and ND-AHA2 had overall Stokes diameters of 10.4 and 11.8 nm, respectively. The calculated Stokes diameter for empty nanodiscs are in reasonable agreement with earlier results on MSP1D1 assembled nanodiscs, reporting a Stokes diameter of 9.6 nm for nanodiscs assembled using dipalmitoyl-PC lipids and MSP1D1(Ϫ) (24, 25). As illustrated in Fig. 5D, the largest dimension on a model of ND- AHA2 results in values of about 11 nm, consistent with the measured Stokes diameter from SEC analysis.
The dimension and homogeneity of empty nanodiscs and ND-AHA2 were further assayed from analysis of transmission electron microscopy images (Fig. 5, A1 and B1). This revealed most of the discs, whether full or empty, to exhibit diameters in the range of 10 -11 nm (Fig. 5C) and to be lying individually on the carbon film. However, within the samples of ND-AHA2, we observed a number of paired or even stacked discs (Fig. 5, B2  and B3). Because such side view positioned paired discs were only found in samples from full discs, it is likely that this phenomenon was due to interaction between the inserted pumps and probably not the process of preparation.
Homogeneity of the assembled and affinity-purified nanodiscs was further investigated by BN-PAGE analysis. Under these conditions, empty discs migrated at ϳ146 kDa (Fig. 6A,  lane 1), whereas ND-AHA2 migrated as a single band at ϳ361 kDa (Fig. 6A, lane 2). Upon solubilization of ND-AHA2 by SDS, single bands at ϳ200 and ϳ20 kDa were observed corresponding to free PM H ϩ -ATPase and MSP (Fig. 6A, lane 4), whereas solubilized empty nanodiscs displayed only a single band for MSP at ϳ20 kDa (Fig. 6A, lane 5). Consistent with previous BN-PAGE analysis, purified DDM-solubilized PM H ϩ -ATPase was detected at ϳ200 kDa (Fig. 6A, lane 3). Based on simple geometrical assumptions when estimating the amount of lipids in the discs, the apparent molecular mass of empty nanodiscs and ND-AHA2 was estimated to be 129 and 213 kDa, respec-  tively. The discrepancy between calculated and apparent molecular weights in BN-PAGE analysis may be explained by the difference in Coomassie staining between the soluble protein standards and nanodiscs, with the latter representing a complex of scaffold proteins, membrane protein, and lipids (51). This is further evident from the relatively low degree of Coomassie Brilliant Blue staining of aha2⌬73 observed in densitometry analysis of SDS-polyacrylamide gels (supplemental Fig. S2). The Stokes diameters estimated from the protein standards are in agreement with the values found from SEC and transmission electron microscopy analysis, with ϳ9.7 nd ϳ11.6 nm for empty nanodiscs and ND-AHA2 respectively.
PM H ϩ -ATPase Reconstitutes into Nanodiscs as a Monomer-The MSP/aha2⌬73 stoichiometry was found to be 2:1 from densitometry analysis of Coomassie-stained gels, including a standard curve of purified proteins for which concentrations had been determined by quantitative amino acid analysis (supplemental Fig. S2). Collectively, these results demonstrate predominant formation of stable nanodiscs with a single PM H ϩ -ATPase monomer. To confirm the monomeric state of the reconstituted PM H ϩ -ATPase, cross-linking was performed using the amine-specific cross-linker DMS, previously shown to cross-link DDM-solubilized aha2⌬73 dimers (Fig. 1C). Cross-linking of empty nanodiscs resulted in two additional bands migrating to ϳ40 kDa (D1) and ϳ55 kDa (D2) in SDS-PAGE (Fig. 6B, lanes 2 and 3). This pattern is consistent with previous reports of cross-linked ApoAI in reconstituted high density lipoprotein particles, resulting in two distinct dimer forms (52). The two dimer forms of MSP1D1(Ϫ) were also observed from cross-linking of ND-AHA2 on both Coomassiestained SDS-PAGE and Western blotting using an ApoA-I Cterm antibody (Fig. 6C, lanes 4 and 4*), confirming a similar conformation of MSPs in full and empty nanodisc complexes. From cross-linking of ND-AHA2, three additional bands at ϳ115 kDa (C1), ϳ150 kDa (C2), and above the 170-kDa marker (C3) were observed in Coomassie-stained SDS-PAGE (Fig. 6C, lane  4 left panel). Only the bands at C1 and C2 are clearly visible from Western blotting using AHA2 cat and ApoA-I Cterm antibodies (Fig. 6C, lanes 4 and 4*, right panel). Comparison of C1, C2, and C3 to the dimer product of cross-linked DDM-solubilized aha2⌬73 (Fig. 6C, lane 2) reveals all to migrate to lower molecular weights than the apparent aha2⌬73 dimer. Based on these observations, we conclude that C1, C2, and C3 are products of cross-linking between the MSPs and aha2⌬73 and that aha2⌬73 is reconstituted as a monomer in the nanodiscs.
Enzymatic Properties of Plant Plasma Membrane H ϩ -ATPase in Nanodiscs-To ascertain that the PM H ϩ -ATPase reconstituted in nanodiscs had maintained its enzymatic properties, we analyzed the ATPase activity of purified and reconstituted PM H ϩ -ATPase. Consistent with previous results, purified PM H ϩ -ATPase in DDM-containing buffer displayed a low basal ATPase activity (about 1.7 mol P i /min/mg) that increased ϳ10-fold upon addition of lipids in the form of asolectin or POPC/POPG (3:2 ratio). For PM H ϩ -ATPase embed- The structure of aha2⌬73 is colored according to the different domains (7); the 10 transmembrane segments are brown, and the N, P, and A domains are red, blue, and yellow, respectively. The model was prepared using Visual Molecular Dynamics (62). ded in nanodiscs with POPC/POPG (3:2), a specific ATPase activity in the range of 10.5 mol P i /min/mg was estimated. Kinetic analyses revealed K m(ATP) of ϳ180 and ϳ60 M for lipid-activated and nanodisc-reconstituted PM H ϩ -ATPase, respectively. These values are within the range of previous studies on C-terminally truncated deletion mutants, reporting values of 50 -500 M for aha2⌬77 in endoplasmic reticulum vesicles from yeast (39) and 76 Ϯ 5 M for lipid-activated aha2⌬73 (37). Taken together, our kinetic analysis demonstrated that reconstitution of PM H ϩ -ATPase into nanodiscs did not compromise with its enzymatic properties.
It is an intrinsic property of the nanodisc system that transport of reconstituted transporters cannot be readily assayed. Thus, in contrast to reconstitution of PM H ϩ -ATPase into liposome vesicles, where transported protons can be made to accumulate from the extravesicular medium into the vesicle lumen, both sides of the nanodisc membrane are freely accessible to the assay medium, and transported protons cannot be trapped and assayed in the same way. In P-type ATPases, transport of ligands from the cytoplasmic to the extra-cytoplasmic side of the membrane takes place during the transition from the E1 conformational state, which hydrolyzes ATP, to the E2 conformational state, more specifically between the phosphorylated E1P and E2P states (53). Vanadate is a transition state analog of inorganic orthophosphate, which inhibits P-type ATPases by binding specifically to the E2 conformation (53). The vanadate sensitivity of PM H ϩ -ATPase was measured following reconstitution of the enzyme in vesicles as well as in nanodiscs (Fig.  7). Prior to the assay, it was ascertained that the H ϩ -ATPase reconstituted into liposomes actually pumps protons (Fig. 7,  inset). PM H ϩ -ATPase reconstituted in both membrane types was found to exhibit the same vanadate sensitivity (Fig. 7). This confirms that PM H ϩ -ATPase reconstituted in both experimental systems has a comparable conformational equilibrium and when reconstituted in both systems are capable of entering the E2 state of the catalytic cycle. Although not direct evidence for proton pumping, this strongly indicates that the H ϩ -ATPase reconstituted into nanodiscs is an active proton pump like the enzyme reconstituted into liposomes.
Immobilization of PM H ϩ -ATPase-containing Nanodiscs on Penta-His Surface-We further investigated the potential of nanodiscs for future SPR-mediated investigation of the PM H ϩ -ATPase. Because direct immobilization of proteins and nonspecific cross-linking of the protein to surfaces might impair protein function (54), the samples are immobilized on CM5 sensor chips covalently modified with antibodies specifically recognizing His tags (penta-His antibody). A high density of penta-His antibody (15,000 -16,000 response units) was covalently bound to the sensor chip surface to facilitate stable binding of nanodiscs containing aha2⌬73 by increasing rebinding of the complex to the antibody.
Two different reconstitution samples were tested for immobilization onto the surfaces as follows: 1) nanodiscs containing aha2⌬73 assembled from MSP1D1 and purified using a single SEC step (Fig. 8A), and 2) nanodiscs containing aha2⌬73 assembled from MSP1D1(Ϫ) purified by nickel affinity purification and SEC to remove excess empty nanodiscs (Fig. 8B,  black trace). Comparison with sensorgrams obtained for empty nanodiscs assembled using MSP1D1(Ϫ) (Fig. 8B, red trace)  confirms that the observed immobilization is due to interaction with the penta-His antibody. The sample containing empty nanodiscs and ND-AHA2 assembled with MSP1D1 has two to three His tags per complex, with one His tag on the PM H ϩ -ATPase and two His tags from the two MSPs. From the avidity effect, the presence of several His tags per complex can promote stable immobilization of the nanodiscs to the surface, although this approach does not allow us to conclude specific immobilization of ND-AHA2 to the surface. For this purpose we applied nickel affinity-purified ND-AHA2 assembled from MSP1D1(Ϫ), which confirmed the specific immobilization of ND-AHA2 through the His 6 tag on aha2⌬73.

Nanodisc-embedded Plasma Membrane H ϩ -ATPase Is an
Active Monomer-In this study, we demonstrate that the functional unit of the PM H ϩ -ATPase is a monomer. To do this, we worked out a reconstitution procedure for the plant PM H ϩ -ATPase at the single molecule level into nanodiscs. Expression of His-tagged PM H ϩ -ATPase from a multicopy plasmid under the control of the strong constitutive promoter (PMA1) allowed for effective purification to homogeneity of an C-terminally truncated version of the PM H ϩ -ATPase. To facilitate the formation of nanodiscs with single PM H ϩ -ATPase monomers, a 4-fold excess of nanodiscs to PM H ϩ -ATPase was applied during the reconstitution procedure. Under this condition, the purified, solubilized ATPases efficiently incorporated in nanodiscs. BN-PAGE analysis of ND-AHA2 revealed a single slowly migrating band confirming the homogeneity of the assembled and affinity-purified nanodiscs.
Despite the oligomeric state of the initially purified detergent-solubilized H ϩ -ATPase, our data show that the pump reconstitutes in nanodiscs as a monomer. This conclusion is based on two independent approaches as follows. 1) Isolation of ND-AHA2 allowed us to estimate the MSP/H ϩ -ATPase stoichiometry by densitometry analysis of SDS-polyacrylamide gels. 2) Cross-linking experiments excluded the presence of dimeric H ϩ -ATPase. Cross-linking experiments also verified a similar conformation of the MSPs in both ND-AHA2 and empty nanodiscs, with a cross-linking product of two distinct dimers in both cases (Fig. 6, B, C, D1, and D2). Furthermore, these experiments confirmed the close proximity of PM H ϩ -ATPase and MSP in the ND-AHA2 complex, giving three different cross-linking complexes after incubation with the DMS cross-linker (Fig. 6C). The three distinct bands with equal spacing likely represent cross-linking products of PM H ϩ -ATPase and one, two, and two in dimer D2 form of MSPs, as also verified from Western blotting against aha2⌬73 and MSP.
Notably, nanodisc-reconstituted PM H ϩ -ATPase displayed a robust ATPase activity, demonstrating that the pump monomer remains folded in an active conformation. This observation is in line with earlier studies on vesicle-reconstituted N. crassa PM H ϩ -ATPases demonstrating that the pump monomer is likely to be the minimal functional unit (17). The vanadate sensitivity of the nanodisc-reconstituted PM H ϩ -ATPase confirmed that it is able to alternate between the E1 conformation, which hydrolyzes ATP, to the E2 conformation, which binds vanadate. This indicates that the nanodisc-embedded PM H ϩ -ATPase carries out the complete catalytic cycle, which is associated with proton pumping. Furthermore, the similar vanadate sensitivity of PM H ϩ -ATPase reconstituted in vesicles and nanodiscs suggests that no significant change in conformational equilibrium is introduced when reconstituting the enzyme in nanodiscs as compared with vesicles, where the PM H ϩ -ATPase is shown to pump protons.
Nanodiscs as a Tool to Study P-type ATPase Function-In vitro studies involving the H ϩ -ATPases have so far been based on either detergent solubilization or membrane vesicles. The presence of detergent is required throughout purification, manipulation, assays, and structural studies to prevent aggregation. Detergents often perturb interactions and strip the membrane proteins of their endogenous lipids, which may be required to retain their functional state. Reconstitution into vesicles is a powerful model for transport studies and molecular characterization. However, drawbacks in using this system include the heterogeneous nature of the prepared vesicles, orientation of the proteins effectively giving rise to two populations exposed to different solvent phases, and the barrier of the membrane to soluble proteins in membrane protein-protein interaction studies.
Nanodiscs provide a complementary system for studying membrane pumps under defined experimental conditions. We found that the V max values of nanodisc-reconstituted PM H ϩ -ATPase were similar to those measured for asolectin-activated, DDM-solubilized PM H ϩ -ATPase. However, we observed an ϳ10-fold higher affinity for ATP of aha2⌬73 in nanodiscs compared with lipid-activated aha2⌬73. This could be due to both the MSPs interacting directly with the PM H ϩ -ATPase or the different lipid environment compared with lipid/DDM micelles. A recent study on empty nanodiscs using small angle x-ray scattering reported area per lipid headgroup values higher for POPC and lower for dilauroyl-PC in nanodiscs compared with free bilayers (55). Differences in mean bilayer height and phase transition temperature between nanodisc and lamellar dipalmitoyl-PC and dimyristoyl-PC have also been reported using a combined small angle x-ray scattering and differential scanning calorimetry analysis (56). Here, the increase in transition temperature and altered mean bilayer height was explained by the additional lateral pressure and structural perturbations from the MSP. Evidently, the variations in these parameters depend on the applied lipid, although it clearly demonstrates a different lipid environment in nanodiscs compared with other lipid systems, which could explain the different apparent affinity of ATP for PM H ϩ -ATPase in nanodiscs compared with lipid-activated PM H ϩ -ATPase. Natural lipid membranes consist of a large variety of lipids and proteins, with membrane proteins constituting up to 80% of the total weight of membranes (57). As a result, a large fraction of the lipids in a membrane are in contact with proteins. In this light, it could be speculated that the lipid bilayer in a nanodisc more accurately reflects the natural environment found in membranes, as compared with vesicles of pure lipids and lipid/ detergent micelles (58). Naturally, the absence of a proton gradient in the nanodisc system is another important difference compared with vesicle-reconstituted PM H ϩ -ATPases, which may influence the kinetics and activity.

Immobilization of Nanodiscs to Study Regulatory Pump
Interactions-Surface-sensitive techniques allow for real time measurement of biological events and thereby provide thermodynamics and kinetics information on the interaction. For this purpose, SPR is one of the most used techniques. Based on changes of refractive indexes at the surface, SPR does not require any labeling of the proteins. A requirement for the use of SPR is the immobilization of functional proteins onto surfaces. Furthermore, the improved stability of isolated membrane proteins in nanodiscs compared with detergent is a considerable advantage for investigation by surface-sensitive techniques (59 -61). Using both a mixture of empty nanodiscs and ND-AHA2 assembled with MSP1D1 and nickel affinitypurified ND-AHA2, we found stable and specific immobilization to the sensor chip. The stability of the immobilization observed in all cases and the possible regeneration of the surface after treatment at low pH (pH 2) makes thus the proposed strategy suitable for further monitoring protein-protein interactions by SPR.
Conclusion-We have demonstrated that the PM H ϩ -ATPases can be embedded into lipid bilayer nanodiscs as a functional monomer. The purification and reconstitution procedures developed here should be useful for the analysis of PM H ϩ -ATPases and other P-type ATPases at the molecular level. The ability to study these pumps embedded in a nanodisc rather than in detergent micelles or vesicles enables the use of a broad arsenal of biochemical and biophysical tools to quantitatively characterize the embedded protein and protein-protein interaction studies between the pump and its soluble interaction partners.