(Received for publication, February 3, 1997, and in revised form, May 18, 1997)
From the Biozentrum, Universität Basel,
Klingelbergstrasse 70, CH-4056 Basel, Switzerland and
¶ Biochemisches Institut, Universität Zürich,
Winterthurerstrasse 190, CH-8057 Zürich, Switzerland
Mitochondrial hsp70 (mhsp70) is a key component in the import and folding of mitochondrial proteins. In both processes, mhsp70 cooperates with the mitochondrial nucleotide exchange factor mGrpE (also termed Mge1p). In this work we have characterized the self-association of purified mhsp70, the interaction of mhsp70 with isolated mGrpE and protein substrate, and the effect of nucleotides on these interactions. mhsp70 can form oligomers that are dissociated by ATP or by a nonhydrolyzable ATP analog. A substrate peptide binds to mhsp70 in the absence of added nucleotides and is released by ATP but not by ADP. Binding of the peptide causes nucleotide-independent dissociation of the mhsp70 oligomers and enhances the mhsp70 ATPase. Purified mGrpE forms a homodimer. In the absence of added nucleotides, one mGrpE dimer binds to one molecule of mhsp70, forming a stable 122 kDa hetero-oligomer. This complex is weakened by ADP and completely dissociated by ATP.
Members of the 70-kDa heat shock protein (hsp70) family play an essential role in protein folding, transport of proteins into different cellular compartments, and regulation of the heat shock response (1, 2). The members of this protein family are highly conserved from bacteria to man (3, 4). In eukaryotes, at least one hsp70 is uniquely located in the mitochondrial matrix. In the yeast Saccharomyces cerevisiae the product of the SSC1 gene, mitochondrial hsp70 (mhsp70),1 is essential for growth and plays a major role in the import and folding of mitochondrial proteins (5-9). Recently it was shown that mhsp70 forms a transient complex with its membrane anchor Tim44 and mitochondrial GrpE (mGrpE; Refs. 10-12). As the translocating polypeptide chain emerges in the matrix, mhsp70 binds to it in an ATP-dependent manner, and ATP hydrolysis by mhsp70 then drives the inward movement of the polypeptide chain (13, 14).
Although the different functions of mhsp70 have been extensively studied, very little is known about the properties of the purified protein, its interaction with nucleotides and protein substrates, and the effect of co-chaperones on these interactions. Information on these points will be essential for understanding how mhsp70 mediates the translocation of polypeptides across the mitochondrial inner membrane as well as the folding of some of these polypeptides in the mitochondrial matrix. To address these questions, we purified mhsp70 from yeast and studied its oligomeric state; we also explored the effect of nucleotides and a peptide substrate on mhsp70 oligomerization. Finally, we determined the N terminus of mature mGrpE, expressed the mature protein in Escherichia coli, and assayed its interaction with mhsp70.
Two hundred mg of purified yeast mitochondria containing hexahistidine-tagged mhsp70 were solubilized in 40 ml of Buffer A (50 mM K-HEPES, pH 7.4, 0.5% nonionic detergent octyl-polyoxyethylene, 200 mM NaCl, 20 mM imidazole, and 10 units/ml apyrase). After a 20-min incubation at 4 °C, the solution was cleared by centrifugation for 20 min at 20,000 × g, and the supernatant was incubated with 1.25 ml of Ni-NTA-agarose for 90 min. The beads were washed three times with 100 ml of apyrase-free Buffer A. mGrpE was eluted by incubation for 20 min in 10 ml of Buffer A that lacked apyrase and NaCl but contained instead 10 mM MgCl2 and 2 mM ATP. For further purification of mGrpE, the eluate was applied to a Mono Q column (Pharmacia Corp.) equilibrated with 50 mM K-HEPES, pH 7.4, 0.5% octyl-polyoxyethylene and eluted from the column with a linear salt gradient (0-500 mM NaCl). mGrpE eluted at about 200 mM NaCl. The peak fractions were pooled, concentrated in Centricon tubes, washed three times with 0.1% acetic acid, and applied to a reverse phase C4 HPLC column (Aquapore Butyl, 2.1-30 mm) equilibrated with 0.1% trifluoroacetic acid. mGrpE was eluted by a linear gradient of 0.1% trifluoroacetic acid to 60% acetonitrile in 30 min. The protein that eluted as a single peak was collected and subjected to N-terminal sequencing as described (15).
Construction of Hexahistidine-tagged mGrpETwo mGrpE
constructs carrying a C-terminal hexahistidine extension were made. One
was 24mGrpE, in which only part of the mitochondrial targeting
sequence (residues 1-23) was replaced by Met-Ala (16). The other was
45mGrpE, in which the entire mitochondrial targeting sequence
(residues 1-44) was replaced by Met-Gly. Both constructs carried the
C-terminal extension -Gly-Ser-Arg-Ser-His6. The
purification of both proteins was carried out as described (16).
Cross-linking was carried out with 10 mM glutaraldehyde at 30 °C for 30 min in 50 mM K-HEPES, pH 7.5, 100 mM KCl, 5 mM MgCl2. The cross-linking reaction was stopped by the addition of SDS-containing sample buffer containing 200 mM glycine. The cross-linked products were analyzed by SDS-PAGE using either linear 4-18% slab gels (17) or 4% tube gels (18).
Stopped-flow MeasurementsStopped-flow measurements were done at 25 °C with an SX-17MV spectrofluorimeter (Applied Photophysics, Leatherhead, UK). Excitation was at 370 nm, and emission light was passed through a GG455 cutoff filter (Jena, Germany).
The dissociation constant of the p5·mhsp70 complex was determined as described (30). A solution of 50 nM p5 peptide was placed in a cuvette and the base line was recorded. Increasing amounts of mhsp70 were then added. After each addition, the fluorescence was recorded until equilibrium was reached. The dissociation constant was determined assuming one binding site and a right-angled hyperbola as the saturation curve.
ATPase MeasurementsThe ATPase activity of mhsp70 was
measured as described by Viitanen et al. (19) at room
temperature in 50 mM K-HEPES, pH 7.4, 5 mM
MgCl2, 100 mM KCl, 0.1 mM
[-32P]ATP (0.04 Ci/mmol). Aliquots of 25 µl were
withdrawn at different intervals and mixed with 7 volumes of 1 M perchloric acid, 1 mM sodium phosphate. After
adding 16 volumes of 20 mM ammonium molybdate and 16 volumes of isopropyl acetate, the samples were mixed vigorously for
15 s. The inorganic and organic phases were separated by
centrifugation for 15 s in a microcentrifuge, and an aliquot of
the organic phase containing the radioactive phospho-molybdate complex
was counted in a liquid scintillation counter. All ATPase values were
calculated from initial rates during which the amount of ATP hydrolyzed
was a linear function of time. In all experiments, the background resulting from contamination of ATP by inorganic phosphate or from
spontaneous hydrolysis of ATP was <1.5% that of the measured values.
Sedimentation velocity and sedimentation equilibrium measurements were performed at 20 °C in an analytical ultracentrifuge (Beckman XLA) equipped with an optical absorption system. Sedimentation velocity was carried out at 56,000 rpm in a 12-mm double-sectored Epon cell (500 µl). Sedimentation velocity was measured at 20 °C with 100-µl samples at either 10,000 rpm (mhsp70 or mGrpE·mhsp70 complexes) or at 14,000 rpm (mGrpE alone). The molecular masses were determined using a linear regression computer program that adjusts the base-line absorption to obtain the best linear fit of ln A versus r2 (A = absorption, r2 = radial distance from the rotor center). Measurements of sedimentation equilibrium were carried out in the presence of 0.1 mM adenine nucleotide.
MiscellaneousMhsp70 and nucleotides were purified as described (20). The concentration of proteins was determined using the bicinchoninic acid procedure (Pierce) with bovine serum albumin as the standard. DnaK and bacterial GrpE were gifts from Dr. H.-J. Schönfeld (Roche Co, Basel, Switzerland). Hexahistidine-tagged Mdj1 was purified on Ni-NTA beads as described (16).
mGrpE is made as a precursor that is cleaved upon import into the mitochondrial matrix (21-23). Our previous attempt to identify the cleavage site had been unsuccessful, perhaps because the N terminus of the purified protein was blocked (21). Sequence alignment of the mGrpE precursor with bacterial GrpE suggested to us that cleavage of the presequence occurs after amino acid 23 of the yeast precursor (21). In contrast, Laloraya et al. (22) concluded that a cleavage site occurs after amino acid 42. As it seemed possible that blockage of the mGrpE N terminus was an artifact caused by purification of the protein on SDS-PAGE, we used a different purification strategy. First, we isolated mitochondria from a yeast strain that overexpressed mhsp70 carrying a C-terminal hexahistidine tag. Second, we solubilized the mitochondria with nonionic detergent in the absence of nucleotides, bound the mhsp70·mGrpE complex to Ni-NTA agarose beads, and removed unspecifically bound proteins by washing. Third, we specifically released mGrpE from the mGrpE·mhsp70·Ni-NTA-agarose complex by MgATP. Fourth, we purified the eluted mGrpE to homogeneity by anion exchange chromatography followed by reverse phase HPLC. The N-terminal sequence of the pure mGrpE was Ser-Asp-Glu-Ala-Lys. Comparison of this sequence with that of the predicted mGrpE presequence indicated that the cleavage occurs after amino acid 44.
In Vitro Function of mGrpE and Mdj1 with a C-terminal Hexahistidine TagTo obtain large amounts of mGrpE and Mdj1, both proteins were tagged with a C-terminal hexahistidine extension, expressed in E. coli, bound to Ni-NTA-agarose beads, and eluted with imidazole (16). To check whether the C-terminal tag altered the structure of mGrpE, we analyzed the tagged protein by analytical equilibrium sedimentation. We determined a native molecular mass of 44 kDa (Table I). As the monomer has a mass of about 22 kDa, the tagged mGrpE seems to exist as a dimer, as was also shown for bacterial GrpE (24, 25). Sedimentation velocity centrifugation yielded a sedimentation coefficient of 2.3 S (Table I). This is a rather low value for a 44-kDa globular protein, suggesting that the mGrpE dimer has an elongated shape.
|
In the corresponding bacterial system, DnaJ alone (which accelerates the ATP hydrolysis of DnaK) or GrpE alone (which acts as a nucleotide exchange factor) stimulates the ATPase of DnaK only slightly, whereas the combination of both co-chaperones stimulates strongly (26-29). Stimulation of the ATPase activity of mhsp70 by mGrpE and Mdj1 can thus be used as an indicator for specific interaction between these proteins. Table II presents an analysis of the mhsp70 ATPase in the presence of various cofactors. Mhsp70 alone only hydrolyzed 0.11 mol of ATP/mol of mhsp70/min (Table II). This value corresponds to one turnover of the ATPase in 9 min. mGrpE alone or Mdj1 alone each caused slight, but significant stimulation of the mhsp70 ATPase (to 0.21 mol and 0.8 mol of ATP hydrolyzed/mol of mhsp70/min, respectively). The combination of mGrpE and Mdj1 strongly stimulated the mhsp70 ATPase to 2.58 mol of ATP hydrolyzed/mol of mhsp70/min.
|
The oligomeric state of mhsp70
was studied by cross-linking with glutaraldehyde and analytical
ultracentrifugation. In our standard cross-linking protocol, we
incubated the protein samples with 10 mM glutaraldehyde at
30 °C for 30 min, as preliminary experiments had shown these
conditions to give maximal cross-linking (data not shown). In the
absence of added nucleotides, mhsp70 formed oligomers (Fig.
1, lane 1; Table I). ADP did
not affect the type of oligomers detected by cross-linking (Fig. 1,
lane 2) but increased the percentage of monomer 2-fold. In
the absence of nucleotides, some of the cross-linked oligomers were
apparently too large to enter the gel; this effect probably explains
the lower combined intensity of the cross-linked bands seen in
lane 1.
ATP or the nonhydrolyzable ATP analog AMP-PNP almost completely prevented oligomer formation; only a minor fraction of mhsp70 existed as dimers and trimers (Fig. 1, lanes 3 and 4). These results were also confirmed by analytical ultracentrifugation (Table I). Disoligomerization of mhsp70 in vitro thus appears to require only binding but not hydrolysis of ATP.
The Interaction of a Peptide Substrate with mhsp70To test
the interaction of mhsp70 with a peptide substrate, we used the
decapeptide CALLLSAPRR (termed p5) that represents part of the
mitochondrial targeting sequence of aspartate aminotransferase from
chicken. The peptide was labeled at its single cysteine residue with
the fluorescent dye acrylodan (30). Binding of fluorescent p5 to mhsp70
slightly shifted the emission maximum from 524 to 518 nm and increased
the emission intensity at 500 nm by 50% (Fig. 2). The dissociation constant for the
interaction of p5 (50 nM) with mhsp70 in the absence of
added nucleotides was 43 nM (Fig. 2, inset).
This high affinity suggests a specific interaction. Addition of ATP
caused the peptide to be released from mhsp70 with an apparent
half-time of 40 ms. ADP did not cause any release (Fig.
3).
The ATPase activity of mhsp70 alone or in the presence of mGrpE was enhanced 3-fold in the presence of p5 peptide. In the presence of either Mdj1 alone or of both mGrpE and Mdj1, the p5 peptide did not provide any additional stimulation (Table II).
The p5 peptide caused complete monomerization of mhsp70 (Fig. 1, lane 5; Table I), perhaps because it competes for hydrophobic surfaces on mhsp70 with other mhsp70 molecules or because it alters the structure of mhsp70 in a way that decreases the exposure of hydrophobic surfaces.
Interaction of mGrpE with mhsp70To study the interaction of
mGrpE with mhsp70, we stabilized the hetero-oligomer formed between
these two proteins by cross-linking with glutaraldehyde (24). Fig.
4 shows the cross-linking pattern observed with different mhsp70/mGrpE mixtures. As already shown in Fig.
1, mhsp70 alone yielded only three cross-linked mhsp70 species, the
monomer being dominant. As the mhsp70 concentration in this experiment
was half that used in the experiment shown in Fig. 1, oligomer
formation was less pronounced. Addition of mGrpE yielded a new species
whose migration on SDS-PAGE suggested that it is composed of two
molecules of mGrpE and one molecule of mhsp70 (Fig. 4). At mGrpE:mhsp70
ratios in excess of 1.5, this species was the predominant one. The
existence of such an mhsp70·(mGrpE)2 complex was
confirmed by analytical ultracentrifugation. In a solution containing
mGrpE and mhsp70 at a molecular ratio of 2:1, a homogeneous species
with an apparent mass of about 122 kDa was detected (Table I). In
addition, cross-linking revealed a minor species with a mobility less
than that of the mhsp70(mGrpE)2 hetero-oligomer (Fig.
4A, lanes 5-7, asterisk). This form
has an apparent mass of ~240 kDa as determined by SDS-PAGE,
consistent with a hetero-oligomer composed of two molecules of mhsp70
and four molecules of mGrpE.
Recently it was suggested that at bacterial GrpE concentrations of
above 0.32 mg/ml, six molecules of bacterial GrpE form a
physiologically relevant complex with one molecule of DnaK (31). To
study the formation of mGrpE·mhsp70 complexes at high protein concentrations, we repeated the experiment shown in Fig. 4 with higher
concentrations of mGrpE and mhsp70. We then separated the cross-linked
products on SDS-PAGE and characterized them by immunoblotting with
antibodies against mhsp70 or mGrpE (Fig.
5). Cross-linking of mGrpE alone showed
that at low concentration (1.0 µM) the protein exists
mainly as a dimer (Fig. 5B, lane 8), but that at
13 µM it forms larger oligomers (Fig.
6B, lane 9).
Surprisingly, when mGrpE was tested by analytical ultracentrifugation
at concentrations of up to 20 µM, only mGrpE dimers were
detected (Table I), suggesting that the larger oligomers are rather
labile.
We then tested the effect of three different mhsp70 concentrations on the oligomeric state of mGrpE·mhsp70. At 0.5 µM mhsp70 and 1 µM mGrpE, only the usual mhsp70·(mGrpE)2 complex was detected (Fig. 5B, lane 3). When the mGrpE concentration was raised to 3 µM, all of the mhsp70 was immunodetected as the mhsp70·(mGrpE)2 complex (Fig. 5A, lane 4). Even at 13 µM mGrpE, most of the mhsp70 was observed as mhsp70·(mGrpE)2 complex, although larger mhsp70-containing complexes could be detected. Some of these might not have entered the gel, thereby lowering the intensity of the mhsp70·(mGrpE)2 band (Fig. 5A, lane 5). At the high mGrpE concentrations at which mGrpE was in large excess over mhsp70, the excess mGrpE formed higher oligomers that lacked mhsp70 (Fig. 5B, lanes 4 and 5).
When the concentration of mhsp70 was raised so as to saturate all of the mGrpE (mhsp70/mGrpE = 0.5), most of the large mGrpE oligomers disappeared, and the mhsp70·(mGrpE)2 complex again became the predominant species (Fig. 5A and B, lanes 6 and 7). Similarly, when solutions containing up to 20 µM mGrpE were examined by analytical ultracentrifugation at an mhsp70/mGrpE ratio of 0.5, only the 122-kDa hetero-oligomer was detected.
In the absence of nucleotides, mGrpE seemed to bind to mhsp70 with high affinity, as free mGrpE was detected only when no free mhsp70 molecules were available (Fig. 4, lane 6; see also Fig. 6). To study the effect of nucleotides on the mGrpE-mhsp70 interaction, we incubated 0.5 µM mhsp70 and 1.5 µM mGrpE with 1 mM of either ADP, ATP, or AMP-PNP before cross-linking (Fig. 6). The cross-linking products were analyzed on SDS-PAGE and characterized by immunoblotting with antibodies against mhsp70 (Fig. 6A) and mGrpE (Fig. 6B). In the absence of added nucleotides, we mainly detected the mhsp70·(mGrpE)2 hetero-oligomer as well as the (mGrpE)2 homodimer (Fig. 5, lane 1). A similar result was obtained in the presence of the nonhydrolyzable ATP analog AMP-PNP (Fig. 6, lane 4). Very little mhsp70·(mGrpE)2 complex was detected in the presence of ADP (Fig. 6, lane 2), whereas upon the addition of ATP the mhsp70·(mGrpE)2 complex dissociated completely (Fig. 6, lane 3). Complete dissociation of the hetero-oligomeric complex thus seems to occur in the presence of ATP.
The hsp70 chaperone system of E. coli consists of DnaK, DnaJ, and GrpE (28, 29, 32). Whereas homologs of DnaK and DnaJ have been found in the eukaryotic cytosol and in chloroplasts (32, 33), eukaryotic homologs of all three proteins have only been detected in mitochondria. The mitochondrial hsp70 system thus appears to be the closest eukaryotic counterpart of the bacterial chaperone triad (16, 21, 22, 34-36). Although the function of the mitochondrial system has been studied in intact cells, isolated mitochondria, and mitochondrial extracts, little is known on the properties of the purified proteins and their interaction with each other. In the work described here, we have studied the physical and functional interaction of purified mhsp70 with purified mGrpE in the presence and absence of cofactors using enzyme assays, analytical ultracentrifugation, and chemical cross-linking.
For these studies, we first had to determine the N terminus of mature mGrpE. It is well established that mGrpE is made with a cleaved N-terminal matrix targeting signal, but the length of this signal has been controversial (21, 37). We now show that mature protein starts at amino acid 45 and that neither of the earlier predictions had been correct. Still, the recombinant mature mGrpE we have studied here behaved quite similarly to an earlier construct that lacked only the first 24 residues of the cleaved targeting signal (16). Independently, Deloche and Georgopoulos (37) showed that two mGrpE constructs similar to those used by us in this study could complement the growth defect of a bacterial temperature-sensitive GrpE mutant.
Our study used hexahistidine-tagged mGrpE and Mdj1, and it might be argued that the properties of the tagged proteins are different from those of the authentic proteins. It is difficult to exclude such an artifact completely, but analytical ultracentrifugation and chemical cross-linking showed that the quaternary structure of the tagged mGrpE closely resembles that of untagged bacterial GrpE. Also, the physical and functional interactions of the tagged mGrpE with mhsp70 or DnaK were essentially indistinguishable from those reported for bacterial GrpE or untagged mGrpE (16, 25, 27, 38; compare also Fig. 4B of this study with Fig. 3 of Ref. 37). Thus, the tag does not seem to profoundly alter the structure or function of mGrpE.
Our results show that the mhsp70 chaperone system shares many structural features with its counterparts in E. coli and the eukaryotic cytosol. First, mhsp70 self-associates in the absence of added nucleotides or in the presence of ADP, and the oligomers are dissociated by ATP or a nonhydrolyzable ATP analog (24, 25, 39-41). This effect has been proposed to be mechanistically important for the activity of hsp70 s. The ATP-induced conformation of hsp70 may have fewer exposed hydrophobic surfaces that lead to self-association. Second, mhsp70 interacts with both Mdj1 and mGrpE (16), and these interactions stimulate the ATPase activity of mhsp70 (26, 27). Third, mhsp70 binds a peptide substrate in the absence of nucleotides or in the presence of ADP but not in the presence of ATP, and peptide binding dissociates the mhsp70 oligomers and stimulates the mhsp70 ATPase (27, 30, 39). Fourth, mGrpE and Mdj1 can bind to DnaK and replace their bacterial homologs in stimulating the DnaK ATPase (Fig. 4B of this paper and Refs. 16 and 38).
Recently it was suggested that the complex formed at high concentrations of bacterial GrpE and DnaK is composed of six molecules of GrpE and of one molecule of DnaK (31). This conclusion was reached on the basis of experiments employing cross-linking with glutaraldehyde and gel filtration. We investigated the oligomeric state and the interaction of mGrpE with mhsp70 at high protein concentration by cross-linking with glutaraldehyde and analytical ultracentrifugation. Whereas cross-linking revealed that mGrpE alone at concentrations >3 µM oligomerizes to tetramers and hexamers (Fig. 5, lane 9), analytical ultracentrifugation only detected mGrpE dimers, even at mGrpE concentrations of up to 20 µM (Table I). This discrepancy might either mean that glutaraldehyde induces oligomerization of mGrpE or that the higher oligomers are too labile to be detected without covalent stabilization.
In contrast, analytical ultracentrifugation and cross-linking with glutaraldehyde yielded identical results for the structure of the mhsp70·mGrpE complex. At high concentrations of mGrpE and mhsp70, cross-linking detected mostly the mhsp70·(mGrpE)2 complex, and analytical ultracentrifugation detected the corresponding 122-kDa species. We conclude that a complex of two molecules of mGrpE and one molecule of mhsp70 is the major hetero-oligomer both at low and at high protein concentrations.
The mhsp70·(mGrpE)2 oligomer is stable in the absence of
externally added nucleotides or in the presence of the nonhydrolyzable ATP analog AMP-PNP but dissociates in the presence of ATP. This result
appears to conflict with the recent report that ATPS, another
nonhydrolyzable ATP analog, dissociates the mGrpE·DnaK complex (37).
This discrepancy could be explained by the different methods used. The
dissociating effect of ATP
S was detected by an enzyme-linked
immunosorbent assay, which measures the steady state interaction
between DnaK and mGrpE. In contrast, cross-linking by glutaraldehyde is
much faster and irreversible, thereby allowing detection of weak and
transient interactions (31).
We thank M. Jaeggi and L. Müller for the artwork, H.-J. Schönfeld for samples of DnaK and bacterial GrpE, and N. G. Kronidou, J. Rötter, and members of our laboratory for helpful discussions.