Spectroscopical identification of residues of subunit G of the yeast V-ATPase in its connection with subunit E

Abstract

A critical point in the V₁ sector and entire V₁V_O complex is the interaction of stalk subunits G (Vma10p) and E (Vma4p). Previous work, using precipitation assays, has shown that both subunits form a complex. In this work, we have analysed the N-terminal segment of subunit G (G_1–59) of the V₁V_O ATPase from Saccharomyces cerevisiae by using nuclear magnetic resonance (NMR) spectroscopy. Analyses of ¹H-¹⁵N heteronuclear single quantum coherence (HSQC) spectra of G_1–59 in the absence and presence of the N-terminal peptides E_1–18 and E_18–38 as well as the produced and purified C-terminal segment (E_39–233) shows specific interactions only with the peptide fragment E_18–38. The binding of this peptide occurs via the residues M₁, V₂, S₃, and K₅ as well for V₂₂, S₂₃, K₂₄, A₂₅ and R₂₆ of G_1–59. The specific E_18–38/G_1–59 binding has been confirmed by fluorescence correlation spectroscopy data. The E_18–38 peptide has been studied by CD spectroscopy and NMR. The 3D structure of this peptide adopts a stable helix-hinge-helix formation in solution. A model structure of the E_18–38/G_1–59 complex reveals the orientation of E_18–38 relative to G_1–59 via salt-bridges of the polar residues and van der Waals forces at the very N-terminus of both segments.

• Vacuolar H⁺-ATPase
• V₁V_O ATP
• V₁ ATPase
• subunit G (Vma10p)
• subunit E (Vma4p)

Abbreviations
CSI	=	chemical shift index
DSS	=	2,2-Dimethyl-2-silapentane-5-sulfonate
DTT	=	dithiothreitol
EDC	=	1-ethyl-3-(dimethylaminopropyl)-carbodiimide
FCS	=	fluorescence correlation spectroscopy
GST	=	glutathione-S-transferase
HSQC	=	heteronuclear single quantum coherence
HPLC	=	high pressure liquid chromatography
IPTG	=	isopropyl-β-D-thio-galactoside
NMR	=	nuclear magnetic resonance
NOESY	=	nuclear Overhauser effect spectroscopy
NTA	=	nitrilotriacetic acid
PAGE	=	polyacrylamide gel electrophoresis
PCR	=	polymerase chain reaction
SAXS	=	small angle X-ray scattering
SDS	=	sodium dodecyl sulfate
TFE	=	trifluorethanol
TOCSY	=	total correlation spectroscopy
Tris	=	Tris-(hydroxymethyl)aminomethane

Introduction

The vacuolar H⁺-ATPase (V₁V_O ATPase) functions in almost every eukaryotic cell and energizes a wide variety of organelles and membranes [1]. Besides this function the enzyme is also described to be responsible for sensing and/or measuring the intravesicular pH, and subsequently transmitting this information from the vesicle lumen to the cytosolic side of the membrane [2]. The V₁V_O ATPase is composed of a water-soluble V₁ ATPase and an integral membrane subcomplex, V_O [3–5]. Under certain physiological conditions both domains can undergo a reversible disassembly, a phenomenon unique for eukaryotic vacuolar H⁺-ATPases [6–8]. The integral and proton transducting V_O domain contains six different subunits in a stoichiometry of a₁:d₁:c_4–5:c’₁:c”₁:e₁. The energy, needed for proton translocation, is provided from the cleavage of adenosine triphosphate into adenosine diphosphate and inorganic phosphate, catalyzed in the A₃:B₃ hexamer of the V₁ headpiece. The V₁ sector is composed of eight subunits in a proposed stoichiometry of A₃:B₃:C₁:D₁:E_X:F₁:G₂:H_X [3], [4]. The A and B subunits are arranged hexagonally and alternate around a central cavity in which a seventh mass is located [9–11]. A comparison of independently identified V₁ structures reveals that a compact stalk protrudes from the bottom of the A₃:B₃ hexamer with a length of 11 nm [9], [11], [12]. The stalk subunits undergo rearrangements when the V₁ ATPase reassembles with the V_O sector to form the entire V₁V_O complex as shown by crosslinking experiments, in which H-E, H-F crosslinks are observed in the V₁ sector and C-E crosslink in the V₁V_O ATPase [13], [14], and electron microscopy images [15]. In addition, the arrangement of the stalk subunits, which are involved in the coupling process of ATP-hydrolysis and proton pumping, is nucleotide dependent. While binding of CaATP to the V₁ part results in the zero-length crosslink products E-F and E-G, a strong D-F formation is observed in the presence of CaADP [16], [17]. Since the assembled V₁V_O ATPase and the dissociated V₁ sector as well as the coupling process are of physiological relevance, the topology of these stalk subunits are of great interest. Insight into the structural traits of stalk subunits came from the low resolution- [18] and crystallographic structure of subunit C [19], and H [20]. Subunit C has an elongated boot-shaped feature [18] with an upper head domain, a large globular foot and an elongated neck domain [19]. Stalk subunit H is characterized by a large, primarily α-helical N-terminal domain, forming a shallow groove, and a C-terminal domain, both connected by a four-residue loop [20]. In contrast, no structural details are known for the stalk subunits E and G, which have been proposed to be essential for the assembly of the V₁V_O ATPase [21–23] and activity of the V₁- [21], [24–27] as well as the V₁V_O ATPase [22–24]. Subunit G of the V-ATPase from yeast has been shown to be elongated with about 8.0 nm in length [28], and to interact with the N-terminus of subunit E [29], [30].

Here, we describe the spectoscopical studies of the N- and C-terminal part of G subunit, G_1–59 and G_61–114, respectively, of the V₁V_O ATPase from Saccharomyces cerevisiae in solution using nuclear magnetic resonance spectroscopy and fluorescence correlation spectroscopy. The 2D ¹H-¹⁵N heteronuclear single quantum correlation (HSQC) spectra provided a unique opportunity to assign the amino acids involved in the E-G assembly.

Materials and methods

Materials

ProofStart^™ DNA Polymerase and Ni^2 +-NTA-chromatography resin were received from Qiagen (Hilden, Germany); restriction enzymes were purchased from Fermentas (St Leon-Rot, Germany). Chemicals for gel electrophoresis were received from Serva (Heidelberg, Germany). Bovine serum albumin was purchased from GERBU Biochemicals (Heidelberg, Germany). (¹⁵NH₄)₂SO₄ and (¹³C) glucose were purchased from Cambridge Isotope Laboratories (Andover, USA). All other chemicals were at least of analytical grade and received from BIOMOL (Hamburg, Germany), Merck (Darmstadt, Germany), Roth (Karlsruhe, Germany), Sigma (Deisenhofen, Germany), or Serva (Heidelberg, Germany).

Expression and purification of proteins

The constructs of subunit E (E_39–233), and G (G_1–59, G_61–114) were cloned from S. cerevisiae genome using PCR cloning method. PCR products incorporating NcoI and SacI restriction sites were digested and ligated to pET9d(+)-His₃ vector. Forward Primer F1 5′ TCATGCCATGGTAGCTGACCAAGAGT 3′ and reverser primer R1 5′ CTTCGAGCTCTCAATCAAAGAACTTTCTTG 3′ were used to amplify and clone E_39–233 (Vma4p_39–233). G_1–59 (Vma10p_1–59) was cloned using forward primer F2 5′ TTTCCATGGCCTCCGCTATTACTG 3′ and reverser primer R2 5′ TTTGAGCTCTTAGCAGGCGATTCC 3′. G_61–114 (Vma10p_61–114) was cloned using forward primer F3 5′ATGGATAATGCCGGTGGTGTTG3′ and reverse primer R3 5′ CGTTCGAGCTCTTAGCACAAGGCATTGATATG 3′.

To express the respective proteins, liquid cultures were shaken in LB medium containing kanamycin (30 µg ml^-1) for about 6 h at 37°C until an optical density OD₆₀₀ of 0.6–0.7 was reached. The uniformly ¹⁵N- and ¹⁵N/¹³C-labeled G_1–59 and G_61–114 were expressed in Escherichia coli BL21 (DE3) cells using M9 minimal media containing ¹⁵NH₄Cl, or ¹⁵NH₄Cl plus [U-¹³C]-glucose. To induce production of proteins, cultures were supplemented with isopropyl-β-D-thio-galactoside (IPTG) to a final concentration of 1 mM. Following incubation for another 4 h at 30°C, the cells were harvested at 8000 g for 15 min at 4°C. Subsequently, all cells were lysed on ice by sonication for 3×1 min at 50% power in appropriate buffer, G_1–59 in buffer A (50 mM HEPES, pH 7.0, 150 mM NaCl, 1 mM DTT 1 mM PMSF), G_61–114 in buffer B (50 mM HEPES, pH 7.5, 150 mM NaCl, 1 mM DTT, 1 mM PMSF), and E_39–233 in buffer C (50 mM Tris/HCl, pH 8.5, 200 mM NaCl, 4 mM Pefabloc^SC (BIOMOL) and 2 mM PMSF). The lysate was cleared by centrifugation at 10,000 g for 30 min at 4°C, the supernatant was passed through a filter (0.45 µm pore-size) and supplemented with Ni^2 +-NTA resin. The His-tagged protein was allowed to bind to the matrix for 2 h at 4°C by mixing on a sample rotator (Neolab), and eluted with an imidazole-gradient (25–250 mM) in buffer A, B and C, respectively. Fractions containing His-tagged proteins were identified by SDS-PAGE¹ (31), pooled and concentrated using Centriprep YM-3 or YM-5 (5 kDa molecular mass [MM] cut-off) spin concentrators (Millipore) and subsequently applied on a gel filtration column Superdex HR75 (10/30, Amersham Biosciences). Selected fractions were concentrated in 3 kDa cut-off centricon. The purity of the protein sample was analysed by SDS-PAGE [31]. The SDS-gels were stained with Coomassie Brilliant Blue R250. Protein concentrations were determined by the bicinchoninic acid assay (BCA; Pierce, Rockford, IL, USA).

Peptide synthesis

The peptides E_1–18 and E_18–38 were synthesized at the peptide core facility, School of Biological Sciences, NTU on Liberty Automatic Microwave Peptide Synthesizer (CEM, USA) using N-(9-fluorenyl) methoxycarbonyl (Fmoc) chemistry on a Rink amide MBHA resin (Novabiochem, Germany). The C-terminal amidinated peptides were purified by reverse phase high pressure liquid chromatography (HPLC) on a Dynamax C-18 column (Varian Inc., USA) eluted with a linear 5–100% gradient of acetonitrile in 0.04% aqueous trifluoroacetic acid. The identity of the purified peptide was confirmed by MALDI-TOF mass spectrometry (4800 MALDI TOF/TOF, AB Applied Biosystems MDS Sciex, USA). The final product was judged to be 99% pure based on analytical high pressure liquid chromatography.

Circular dichroism spectroscopy

Steady state CD spectra were measured in the far UV-light (185–260 nm) using a CHIRASCAN spectropolarimeter (Applied Photophysics, UK). Spectra were collected in a 60 µl quartz cell (Hellma, Germany) with a path length of 0.1 mm, at 20°C and a step resolution of 1 nm. The readings were the average of 2 sec at each wavelength and the recorded ellipticity values were the average of three determinations for each sample. CD spectroscopy of the E_39–233 (2.0 mg/ml), E_1–18 and E_18–38 was performed in a buffer of 25 mM Tris/HCl, pH 8.5, 200 mM NaCl, 1 mM DTT for E_39–233 and 50 mM phosphate buffer (pH 6.8) and 30% trifluorethanol (TFE) for E_1–18 and E_18–38. The spectrum for the buffer was subtracted from the spectrum of the protein/peptide. CD values were converted to mean residue ellipticity (κ) in units of degree cm² dmol^–1 using the software Chirascan Version 1.2, Applied Photophysics. This baseline corrected spectrum was used as input for computer methods to obtain predictions of secondary structure. The CD spectra were analysed as described recently [32].

NMR data collection and processing

The NMR sample, G_1–59 and G_61–114 was prepared in 90% H₂O/10% D₂O containing 25 mM NaH₂PO₄ (pH 6.8), 0.1% NaN₃ and 50 mM Tris/HCl (pH 7.5) 200 mM NaCl, respectively. All NMR experiments were performed at 288 K on a Bruker Avance 600 MHz spectrometer. The experiments recorded on ¹⁵N/¹³C-labelled sample were HNCO, HNCACO, CBCA(CO)NH, HNCACB and 3D ¹⁵N-NOESY-HSQC (τm = 200 ms). The 3D ¹⁵N-NOESY-HSQC used mixing time of 200 ms. All the two- and three-dimensional experiments made use of pulsed-field gradients for coherence selection and artifact suppression, and utilized gradient sensitivity enhancement schemes. Quadrature detection in the indirectly detected dimensions was achieved using either the States/TPPI (time-proportional phase incrementation) or the echo/anti-echo method. Baseline corrections were applied wherever necessary. The proton chemical shift was referenced to the methyl signal of DSS (2, 2-dimethyl-2-silapentane-5-sulphonate [Cambridge Isotope Laboratories]) as an external reference to 0 ppm. The ¹³C and ¹⁵N chemical shifts were referenced indirectly to DSS. All the NMR data were processed using Bruker Avance spectrometer in-built software Topspin. Peak-picking and data analysis of the Fourier transformed spectra were performed with the SPARKY program [33].

Peptide E_18–38 was prepared by dissolving appropriate amount in 50 mM phosphate buffer pH 6.8 and 30% TFE-d3. The one dimensional (1D) and two dimensional (2D) ¹H NMR spectra including total correlation spectroscopy (TOCSY) and nuclear overhauser enhancement spectroscopy (NOESY) were obtained at the temperature of 288 K on a Avance Bruker NMR spectrometer at 600 MHz proton frequency. TOCSY and NOESY spectra of the peptide were recorded with mixing times of 80 and 300 ms, respectively. All the NMR data were processed using Bruker Avance spectrometer in-built software Topspin. Peak-picking and data analysis of the Fourier-transformed spectra were performed with SPARKY program [33]. Assignments were carried out according to classical procedures including spin-system identification and sequential assignment [34].

Structure calculation

Three-dimensional structures of peptide E_18–38 was calculated based on both distance and angle restraints by using CYANA 2.1 program package [35]. Dihedral angle restraints were calculated from chemical shifts using torsion angle likelihood obtained from shift and sequence similarity (TALOS) [34]. In total 100 structures were calculated. An ensemble of five structures with lowest total energy was chosen for structural analysis. This final ensemble of accepted structures satisfies the following criteria: no NOE violations greater than 0.3 Å, rmsd of 0.825±0.29 Å.

NMR binding experiments

To analyze the binding between subunits G_1–59-E_1–18, G_1-59-E_18–38, G_1-59-E_39–233 and G_61-114-E_39–233 a series of ¹H-¹⁵N heteronuclear single quantum coherence (HSQC) spectra were recorded at 288 K for the fixed concentration of 100 µM of G_1–59, titrated with increasing amounts (up to 1.5 equivalents) of E_1–18, E_18–38 and E_39–233 separately or with 100 µM of G_61–114 mixed with increasing amounts of E_39–233. The proteins were incubated for 30 min for binding during each step of the experiment. The change of chemical shift was monitored in the HSQC spectra. All the samples used were either finally dissolved in or exchanged with the 25 mM sodium phosphate (pH 6.8) buffer prior to the binding experiments.

Fluorescence correlation spectroscopy

Fluorescence correlation spectroscopy (FCS) was performed on a LSM-FCS system (ConfoCor 3, Zeiss, Jena/Germany) by using an Atto655-NHS ester (ATTO-TEC, Siegen/Germany) to label the E_18–38 peptide. The labeling and FCS experiments were performed in 25 mM sodium phosphate buffer, pH 6.8 for 10 min at room temperature. The excess of the unbound dye was removed by using a ZipTip^® P-10 Pipette Tip with C4 resin (Millipore, USA) for seven times replacing the solution with 5% acetonitrile/water, 0.1% trifluoroacetic acid which was verified by FCS. The fluorescent labeled peptide was subsequently eluted by 60% acetonitrile/water, 0.1% trifluoroacetic acid.

The 633 nm laser line of 5 mW HeNe laser was focused into the aqueous solution by a water immersion objective (40×/1,2 W Korr UL-VIS-IR, Zeiss). FCS was measured in 15 µl droplets, which were placed on gelatin treated Nunc 8 well chambered cover glass according to [36]. The following filter sets were used: MBS: HFT 514/633, EF: None, DBS: None, EF2: BP 655–710 IR. Out-of-focus fluorescence was rejected by a 90 µm pinhole in the detection pathway, resulting in a confocal detection volume dimensions of L: 1.5 µm and ∅: 0.3 µm. Fluorescence autocorrelation functions were measured for 30 sec each with 12 repetitions. Solutions of Atto655 dye in buffer were used as references for the calibration of the ConfoCor 3 beam path. To analyze the autocorrelation functions of E_18–38 bound to subunit G_1–59, a standard autocorrelation two diffusion coefficients normalized triplet model (FCS-LSM software, ConfoCor 3, Zeiss) was used for fitting. The diffusion time of fluorescently labeled peptide E_18–38 was measured independently, and kept fixed during the fitting of the FCS data. Therefore, the determination of the binding constants were made by the calculation of the relative amount of free labeled E_18–38 that displays a shorter diffusion time in comparison with G_1–59 bound to E_18–38 with rising diffusion times. The calculations were done by the ConfoCor 3-software 4.2, Microsoft Excel 2003, and Origin 7.5.

Modeling of the interacting domain of E_18–38 with G_1–59

Molecular modeling was performed using a HP-Linux workstation. The three-dimensional model of subunit G_1–59 was constructed based on the secondary structural information obtained from NMR data, using the programs Swiss-PdbViewer 3.7 [37] and COOT [38]. The model was optimized by carrying out energy minimization in vacuo with the GROMOS96 43B1 parameter set, without reaction field, implemented in Swiss-PdbViewer program. Initially 200 steps of Steepest Descent followed by multiple steps of Conjugate Gradient protocols were adopted until the derivative convergences to 0.050 kJ/mol. This model was then docked with the average NMR structure of E_18–38 using the web interface of the protein docking software, GRAMM [39] that uses Fast Fourier Transform methodology for the global search of the best rigid body conformations. The top 100 complex conformations which were scored based on soft Lennard-Jones potential, evolutionary conservation of predicted interface, statistical residue-residue preference, volume of the minimum, empirical binding free energy and atomic contact energy were analyzed and the one that best satisfies the NMR HSQC titration study was chosen. The best docking model was chosen in such a way that the model should have almost all residues that show shift in NMR HSQC spectrum, and has some kind of interaction in the complex. The subunits G_1–59, E_18–38 and the docked complex where optimized as described above and the docking energy was calculated to be −309.73 kJ/mol using the formula,

Docking energy = Energy of the G_1–59 E_18–38 Complex − (Energy of the subunit G_1–59+Energy of the subunit E_18–38); where the final minimized energy of the G_1–59 E_18–38 Complex, subunit G_1–59 model and the subunit E_18–38 are −4462.21, −3034.05 and −1118.44 kJ/mol, respectively.

Results

Resonance assignments of G_1–59 of yeast V-ATPase

From mutational analysis it is proposed that subunit G (Vma10p) of the yeast V-ATPase might interact via its N-terminal segment to subunit E (Vma4p) [23]. To get further inside into this phenomenon and the question about the amino acids involved in the proposed binding segments, NMR studies have been performed with the N-terminal segment G_1–59, which has been produced in E. coli cells and purified to high quality and quantity (Figure 1A). A 2D ¹H-¹⁵N heteronuclear single quantum coherence (HSQC) spectrum (Figure 1B) shows that the protein is properly folded. Besides the residues L₁₁, I₂₁, K₂₇ and K₅₃ all residues of G_1–59 could be assigned by backbone ¹⁵N and ¹³C signals using ¹⁵N NOESY-HSQC as well as triple resonance backbone experiments (HNCACB, CBCA(CO)NH).

Figure 1.  (A) SDS-gel of the purified G_1–59 (lane 1) and a protein ladder (lane 2) with molecular weights as indicated. (B) 2D ¹H-¹⁵N-HSQC spectrum of G_1–59 in 25 mM sodium phosphate buffer (pH 6.5) at 288 K. Backbone and amide assignments (N and Q) are shown for each residue. Signals from side-chain NH₂ groups are connected by horizontal lines.

NMR-titration of G_1–59 with E_1–18, E_18–38 and E_39–233 as well as G_61–114 with E_39–233

Recently, an N-terminal truncated and GST-fused form of subunit E (E_39–233) from the yeast V-ATPase shows no interaction with subunit G [29]. Here, we produced and purified the truncated protein E_39–233, encoding the amino acids 39–233 of subunit E. The CD-spectrum of E_39–233 shows a protein with a content of 41% of α-helix and 23% of β-sheet structure (data not shown). This protein was used in NMR titration experiments using ¹H-¹⁵N HSQC spectra to study possible interactions between subunit G_1–59 and the N-terminal truncated segment of E_39–233. HSQC spectra of ¹⁵N labeled G_1–59 were recorded in the presence of increasing amounts of unlabeled E_39–233. As demonstrated in Supplementary Figure S1A (online version only) no obvious chemical shift of the residues of G_1–59 (shown in red) could be detected in the presence E_39–233 (shown in blue), indicating that no interaction of both segments occurs. This was similar, when E_39–233 was added to the C-terminal and ¹⁵N labeled domain of G_61–114 (Supplementary Figure S1B; online version only). Furthermore, no chemical shift could be detected, when the N-terminal peptide E_1–18, with the amino acid sequence ₁MSSAITALTPNQVNDELN₁₈, has been titrated in increasing amounts to G_1–59 (see Supplementary Figure S1C; online version only). On the contrary, when equimolar amounts of the peptide E_18–38 (₁₈NKMQAFIRKEAEEKAKEIQLK₃₈) of subunit E was added to G_1–59 significant chemical shift perturbations could be observed. Figure 2A shows superpositions of the HSQC spectra of the ¹⁵N-labeled G_1–59 (shown in red) and the G_1–59-E_18–38 complex (shown in green), recorded in the presence or absence of E_18–38. The HSQC spectra of the ¹⁵N-labeled G_1–59 in absence and presence of E_18–38 yield noticeable differences in terms of changes in chemical shift specifically for residues M₁, V₂, S₃, and K₅ as well for V₂₂, S₂₃, K₂₄, A₂₅ and R₂₆ (Figure 2B).

Figure 2.  (A) Sections from the overlaid 2D ¹H-¹⁵N-HSQC spectra of G_1–59 in the absence (red) and presence (green) of 1.5 equivalent of unlabeled E_18–39. (B) Combined amide (¹H) and nitrogen (¹⁵N) chemical shift changes ([(Δ¹H_N)²+(Δ¹⁵N_ppm/6.51)²]^0.5) for G_1–59 and E_18–38 binding as a function of the amino acid sequence.

G_1–59/E_18–38-binding characterized by fluorescence correlation spectroscopy

To confirm and quantify the G_1–59/E_18–38-binding, fluorescence correlation spectroscopy (FCS) has been used. The fit of the autocorrelation function of Atto655 labeled peptide E_18–38 resulted in a characteristic diffusion time of 53.5 µs with a mean count rate of 61.1 kHz. Figure 3A displays the plotted autocorrelation curves of the fluorescent labeled E_18–38 in the absence and presence of non-fluorescent-labeled subunit G_1–59. The increase of the diffusion time was due to the increase in the mass of the diffusing particle, when Atto655-E_18–38 was interacting with subunit G_1–59. A binding constant of about 10 µM bound E_18–38 was calculated (Figure 3B).

Figure 3.  G_1–59/E_18–38–binding studied by fluorescence correlation spectroscopy. (A) Normalized autocorrelation functions of Atto655 labeled peptide E_18–38 by increasing the quantity of subunit G_1–59. (B) Concentration-dependent binding of E_18–38 to subunit G_1–59. The binding constant was calculated by determining the relative bound fraction using a standard autocorrelation two diffusion coefficients normalized triplet model (FCS-LSM software, ConfoCor 3, Zeiss). Best fits yielding the binding constants are represented as fitted line by a non-linear, asymptotic curve fit (see the text).

Assignment and structure calculation

To better understand the feature of the binding peptide E_18–38, we determined its structural traits. The CD spectrum of E_18–38 shows an α-helical structure (69%), indicated by minima at 208 and 222 nm (Figure 4A). Using the standard procedure for sequential assignment based in homonuclear TOCSY and NOESY experiments, besides K₃₈ all the residues of E_18–38 were assigned (Figure 4B). The connectivity diagram of E_18–38 is indicative of a helical conformation with the sequential HN—HN, Hα—HN(i, i + 3), Hα—HN(i, i + 4), and Hα—Hβ(i, i + 3) connectivities (Figure 4C). Figure 5A shows an overlay of the five lowest energy structures of E_18–38, and the statistics are given in Table I. These structures have an overall root mean square deviation (RMSD) of 0.825±0.29 Å. All these structures have energies lower than −100 kcal mol^-1, no NOE violations greater than 0.3 Å and no dihedral violations greater than 5°. The peptide displays a helix-hinge-helix structural motif (Figure 5B). Both helices are separated by 119° and the C-terminal helix is bent in the z-axis relative to the N-terminal one by about 67° (Figure 5C). It should be noted that in all the structures it is the hinge region which show least deviation. The peptide exhibits amphiphilic character such that the positive charge spread from the N-terminal and fades away at the C-terminal while the negative charge starts from the C-terminal and extends till the hinge region (Supplementary Figure S2A, online version only). The exterior region of the helix-hinge-helix motif is distributed evenly with positive charge which is attributed to the residues K₁₉, R₂₅, K₂₆, K₃₁, and K₃₈, whereas the interior region of this motif is distributed by negative charge coming from residues E₂₇, E₂₉, E₃₀ and E₃₄, where the hinge region is covered by E₂₉ and E₃₀ (Supplementary Figure S2B, online version only).

Figure 4.  (A) Far UV-CD spectrum of the peptide E_18–38. (B) HN-HN region of the NOESY spectra of E_18–38. (C) Connectivity diagram of E_18–38. The line thickness reflects the intensities of the NOE connectivities.

Figure 5.  Cartoon representation for the NMR structure of the peptide E_18–38. (A) Superposition of the five lowest energy structures, (B) average structure showing the angle between the helices and (C) average structure indicating the bend angle of the C-terminal helix with respect the N-terminal helix.

Table I.  Summary of structural statistics for five selected structures of E_18–38.

Structural statistics for E18–38
Distance restraints
Total	180
Intraresidue (i–j = 0)	69
Sequential (∣i-j∣ = 1)	72
Medium-range (2 ≤ ∣i − j∣ ≤ 4)	39
Long-range (∣i − j∣≥5)	0
Average number of violations
Distance violations > 5 Å	0
Ramachandran plot^2 (%)
Residues in most favoured regions	94.4
Residues in additionally allowed regions	5.6
Residues in generously allowed regions	0.0
Residues in disallowed regions	0.0
Average RMSD to Mean (Å)
Backbone (C^α, C′, and N)	0.825±0.29

Model for the interface of G_1–59/E_18–38 complex

The NMR data presented demonstrate that residues M₁, V₂, S₃, and K₅ as well for V₂₂, S₂₃, K₂₄, A₂₅ and R₂₆ contribute largely to the binding region of both subunits. While building the G_1–59/E_18–38 interacting domain, we have docked the helical peptide, E_18–38, with G_1–59 iteratively maximizing favorable interactions between the exposed sidechains (see Material and methods). Figure 6 shows a plausible model of the connected segments E_18–38 and G_1–59. The interface of the interacting segments turns out to be rich in polar residues as expected, since the E_18–38 peptide are rich in polar residues, wherein the charged residues alone contributes 47.6%, other polar residues 14.3% and the non-polar residues 38.1%. Residues, E₂₉, E₃₀ of E_18–38 may form potential salt bridges with R₂₆ of G_1–59 whereas the main chain nitrogen of N₁₈ of E_18–38 forms a hydrogen bond with mainchain carbonyl of M₁ of G_1–59. These strong interactions are also evidenced from the relatively large chemical shift in the HSQC spectrum (Figure 2A) for the R₂₆ and M₁ residues. The shift in the HSQC spectrum for the residues S₃ and K₅ of G_1–59 may be attributed to the weak hydrophilic interactions that they are involved in. Such as, the sidechain hydroxyl of S₃ shows weak interaction with the sulphur of M₂₀ (E_18–38) and the sidechain nitrogen of K₅ (G_1–59) shows cation-Π interaction with the phenyl ring of the F₂₃ (E_18–38). The other binding residues, V₂₂, S₂₃, K₂₄, A₂₅ and V₂ may confer hydrophobic interaction through van der Waals forces for the stability of the complex. In this model, in addition to these interactions with the binding region identified through NMR titration experiments, R₂₅ and K₂₆ of E_18–38 forms hydrogen bond interaction with E₂₀ and E₁₇ of G_1–59 respectively. The NMR HSQC spectrum did not show any marked shift for residues E₂₀ and E₁₇, probably because they are found near the N-terminal region which is flexible.

Figure 6.  The structure of the modeled segment G_1–59 (green). The residues shown in stick representation encircled with dotted spheres are the ones that are shown to have binding (M₁, V₂, S₃ and K₅; V₂₂, S₂₃, K₂₄, A₂₅ and R₂₆) with the E_18–38 peptide. Residues shown in line representation are the ones proposed to be in the interacting region in the docked complex.

Discussion

Subunit G is required for assembly [21–23], ATPase activity [24–27] as well as for proton pumping catalyzed by the V₁V_O complex [23], [24], [41]. A vma10 null mutation resulted in loss of ATP-dependent proton uptake activity [23]. For the H⁺-ATPase of clathrin-coated vesicles it has been demonstrated that the addition of recombinant subunit G to a recombinant A₃B₃CE-complex significantly increased the ATPase activity [24]. Recent work on the relationship of nucleotide binding and interaction of subunit G with other stalk subunits showed a trapped CaATP conformation of the V₁ ATPase after CuCl₂ addition in which a zero-length crosslink E-G was formed [16]. By comparison, no crosslink between these subunits can be obtained in the trapped CaADP form, implying a rearrangement of both subunits during coupling [16]. Using precipitation assays with truncated forms of subunit E of the yeast V-ATPase, Jones et al. [29] have shown that deletions of the N-terminal 38 (E_39–233) and 110 amino acids (E_110–233) results in a complete loss of E-G association, whereby the truncations of the first 19 residues or the C-terminal amino acids 112–233 (E_1–112) did not alter the binding traits of both subunits, suggesting that the N-terminal region 19–38 of subunit E might interact with subunit G. The presented NMR and FCS titration data support the view that the association of subunit E and G occurs only via the N-terminal regions of both subunits, and allows the clear identification of amino acids M₁, V₂, S₃, and K₅ as well for V₂₂, S₂₃, K₂₄, A₂₅ and R₂₆ of subunit G in the assembly of E_18–38 (Figure 2B). The importance of the positively charged amino acids among K₂₄ to R₂₆ is supported by mutational analysis [23], in which the residue R₂₆ has been substituted to Ala or Leu, resulting in an inhibited disassembly of V₁ from V_O in response to glucose deprivation. As pointed out above (see Results) the interface of the interacting subunits are rich in polar residues in G_1–59 as well as in the peptides E_18–38, with the latter being almost universally conserved among eukaryotic V-ATPases [29], and allows thereby salt bridge formations between both proteins (Figure 6). Importantly, no binding could be detected, when the truncated protein E_39–233 (Figure S1A, online version only) or the peptide E_1–18, shown to interact with subunit C [30], has been used. Furthermore, when ¹⁵N labeled G_61–114 has been titrated with E_39–233 (Figure S1B, online version only) no chemical shift could be observed either, indicating that the E-G association occurs only via the N-terminal segment of both subunits. Together with the data published by Jones et al. [29] the present results exclude that E and G might form a coiled-coiled interaction over the length of both molecules as proposed [40].

Figure S1.  (A) Overlaid ¹H/¹⁵N-HSQC spectra of G_1–59 in the absence (red) and presence (blue) of 1.5 equivalents of E_39–233 in 25 mM sodium phosphate buffer (pH 6.8) at 288 K. (B) Overlaid ¹H/¹⁵N-HSQC spectra of G_61–114 in the absence (red) and presence (blue) of 1.5 equivalents of E_39–233 in 25 mM sodium phosphate buffer (pH 6.8) at 288 K. (C) Overlaid ¹H/¹⁵N-HSQC spectra of G_1–59 in the absence (red) and presence (yellow) of 1.5 equivalents of E_1–18 in 25 mM sodium phosphate buffer (pH 6.8) at 288 K.

Figure S2.  The molecular surface showing the electrostatic potential of the peptide E_18–38 drawn in pyMol (DeLano WL [2001] The pyMol Molecular Graphics SystemSan Carlos, CA: DeLano Scientific) that uses Poisson-Boltzmann equation, where the positive potentials are drawn in blue and negative in red. (A) Front view, (B) Back view.

The 3D structure of E_18–38 shows a helix-hinge-helix formation in which both helices are separated by 119° and the C-terminal helix is bent in the z-axis relative to the N-terminal one by about 67°. This arrangement brings both helices in close proximity with the detected amino acid regions inside the N-terminus of subunit G and allows the plausible docking of this peptide to G_1–59. The helix-hinge-helix formation also indicates that the N- and C-termini of subunit E might at least partially be separate, which allows subunit C and E to associate [29], [30] via the very N-terminal amino acids 1–18 of E [29]. Point mutations in the N-terminal half of subunit G (E14A) resulted in destabilization of subunit E and C [23], which could be caused by an alteration of the flexible part between Lys₅ and Lys₁₆ of subunit G and thereby weaken interaction between subunit E and G. Based on crosslinking by zero-length crosslink reagents EDC and CuCl₂, subunit E is in close neighborhood to subunit D in both the V₁ complex [17] and the V₁V_O ATPase [13]. As mapped for the D-E product inside the Manduca sexta V₁ ATPase the CuCl₂ induced crosslink is formed by the cysteine residue 178 of the C-terminus of subunit E [17]. In addition, subunit E forms also a zero-length crosslink with subunit F in the V₁ complex, when CaATP is bound to the enzyme [16]. As mentioned above the crosslink formation and/or the strength of such formed products inside the V₁ sector are nucleotide dependent. In the way as CaATP induces E-F and E-G crosslinks [16], the D-E formation is strong after addition of MgADP whereas a lower amount is produced in the presence of MgATP or -AMP·PNP [17]. Although E_18–38 makes up only a segment of subunit E, the 3D structure of this peptide gives a first insight into how the neighboring subunit(s) of E might bind and that this subunit has at least one hinge region which might translate conformational changes in the V₁ ATPase during coupling as well as rearrangements combined with the process of reversible V₁ and V_O disassembly. The binding constant of G_1–59/E_18–38 determined, which might be lower in the entire enzyme, is in line with two subunits, undergoing rearrangements during catalysis, assembly and regulative processes [41], [42].

Acknowledgements

We are grateful to Dr C. F. Liu and Ms I. Toh Mui Jin (School of Biological Sciences, NTU) for synthesizing the peptides. S. Rishikesan and S. Gayen are grateful to the authority of Nanyang Technological University (NTU) for awarding research scholarship. This research was supported by the School of Biological Sciences, NTU. Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper. A Declaration of interest statement reporting no conflict of interest has been inserted.