Intramitochondrial Folding and Assembly of Medium-chain Acyl-CoA Dehydrogenase (MCAD) DEMONSTRATION OF IMPAIRED TRANSFER OF K304E-VARIANT MCAD FROM ITS COMPLEX WITH hsp6O TO THE NATIVE TETRAMER*

We incubated in vitro translated precursor of me- dium-chain acyl-CoA dehydrogenase (MCAD) with isolated rat liver mitochondria and fractionated the solu- bilized mitochondria on gel filtration. After a 6-min import into mitochondria, MCAD was recovered exclusively as a high molecular weight (hM,) complex (700,000), while after a 10-min import, it was recovered mainly in the hM, complex and mature tetramer, with a small amount in monomer. Either a further 15-min chase or exposure to ATP caused a marked decrease of MCAD in the hM, complex and an increase in the mature tetramer in comparable amounts, suggesting that the hM, complex was the precursor of tetramer. No monomer was detected in either case. Using specific antibodies, we have shown that the hM, complex represented a complex of MCAD and heat-shock protein 60 (hspBO), and, that upon import into mitochondria, un- folded MCAD first formed a transient complex with mitochondrial heat-shock protein 70 (hsp70mit) and then transferred to hsp6O to complete its folding into an as-sembly-competent conformation. We also examined the assembly of K304E MCAD, which is a prevalent variant enzyme among patients with labeled MCAD precipitates


203-785-3363.
peptide is proteolytically cleaved, producing the 396-amino acid (43.6 kDa) mature protein. The monomeric enzyme is then assembled into the native and biologically active homotetrameric form (2). Recent x-ray crystallographic studies indicate that MCAD-tetramer is actually a dimer of two dimers. Domains involved in the monomer-monomer interaction are different from those involved in the dimer-dimer interaction (5).
Hereditary MCAD deficiency is mainly detected among Caucasian children of northwestern European origin (6). It causes episodic vomiting and hypoketotic hypoglycemia, and, if not treated, the patients may die. The incidence of MCAD deficiency is relatively high for a genetic metabolic disorder. Among MCAD-deficient patients, an A to G transition a t position 985 in the coding region of the gene is a highly prevalent mutation, which is found in 89% of all variant MCAD alleles. This mutation results in a glutamate being substituted for the normal lysine at position 304 (K304E) of the mature MCAD. This K304E-variant of MCAD is unstable and it is undetectable in patients' tissues (7). Since lysine 304 is located in the critical domain involved in the normal dimer-dimer interaction in the tetramer assembly (5,8), it has been suggested that this change of the charge at the interface of subunits causes an impairment of the normal assembly of variant subunits into a tetramer, thereby causing instability of the protein (9)(10)(11).
We have recently studied the tetrameric assembly process of normal human MCAD and three of its variants, each of the latter containing a n acidic or basic substitution of lysine 304 including K304E (12). Each of the variants was produced via site-directed mutagenesis and then examined by in vitro expression in the presence of isolated rat liver mitochondria. For each protein, we analyzed the molecular forms of MCAD that were imported into the isolated mitochondria using gel filtration at different times following the import reaction. Three forms of MCAD-related proteins of different size were detected. These included the monomer (44 kDa), the mature tetramer (176 kDa), and, finally, a high molecular weight complex ( h M r complex; -700,000). No dimer was detectable. The precise nature of this h M r complex was unknown. Time course studies demonstrated that tetrameric assembly of acidic variants was impaired. Furthermore, our kinetic analysis indicated that the hM, complex, and not the free monomer, represented the immediate precursor of the mature tetramer, both in the wild-type and variant forms of MCAD.
Recent studies have demonstrated a class of proteins referred to as molecular chaperones, which facilitate various steps of protein synthesis (13,14). Relevant examples include members of the hsp7O and hsp60 families of stress proteins which may work in tandem to achieve protein folding andor assembly (15)(16)(17). In the case of mitochondria, proteins entering into the organelle appear to first bind to mitochondrial form 4402 Folding and Assembly of Wild-type and K304E MCAD of hsp7O (hsp70mit). Such an interaction is believed to prevent the premature folding of the incoming polypeptide. Once entirely translocated into the organelle, the newly imported protein is transferred to the large oligomeric hsp6O complex which presumably functions to facilitate the folding of the protein and perhaps orchestrate oligomeric assembly.
Our observation that a portion of newly imported MCAD was found within a hMr complex, along with the known characteristics of molecular chaperones like h~p 7 0~~ and hsp60, prompted our interest in the nature and role of the hMr complex in the assembly process of MCAD. In the present study, the identity of the hMr complex and its potential role in the assembly of tetrameric MCAD, be it the wild-type or K304E-variant, were studied.
EXPERIMENTAL PROCEDURES Materials-The entire coding region of normal human pMCAD cDNA was amplified using the polymerase chain reaction (18) and cloned into pBluescript vector (Stratagene). Variant pMCAD cDNA with Ggm, encoding Kp329E2 (K304E), was created by site-directed mutagenesis as previously described (12). Antibodies against the mitochondrial hsp7O and hsp60 proteins were prepared and characterized as described previously (15,19).
In Vitro lFawcriptionllFanslation of pMCAD cDNA-In vitro transcription of the wild-type and variant pMCAD cDNAs in pBluescript was carried out using T7 RNA polymerase (Pharmacia LKB Biotechnology Inc.) according to the protocol of the manufacturer. In vitro translation was performed using a rabbit reticulocyte lysate translation system (Promega) and [36Slmethionine (Amersham) according to the manufacturer's instructions.
Gel Filtration Analysis of Newly Imported Wild-type and K304Evariant MCADs-Rat liver mitochondria were prepared from male Wister rats as previously described (20) and suspended in HMS buffer (2 m~ Hepes, pH 7.4,220 m~ D-mannitol, and 70 m~ sucrose) at 20 mg of proteidml. Import mixture containing 200 pl of in vitro translation product and 100 p1 of rat liver mitochondria was incubated at 30 "C for a period of 5 or 10 min. The mixture was then treated with 2.5 pl of 4 m g / d trypsin for 10 min at 4 "C, followed by the incubation with 6 pl of 10 mdml trypsin inhibitor for 10 min. The mitochondrial pellet was washed twice with HMS buffer and dissolved in 100 pl of solubilization buffer (10 m~ potassium phosphate buffer, pH 8.0, 0.5 m~ EDTA, 1% Triton X-100, and 0.4 mg/ml trypsin inhibitor). Solubilized mitochondria were centrifuged at 15,000 x g for 20 min, and the supernatant was recovered.
For a pulse-chase study of the newly imported MCAD, the import mixture was first incubated at 30 "C for 10 min (pulsed). Trypsintreated mitochondria were washed twice and suspended in 100 pl of HMS buffer. The mitochondrial suspension was further incubated at 30 "C for 15 min (chased). The mitochondrial pellet was solubilized as described and clarified by centrifugation.
For gel filtration, a Sephacryl S-300 HR (Pharmacia) column (0.7 x 16 cm) was equilibrated with running buffer (10 m~ potassium phosphate buffer, pH 8.0, and 0.5 m~ EDTA) and was calibrated with protein standards thyroglobulin (670 m a ) , y-globulin (158 m a ) , and ovalbumin (44 m a ) . The void volume of the column was determined via the elution of blue dextran. The supernatant of the solubilized mitochondria was applied to the column after centrifugation, and eluate was collected in 300-111 fractions. Each fraction was precipitated with the addition of an equal volume of acetone, and the precipitates were redissolved in 40 pl of sample buffer (0.6 M Tris, pH 8.85, 2% sodium dodecyl sulfate (SDS), 5% P-mercaptoethanol, 10% glycerol, and bromphenol blue). A 20-pl aliquot was analyzed using SDS-polyacrylamide gel electrophoresis (SDS-PAGE) according to the method of Laemmli (21). Labeled MCAD bands were visualized by autofluorography, and the intensity of the MCAD bands on the x-ray film was determined densitometrically using a Bio-Image system (Millipore).
Effects of Apyrase and ATP on Tetramer Assembly-200 pl of in vitro translation product were incubated with 100 pl of rat liver mitochondria at 30 "C for 10 min. Trypsin-treated mitochondria were solubilized in 100 pl of the solubilization buffer and centrifuged at 15,000 x g for 20 min. The supernatant was diluted with 100 pl of 50 m~ potassium phosphate buffer, pH 8.0,5 m~ MgC12, and 2 m~ NADH and incubated The number with the prefix p indicates the position in the precursor sequence.
at 30 "C for 15 min with either 20 unitdml apyrase (Sigma, grade VIII) or 2.5 m~ Am. The sample was applied to the gel filtration column, and the eluate in each fraction was analyzed using SDS-PAGE and fluorography.
Immunochemical Analysis of Gel Filtration Eluate for MCAD.hsp60 Complex Using Anti-hsp6O Antibody--ARer a 10-min import, solubilized mitochondria were fractionated by gel filtration. Fractions, each corresponding to the peak of hM, complex or tetramer (fraction 3 or 6, respectively), were adjusted to a final volume of 1.5 ml with TETN buffer (25 m~ Tris-HC1, pH 7.5, 5 m~ EDTA, 250 m~ NaCl, and 0.1% Triton X-loo), and each fraction was divided into three aliquots. To each aliquot 0, 5, or 10 pl of anti-hsp60 antiserum were added, and the difference in the amount of added antiserum in the first two tubes was made up to the total of 10 pl with normal rabbit serum. The samples were incubated at 4 "C for 30 min by rotating end-over-end. 10 pl of 10% Staph A (Life Technologies Inc.) were then added. After a 15 min incubation at 4 "C, Staph A cells were sedimented by centrifugation and washed twice with TETN buffer, followed by two washes with 10 m~ Tris-HC1, pH 7.5, and 5 m~ EDTA. Staph A-immunoglobulin-antigen complexes were dissociated by boiling for 3 min in 30 pl of the sample buffer. Proteins in the supernatant after the first centrifugation were acetone-precipitated and redissolved in 30 pl of the sample buffer. For radiolabeled MCAD, 25 pl of the immunoprecipitated proteins or 12.5 pl of the acetone-precipitated proteins from the supernatant were analyzed by SDS-PAGE and fluorography. To quantlfy hsp60, l pl of each sample was applied to SDS-PAGE, and proteins were electroblotted to an Immobilon-P membrane (Millipore) according to the method of Towbin et al. (22) and probed with anti-hsp6O antiserum using a Protoblot Western Blot AP system (Promega). Intensity of MCAD band on the autofluorogram and hsp6O band on the immunoblot was measured using a densitometer, and was expressed in densitometric units (unit, and unitI, respectively). The quantity of MCAD bound to hsp6O was calculated from the amount of MCAD (unit,) bound to per unit, of hsp6O in the immunoprecipitated fraction.
Immunochemical Analysis of MCAD Imported into Mitochondria and Complexed with hsp60 or hsp'lO,;, Using Specific Antibodies-For the analysis of mitochondrial MCAD complexed with hsp60, 100 pl of in vitro translation product were incubated with 50 pl of rat liver mitochondria at 30 "C for 10 min. The mixture was treated with trypsin followed by the addition of trypsin inhibitor as described above. Isolated mitochondria were suspended i n 200 pl of HMS buffer and were divided into two equal aliquots. Mitochondria in one aliquot were immediately pelleted and solubilized in 100 p l of the solubilization buffer, and the other aliquot was further incubated at 30 "C for 15 min before centrifugation and solubilization. Solubilized mitochondria were centrifuged at 15,000 x g for 20 min. The supernatant was divided into two 50-pl aliquots, and the volume was adjusted to 0.5 ml with TETN buffer. To each tube, either 5 pl of normal rabbit serum or 5 pl of anti-hsp60 antiserum were added. Immunoprecipitation and detection of MCAD and hsp60 were performed as detailed above.
In order to study the sequential action of hsp70mit and hsp60 in the process of intramitochondrial folding and assembly of MCAD, 50 pl of in vitro translation product were mixed with 25 pl of rat liver mitochondria in each of 5 tubes, and they were incubated for varying periods of time up to 25 min. From 0 to 10 min, the mixture was incubated at 16 "C, and, thereafter, the temperature was shifted to 30 "C. At the end of each incubation, mitochondria were isolated after the mixture was treated with trypsin. The isolated mitochondria were solubilized in 100 pl of the solubilization buffer containing 20 unitdml apyrase. &r centrifugation, the supernatant of each sample was divided into two 50-pl aliquots, and the volume was made up to 0.5 ml with TETN buffer. To one tube 5 pl of anti-hsp70mit antiserum were added. 5 pl of anti-hsp60 antiserum were added to the other tube. Immunoprecipitation and detection of labeled MCAD in the resulting precipitates were performed as described above.

RESULTS
Time Course of Tetramer Assembly of Wild-type MCAD in Mitochondria-We first incubated in vitro translated, radiolabeled wild-type pMCAD with isolated mitochondria at 30 "C for 5 min. After removing all the pMCAD attached to the outside surface by treating with trypsin, mitochondria were solubilized by addition of Triton X-100. After centrifugation, the supernatant was analyzed by gel filtration. Radiolabeled wild-type MCAD eluted predominantly as hMr complex from the gel filtration column. The size of this h M r complex was approxi- of the MCAD present within the hM, complex was the largest (55% of the total), followed by that of tetramer (33%). The amount of monomer was small (12%). These results suggested that the hM, complex is the first product formed in the intramitochondrial processing of MCAD into its final tetrameric form. In order to more carefully characterize the transition of these three species of MCAD, in vitro translated pMCAD was incubated with isolated mitochondria at 30 "C for 10 min (pulsed), and, after two washes, mitochondria were further incubated for 15 min (chased) at the same temperature. After the 15-min chase period (Fig. 1, shown in solid squares with dotted line), the majority (78%) of the imported MCAD was present as its final and mature tetrameric form. The amount of hM, complex greatly decreased (17%), and monomer was almost undetectable after 15-min chase.
Effects of ATP and Apyrase on the hM, Complex-Previous studies have shown that the mitochondrial hsp60 proteins are intimately involved in the folding and assembly of newly imported proteins within the mitochondria (23,24). Since the subunit size of hsp6O protein is 60 kDa and hsp60 exists as single or double heptamer-toroidal rings (25), the size of the hM, complex was close to that expected of a hsp60 complex containing MCAD. Thus, we investigated whether the observed hM, complex of newly imported MCAD might represent a MCAD-hsp6O intermediate. Since the release of proteins from hsp60 is known to be caused by ATP (24) effects of either ATP depletion (via the enzyme apyrase) or ATP addition on MCAD assembly in mitochondria. After a 10-min import of MCAD followed by solubilization, mitochondria were further incubated for 15 min at 30 "C with either apyrase or ATP, and the reaction mixture was analyzed by gel filtration. In the mitochondrial lysate incubated with apyrase, 55% and 39% of MCAD eluted as hM, complex and tetramer, respectively tetramer was ATP-dependent. Note that essentially no monomer was observed after exposure of the mitochondrial lysate to ATP.
Identification of the hM, Complex as a Complex of MCAD and hsp60-In order to test whether the hM, complex represented a complex of MCAD and hsp60, the fraction containing hM, complex (tube 3: -700,000) was examined by immunoprecipitation with an antibody specific to hsp60 (Fig. 3). We also tested the tetramer fraction (tube 6: 160,000) with the same antibody as a negative control. In the hM, complex fraction, MCAD coprecipitated with hsp6O. As the amount of anti-hsp60 antibody increased, the amount of residual MCAD in the supernatant decreased, while that of MCAD in the precipitates increased. With the highest amount (10 111) of anti-hsp60 antibody used, the amount of MCAD in the supernatant decreased by 57%, while 40% was recovered as isolated precipitate. When the tetramer fraction (tube 6) was treated with the same anti-hsp6O antibody in a similar manner, essentially all of MCAD remained in the supernatant, while no MCAD precipitated at all regardless of the amount of the antibody used.
We also determined the relative amount of hsp60 in tube 3 (-700,000; containing the hM, complex) and tube 6 (160,000; containing the tetramer) using immunoblot and densitometry.
The ratio of the amount of hsp60 in tube 3 and that in tube 6 was 1.0 to 0.59, indicating that a considerable portion of hsp6O in mitochondria was in forms other than single or double heptamer-toroidal rings. When the amount of hsp6O was quantified using immunoblot analysis with and without treatment with 10 pl of anti-hsp6O antibody, 97% and 92% of hsp6O protein were precipitated in tubes 3 and 6, respectively (data not shown), indicating that with the use of 10 pl of the antibody, hsp6O was almost quantitatively precipitated. Thus, these data suggested that the MCAD in the hM, complex fraction (43% of the total), that was not precipitated with anti-hsp60 antibody, represented macromolecular complex of MCAD other than that with hsp60, such as aggregates of itself or complexes with other proteins.
Interaction of K304E-variant MCAD and hsp60 in Mitochondria-In order to examine the maturation of the K304E-vana n t MCAD, a pulse-chase experiment was performed exactly as described in Fig. 1 for the wild-type MCAD. Following a 5-min import reaction, the vast majority (72%) of the K304E MCAD variant was found within the hM, complex (Fig. 4, shown with open circles and solid line), similar to that observed for the wild-type MCAD protein. Unlike the wild-type, however, a small but significant amount (11%) of monomer of the variant was detectable. After 10 min of import, the amount of the K304E-variant within the monomer fraction (28%) considerably increased, but most of the protein was still present within hM, complex (60%) (Fig. 4, shown with solid  circles and solid line). Only a small amount, if any, of tetramer appeared. After a 15-min chase, 49% of K304E still remained within the hM, complex. In addition, a broad peak encompassing tubes 5 to 10 was observed (Fig. 4, shown with pMCAD was incubated with rat liver mitochondria a t 30 "C for 5 or 10 min. After trypsin treatment, mitochondria were immediately solubilized or further incubated for 15 min a t 30 "C before solubilization. The following procedures were performed exactly as detailed in the legend to Fig. 1. Symbols are: W, 5-min import, U , 10-min import, C -4, 15-min chase aRer 10-min import. solid squares and dotted line), which is likely due to unresolved peaks of the tetramer and monomer. Thus, as compared to the wild-type MCAD, the assembly of the variant K304E MCAD protein appeared to be severely impaired. It was also notable that a significant amount of K304E monomer was detectable not only after a 10-min import alone, but also with a subsequent 15-min chase.
Next, we tested the effects of apyrase and ATP on the K304Evariant MCAD.hsp6O complex as was done earlier to the wildtype MCAD-hsp6O complex. In the mitochondrial lysate incubated for 15  Our previous studies indicated that both the rate and overall amount of transfer of the K304E-variant MCAD from the hM, complex into its final and mature tetrameric form were markedly less than those observed for the wild-type protein. This observation, together with the consideration for the molecular size of the complex, provided the first clue that the h M r complex represents the complex of MCAD and hsp60 (12). In this study, we accurately compared the amount of the wild-type MCAD associated with hsp60 in the hM, complex and that of K304Evariant before and after the chase reaction using anti-hsp60 antibody. After the in vitro translated pMCAD protein was incubated with mitochondria, the mitochondria were solubilized and treated with anti-hsp6O antibody. After a 10-min incubation, both the wild-type and variant MCAD proteins in the mitochondria co-precipitated with hsp60 in similar amounts, 3.9 and 4.5 unitR per unitI of hsp60, respectively (Fig. 6). After a 10-min import and subsequent 15-min chase period, only 0.2 unitR of the wild-type MCAD co-precipitated with a unit1 of hsp60, indicating that the majority of the wild-type was released (transferred) from the complex during the 15-min chase period. In contrast, even after the 15-min chase period, 2.0 unitR of the variant MCAD still co-precipitated per unit1 of hsp60, indicating that the rate of release (transfer) of the variant from the complex was severely retarded.
Role of Mitochondrial hsp70 (hsp70miJ in the Zntramitochondrial Folding and Assembly of MCAD--Recent studies have shown that some proteins, which are newly imported into mitochondria, first interact with hsp70mi,. The interaction appears to occur in the course of translocation into mitochondria (26,27). The loosely folded mature subunits are released from hsp70mit in an ATP-dependent manner, and the released subunits form a complex with hsp60 prior to complete folding and oligomer assembly (15,16,28). Therefore, we studied the interaction of the wild-type and K304E-variant MCADs with hsp70mit and their transfer from hsp70mit to hsp6O. After varying periods of import, mitochondria were treated with trypsin removing pMCAD attached to the outside surface. Mitochondria were then solubilized in a buffer containing apyrase and divided into two equal aliquots. Each aliquot was immunoprecipitated with either anti-hsp70mit antibody or anti-hsp6O antibody (Fig. 7). After a 10-min incubation at 30 "C, both the wild-type and the variant co-precipitated with hsp60, but not with hsp70mit (data not shown). After a 5-min or 10-min incubation a t 16 "C, however, a large amount of wild-type MCAD was complexed with hsp70mit, presumably because at the low temperature the translocation of the precursor MCAD is arrested a t the mitochondrial contact site. When the temperature was raised to 30 "C, wild-type MCAD was completely transferred from hsp70mit to hsp6O within 10 min. The interaction of K304E-variant MCAD with hsp70mit was similar to that of the wild-type: after a 5-10 min incubation a t 16 "C, the variant was complexed with hsp70mit in a large amount similar to that of the wild type. When the temperature was raised to 30 "C, K304E-variant MCAD was also completely transferred from hsp70mit to hsp6O within 10 min, suggesting that its binding to and release from hsp70mit and transfer to hsp6O were normal. ThereaRer, however, the amount of the variant MCAD bound to hsp60 was considerably greater than that of the wild-type for an extended period of time as we have previously demonstrated.

body.
In vitro translation product containing either wild-type or Kp329E pMCAD was incubated with rat liver mitochondria a t 30 "C for 10 min. Trypsin-treated mitochondria were suspended in HMS buffer and a small aliquot (one-twentieth) was taken to determine the total amount of MCAD imported. The rest of the sample was divided into two halves. Mitochondria in one half were immediately pelleted and solubilized, and the other half was further incubated a t 30 "C for 15 min before solubilization. Solubilized mitochondria were divided into two and were immunoprecipitated either with 5 pl of anti-hsp6O antiserum radiolabeled MCAD and hsp6O were detected by autofluorography and (hsp6OAb) or 5 pl of normal rabbit serum (NRS). Immunoprecipitated immunoblot analysis, respectively. Intensity of both MCAD band on autofluorogram and hsp6O band on immunoblot was measured by a densitometer and was expressed in densitometric units (unitR and unit,, respectively). A, SDS-PAGE fluorogram of wild-type and K304E MCADs. B, quantitative determination of MCAD bound to hsp60. The relative amount of MCAD bound to a unit of hsp6O is shown on the ordinate. The value is expressed as the ratio of the intensity of MCAD band (unit,) to that of hsp60 (unitI).

DISCUSSION
In the previous study, we had demonstrated that in mitochondria newly imported MCAD formed a high molecular weight complex (hMr complex) which acted as the precursor for tetramer (12). This kinetic behavior, together with the molecular size of the hMr complex and the consideration for the role of hsp6O in the folding and assembly of mitochondrial proteins in general, strongly suggested that the h M r complex represented a complex of MCAD and hsp60. In the present report, we have demonstrated using immunological and biochemical methods that this was indeed the case. Sixty percent of the wild-type FIG. 7. Immunoprecipitation of solubilized mitochondria containing wild-type or K304E-variant MCAD with anti-hsp70d, or anti-hsp60 antibody. In vitro translation product (50 pl) containing either wild-type or K304E pMCAD was mixed with 25 pl of rat liver mitochondria in each of 5 tubes, with each of them incubated for different periods of time up to 25 min. For the first 10 min, the incubation was performed at 16 "C, and, thereafter, the temperature was shifted to 30 "C. At the end of each incubation, a small aliquot (1.5 pl) of the import mixture was taken and analyzed by SDS-PAGE (Import). The rest of the sample was treated with trypsin, and isolated mitochondria were solubilized in the solubilization buffer containing 20 unitdml apyrase to prevent further ATP-dependent release of MCAD subunit from its hsp7O or hsp60 complex. Solubilized mitochondria were divided into two equal aliquots and were immunoprecipitated with either 5 pl of anti-hsp70mi, antiserum (hsp70Ab) or 5 pl of anti-hsp60 antiserum (hsp6OAb). Immunoprecipitated MCAD was detected and quantified as described in the legend to Fig. 3.

MCAD in the hM, complex co-immunoprecipitated with hsp6O
when treated with anti-hsp60 antibody. Upon exposure to 2.5 m ATP, a similar quantity of MCAD was released from the complex. These results indicated that at least 60% of the hM, complex was the complex of MCAD and hsp6O. The exact cause for the incomplete immunoprecipitation of MCAD in the M.l, complex with anti-hsp6O antibody is currently unknown. One possibility is that the entire MI, complex represented the MCAD.hsp60 complex, but the MCAD moiety was lost from some of the complex during a few cycles of washing after gel filtration. Another possibility is that the hM, complex fraction included MCAD-containing macromolecular complexes other than the MCAD.hsp60 complex. Other macromolecular complexes may include the aggregates of unfolded MCAD and the complexes with other proteins. Because of the small size of the gel filtration column used in this study, it was difficult to completely resolve the void volume and the hM, complex fraction. The hM, complex eluted only a single tube after the void volume. The possibility that MCAD in the hM, complex fraction contained MCAD aggregates is unlikely, however, since in this experiment the supernatant of the mitochondrial lysate was applied to the gel filtration column.
Recently, there have been accumulating evidences that after import into mitochondria as unfolded peptide, nuclear-coded proteins interact sequentially, first with hsp70mit and then transferred to hsp6O before completing the process of folding and assembly (15)(16)(17). We have shown in this paper that both hsp70mit and hsp60 were indeed involved in the folding and assembly of MCAD in mitochondria. When treated with anti-hsp70mit antibody, MCAD co-precipitated with hsp70mit only at a low temperature (16 "C) for a relatively short period of time immediately following the import, indicating that MCAD and hsp70mit formed a transient complex in the early step of mitochondrial import. MCAD protein was quickly transferred to hsp60 when the temperature was raised to 30 "C. MCAD-hsp6O complex survived for a relatively longer period of time even at 30 "C. These data suggest that the binding of MCAD to, and its release from, hsp70mit are very fast at physiological temperature as in the case of Maslp, the larger of the two subunits of the MAS-encoded processing protease (16), and that the folding of MCAD on hsp6O scaffold is the rate-limiting step in the biogenesis pathway of MCAD tetramer.
In the present study, we have shown that after a 15-min chase in intact mitochondria or after ATP addition to the mitochondrial lysate, the amount of the wild-type MCAD-hsp6O complex greatly decreased while the quantity of tetramer inversely increased. Since no monomer was detectable in either experiment, this appeared to indicate that the release of the wild-type MCAD from the MCAD.hsp6O complex and its assembly into tetramer were tightly coupled, as if suggesting that hsp6O assists not only folding of loosely folded proteins into appropriate conformation but also assembly of the correctly folded proteins. However, the currently prevailing view in the study of chaperonins including hsp6O and its bacterial counterpart, groEL, is that while in the process of folding, chaperonins protect loosely folded proteins from forming aggregates via hydrophobic interactions (14,29,30). Recently, two groups of investigators have demonstrated that in a proteidgroELS complex, protein is contained in the central cavity of groEL double heptamer-toroidal rings with one end closed with groES (31,32). GroES is the bacterial counterpart of hspl0. It has been considered that once the proteins are correctly folded, monomers are released and they can be automatically assembled without assistance of chaperonin (14). In fact, Zheng et al. (33) have recently shown that upon addition of groES and ATP to the reconstituted ornithine transcarbamylase-groEL complex, ornithine transcarbamylase monomers were first released within 30 s, and, after 90 s, approximately 40% were assembled into the native trimer. At 20 min, no monomers were remaining. In our present experiments, the analysis of the product was done 15 min after chase or after ATP addition. Therefore, it is possible that our observation of the apparent tight coupling of the release of MCAD from its complex with hsp60 and the tetramer assembly could have been the result of a quick sequence of release of monomers, followed by efficient assembly of the resulting monomers into tetramer. It is unknown a t present, however, whether or not the behavior of the in vitro reconstituted ornithine transcarbamylase-groEL complex exactly mimics that of the native mammalian enzyme-hsp60 complex in mitochondria. For instance, Viitanen et al. (25) suggested that unlike groEL which is a double heptamer-toroidal ring, mammalian mitochondrial hsp60 is a single heptamer-toroidal ring. Also, MCAD may behave differently from the way ornithine transcarbamylase did in the reconstituted system. There has so far been no conclusive evidence for or against the role of chaperonin in the assembly of correctly folded protein. Thus, the mechanisms of MCAD transfer from the complex with hsp6O to the final tetramer needs to be studied further. In this regard, the findings from the study of K304E-variant MCAD provides important information concerning the mechanisms for the release of appropriately folded MCAD protein from the complex with hsp6O and subsequent assembly into active tetramer as discussed below.
We had previously shown that overall foldinglassembly process of the K304E-variant protein was severely impaired, probably at the transfer of the variant MCAD moiety from the hM, complex to the tetramer (12). However, the precise mechanism for the impairment was unknown, as the nature of the hM, complex and the detailed mechanism of the folding and assembly of MCAD were yet to be studied at the time. In the present paper, we have shown that following the import into mitochondria, the binding of K304E-variant protein to hsp70mit and its transfer to hsp6O were normal. However, the stability of the complex of the variant with hsp6O was markedly different from that of the wild-type. The accurate quantitation of the amount of the wild-type MCAD associated with hsp60 in the hM, complex and that of K304E-variant using anti-hsp6O antibody revealed that after a 15-min chase period, only 2% of the wildtype MCAD was remaining in the complex, whereas 4 4 % of the variant MCAD was still found with the complex. Likewise, when the mitochondrial lysate was incubated with ATP, 21% of the wild-type MCAD was found with the hM, complex, whereas 64% of K304E still remained with the h M r complex. These data indicated that the rate of release of the variant from the complex was severely retarded. K304E involves the change of the charge in the interface in the dimer-dimer interaction, not that of monomer-monomer interaction. Since a short gel filtration column was used in the present study, dimer peak was not well resolved from those of monomer and tetramer. In the previous study, however, a longer column that was capable of separating these three peaks was used for the analysis, but at no time point after mitochondrial import was dimer detectable. Since there were no gross accumulation of dimer, an impairment in the assembly of properly folded variant protein into tetramer is unlikely as the cause of the stable nature of the MI, complex of the K304E-variant. Such a stable nature of the variant complex is likely to be caused by a hindrance in the folding of the variant protein into a proper monomeric conformation while bound to hsp60.
Our recent study on the MCAD biogenesis in riboflavin deficiency provides a clue which is relevant in this regard. We had previously demonstrated that the wild-type MCAD, as a FADless apoprotein, was labile in riboflavin-deficient rats (34). More recently, we have observed that in riboflavin-deficient mitochondria, the assembly of the wild-type MCAD into the native tetramer was impaired, and that upon exposure to ATP, the wild-type MCAD was not efficiently released from the complex with hsp60 as was the K304E-variant in normal mitochondria.3 These data on riboflavin-deficient mitochondria indicated that MCAD molecule cannot be properly folded in the absence of FAD and that unfolded MCAD protein is not released efficiently from the complex with hsp60. The similarity in response to ATP of the wild-type MCAD in riboflavin-deficient mitochondria and that of K304E in normal mitochondria supports the possibility that the K304E-variant MCAD, too, was not efficiently released from its complex with hsp60 upon exposure to ATP because folding of K304E was impaired.
Thus, in contrast to the initial hypothesis (9-12), the available evidence suggests that the hindrance of folding into a proper conformation is the major cause of the impairment of assembly of K304E into the native tetramer. However, as we have previously shown, a small amount of tetramer of the variant was formed (12). Thus, it is still possible that the change of the charge in the interface interferes with the dimer-dimer interaction, causing instability of the resulting variant tetramer. In addition, the possibility that folding of K304E proceeds normally, as initially considered, still cannot be completely excluded. Although the native MCAD tetramer is regarded as a dimer of two dimers based on the x-ray crystallographic evidence (51, this concept may not reflect the sequence of events in the assembly of properly folded MCAD monomers. In this hypothesis, the properly folded K304E cannot leave, or can leave but goes back to, the central cavity of hsp60 because of its failure to interact with another K304E.hsp60 complex via the externally exposed interface due to the change of the charge at the site. This hypothesis assumes that MCAD monomer, including the wild-type and K304E, is protected by hsp6O until immediately before assembly. Such protection of folded protein by chaperonin until the protein interacts with its counterpart for forming an oligomer has been suggested by the recent study of the El component of mammalian branched chain a-ketoacid dehydrogenase. El is a heterotetramer with an azPz structure. Chuang and associates observed that while co-expression of Ela and E1p in Escherichia coli produced biologically active El, in vitro mixing of individu-T. Saijo and K. Tanaka, unpublished observations. ally expressed Ela and E1P did not result in assembly or produce El activity (35).
The nature of the wild-type and K304E-variant monomers detected after a 10-min import needs to be discussed. In the folding/assembly pathway of MCAD in the mitochondria, three possible types of monomer can be theoretically predicted. These are loosely folded monomer, which is released from hsp70mit before making a complex with hsp60, properly folded monomer released from hsp60, and aged monomer dissociated from the tetramer. It is unlikely that the wild-type and K304E monomers observed at a 10-min import represents either of the latter two types, since no monomer was detected at other time points nor under any other experimental conditions, such as exposure to ATP. Hence, it is likely that the monomers of the wild-type and those of K304E, detected after 10 min of import, may represent the loosely folded monomeric subunits that were released from hsp70mit and overtlowed the binding capacity of hsp60 as the amount of imported monomer increased from 5 min to 10 min. This hypothesis is firther supported by the finding that the amount of the K304E monomer and that in h M r complex after 10-min import are both considerably greater than the wild-type counterparts.