Portal de Periódicos da CAPES

The ATP Hydrolysis Cycle of the Nucleotide-binding Domain of the Mitochondrial ATP-binding Cassette Transporter Mdl1p

2003; Elsevier BV; Volume: 278; Issue: 29 Linguagem: Inglês

10.1074/jbc.m301227200

1083-351X

Eva Janas, Matthias Hofacker, Min Chen, Simone Gompf, Chris van der Does, Robert Tampé,

Trace Elements in Health

The ABC transporter Mdl1p, a structural and functional homologue of the transporter associated with antigen processing (TAP) plays an important role in intracellular peptide transport from the mitochondrial matrix of Saccharomyces cerevisiae. To characterize the ATP hydrolysis cycle of Mdl1p, the nucleotide-binding domain (NBD) was overexpressed in Escherichia coli and purified to homogeneity. The isolated NBD was active in ATP binding and hydrolysis with a turnover of 25 ATP per minute and a K m of 0.6 mm and did not show cooperativity in ATPase activity. However, the ATPase activity was non-linearly dependent on protein concentration (Hill coefficient of 1.7), indicating that the functional state is a dimer. Dimeric catalytic transition states could be trapped either by incubation with orthovanadate or beryllium fluoride, or by mutagenesis of the NBD. The nucleotide composition of trapped intermediate states was determined using [α-32P]ATP and [γ-32P]ATP. Three different dimeric intermediate states were isolated, containing either two ATPs, one ATP and one ADP, or two ADPs. Based on these experiments, it was shown that: (i) ATP binding to two NBDs induces dimerization, (ii) in all isolated dimeric states, two nucleotides are present, (iii) phosphate can dissociate from the dimer, (iv) both nucleotides are hydrolyzed, and (v) hydrolysis occurs in a sequential mode. Based on these data, we propose a processive-clamp model for the catalytic cycle in which association and dissociation of the NBDs depends on the status of bound nucleotides. The ABC transporter Mdl1p, a structural and functional homologue of the transporter associated with antigen processing (TAP) plays an important role in intracellular peptide transport from the mitochondrial matrix of Saccharomyces cerevisiae. To characterize the ATP hydrolysis cycle of Mdl1p, the nucleotide-binding domain (NBD) was overexpressed in Escherichia coli and purified to homogeneity. The isolated NBD was active in ATP binding and hydrolysis with a turnover of 25 ATP per minute and a K m of 0.6 mm and did not show cooperativity in ATPase activity. However, the ATPase activity was non-linearly dependent on protein concentration (Hill coefficient of 1.7), indicating that the functional state is a dimer. Dimeric catalytic transition states could be trapped either by incubation with orthovanadate or beryllium fluoride, or by mutagenesis of the NBD. The nucleotide composition of trapped intermediate states was determined using [α-32P]ATP and [γ-32P]ATP. Three different dimeric intermediate states were isolated, containing either two ATPs, one ATP and one ADP, or two ADPs. Based on these experiments, it was shown that: (i) ATP binding to two NBDs induces dimerization, (ii) in all isolated dimeric states, two nucleotides are present, (iii) phosphate can dissociate from the dimer, (iv) both nucleotides are hydrolyzed, and (v) hydrolysis occurs in a sequential mode. Based on these data, we propose a processive-clamp model for the catalytic cycle in which association and dissociation of the NBDs depends on the status of bound nucleotides. ATP-binding cassette (ABC) 1The abbreviations used are: ABC, ATP-binding cassette; AMP-PNP, 5′-adenylyl-β,γ-imidodiphosphate; ATPγS, adenosine 5′-O-(3-thio)triphosphate; ER, endoplasmic reticulum; IC50, 50% inhibitory concentration; mAAA, matrix-oriented ATPases associated with a variety of cellular activities; MALDI-TOF MS, matrix-assisted laser desorption/ionization-time of flight mass spectrometry; NBD, nucleotide-binding domain; PBS, phosphate-buffered saline; P-gp, P-glycoprotein; TMD, transmembrane domain; TLC, thin layer chromatography; TAP, transporter associated with antigen processing.1The abbreviations used are: ABC, ATP-binding cassette; AMP-PNP, 5′-adenylyl-β,γ-imidodiphosphate; ATPγS, adenosine 5′-O-(3-thio)triphosphate; ER, endoplasmic reticulum; IC50, 50% inhibitory concentration; mAAA, matrix-oriented ATPases associated with a variety of cellular activities; MALDI-TOF MS, matrix-assisted laser desorption/ionization-time of flight mass spectrometry; NBD, nucleotide-binding domain; PBS, phosphate-buffered saline; P-gp, P-glycoprotein; TMD, transmembrane domain; TLC, thin layer chromatography; TAP, transporter associated with antigen processing. transporters comprise a large family of membrane proteins that catalyze the active transfer of a variety of solutes across biological membranes (1Higgins C.F. Annu. Rev. Cell Biol. 1992; 8: 67-113Crossref PubMed Scopus (3364) Google Scholar). The function of ABC transporters is central to various human pathologies such as cystic fibrosis, adrenoleukodystrophy, retinal dystrophies, and multidrug resistance. The transporter associated with antigen processing (TAP) is an ABC transporter in vertebrates, which translocates peptides from the cytosol into the ER and performs a key function in the antigen presentation and adaptive immune response (2Lankat-Buttgereit B. Tampé R. Physiol. Rev. 2002; 82: 187-204Crossref PubMed Scopus (153) Google Scholar). Recently, a close homologue, Mdl1p (multidrug resistance like), localized in the inner mitochondrial membrane of Saccharomyces cerevisiae, has been identified as an intracellular peptide transporter (3Young L. Leonhard K. Tatsuta T. Trowsdale J. Langer T. Science. 2001; 291: 2135-2138Crossref PubMed Scopus (178) Google Scholar). This transporter exports peptides derived from the degradation of non-assembled membrane proteins. These peptides are generated by ATP-dependent m-AAA (matrix-oriented ATPases associated with a variety of cellular activities) proteases, which mediate the degradation and turnover of inner mitochondrial membrane proteins and short-lived regulatory proteins in an ubiquitin/proteasome-independent manner (4Arnold I. Langer T. Biochim. Biophys. Acta. 2002; 1592: 89Crossref PubMed Scopus (129) Google Scholar). Protein fragments with a length of 6–21 amino acids are released by Mdl1p into the intermembrane space (3Young L. Leonhard K. Tatsuta T. Trowsdale J. Langer T. Science. 2001; 291: 2135-2138Crossref PubMed Scopus (178) Google Scholar). Half-size ABC transporters, like the heterodimeric TAP and homodimeric Mdl1p, have a common molecular architecture consisting of two polytopic transmembrane domains (TMD) and two nucleotide-binding domains (NBD). The transmembrane domains interact with the substrates and form the substrate translocation pore across the membrane. The TMDs generally share little homology (5Holland I.B. Blight M.A. J. Mol. Biol. 1999; 293: 381-399Crossref PubMed Scopus (489) Google Scholar), probably caused by the broad substrate spectrum of the ABC transporter family. Binding and hydrolysis of nucleotides drive the transport process by transducing conformational changes from the NBDs to the TMDs. The similarity of different NBDs is significantly higher compared with the TMDs, suggesting that even in transporters of unrelated function the structure and function of the NBDs be highly conserved. Each NBD contains a highly conserved Walker A and Walker B motif (6Walker J.E. Saraste M. Gay N.J. Nature. 1982; 298: 867-869Crossref PubMed Scopus (73) Google Scholar) characteristic of ATP-binding P-loop proteins, as well as the C-loop motif (LSGGQ) unique to ABC proteins, which is also known as the ABC signature motif. The crystal structures of bacterial ABC transporters (e.g. MsbA, BtuCD) and of isolated NBDs (e.g. HisP, MalK, MJ1276, and TAP1) show a consensus fold for the monomer (7Hung L.W. Wang I.X. Nikaido K. Liu P.Q. Ames G.F. Kim S.H. Nature. 1998; 396: 703-707Crossref PubMed Scopus (618) Google Scholar, 8Diederichs K. Diez J. Greller G. Muller C. Breed J. Schnell C. Vonrhein C. Boos W. Welte W. EMBO J. 2000; 19: 5951-5961Crossref PubMed Scopus (276) Google Scholar, 9Karpowich N. Martsinkevich O. Millen L. Yuan Y.R. Dai P.L. MacVey K. Thomas P.J. Hunt J.F. Structure (Camb). 2001; 9: 571-586Abstract Full Text Full Text PDF PubMed Scopus (263) Google Scholar, 10Yuan Y.R. Blecker S. Martsinkevich O. Millen L. Thomas P.J. Hunt J.F. J. Biol. Chem. 2001; 276: 32313-32321Abstract Full Text Full Text PDF PubMed Scopus (224) Google Scholar, 11Gaudet R. Wiley D.C. EMBO J. 2001; 20: 4964-4972Crossref PubMed Scopus (245) Google Scholar, 12Chang G. Roth C.B. Science. 2001; 293: 1793-1800Crossref PubMed Scopus (584) Google Scholar, 13Locher K.P. Lee A.T. Rees D.C. Science. 2002; 296: 1091-1098Crossref PubMed Scopus (932) Google Scholar, 14Smith P.C. Karpowich N. Millen L. Moody J.E. Rosen J. Thomas P.J. Hunt J.F. Mol. Cell. 2002; 10: 139-149Abstract Full Text Full Text PDF PubMed Scopus (678) Google Scholar). This fold shows an L-shaped molecule with two arms, one including the ATP-binding domain containing the Walker A and B motifs, and the other including the ABC signature motif (7Hung L.W. Wang I.X. Nikaido K. Liu P.Q. Ames G.F. Kim S.H. Nature. 1998; 396: 703-707Crossref PubMed Scopus (618) Google Scholar). The crystal structures of several NBD dimers have been solved, but they differ significantly in the manner the monomers are associated (15Schmitt L. Tampé R. Curr. Opin. Struct. Biol. 2002; 12: 754-760Crossref PubMed Scopus (276) Google Scholar, 16Kerr I.D. Biochim. Biophys. Acta. 2002; 1561: 47-64Crossref PubMed Scopus (118) Google Scholar). The crystal structure of the ATP-bound Rad50 dimer, a DNA repair enzyme that shares homology with the NBDs of ABC proteins is assumed to be the physiologically relevant dimer. This dimer contains two nucleotides, which are clamped at the interface between two monomers causing them to dimerize in a head-to-tail manner (17Hopfner K.P. Karcher A. Shin D.S. Craig L. Arthur L.M. Carney J.P. Tainer J.A. Cell. 2000; 101: 789-800Abstract Full Text Full Text PDF PubMed Scopus (810) Google Scholar). The ATP molecules are sandwiched between the Walker A and B motifs from one monomer and the C-loop from the other monomer. This proposed functional dimer configuration was also found in the crystal structure of a mutant NBD from Methanococcus jannaschii MJ0796 (14Smith P.C. Karpowich N. Millen L. Moody J.E. Rosen J. Thomas P.J. Hunt J.F. Mol. Cell. 2002; 10: 139-149Abstract Full Text Full Text PDF PubMed Scopus (678) Google Scholar) and in the NBD of the full-length bacterial vitamin B12 transporter BtuCD (13Locher K.P. Lee A.T. Rees D.C. Science. 2002; 296: 1091-1098Crossref PubMed Scopus (932) Google Scholar). Biochemical data on NBDs like MalK or of P-gp further confirmed the involvement of the C-loop in ATP binding (18Fetsch E.E. Davidson A.L. Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 9685-9690Crossref PubMed Scopus (118) Google Scholar, 19Loo T.W. Bartlett M.C. Clarke D.M. J. Biol. Chem. 2002; 277: 41303-41306Abstract Full Text Full Text PDF PubMed Scopus (142) Google Scholar). The dimer is generally regarded as the ATPase active form of ABC transporters, but it remains controversial at which point of the transport cycle the dimer is formed and how both NBDs cooperate. Whether the motor domains work as equivalent modules, hydrolyzing both ATP synchronously, sequentially or alternatively is still under discussion. It is also not clarified how ATP hydrolysis powers the substrate translocation. Based on the P-gp ATPase cycle, it was suggested that hydrolysis of one ATP provides energy for the translocation of the substrate whereas hydrolysis of the ATP in the second binding pocket might have regulatory functions, possibly returning the complex to the starting point of the ATPase cycle (20Hrycyna C.A. Ramachandra M. Germann U.A. Cheng P.W. Pastan I. Gottesman M.M. Biochemistry. 1999; 38: 13887-13899Crossref PubMed Scopus (128) Google Scholar, 21Sauna Z.E. Muller M. Peng X.H. Ambudkar S.V. Biochemistry. 2002; 41: 13989-14000Crossref PubMed Scopus (94) Google Scholar). To obtain insight into the mechanism of the NBDs, we expressed and purified the NBD of Mdl1p. Either by incubation with orthovanadate or beryllium fluoride, or by mutagenesis of the NBD, catalytic transition states were trapped and the nucleotide composition of these states was analyzed. Based on these results, we propose a new model for the ATPase cycle combining structure and function of the ABC dimer. Heterologous Expression of Mdl1p-NBD in Escherichia coli—The C-terminal domain (amino acids 423–695) of Mdl1p (ORF YLR188W) was amplified by PCR on genomic DNA of S. cerevisiae using the following primers: 5′-TCCGACTATTGGAAAGGATCCTGTGTC-3′; 5′-TACTCGAGAAGCTTTATACTTCC CGGGCAACACTAT-3′ and cloned into the BamHI and HindIII sites of pET28b (Novagen). The resulting plasmid (pET-NBD) includes the NBD with an N-terminal His6 tag, a thrombin cleavage site and a T7 tag encoded by the vector. A conserved glutamate (E599Q, GAA->CAA) one amino acid downstream of the Walker B motif was changed to glutamine via PCR mutagenesis resulting in pET-NBD(E599Q). E. coli strain BL21(DE3) (Novagen) was transformed with the plasmids and grown in LB medium with 50 μg/ml kanamycin at 30 °C. At an OD600 of 0.5 the cells were induced with 0.2 mm isopropyl-β-d-thiogalactopyranoside, and after 3 h the cells were harvested by centrifugation. Purification of Mdl1p-NBD—The cell pellet was resuspended in lysis buffer (PBS: 10 mm Na2HPO4/NaH2PO4, pH 7.6, 150 mm NaCl, 5 mm KCl), 20 mm imidazole containing 1 mm phenylmethylsulfonylfluoride, 1% lysozyme, and 5–10 units of benzonase (Merck) and disrupted by French press treatment at 1,000 p.s.i. After centrifugation for 30 min at 100,000 × g, the supernatant was applied on a Ni-IDA column (Amersham Biosciences) preequilibrated in PBS supplemented with 20 mm imidazole. The column was washed with step gradients of 80 and 100 mm imidazole in PBS, and the protein was eluted with 250 mm imidazole. Fractions containing the NBD were pooled, concentrated by Centricon 10 (Millipore), and applied to a Superdex 200 HR 26/60 prep grade gel filtration column (Amersham Biosciences) in 20 mm Tris, pH 8.0, 100 mm NaCl. The NBD eluted with an apparent molecular weight of 33 kDa corresponding to the size of the monomer. Fractions were concentrated to 10 mg/ml and stored at 4 °C. Wild-type and mutant proteins were purified to greater than 99% homogeneity as assessed by Coomassie-stained protein on SDS-PAGE (15%) and MALDI-TOF MS. 8-Azido-ATP Photolabeling—After preincubation of the purified NBD (0.3 μm) with different concentrations of nucleoside tri/diphosphates for 5 min in binding buffer (20 mm Tris, pH 8.0, 100 mm NaCl, 5 mm MgCl2, 1 mm MnCl2) on ice, 8-azido-[α-32P]ATP was added to a final concentration of 0.5 μm and incubation was continued on ice for 5 min. Samples were irradiated by UV (254 nm) for 5 min, directly resuspended in SDS loading buffer, and separated by 15% SDS-PAGE. The gel was dried and quantified by phosphor imaging. The data were fitted to the Hill equation y = (axn/(bn + xn)), resulting both for ATP and ADP in a best fit (R 2 = 0.98) with a Hill coefficient (n) of 1. ATPase Assay—ATPase activities were measured by the malachite green assay as described previously (22Morbach S. Tebbe S. Schneider E. J. Biol. Chem. 1993; 268: 18617-18621Abstract Full Text PDF PubMed Google Scholar). The ATPase assay was performed in 20 mm Tris, pH 9.0, 100 mm NaCl and 15 mm MgCl2 with 0–150 μm NBD for either 2 min at 16 °C (high NBD concentrations) or 4 min at 30 °C. The reaction was started by the addition of ATP to final concentrations of 0 to 10 mm. Data were fitted to the Hill equation y = (axn /(bn + xn)). The ATPase activity of the E599Q mutant was determined using radiolabeled ATP. Shortly, E599Q NBD (250 μm) was incubated in 20 mm Tris, pH 9.0, 100 mm NaCl, and 15 mm MgCl2 at 30 °C. The reaction was started by addition of 2 mm MgATP supplemented with 0.04 μm [γ-32P]ATP (4500 Ci/mmol, ICN). Samples were taken at different time points (0–200 min), separated by thin layer chromatography (see below) and the rate of released radioactive phosphate was quantified by a phosphorimager. As a negative control, spontaneous ATP hydrolysis was tested in the presence of heat-inactivated NBD. Dimerization Assays—Before use, orthovanadate stock solution adjusted to pH 8.0 was boiled for 5 min to break polymeric species (23Urbatsch I.L. Sankaran B. Weber J. Senior A.E. J. Biol. Chem. 1995; 270: 19383-19390Abstract Full Text Full Text PDF PubMed Scopus (364) Google Scholar). Before use, beryllium fluoride solution was prepared by mixing BeCl2 and NaF in a ratio of 1:100. For non-radioactive assays, purified wild-type NBD (30 μm) was incubated with different concentrations of beryllium fluoride (BeFx) or orthovanadate in the presence or absence of nucleotides and nucleotide analogues for 5 min at 30 °C in binding buffer. The mutant NBD (E599Q, 30 μm) was incubated for 5 min under non-hydrolyzing conditions (in the absence of Mg2+ and on ice) or ATP hydrolysis conditions (in the presence of 5 mm Mg2+ and at 30 °C) with different concentrations of nucleotides and nucleotide analogues. The samples were applied to a Superdex 75 PC 3.2 (Amersham Biosciences) gel filtration column at a flow rate of 50 μl/min at 4 °C in 20 mm, Tris pH 8.0, 150 mm NaCl, and 5 mm MgCl2 if indicated. Determination of the Nucleotide Stoichiometry—Radioactive trapping experiments, were performed in the same buffer as non-radioactive trapping experiments, and separated by gel filtration. Fractions were analyzed by β-counting, thin layer chromatography, and for protein concentration by the BCA assay (Pierce). Over the range of protein concentrations used, the A 280 measured with the UV detector of the SMART system (Amersham Biosciences) was linearly related to the protein concentration determined by the BCA assay. In particular, wild-type NBD (250 μm) was incubated with 500 μm MgATP supplemented with 0.1 μm of either [α-32P]ATP or [γ-32P]ATP (4500 Ci/mmol, ICN) in the presence of BeFx (500 μm) for 5 min at 30 °C. The E599Q mutant (250 μm) was incubated for 5 min in the absence of Mg2+ and on ice. To follow one to two turnovers of the E599Q mutant, 250 μm NBD were incubated for prolonged time periods (0–500 min) at 30 °C with limiting concentration of MgATP (250–500 μm) and tracer amounts of radioactive ATP. To test whether ADP is incorporated in the dimer at a high ADP concentration, the E599Q mutant was incubated with 500 μm ADP and 0.1 μm [α-32P]ADP in the presence of various concentration of unlabeled ATP (0, 5, 50, 100, and 500 μm) for 10 min at 30 °C. Alternatively, incorporation of ADP under limiting concentrations of MgATP was tested by incubating the E599Q mutant (250 μm) with MgATP (250–500 μm) supplemented with tracer amounts of [α-32P]ADP at 30 °C for 400 min. The [α-32P]ADP was formed by incubation of [α-32P]ATP with hexokinase (Sigma Aldrich) according to the manufacturer's instructions, separated from hexokinase by Centricon 10 centrifugation, and full conversion to [α-32P]ADP was confirmed by TLC. Thin Layer Chromatography—The nucleotide composition of the dimer was analyzed by TLC. Immediately after gel filtration the fraction containing the dimer was incubated with 15 mm EDTA for 30 min and precipitated with trichloroacetic acid (10% w/v) for 30 min at 4 °C. After centrifugation at 20,000 × g for 5 min, the sample was neutralized with 0.5 m KHCO3, pH 8.3 and applied onto polyethylenimine cellulose plates (Merck) in 2 m formate and 250 mm LiCl. [α-32P]ATP and [γ-32P]ATP treated with hexokinase were used as references. Purification and Activity of NBD of Mdl1p—ABC transporters undergo a series of conformational changes in response to ATP binding and hydrolysis at the NBDs, and binding, transport and release of substrate at the TMDs. Various aspects of the mechanism how the motor domains of ABC transporters work can be learned from the isolated NBDs. Therefore, the NBD of the mitochondrial half-size ABC transporter Mdl1p was cloned, expressed in E. coli and purified to homogeneity (Fig. 1, left panel). About 50 mg of soluble and purified protein at 10 mg/ml was obtained from a 1-liter culture. The identity of the sample was confirmed by MALDI-TOF MS (data not shown). The activity of the isolated wild-type NBD was analyzed with respect to ATP and ADP binding using 8-azido-[α-32P]ATP photolabeling experiments. Photolabeling with 8-azido-[α-32P]ATP was specifically competed by excess of unlabeled nucleoside di/triphosphates (Fig. 2A). The IC50, roughly reflecting the K d value, was determined to be 2 μm for MgATP and 64 μm for MgADP (Fig. 2B). The absence of Mg2+ dramatically reduced the affinity of nucleotides to the NBD. The affinity for MgCTP, MgGTP, MgUTP was ∼30-fold lower compared with MgATP (data not shown). The isolated NBD was active in ATP hydrolysis (30 μm NBD, Fig. 3A) with a K m, ATP of 0.6 mm and showed no cooperativity with increasing ATP concentration (Hill coefficient of 1.0, Fig. 3A, inset). However, the ATPase activity was observed to be non-linearly dependent on protein concentration. The specific ATPase activity increased over two orders of magnitude at higher protein concentrations. At concentrations higher than 250 μm the NBD precipitated. Based on a V max of 25 ATP/NBD/min, a K 0.5 NBD of 75 μm and a Hill coefficient of 1.7 (Fig. 3B, inset) were determined, suggesting that the active form of the enzyme is a dimer. Also at low (5 μm) and high (150 μm) NBD concentrations, the K m, ATP was 0.6 mm and showed no cooperativity with increasing ATP concentration (data not shown). The NBD had a maximal ATPase activity at a pH of 8–9 (data not shown), comparable to the pH of the mitochondrial matrix.Fig. 28-Azido-[α-32P]ATP photolabeling of purified wild-type NBD of Mdl1p. A, competition of 8-azido-[α-32P]ATP (0.5 μm) photolabeling of wild-type NBD with increasing amounts of MgATP (closed circles) or MgADP (open circles). B, quantification of photolabeling efficiency by phosphorimager. The data were fitted to the Hill equation y = (axn/(bn + xn)) resulting both for ATP and ADP in a best fit (R 2 = 0.98) with a Hill coefficient (n) of 1.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Fig. 3ATPase activity of wild-type NBD of Mdl1p. A, ATPase activity as a function of the ATP concentration measured at 30 °C with 30 μm NBD. The inset shows the same data plotted according to the Hill equation (Hill coefficient = 1.0). B, ATPase activity as a function of the NBD concentration (0–150 μm) measured in the presence of 10 mm ATP. The inset shows the same data plotted according to the Hill equation (Hill coefficient = 1.7).View Large Image Figure ViewerDownload Hi-res image Download (PPT) The Active Form of NBD Is a Dimer—It is postulated that the interaction of both NBDs plays a central role in the catalytic cycle of ABC transporters (see review in Ref. 24Nikaido H. Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 9609-9610Crossref PubMed Scopus (17) Google Scholar). So far, gel filtration experiments with isolated NBDs (e.g. HisP or NBDs of P-gp) could not demonstrate ATP-dependent dimerization, independent of the presence of nucleotides or non-hydrolyzable analogues such as AMP-PNP (25Nikaido K. Liu P.Q. Ames G.F. J. Biol. Chem. 1997; 272: 27745-27752Abstract Full Text Full Text PDF PubMed Scopus (105) Google Scholar, 26Booth C.L. Pulaski L. Gottesman M.M. Pastan I. Biochemistry. 2000; 39: 5518-5526Crossref PubMed Scopus (27) Google Scholar). The wild-type NBD of Mdl1p also behaved as a monomer in gel filtration studies independent of the presence of nucleotides. To analyze intermediate states of the ATPase cycle, we tried to trap the NBD in the conformation directly after ATP hydrolysis using orthovanadate (Vi) or beryllium fluoride (BeFx). As shown on myosin x-ray structures the vanadate-inhibited complex resembles a post-hydrolysis state while the complex trapped by BeFx represents a pre-hydrolysis state (27Fisher A.J. Smith C.A. Thoden J.B. Smith R. Sutoh K. Holden H.M. Rayment I. Biochemistry. 1995; 34: 8960-8972Crossref PubMed Scopus (636) Google Scholar). After 5 min of preincubation of the NBD with BeFx in the presence of MgATP at 30 °C, a stable dimer corresponding to a molecular mass of 66 kDa was observed by gel filtration even in the absence of MgATP and BeFx in the mobile phase (Fig. 4A). The formation of the dimer was dependent on the concentration of BeFx. At a concentration of 1 mm BeFx almost 80% of the protein was detected as a dimer. The same result was observed with orthovanadate (Fig. 4B). Importantly, neither preincubation of the NBD with MgADP or MgAMP-PNP in the presence of BeFx/vanadate at 30 °C, nor incubation with MgATP and BeFx/vanadate at 4 °C induced dimerization of NBD. Omitting Mg2+ from the reaction inhibited the formation of dimers. This demonstrates that ATP hydrolysis is essential for the formation of the BeFx/vanadate-stabilized dimer. ATP-dependent Dimerization of the E599Q Mutant NBD— Exchange of the conserved glutamate downstream of the Walker B motif (E552Q/E1197Q) of mouse P-gp produced an ABC transporter with severely impaired biological activity and no substrate-stimulated ATPase activity (28Urbatsch I.L. Julien M. Carrier I. Rousseau M.E. Cayrol R. Gros P. Biochemistry. 2000; 39: 14138-14149Crossref PubMed Scopus (79) Google Scholar). Equivalent mutation of the archaeal NBDs of MJ0796 and MJ1267 (E171Q/E179Q) induced an ATP-dependent dimerization of the NBDs (29Moody J.E. Millen L. Binns D. Hunt J.F. Thomas P.J. J. Biol. Chem. 2002; 277: 21111-21114Abstract Full Text Full Text PDF PubMed Scopus (280) Google Scholar). We introduced an equivalent mutation in the conserved glutamate (E599Q) in the NBD of Mdl1p, which was expressed and purified in equal amounts as the wild-type protein (Fig. 1). The mutant NBD bound ATP to the same extent as the wild-type as analyzed by 8-azido-ATP photolabeling experiments (data not shown). As expected, no steady-state ATP hydrolysis was observed by measuring the release of Pi using the malachite green assay (data not shown). To assay the ability of the mutant to form nucleotide-dependent dimers, we incubated the mutant NBD with MgATP on ice for 5 min. Gel filtration experiments showed that MgATP induced dimerization in a concentration-dependent manner (Fig. 5). The E599Q NBD dimer was stable during gel filtration even without nucleotides in the mobile phase, indicating that the ATP-induced dimer only slowly dissociates. Importantly, ADP did not induce dimerization, and the addition of ADP to the monomers impaired ATP-dependent dimer formation. Addition of Mg2+ (0–5 mm) did not influence the dimerization. Furthermore, MgAMP-PNP or MgATPγS could bind to the wild-type and mutant NBD of Mdl1p to the same extent as MgATP, but did neither induce dimerization of the wild-type NBD nor of the E599Q mutant (data not shown). Based on these results we can conclude that presence of ATP is absolutely required for the formation of both wild-type and mutant dimer. Nucleotide Composition of the Dimer—To analyze the ATP/ADP composition of the nucleotide-induced dimer, wild-type NBD was incubated with tracer amount of either [α-32P]ATP or [γ-32P]ATP under BeFx-trapping conditions and applied to the gel filtration column. Radioactive nucleotides exclusively coeluted with the dimer, whereas no radioactive nucleotides were observed in fractions corresponding to the monomer (Fig. 6A). The amount of [α-32P]ATP bound per wild-type dimer was determined and a stoichiometry of two nucleotides per dimer was obtained (Fig. 6A, open circles). Trapping the NBD by BeFx in the presence of [γ-32P]ATP resulted in a dimer with almost no radioactive nucleotides incorporated (Fig. 6A, filled circles, Table I). These results demonstrate that two ATP molecules are already hydrolyzed in the BcFx-trapped wild-type dimer and the [γ-32P]i is released from the complex. This was confirmed by TLC analysis (Fig. 6A). Under the same conditions, incorporation of [α-32P]ADP into the dimer was not observed, even in the presence of different concentrations of ATP (data not shown).Table INucleotide composition of the dimer (average of 5 experiments)Nucleotide/dimerWild-type 30°C, BeFx-trappedE599Q 4°C, no Mg2+, ATPE599Q 30°C, Mg2+, limiting ATPmol/mol[γ-32P]ATP0.01 ± 0.012.50 ± 0.181.16 ± 0.06[α-32P]ATP1.76 ± 0.062.27 ± 0.152.18 ± 0.14[α-32P]ADP0.02 ± 0.02 Open table in a new tab In experiments with the E599Q mutant, performed at 4 °C and in the absence of Mg2+, the nucleotides exclusively coeluted with the dimer as well. In this experiment, the amount of both [α-32P]ATP and [γ-32P]ATP bound per dimer was equal and a stoichiometry of two ATP molecules per dimer was obtained (Fig. 6B, Table I). TLC clearly demonstrated that ATP and not ADP and Pi were incorporated into the dimer. As mentioned above, incubation with non-hydrolyzable ATP analogues did not trigger dimer formation. Our data from both the BeFx-trapped wild-type NBD and the E599Q mutant suggest that during the ATP hydrolysis cycle at least two different intermediate states can be isolated. Importantly, both states contain two nucleotides. In one state the nucleotides are bound as ATP as shown for the E599Q mutant. In the other state both nucleotides have been hydrolyzed to ADP and Pi as demonstrated for the trapped wild-type NBD. This suggests that ATP binding on both NBDs induces formation of the dimer and that after hydrolysis of both ATP molecules to ADP, the dimeric complex dissociates and ADP is released. Hydrolysis Cycle of the E599Q Mutant—An important remaining question is how the ATPs are hydrolyzed during one cycle. Using radiolabeled ATP, it was determined that the E599Q mutant, a