Binding of subunit E into the A–B interface of the A1AO ATP synthase

Biochimica et Biophysica Acta 1808 (2011) 2111–2118

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / b b a m e m

Binding of subunit E into the A–B interface of the A1AO ATP synthase Cornelia Hunke a, Martin Antosch b, 1, Volker Müller b, Gerhard Grüber a,⁎ a b

Nanyang Technological University, School of Biological Sciences, 60 Nanyang Drive, Singapore 637551, Republic of Singapore Molecular Microbiology and Bioenergetics, Institute of Molecular Biosciences, Goethe University, Max-von-Laue Strasse 9, D-60438 Frankfurt am Main, Germany

a r t i c l e

i n f o

Article history: Received 20 October 2010 Received in revised form 27 May 2011 Accepted 31 May 2011 Available online 7 June 2011 Keywords: Archaeal ATP synthase A1AO ATP synthase Methanosarcina mazei Gö1 Peripheral stalk Fluorescence correlation spectroscopy F1FO ATP synthase

a b s t r a c t Two of the distinct diversities of the engines A1AO ATP synthase and F1FO ATP synthase are the existence of two peripheral stalks and the 24 kDa stalk subunit E inside the A1AO ATP synthase. Crystallographic structures of subunit E have been determined recently, but the epitope(s) and the strength to which this subunit does bind in the enzyme complex are still a puzzle. Using the recombinant A3B3D complex and the major subunits A and B of the methanogenic A1AO ATP synthase in combination with fluorescence correlation spectroscopy (FCS) we demonstrate, that the stalk subunit E does bind to the catalytic headpiece formed by the A3B3 hexamer with an affinity (Kd) of 6.1 ± 0.2 μM. FCS experiments with single A and B, respectively, demonstrated unequivocally that subunit E binds stronger to subunit B (Kd = 18.9 ± 3.7 μM) than to the catalytic A subunit (Kd = 53.1 ± 4.4). Based on the crystallographic structures of the three subunits A, B and E available, the arrangement of the peripheral stalk subunit E in the A–B interface has been modeled, shining light into the A–B–E assembly of this enzyme. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Archaea synthesize adenosine triphosphate (ATP) from adenosine diphosphate (ADP) and inorganic phosphate by a so-called A1AO ATP synthase that is distinct from the known F1FO ATP synthases occurring in bacteria, mitochondria, and chloroplasts. The overall structure of A1AO ATP synthase can be divided into a cytoplasmic A1 domain that catalyzes ATP synthesis as well as — hydrolysis, and an integral membrane part, AO, which serves as an ion channel. The A1 sector consists of the alternating, nucleotide-binding subunits A and B, arranged as an A3B3 hexamer, in which catalysis takes place. Subunits D and F are directly linked to the catalytic hexamer and form together with subunit C the central stalk and thereby the coupling element for ion-transport in the AO sector and ATP synthesis in the A3B3 sector [1,2]. Besides a central stalk, the A1AO ATP synthase comprises two peripheral stalks, and a collar-like structure, proposed to be composed of the subunits H, E and a, respectively [3,4]. Both the peripheral stalks and the collar-like structure are believed to be involved in constituting the stator region of the enzyme, which is predicted to have important roles in enzyme assembly as well as energy coupling between A1 and AO [5]. The peripheral stalk subunit H of the

⁎ Corresponding author. Tel.: + 65 6316 2989; fax: + 65 6791 3856. E-mail address: [email protected] (G. Grüber). 1 Present address: Cell Biology & Plant Biochemistry, University of Regensburg, Universitätsstrasse, 31, 93503 Regensburg, Germany. 0005-2736/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.bbamem.2011.05.022

methanogenic A1AO ATP synthase binds with its C-terminus (residues 98–104) to the N-terminal region of the catalytic A subunit (Kd = 206 nM [6]) and is associated with the N-terminal domain of subunit E via its residues 1–24 [7] as demonstrated by intrinsic fluorescence spectroscopy and nuclear magnetic resonance spectroscopy (NMR), respectively. These data reflect that subunit H provides a critical link between the so-called collar-like domain and the catalytic A subunit. The same function is proposed for subunit E, whose structural traits have been solved by X-ray crystallography [8–10] and NMR spectroscopy (10). The protein consists of two segments, the Cterminal α–β globular head part, with three α-helices and four antiparallel β-strands as well as the very C-terminal single α-helix [8– 10]. The N-terminus of subunit E adopts a regular α-helix, which is not straight but slightly bent at opposite directions [9,11]. The crosier-like shaped and more than 140 Å long subunit E [9,10] is proposed to be linked to the A3B3 headpiece via its C-terminal globular domain, which has been discussed to fit into the knobs at the top of the 3Dreconstructions of A1AO ATP synthases [4,9]. First binding experiments, using a 23 residue peptide of the N-terminus of subunit B and truncated forms of subunit E showed a weak interaction (Kd = 2 mM) of the C-terminal region of subunit E with the N-terminal peptide of B [12]. In the presented studies we have determined the binding constants of entire subunit E to the enzymatic active A3B3D complex as well as to the recombinant subunits A and B of the methanogenic A1AO ATP synthase, respectively, in order to localize and quantify the binding of the peripheral stalk subunit E inside the catalytic A3B3 headpiece.

2112

C. Hunke et al. / Biochimica et Biophysica Acta 1808 (2011) 2111–2118

2. Materials and methods 2.1. Cloning and purification of the A3B3D complex from Methanosarcina mazei Gö1

Eq. (1) is the calculation of the number of fluorescent molecules from the amplitudes G(0) of autocorrelation curves. G(0): correlation amplitude at τ=0, N: average number of fluorescence molecules in the

From chromosomal DNA of M. mazei Gö1 a fragment containing the genes ahaA, ahaB, and ahaD was amplified and a XbaI and KpnI site was introduced into the 5′ and 3′ end, respectively. The fragment was cloned into pGEM-4Z [13], resulting in plasmid pMG3. The construct contains a sequence encoding an in frame hexa-histidine tag at the N-terminus of subunit A. The engineered plasmid was afterwards transformed into E. coli DK8 cells, which were used to produce the A3B3D subcomplex. The bacteria cultures were grown at 37 °C and shaken at 180 rpm until an optical density OD600 was about 0.7. The protein production was induced by the addition of isopropyl-β-D-thiogalactopyranoside (IPTG). The cells were harvested at 8500×g for 10 min, 6 °C, flash frozen by liquid nitrogen and stored at −80 °C. The cells, containing the A3B3D protein complex, were thawed and lysed on ice in buffer A (50 mM Tris, pH 7.5, 350 mM NaCl, 0.8 mM DTT, 1 mM PMSF, 4 mM Pefabloc SC) by sonication with a Ultrasonic Homogenizer (Bandelin, Tip KE76) for 4 × 1 min. After sonication the cell lysate was ultra-centrifuged at 100,000 ×g for 25 min at 4 °C. The resulting supernatant was passed through an amicon cell (0.45 mm pore-size) and combined with Ni2+-NTA resin. The His-tagged protein was allowed to bind to the matrix for 2 h at 4 °C by revolving on a sample rotator (Neolab), followed by an imidazole-gradient (0 to 600 mM) in buffer A. The respective fractions (125–400 mM imidazole) were pooled and concentrated using a Millipore spin concentrator with a molecular mass cut-off of 30 kDa and applied to Superdex 200 HR 10/30 gel filtration column (GE Healthcare). Respectable emerged fractions were further concentrated in Millipore spin concentrators. The purity of the protein sample was analyzed by SDS-PAGE [14] which was stained by Coomassie Brilliant Blue R250. The ATP hydrolysis was performed using Malachite Green Assay Kit (Cayman) to determine the hydrolytic ability of the A3B3D complex [15]. 2.2. Purification of recombinant subunits A, B and E The recombinant subunits A and B have been described previously [16]. Subunit E was purified firstly by affinity chromatography using 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 (0–250 mM) in buffer comprised of 50 mM Tris, pH 7.5, 150 mM NaCl, 1 mM PMSF and 4 mM Pefabloc SC. Subsequently the eluted protein fractions, containing subunit E, were applied on a size exclusion column Superdex HR75 (10/30, GE Healthcare). Pure subunit E was eluted with buffer, comprised of 50 mM Tris, pH 7.5 and 250 mM NaCl, and respective fractions were concentrated in Millipore spin concentrators. 2.3. Determination of MgATP-binding to A3B3D by fluorescence correlation spectroscopy Fluorescence correlation spectroscopy (FCS) was performed on an LSM 510 Meta/ConfoCor 3 (Zeiss, Jena, Germany) using the fluorescently labeled ATP analog EDA-ATP ATTO-647N (ATTO-TEC, Siegen, Germany). The number of fluorescently labeled ATP molecules bound to the A3B3D-complex was determined to be approximately 2.5. The number of molecules was determined by FCS using Eq. (3). To determine the binding degree (number of ATP-molecules bound to the complex) the number of fluorescence molecules bound to the A3B3D–MgATP complex was divided by the number of molecules of Atto647N-ATP alone (Eq. (1)). N=

1 1 = Gð0Þ−1 V Ac

ð1Þ

Fig. 1. (A) Purification of the M. mazei Gö1 A3B3D subcomplex by gel filtration chromatography. The inset in the figure shows an SDS gel of A3B3D after purification via Superdex 200, with indicated fraction from elution peaks. (B) Binding study of A3B3D to MgATP ATTO-647N by fluorescence correlation spectroscopy. Normalized autocorrelation functions of Atto647N-ATP by increased concentrations of A 3 B 3 D. (C) Concentration-dependent binding traits of fluorescently labeled Mg-ATP to A3B3D. Best fits yielding the binding constants are represented as fitted line by a non-linear, Hill curve fit.

C. Hunke et al. / Biochimica et Biophysica Acta 1808 (2011) 2111–2118

focus, V: confocal volume, A: Avogadro number (6.023 1023 mol− 1), c: concentration of excitable molecules [17]. Binding degree =

½A3 B3 D−MgATP complex ½MgATP

ð2Þ

Eq. (2) is the determination of the binding degree. For the FCS experiment 50 mM Tris, pH 7.5, 350 mM NaCl, 1 mM DTT containing buffer were used. The temperature was adjusted to 25 °C in an incubation chamber (Zeiss). The 633 nm laser line of an HeNe633 laser was attenuated to 5 mW and 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 of the diluted fluorescent derivatives of ATP, which were placed on Nunc 8 well chambered cover glass. Before usage, the cover glasses were treated with 3% of gelatin, in order to prevent unspecific binding and removed by washing steps with H2O [18]. The following filter sets were used: MBS: HFT 514/633, EF1: BP 655–710 IR, EF: none, and DBS: none. Outof-focus fluorescence was rejected by a 90 μm pinhole in the detection pathway. The confocal detection volume was calculated to be

2113

approximately 0.93 fl at λ = 655 nm at a numerical aperture NA of 1.2. Solution of Rhodamine 6G (Rho6G) in pure water (Fluka) was used as references for the calibration of the confocal microscope. Variable concentrations of protein solutions were mixed after addition of MgCl2 and fluorescently labeled nucleotide. The drop was incubated on the glass slip surface for 3 min, which was monitored by FCS. The fluorescence autocorrelation functions were determined by measurements of at least 10 repetitions with 30 s each. To analyze the autocorrelation functions of MgATP ATTO-647N bound to the A3B3D complex models with the diffusion time and the triplet state were used for fitting (FCS-LSM software, ConfoCor 3, Zeiss). The diffusion times of fluorescent nucleotide were measured independently and the determined value was kept fixed during the fitting of the FCS data. Based on this, the determination of the binding constants required only the calculation of the relative amounts of free nucleotide with the short diffusion time and of the bound nucleotides with the diffusion time of A3B3D. The calculations of the bound fractions and the dissociation constants were done by the ConfoCor 3software 4.2, Excel2003 and OriginPro 8 SR4. A ‘Standard AC normalized triplet’ model was used to determine the diffusion time

Fig. 2. (A) Subunit E was reacted with an Atto488-NHS ester as described under the Materials and methods section and applied to an SDS-PAGE. Lane 1, the gel stained with Coomassie; lane 2, the fluorogram of the same gel. (B) Binding study of A3B3D to ATTO-488 labeled subunit E by fluorescence correlation spectroscopy. Normalized autocorrelation functions of subunit E by increased concentrations of A3B3D. (C) Concentration dependent binding traits of A3B3D to labeled subunit E. Best fits yielding the binding constants are represented as fitted line by a non-linear, curve fit by a Hill equation.

2114

C. Hunke et al. / Biochimica et Biophysica Acta 1808 (2011) 2111–2118

of the fluorescently labeled subunit E. The general mathematical model for autocorrelations is shown in Eq. (3) derived from Weisshart et al. [19] which is used in the ConfoCor 3 software. “The determined value was fixed for the fitting in the titration experiments, whereby the ‘Standard AC2 diffusion coefficients normalized triplet’ model has been used, which takes two components in consideration.” The second parameter was determined by keeping the second diffusion time free during fitting and verified also in further titrated concentrations. The identified second diffusion time value was subsequently set as the second diffusion time value. The resulting bound fractions were analyzed in Windows Excel and visualized Origin. The standard deviations of the calculated bound-fractionsdata-points were evaluated in Excel. The resulting sigmoidal nonlinear curve fit was used to calculate the binding constant in OriginPro 8. The according autocorrelation functions were normalized in Excel and graphic figures were generated in Origin. "

GðτÞ = 1 +

−τt

#

0

1

1 Fe F B B∑  1+ N 1−F @i = 1 1+ x

τ τD

i

Y  i 1+

τ τD S2

C  C A 1

ð3Þ

2

i

Eq. (3) is the fitting model with free translational 3D-diffusion, x components and triplet fraction. Di: diffusion constant of particle i, F: triplet fraction, F = 0…1, N: total number of particles, S: structure parameter, τ: correlation time, τF: triplet time, τDi: diffusion time of particle i, Yi: relative molecular fraction of particle i, x: number of particle species, i = 1..x.

2.5. A–E and B–E formation and chromatographic separation To reconstitute an A–E or A–B heterodimer the recombinant subunits A and E or B and E were incubated for 2 h at room temperature in a molar ratio of 1:1.5. The separation of the A–E or B–E assembly and the excess of subunit E were performed by gel filtration chromatography using a Superdex 75 HR 10/30 column (GE Healthcare) in a buffer compost of 50 mM Tris/HCl, pH 7.5 and 150 mM NaCl. The observed peak fractions were applied to a SDS-PAGE which was stained by Coomassie Brilliant Blue G250. 2.6. Modeling of the A3B3E complex The model of the A3B3 complex was generated by fitting the A1AO ATP synthase crystal structure of subunit A (PDB 3M4Y, [20]) and subunit B (green, PDB 2C61, [16]) as described recently. This complex was chosen for docking studies with the crystallographic structure of subunit E E101–206 (PDB 3LG8, [21]). The web interface of the protein docking software GRAMM [22] was used for the docking studies. GRAMM does a global search using the Fast Fourier Transform methodology to identify the best rigid body conformations between the two proteins. The best configuration was scored in GRAMM 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.

2.4. Subunit interactions of E with A3B3D and the subunits A and B by FCS Using subunit E, which was fluorescently labeled by an Atto488NHS ester (ATTO-TEC, Siegen/Germany), the binding of E to the A3B3D part and the subunits A as well as B has been investigated by FCS. Prior to the labeling a buffer-exchange was performed three times in a 3 kDa cutoff centrifugal filter device (Millipore). The labeling was performed in 25 mM Hepes buffer, pH 8.1 with 150 mM NaCl for 10 min at room temperature. To remove the excess of dye the diluted sample was purified by a size exclusion column (Superdex 75 HR 10/30; GE Healthcare), using 50 mM Tris, pH 7.5, 200 mM NaCl. The labeled protein was further concentrated. The labeling efficiency determined by Biospec-nano Spectrophotometer (Shimadzu) was about 83%. During the experiments the temperature was set in an incubation chamber (Zeiss) to 25 °C. Increasing amounts of A 3 B 3 D or recombinant subunits A and B have been added, respectively. The 488 nm laser line of a 30 mW argon ion laser with the following filter sets was used: MBS: HFT 488, EF: none, DBS: mirror, EF2: LP505 (pinhole: 90 μm). The 15 μl droplet was equilibrated on gelatin pre-treated Nunc 8-well chambered cover glass for about 3 min. Solutions of Atto488-NHS ester in Tris buffer were used as references for the calibration of the ConfoCor 3 system. Standard autocorrelation two-diffusion-coefficient normalized triplet model was used for fitting (FCS-LSM software, ConfoCor 3, Zeiss) to analyze the autocorrelation functions of subunit E bound to subunits A, B or the A3B3D-complex alternatively. The diffusion time of fluorescently labeled subunit E was measured independently, and kept fixed during the fitting of the FCS titration data. The determination of the binding constants required the calculation of the relative amounts of free labeled subunit E with the short diffusion time versus the diffusion time of the larger complex of the nonlabeled protein with subunit E. The calculations of the bound fractions and the dissociation constants were done as described above.

Fig. 3. Binding of subunit E to subunit B by fluorescence correlation spectroscopy. (A) Normalized autocorrelation functions of Atto488-subunit E by increased concentrations of subunit B. (B) Concentration dependent binding traits of subunit E to subunit B. Best fits yielding the binding constants are represented as fitted line by a non-linear, Boltzmann curve fit.

C. Hunke et al. / Biochimica et Biophysica Acta 1808 (2011) 2111–2118

From these 150 complex configurations 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 [12], and has some kind of interaction in the complex. The A3B3 complex E101–206 and the docked complex were optimized using the program Swiss-PdbViewer 3.7 [23] by carrying out energy minimization in vacuo with the GROMOS96 43B1 parameter set, without reaction field as implemented in the program. For fast energy convergence initially 200 steps of Steepest Descent protocol was carried out followed by multiple steps of Conjugate Gradient technique until the derivative convergences to 0.050 kJ/mol. 3. Results and discussion 3.1. Production and characterization of the M. mazei Gö1 A3B3D subcomplex The recombinant A3B3D part of the A1AO ATP synthase from M. mazei Gö1 was purified by affinity chromatography and size exclusion chromatography, resulting in an highly pure complex with the right stoichiometry of A3:B3:D (Fig. 1A). As A3B3D forms a subcomplex of the hydrolytic A1 ATPase its ability to bind its substrate Mg-ATP with

2115

high affinity was studied by fluorescence correlation spectroscopy using the fluorescent ATP-analog ATP ATTO-647N. As an aspired reference the mean count rate per Rhodamine 6G (Rho6G) fluorophore was determined to be 50.2 ± 3.6 kHz. Compared to Rho6G the value of MgATP ATTO-647N was 16.7 ± 1.6 kHz. Fitting the autocorrelation functions resulted in characteristic times of diffusion τD = 71.5 ± 0.7 μs for Rho6G and τD = 78.5 ± 1.7 μs for MgATP ATTO647N. Fig. 1B shows the fitted autocorrelation curves of MgATP ATTO647N in the absence and presence of protein. The addition of A3B3D resulted in a significant change of the mean diffusion time τD. The diffusion time for Mg-Atto-ATP bound to ABD was determined to be 496.4 ± 42.1 μs. The increase of the diffusion time was due to a raise in the mass of the diffusing particle, when MgATP ATTO-647N bound to the protein complex. The determined bound fractions for rising concentrations of A3B3D versus MgATP ATTO-647N are shown in Fig. 1C. An affinity (Kd) of A3B3D bound to MgATP ATTO-647N of 441 ± 10 nM was determined (Fig. 1C). By comparison, Kd values of MgATP binding to single subunits A and B have been determined to be 2.38 μM [20] and 22 μM [24], respectively. The ATPase activity of A3B3D in the presence of MgCl2 was determined to be 1.3 μmol ATP hydrolyzed/(min·mg).

Fig. 4. Subunit A–E formation. (A) Normalized autocorrelation functions of Atto488-subunit E by increasing the quantity of subunit A. (B) Concentration dependent binding of subunit E to subunit A. The binding constant was calculated by determination of the inflection points of the normalized autocorrelation functions. Best fits yielding the binding constants are represented as fitted line by a non-linear, Boltzmann curve fit. (C) Elution profile and electrophoretic analysis of the assembled A–E heterodimer. Subunits A and E were incubated in a ratio of 1:1.5 for 2 h at room temperature, before applying on a Superdex 75 HR 10/30 column. (D) The samples of peak I and II have been applied on to an SDS-PAGE in lanes 1 and 2, respectively. In comparison, the same procedure has been used for the B–E-assembly, resulting in a similar chromatographic pattern with peak I, including the B–E heterodimer (lane 3).

2116

C. Hunke et al. / Biochimica et Biophysica Acta 1808 (2011) 2111–2118

3.2. Subunit E binding to the A3B3D subcomplex The first 3D reconstruction of the A1 ATPase from M. mazei Gö1 shows a groove at the top of the A3B3 headpiece, which is formed by the alternating subunits A and B [25,26], and proposed to form the binding epitope of one of the peripheral stalks of the A-ATP synthase. Using the enzymatic active A3B3D subcomplex binding of the

peripheral stalk subunit E to the A3B3 headpiece of the enzyme complex has been analyzed by FCS. Increasing amounts of recombinant A3B3D have been added to subunit E, which was labeled by the fluorophore Atto-488 (Fig. 2A). To calibrate the instrument Rhodamine 6G was used as a reference and a mean count rate of 50.2 ± 3.6 kHz with a characteristic diffusion time of 71.5 ± 0.7 μs was determined. A mean count rate of 25.3 ± 1.2 kHz was determined for

Fig. 5. (A) Surface representation of the P. furiosus A1AO ATP synthase modified according to [11]. Reprinted with permission from this article, copyright 2009, the American Society for Biochemistry and Molecular Biology. (B) ABE model structure of the methanogenic A1AO ATP synthase in side and (C) top view. The model was generated by fitting the A1AO ATP synthase crystal structure of subunit A (orange, PDB 3M4Y [20]), subunit B (green, PDB 2C61 [16]) and subunit E (red), with the N-terminal NMR structure E1–52 (PDB 2KK7, [11]) and E101–206 (PDB 3LG8 [27]). The so far structurally unresolved segment of amino acids 53 to 100 of the methanogenic subunit E is highlighted as a mesh representation. Focus on interacting β-sheets of subunits A (orange) and E (red) (D) as well as the amino acid residues of subunit B (green) and E (red) involved in B–E binding (E).

C. Hunke et al. / Biochimica et Biophysica Acta 1808 (2011) 2111–2118

fluorescently labeled E without A3B3D with a characteristic diffusion time of 255.9 ± 14.2 μs. The diffusion time for fluorescently labeled subunit E bound to A3B3D was determined to be 384.2 ± 16.5 μs. The diffusion times we observed for subunit E are indicating an extended molecule, which is supported by NMR data of subunit E [11,27]. The diffusion coefficient of the spherical particles can be calculated by using the theoretical Stokes–Einstein relation [28]. For rod-like molecule on the other hand the diffusion coefficient is different due to the different hydrodynamic properties of elongated molecules in solution [29]. To determine the binding constant the calculated values of the fitted autocorrelation functions were plotted as normalized bound fractions versus the concentration of A3B3D. The addition of protein caused a significant change of the mean diffusion time, which reflects the interaction of Atto488-subunit E with A3B3D (Fig. 2B). To appraise a binding constant, a curve fitting was performed by using a nonlinear Boltzmann function and a binding constant of 6.1 ± 0.16 μM was calculated for the interaction of A3B3D to subunit E (Fig. 2C). 3.3. Binding of subunits E and B Previously the 3D-reconstruction of the A1AO ATP synthase indicates that one of the peripheral stalks fits into the A–B interface, with a closer proximity to subunit B [16]. In order to demonstrate the B–E assembly the binding of fluorescently labeled subunit E to recombinant B has been studied. The measured autocorrelation curves of the Atto488 labeled subunit E in the absence and presence of nonlabeled subunit B are shown in Fig. 3A. The diffusion time for Atto488 labeled subunit E bound to subunit B was determined to be 331.8 ± 12.8 μs. The addition of recombinant B caused a significant change of the mean diffusion time τD, indicating binding of both proteins. The calculated values of the fitted autocorrelation functions were plotted as normalized bound fractions versus the concentration of subunit B (Fig. 3B). A curve was fitted into the data points by using a non-linear Boltzmann function, resulting in a binding constant of 18.9 ± 3.7 μM for the interaction of subunit B with subunit E (Fig. 3B).

2117

complex with the C-terminal peptide of subunit H, H91–111 [12]. Although it has been observed that subunit E of the T. acidophilum as well as the T. thermophilus A1AO ATP synthase would be fragile in the absence of subunit H [8], the recent [27] and the present study show that a single and stable subunit E can be isolated, which does efficiently bind to the A3B3D complex (Kd = 6.1 ± 0.16 μM) as well as to the single subunits A (Kd = 53.1 ± 4.4) and B (Kd = 18.9 ± 3.7 μM), respectively, with a stronger binding to B. The recent NMR solution [11] and crystallographic structure [10] of the N- and C-terminus of subunit E from the methanogenic A1AO ATP synthase reveal this subunit as a crosier-like shape of more than 140 Å in length with a slight twist in the N-terminal and α-helical segment, concordant with the crystallographic structure of subunit E of the T. thermophilus A1AO ATP synthase [9]. As described for the M. jannaschii A1AO ATP synthase [26] and confirmed in 3D-reconstrutions of the Pyrococcus furiosus enzyme [4] one of the peripheral stalks has a kinked feature in the lower part, giving this region the flexibility to bind along the A–B interface and end up as a cap on the top of the A3B3-hexamer (Fig. 5A). Based on the data described above, we have modeled the structure of the methanogenic E subunit relative to the atomic structures of subunits A [20] and B [16], indicating the bend, formed by the Nterminal region of E, which is in close proximity with subunit B. The curvature (Fig. 5B) is nicely in line with the one observed in the 3Dreconstruction shown in Fig. 5A. The curveted peripheral stalk binds with its globular C-terminal region in the A–B interface at the top of the enzyme (Fig. 5C). This brings the two N-terminal β-sheets of subunit A, including Ile177 to Thr179 and Leu187 to Gly182, and the two β-sheets of subunit E (Ala168–Ile172 and Leu179–Asp183) close together, enabling a hydrogen bonding interaction between them (Fig. 5D). At the same time the very C-terminus of subunit E interacts via its positively charged Arg198 and its hydrophobic residue Ile201 to the residues Glu19 and Val18 at the very N-terminus of subunit B, respectively (Fig. 5E). Taking into account that the A3B3 headpiece of the methanogenic A1AO ATP synthase is about 94 Å high and the collar domain 45 Å below the headpiece [12,26], the subunit arrangement presented would enable the 140 Å long subunit E to connect both, the N-terminal A–B interface and the collar segment (Fig. 5A).

3.4. Assembly of the A1AO ATP synthase subunits A and E As subunit E is proposed to be allocated in the A–B interface, we looked for the interaction of subunits A to E, using Atto488 labeled subunit E. The diffusion time for Atto488 labeled subunit E bound to subunit A was determined to be 297.0 ± 26.9 μs. As revealed in Fig. 4A the addition of recombinant A to labeled subunit E caused a strong increase of the mean diffusion time τD, correlating with the increase in the mass of the diffusing particle, when Atto488-subunit E was interacting with subunit A. The calculated values of the fitted autocorrelation functions were plotted as normalized bound fractions versus the concentration of subunit A. A Kd of 53.1 ± 4.4 μM was determined for the assembly of subunit A with subunit E (Fig. 4B). To confirm the A–E formation both subunits A and E have been incubated for 2 h in a molar ratio of 1:1.5 as described in the Materials and methods section. Size exclusion chromatography of this mixture resulted in an elution diagram with two main peaks I–II, consisting of an A–E assembly (I) and the excess of subunit E (II), indicating the A–E formation observed in the FCS data described above. 3.5. Arrangement of subunit E inside the A3B3 headpiece Previously NMR titration experiments, with a peptide of the Nterminal 23 residues of subunit B and the recombinant C-terminal globular domain of the Thermoplasma acidophilum A1AO ATP synthase, consisting of the residues 83 to 185 of subunit E, have shown that both peptides interact with an affinity of 2 mM [12]. However, these studies also showed that the N-terminal peptide of subunit B binds to the C-terminal domain of subunit E only with the later being in

Abbreviations IPTG NTA PAGE SDS Tris DTT FCS

isopropyl-β-D-thio-galactoside nitrilotriacetic acid polyacrylamide gel electrophoresis sodium dodecyl sulfate Tris-(hydroxymethyl)aminomethane dithiothreitol fluorescence correlation spectroscopy

Acknowledgements We thank Dr. M. S. S. Manimekalai (SBS, NTU) for art work. This research was supported by A*STAR BMRC (08/1/22/19/576; G. Grüber) and SFB 807 (V. Müller). References [1] E.J. Boekema, J.F.L. van Bremen, A. Brisson, T. Ubbink-Kok, W.N. Konings, J.S. Lolkema, Connecting stalks in V-type ATPase, Nature 401 (1999) 37–38. [2] V. Müller, G. Grüber, ATP synthases: structure function and evolution of unique secondary energy converters, Cell. Mol. Life Sci. 60 (2003) 474–494. [3] G. Grüber, V. Marshansky, New insights into structure–function relationships between archeal ATP synthase (A1AO) and vacuolar type ATPase (V1VO), BioEssays 30 (2008) 1096–1109. [4] J. Vonck, K.Y. Pisa, N. Morgner, B. Brutschy, V. Müller, Three-dimensional structure of A1AO ATP synthase from the hyperthermophilic archaeon Pyrococcus furiosus by electron microscopy, J. Biol. Chem. 284 (2009) 10110–10119. [5] I. Schäfer, M. Rössle, G. Biuković, V. Müller, G. Grüber, Structural and functional analysis of the coupling subunit F in solution and topological arrangement of the

2118

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14] [15]

[16]

C. Hunke et al. / Biochimica et Biophysica Acta 1808 (2011) 2111–2118 stalk domains of the methanogenic A1AO ATP synthase, J. Bioenerg. Biomembr. 38 (2006) 83–92. G. Biuković, M. Rössle, S. Gayen, Y. Mu, G. Grüber, Small-angle X-ray scattering reveals the solution structure of the peripheral stalk subunit H of the A1AO ATP synthase from Methanocaldococcus jannaschii and its binding to the catalytic A subunit, Biochemistry 46 (2007) 2070–2078. S. Gayen, A.M. Balakrishna, G. Biuković, W. Yulei, G. Grüber, Identification of critical residues of subunit H in its interaction with subunit E of the A-ATP synthase from Methanocaldococcus jannaschii, FEBS J. 275 (2008) 1803–1812. N.K. Lokanath, Y. Matsuura, C. Kuroishi, N. Takahashi, N. Kunishima, Dimeric core structure of modular stator subunit E of archaeal H+-ATPase, J. Mol. Biol. 366 (2006) 933–944. L.K. Lee, A.G. Stewart, M. Donohoe, R.A. Bernal, D. Stock, The structure of the peripheral stalk of Thermus thermophilus H+-ATPase/synthase, Nat. Struct. Mol. Biol. 17 (2010) 373–378. A.M. Balakrishna, M.S.S. Manimekalai, C. Hunke, S. Gayen, M. Rössle, J. Jeyakanthan, G. Grüber, Crystal and solution structure of the C-terminal part of the Methanocaldococcus jannaschii A1AO ATP synthase subunit E revealed by X-ray diffraction and small-angle X-ray scattering, J. Bioenerg. Biomembr. 42 (2010) 311–320. S. Gayen, A.M. Balakrishna, G. Grüber, NMR solution structure of the N-terminal domain of subunit E (E1–52) of A1AO ATP synthase from Methanocaldococcus jannaschii, J. Bioenerg. Biomembr. 41 (2009) 343–348. E. Kish-Trier, S. Wilkens, Interaction of the Thermoplasma acidophilum A1A0-ATP synthase peripheral stalk with the catalytic domain, FEBS Lett. 583 (2009) 3121–3126. T. Lemker, G. Grüber, R. Schmidt, V. Müller, Defined subcomplexes of the A1 ATPase from the archaeon Methanosarcina mazei Gö1: biochemical properties and redox regulation, FEBS Lett. 544 (2003) 206–209. U.K. Laemmli, Cleavage of structural proteins during the assembly of the head of bacteriophage T4, Nature 227 (1970) 680–685. G. Grüber, J. Godovac-Zimmermann, T. Nawroth, ATP-synthesis and -hydolysis of the ATP-synthase from Micrococcus luteus regulated by an inhibitor subunit and membrane energization, Biochim. Biophys. Acta Bioenerg. 1186 (1994) 43–51. I. Schäfer, S.M. Bailer, M.G. Düser, M. Börsch, A.B. Ricardo, D. Stock, G. Grüber, Crystal structure of the archaeal A1AO ATP synthase subunit B from Methanosarcina mazei Gö1: implications of nucleotide-binding differences in the major A1AO subunits A and B, J. Mol. Biol. 358 (2006) 725–740.

[17] A. Pramanik, J. Widengren, Fluorescence Correlation Spectroscopy (FCS), Encyclopedia of Molecular Cell Biology and Molecular Medicine, Wiley-VCH, Germany, 2006. [18] C. Hunke, W.-Y. Chen, H.-J. Schäfer, G. Grüber, Cloning, purification, and nucleotide-binding traits of the catalytic subunit A of the catalytic V1VO ATPase from Aedes albopictus, Protein Expr. Purif. 53 (2007) 378–383. [19] K. Weisshart, V. Jüngel, S.J. Briddon, The LSM 510 META–ConfoCor 2 system: an integrated imaging and spectroscopic platform for single-molecule detection, Curr. Pharm. Biotechnol. 5 (2004) 135–154. [20] A. Kumar, M.S.S. Manimekalai, A. Balakrishna, J. Jeyakanthan, G. Grüber, Nucleotide-binding states of subunit A of the A-ATP synthase and the implication of P-loop switch in evolution, J. Mol. Biol. 396 (2010) 301–320. [21] G. Grüber, D.I. Svergun, Ü. Coskun, T. Lemker, M.H.J. Koch, H. Schägger, V. Müller, Structural insights into the A1 ATPase from the archaeon, Methanosarcina mazei Gö1, Biochemistry 40 (2000) 1890–1896. [22] A. Tovchigrechko, I.A. Vakser, GRAMM-X public web server for protein–protein docking, Nucleic Acids Res. 34 (2006) 310–314. [23] N. Guex, M.C. Peitsch, SWISS-MODEL and the Swiss-PdbViewer: an environment for comparative protein modeling, Electrophoresis 18 (1997) 2714–2723. [24] A. Kumar, M.S.S. Manimekalai, A.M. Balakrishna, C. Hunke, S. Weigelt, N. Sewald, G. Grüber, Spectroscopic and crystallographic studies of the mutant R416W give insight into the nucleotide binding traits of subunit B of the A1AO ATP synthase, Proteins 75 (2009) 807–819. [25] Ü. Coskun, M. Radermacher, V. Müller, R. Ruiz, G. Grüber, Three-dimensional organization of the archaeal A1-ATPase from Methanosarcina mazei Gö1, J. Biol. Chem. 279 (2004) 22759–22764. [26] Ü. Coskun, Y.L. Chaban, A. Lingl, V. Müller, W. Keegstra, E.J. Boekema, G. Grüber, Structure and subunit arrangement of the A-type ATP synthase complex from the archaeon Methanococcus jannaschii visualized by electron microscopy, J. Biol. Chem. 279 (2004) 38644–38648. [27] A.M. Balakrishna, C. Hunke, G. Grüber, Purification and crystallization of the entire recombinant subunit E of the energy producer A1AO ATP synthase, Acta Crystallogr. F66 (2010) 324–326. [28] T. Krouglova, J. Vercammen, Y. Engelborghs, Correct diffusion coefficients of proteins in fluorescence correlation spectroscopy. Application to tubulin oligomers induced by Mg2+ and Paclitaxel, Biophys. J. 87 (2004) 2635–2646. [29] M.M. Tirado, Comparison of theories for the translational and rotational diffusion coefficients of rod-like macromolecules. Application to short DNA fragments, J. Chem. Phys. 81 (1984) 2047–2052.