A strategy for optimizing the monodispersity of fusion proteins: application to purification of recombinant HPV E6 oncoprotein

Recombinant production of HPV oncoprotein E6 is notoriously difficult. The unfused sequence is produced in inclusion bodies. By contrast, fusions of E6 to the C-terminus of carrier proteins such as maltose-binding protein or glutathione-S-transferase are produced soluble. However, it has not yet been possible to purify E6 protein from such fusion constructs. Here, we show that this was due to the biophysical heterogeneity of the fusion preparations. We find that soluble MBP-E6 preparations contain two subpopulations. A major fraction is aggregated and contains exclusively misfolded E6 moieties ('soluble inclusion bodies'). A minor fraction is monodisperse and contains the properly folded E6 moieties. Using monodispersity as a screening criterion, we optimized the expression conditions, the purification process and the sequence of E6, finally obtaining stable monodisperse MBP-E6 preparations. In contrast to aggregated MBP-E6, these preparations yielded fully soluble E6 after proteolytic removal of MBP. Once purified, these E6 proteins are stable, folded and biologically active. The first biophysical measurements on pure E6 were performed. This work shows that solubility is not a sufficient criterion to check that the passenger protein in a fusion construct is properly folded and active. By contrast, monodispersity appears as a better quality criterion. The monodispersity-based strategy presented here constitutes a general method to prepare fusion proteins with optimized folding and biological activity.     

Protein mutagenesis with monodispersity-based quality probing: selective inactivation of p53 degradation and DNA-binding properties of HPV E6 oncoprotein

Interpretation of protein mutagenesis experiments requires the ability to distinguish functionally relevant mutations from mutations affecting the structure. When a protein is expressed soluble in bacteria, properly folded mutants are expected to remain soluble whereas misfolded mutants should form insoluble aggregates. However, this rule may fail for proteins fused to highly soluble carrier proteins. In a previous study, we analysed the biophysical status of HPV oncoprotein E6 fused to the C-terminus of maltose-binding protein (MBP) and found that misfolded E6 moieties fused to MBP formed soluble aggregates of high molecular weight. By contrast, preparations of properly folded E6 fused to MBP were monodisperse. Here, we have used this finding to evaluate the quality of 19 MBP-fused E6 site-directed mutants by using a light scattering assay performed in a fluorimeter. This assay guided us to rule out structurally defective mutants and to obtain functionally relevant E6 mutants selectively altered for two molecular activities: degradation of tumour suppressor p53 and DNA recognition.     

Formation of well-defined soluble aggregates upon fusion to MBP is a generic property of E6 proteins from various human papillomavirus species

Protein aggregation is a main barrier hindering structural and functional studies of a number of interesting biological targets. The E6 oncoprotein of Human Papillomavirus strain 16 (E6(16)) is difficult to express under a native soluble form in bacteria. Produced as an unfused sequence, it forms inclusion bodies. Fused to the C-terminus of MBP, it is mainly produced in the form of soluble high molecular weight aggregates. Here, we produced as MBP-fusions seven E6 proteins from other HPV strains (5, 11, 18, 33, 45, 52, and 58) belonging to four different species, and we compared their aggregation state to that of MBP-E6(16). Using a fast mutagenesis method, we changed most non-conserved cysteines to the isosteric residue serine to minimize disulfide bridge-mediated aggregation during purification. Static and dynamic light scattering measurements, ultracentrifugation and electron microscopy demonstrated the presence in all MBP-E6 preparations of soluble high-molecular weight aggregates with a well-defined spherical shape. These aggregated particles are relatively monodisperse but their amount and their size vary depending on the conditions of expression and the strain considered. For all strains, minimal aggregate formation occurs when the expression is performed at 15 degrees C. Such observations suggest that the assembly of MBP-E6 aggregates takes place in vivo during protein biosynthesis, rather than occurring during purification. Finally, we show that all MBP-E6 preparations contain two zinc ions per protein monomer, suggesting that E6 domains within the high molecular weight aggregates possess a native-like fold, which enables correct coordination to the metal center.     

Properties of soluble fusions between mammalian aspartic proteinases and bacterial maltose-binding protein

The mammalian aspartic proteinases procathepsin D and pepsinogen form insoluble inclusion bodies when expressed in bacteria. They become soluble but nonnative when synthesized as fusions to the carboxy terminus of E. coli maltose-binding protein (MBP). Since these nonnative states of the two aspartic proteinases showed no tendency to form insoluble aggregates, their biophysical properties were analyzed. The MBP portions were properly folded as shown by binding to amylose, but the aspartic proteinase moieties failed to bind pepstatin and lacked enzymatic activity, indicating that they were not correctly folded. When treated with proteinase K, only the MBP portion of the fusions was resistant to proteolysis. The fusion between MBP and cathepsin D had increased hydrophobic surface exposure compared to the two unfused partners, as determined by bis-ANS binding. Ultracentrifugal sedimentation analysis of MBP–procathepsin D and MBP–pepsinogen revealed species with very large and heterogeneous sedimentation values. Refolding of the fusions from 8 M urea generated proteins no larger than dimers. Refolded MBP–pepsinogen was proteolytically active, while only a few percent of renatured MBP–procathepsin D was obtained. The results suggest that MBP–aspartic proteinase fusions can provide a source of soluble but nonnative folding states of the mammalian polypeptides in the absence of aggregation.     

The N-terminal stem region of bovine and human beta1,4-galactosyltransferase I increases the in vitro folding efficiency of their catalytic domain from inclusion bodies

Many recombinant proteins overexpressed in Escherichia coli are generally misfolded, which then aggregate and accumulate as inclusion bodies. The catalytic domain (CD) of bovine and human beta1,4-galactosyltransferase (beta4Gal-T), expressed in E. coli, it also accumulates as inclusion bodies. We studied the effect of the fusion of the stem region (SR), as an N-terminal extension of the catalytic domain, on the in vitro folding efficiencies of the inclusion bodies. The stem region fused to the catalytic domain (SRCD) increases the folding efficiency of recombinant protein with native fold compared to the protein that contains only the CD. During in vitro folding, also promotes considerably the solubility of the misfolded proteins, which do not bind to UDP-agarose columns and exhibit no galactosyltransferase activity. In contrast, the misfolded proteins that consist of only the CD are insoluble and precipitate out of solution. It is concluded that a protein domain that is produced in a soluble form does not guarantee the presence of the protein molecules in a properly folded and active form. The stem domain has a positive effect on the in vitro folding efficiency of the catalytic domain of both human and bovine beta4Gal-T1, suggesting that the stem region acts as a chaperone during protein folding. Furthermore, investigation of the folding conditions of the sulphonated inclusion bodies resulted in identifying a condition in which the presence of PEG-4000 and L-arginine, compared to their absence, increased the yields of native CD and SRCD 7- and 3-fold, respectively.     

Localization of functional polypeptides in bacterial inclusion bodies

Bacterial inclusion bodies, while showing intriguing amyloid-like features, such as a beta-sheet-based intermolecular organization, binding to amyloid-tropic dyes, and origin in a sequence-selective deposition process, hold an important amount of native-like secondary structure and significant amounts of functional polypeptides. The aggregation mechanics supporting the occurrence of both misfolded and properly folded protein is controversial. Single polypeptide chains might contain both misfolded stretches driving aggregation and properly folded protein domains that, if embracing the active site, would account for the biological activities displayed by inclusion bodies. Alternatively, soluble, functional polypeptides could be surface adsorbed by interactions weaker than those driving the formation of the intermolecular beta-sheet architecture. To explore whether the fraction of properly folded active protein is a natural component or rather a mere contaminant of these aggregates, we have explored their localization by image analysis of inclusion bodies formed by green fluorescent protein. Since the fluorescence distribution is not homogeneous and the core of inclusion bodies is particularly rich in active protein forms, such protein species cannot be passively trapped components and their occurrence might be linked to the reconstruction dynamics steadily endured in vivo by such bacterial aggregates. Intriguingly, even functional protein species in inclusion bodies are not excluded from the interface with the solvent, probably because of the porous structure of these particular protein aggregates.     

Localization of Functional Polypeptides in Bacterial Inclusion Bodies

Bacterial inclusion bodies, while showing intriguing amyloid-like features, such as a β-sheet-based intermolecular organization, binding to amyloid-tropic dyes, and origin in a sequence-selective deposition process, hold an important amount of native-like secondary structure and significant amounts of functional polypeptides. The aggregation mechanics supporting the occurrence of both misfolded and properly folded protein is controversial. Single polypeptide chains might contain both misfolded stretches driving aggregation and properly folded protein domains that, if embracing the active site, would account for the biological activities displayed by inclusion bodies. Alternatively, soluble, functional polypeptides could be surface adsorbed by interactions weaker than those driving the formation of the intermolecular β-sheet architecture. To explore whether the fraction of properly folded active protein is a natural component or rather a mere contaminant of these aggregates, we have explored their localization by image analysis of inclusion bodies formed by green fluorescent protein. Since the fluorescence distribution is not homogeneous and the core of inclusion bodies is particularly rich in active protein forms, such protein species cannot be passively trapped components and their occurrence might be linked to the reconstruction dynamics steadily endured in vivo by such bacterial aggregates. Intriguingly, even functional protein species in inclusion bodies are not excluded from the interface with the solvent, probably because of the porous structure of these particular protein aggregates.     

Strategies for bacterial expression of protein-peptide complexes: application to solubilization of papillomavirus E6

E6 is a small oncoprotein involved in tumorigenesis induced by papillomaviruses (PVs). E6 often recognizes its cellular targets by binding to short motifs presenting the consensus LXXLL. E6 proteins have long resisted structural analysis. We found that bovine papillomavirus type 1 (BPV1) E6 binds the N-terminal LXXLL motif of the cellular protein paxillin with significantly higher affinity as compared to other E6/peptide interactions. Although recombinant BPV1 E6 was poorly soluble in the free state, provision of the paxillin LXXLL peptide during BPV1 E6 biosynthesis greatly enhanced the protein's solubility. Expression of BPV1 E6/LXXLL peptide complexes was carried out in bacteria in the form of triple fusion constructs comprising, from N- to C-terminus, the soluble carrier protein maltose binding protein (MBP), the LXXLL motif and the E6 protein. A TEV protease cleavage site was placed either between MBP and LXXLL motif or between LXXLL motif and E6. These constructs allowed us to produce highly concentrated samples of BPV1 E6, either covalently fused to the C-terminus of the LXXLL motif (intra-molecular complex) or non-covalently bound to it (inter-molecular complex). Heteronuclear NMR measurements were performed and showed that the E6 protein was folded with similar conformations in both covalent and non-covalent complexes. These data open the way to novel structural and functional studies of the BPV1 E6 in complex with its preferential target motif.     

One-step affinity purification of fusion proteins with optimal monodispersity and biological activity: application to aggregation-prone HPV E6 proteins

Background

Bacterial expression and purification of recombinant proteins under homogeneous active form is often challenging. Fusion to highly soluble carrier proteins such as Maltose Binding Protein (MBP) often improves their folding and solubility, but self-association may still occur. For instance, HPV E6 oncoproteins, when produced as MBP-E6 fusions, are expressed as mixtures of biologically inactive oligomers and active monomers. While a protocol was previously developed to isolate MBP-E6 monomers for structural studies, it allows the purification of only one MBP-E6 construct at the time. Here, we explored a parallelizable strategy more adapted for biophysical assays aiming at comparing different E6 proteins.

Results

In this study, we took advantage of the distinct size and diffusion properties of MBP-E6 monomers and oligomers to separate these two species using a rapid batch preparation protocol on affinity resins. We optimized resin reticulation, contact time and elution method in order to maximize the proportion of monomeric MBP-E6 in the final sample. Analytical size-exclusion chromatography was used to quantify the different protein species after purification. Thus, we developed a rapid, single-step protocol for the parallel purification of highly monomeric MBP-E6 samples. MBP-fused HPV16 E6 samples obtained by this approach were validated by testing the binding to their prototypical peptide targets (the LXXLL motif from ubiquitine ligase E6AP) by BIAcore-SPR assay.

Conclusions

We have designed a rapid single-step batch affinity purification approach to isolate biologically active monomers of MBP-fused E6 proteins. This protocol should be generalizable to isolate the monomer (or the minimal biologically active oligomer) of other proteins prone to self-association.

     

Active inclusion body formation using Paenibacillus polymyxa PoxB as a fusion partner in Escherichia coli

Overexpression of Paenibacillus polymyxa PoxB in Escherichia coli induced the formation of inclusion bodies. An enzyme assay showed that the inclusion bodies exhibited PoxB activity, indicating that they were biologically active. Fusion of GFP and Bacillus subtilis AmyE to the C-terminus of the PoxB also induced the formation of biologically active aggregates when they were overexpressed in E. coli. Therefore, P. polymyxa PoxB can be used as a fusion partner to promote the formation of active inclusion bodies in E. coli.     

Divergent genetic control of protein solubility and conformational quality in Escherichia coli

In bacteria, protein overproduction results in the formation of inclusion bodies, sized protein aggregates showing amyloid-like properties such as seeding-driven formation, amyloid-tropic dye binding, intermolecular β-sheet architecture and cytotoxicity on mammalian cells. During protein deposition, exposed hydrophobic patches force intermolecular clustering and aggregation but these aggregation determinants coexist with properly folded stretches, exhibiting native-like secondary structure. Several reports indicate that inclusion bodies formed by different enzymes or fluorescent proteins show detectable biological activity. By using an engineered green fluorescent protein as reporter we have examined how the cell quality control distributes such active but misfolded protein species between the soluble and insoluble cell fractions and how aggregation determinants act in cells deficient in quality control functions. Most of the tested genetic deficiencies in different cytosolic chaperones and proteases (affecting DnaK, GroEL, GroES, ClpB, ClpP and Lon at different extents) resulted in much less soluble but unexpectedly more fluorescent polypeptides. The enrichment of aggregates with fluorescent species results from a dramatic inhibition of ClpP and Lon-mediated, DnaK-surveyed green fluorescent protein degradation, and it does not perturb the amyloid-like architecture of inclusion bodies. Therefore, the Escherichia coli quality control system promotes protein solubility instead of conformational quality through an overcommitted proteolysis of aggregation-prone polypeptides, irrespective of their global conformational status and biological properties.     

Characterization of factors favoring the expression of soluble protozoan tubulin proteins in Escherichia coli

The alpha- and beta-tubulin genes of the parasitic protozoa Giardia duodenalis, Cryptosporidium parvum, and Encephalitozoon intestinalis have been overexpressed in soluble form using Escherichia coli-based expression systems. Several expression systems were compared in terms of the amount of soluble protein produced with different fusion partners, strains of E. coli BL21, and expression temperatures. The cleavability of the fusion partner was also assessed in terms of post-expression applications of the recombinant protein. The maltose-binding protein (MBP) and glutathione S-transferase (GST) fusion partners produced the highest expression levels for all six proteins without the formation of inclusion bodies. The expression system also provided a means of purifying the soluble protein using affinity and anion-exchange chromatography while minimizing protein losses. The yield and purity were therefore very high for both the MBP and GST systems. The tubulin monomers were demonstrated to be assembly-competent using a standard dimerization assay and also retained full antigenicity with monoclonal antibodies. This study presents several methods which are suitable for producing soluble tubulin monomers and, thus, circumventing the formation of inclusion bodies which necessitates re-folding of the tubulin.     

E6-AP promotes misfolded polyglutamine proteins for proteasomal degradation and suppresses polyglutamine protein aggregation and toxicity

The accumulation of intracellular protein deposits as inclusion bodies is the common pathological hallmark of most age-related neurodegenerative disorders including polyglutamine diseases. Appearance of aggregates of the misfolded mutant disease proteins suggest that cells are unable to efficiently degrade them, and failure of clearance leads to the severe disturbances of the cellular quality control system. Recently, the quality control ubiquitin ligase CHIP has been shown to suppress the polyglutamine protein aggregation and toxicity. Here we have identified another ubiquitin ligase, called E6-AP, which is able to promote the proteasomal degradation of misfolded polyglutamine proteins and suppress the polyglutamine protein aggregation and polyglutamine protein-induced cell death. E6-AP interacts with the soluble misfolded polyglutamine protein and associates with their aggregates in both cellular and transgenic mouse models. Partial knockdown of E6-AP enhances the rate of aggregate formation and cell death mediated by the polyglutamine protein. Finally, we have demonstrated the up-regulation of E6-AP in the expanded polyglutamine protein-expressing cells as well as cells exposed to proteasomal stress. These findings suggest that E6-AP is a critical mediator of the neuronal response to misfolded polyglutamine proteins and represents a potential therapeutic target in the polyglutamine diseases.     

Expression and assembly of a functional E1 component (alpha 2 beta 2) of mammalian branched-chain alpha-ketoacid dehydrogenase complex in Escherichia coli.

We have expressed an active recombinant E1 decarboxylase component of the mammalian branched-chain alpha-ketoacid dehydrogenase complex in Escherichia coli by subcloning mature E1 alpha and E1 beta subunit cDNA sequences into a bacterial expression vector. To permit affinity purification under native conditions, the mature E1 alpha subunit was fused with the affinity ligand E. coli maltose-binding protein (MBP) through an endoprotease Factor Xa-specific linker peptide. When co-expressed, the MBP-E1 alpha fusion and E1 beta subunits were shown to co-purify as a MBP-E1 component that exhibited both E1 activity and binding competence for recombinant branched-chain E2 component. In contrast, in vitro mixing of individually expressed MBP-E1 alpha and E1 beta did not result in assembly or produce E1 activity. Following proteolytic removal of the affinity ligand and linker peptide with Factor Xa, a recombinant E1 species was eluted from a Sephacryl S-300HR sizing column as an enzymatically active 160-kDa species. The latter showed 1:1 subunit stoichiometry, which was consistent with an alpha 2 beta 2 structure. The recovery of this 160-kDa recombinant E1 species (estimated at 0.07% of total lysate protein) was low, with the majority of the recombinant protein lost as insoluble aggregates. Our findings suggest that the concurrent expression of both E1 alpha and E1 beta subunits in the same cellular compartment is important for assembly of both subunits into a functional E1 alpha 2 beta 2 heterotetramer. By using this co-expression system, we also find that the E1 alpha missense mutation (Tyr-393----Asn) characterized in Mennonites with maple syrup urine disease prevents the assembly of soluble E1 heterotetramers.     

In vivo increase of solubility of overexpressed Hha protein by tandem expression with interacting protein H-NS

Gene cloning in appropriate vectors followed by protein overexpression in Escherichia coli is the common means for protein purification. This approach has many advantages but also some drawbacks; one of these is that many proteins fail to achieve a soluble conformation when overexpressed in E. coli. Hha protein belongs to a family of nucleoid-associated proteins functionally related to the H-NS family of proteins. Hha-like proteins and H-NS-like proteins are able to semidirectly bind to each other. We show in this work that overexpressed Hha or HisHha protein (a functional derivative of Hha containing a 6x His tag at the amino end) from a T7-polymerase promoter in BL21 DE3 E. coli strains results in the vast majority of the protein accumulated in insoluble aggregates (inclusion bodies). We also show that tandem overexpression of HisHha and H-NS increases the solubility of HisHha and prevents the formation of inclusion bodies. Single amino acid substitutions in the HisHha protein, which impair interaction with H-NS, render insoluble protein even when tandem-expressed with H-NS, tandem expression of an insoluble protein and an interacting partner is an experimental strategy which could be useful to increase the solubility of other overexpressed proteins in E. coli.     

Protein Purification with C-Terminal Fusion of Maltose Binding Protein

For affinity-chromatography-based purification of proteins that are prone to abnormal termination of translation or that may not be modified at their N-termini, affinity tags are needed which can be fused to the C-terminus. In this publication we describe that maltose binding protein (MBP) fused to the C-terminus of the plant photoreceptor phytochrome B allows purification of the fusion protein via amylose affinity chromatography. After overexpression in yeast a 125-fold enrichment could be achieved. The spectral properties of phytochrome B were not impaired by the fusion and purification. These results demonstrate that not only the widely used N-terminal fusions of MBP but also C-terminal fusions can be employed for protein purification.     

Folding of the multidomain human immunodeficiency virus type-I integrase.

Protein folding conditions were established for human immunodeficiency virus integrase (IN) obtained from purified bacterial inclusion bodies. IN was denatured by 6 M guanidine.HCl-5 mM dithiothreitol, purified by gel filtration, and precipitated by ammonium sulfate. The reversible solvation of precipitated IN by 6 M guanidine.HCl allowed for wide variation of protein concentration in the folding reaction. A 6-fold dilution of denatured IN by 1 M NaCl buffer followed by dialysis produced enzymatically active IN capable of 3' OH end processing, strand transfer, and disintegration using various human immunodeficiency virus-1 (HIV-1) long terminal repeat DNA substrates. The specific activities of folded IN preparations for these enzymatic reactions were comparable to those of soluble IN purified directly from bacteria. The subunit composition and enzymatic activities of IN were affected by the folding conditions. Standard folding conditions were defined in which monomers and protein aggregates sedimenting as dimers and tetramers wree produced. These protein aggregates were enzymatically active, whereas monomers had reduced strand transfer activity. Temperature modifications of the folding conditions permitted formation of mainly monomers. Upon assaying, these monomers were efficient for strand transfer and disintegration, but the oligomeric state of IN under the conditions of the assay is determinate. Our results suggest that monomers of the multidomain HIV-1 IN are folded correctly for various catalytic activities, but the conditions for specific oligomerization in the absence of catalytic activity are undefined.     

Escherichia coli small heat shock proteins IbpA/B enhance activity of enzymes sequestered in inclusion bodies

Escherichia coli small heat shock proteins, IbpA/B, function as molecular chaperones and protect misfolded proteins against irreversible aggregation. IbpA/B are induced during overproduction of recombinant proteins and bind to inclusion bodies in E. coli cells. We investigated the effect of DeltaibpA/B mutation on formation of inclusion bodies and biological activity of enzymes sequestered in the aggregates in E. coli cells. Using three different recombinant proteins: Cro-beta-galactosidase, beta-lactamase and rat rHtrA1 we demonstrated that deletion of the ibpA/B operon did not affect the level of produced inclusion bodies. However, in aggregates containing IbpA/B a higher enzymatic activity was detected than in the IbpA/B-deficient inclusion bodies. These results confirm that IbpA/B protect misfolded proteins from inactivation in vivo.     

Probing the Structural Role of an αβ Loop of Maltose-binding Protein by Mutagenesis: Heat-shock Induction by Loop Variants of the Maltose-binding Protein that Form Periplasmic Inclusion Bodies

The maltose-binding protein (MBP) ofEscherichia coliis the periplasmic receptor of the maltose transport system. Previous studies have identified amino acid substitutions in an α/β loop of the structure of MBP that are critical for thein vivofolding. To probe genetically the structural role of this surface loop, we generated a library in which the corresponding codons 32 and 33 ofmalEwere mutagenized. The maltose phenotype, which correlates with a biologically active structure of MBP in the periplasm, indicated a considerable variability in the loop residues compatible with a correctin vivofolding pathway of the protein. By the same genetic screens, we characterized loop-variant MBPs associated with a defective periplasmic folding pathway and aggregated into inclusion bodies. Heat-shock induction with production of misfolded loop variants was examined using bothlon-lacZandhtrA-lacZfusions. We found that the extent of formation of inclusion bodies in the periplasm ofE. coli, from misfolded loop variant MBPs, correlated with the level of heat-shock response regulated by the alternate heat-shock σ factor, σ²⁴.     

Chaperonins groEL and groES promote assembly of heterotetramers (alpha 2 beta 2) of mammalian mitochondrial branched-chain alpha-keto acid decarboxylase in Escherichia coli.

We have investigated the possible role of chaperonins groEL and groES in the folding and assembly of heterotetramers (alpha 2 beta 2) of mammalian mitochondrial branched-chain alpha-keto acid decarboxylase (E1) in Escherichia coli. The mature E1 alpha subunit fused to maltose-binding protein (MBP) was coexpressed with mature E1 beta on the same vector in ES- and EL- mutant strains. Only small or trace amounts of active E1 component were obtained. Cotransformation of the ES- mutant host with a second vector overexpressing groEL and groES resulted in a greater than 500-fold increase in E1-specific activity. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis showed that the content of both MBP-E1 alpha and E1 beta polypeptides was markedly increased in the presence of overexpressed chaperonin proteins. The time course studies showed that the increase in E1-specific activity and subunit levels correlated with the increase in groEL and groES until the concentration of the chaperonins reached a saturating level in the cell. The functional MBP-E1 fusion protein from ES- double transformants were purified by amylose resin affinity chromatography. The MBP moiety was removed by subsequent digestion with Factor Xa endoprotease, followed by Sephacryl S-300HR chromatography. It was found that E1 alpha and E1 beta assembled into an active 160-kDa species, which was consistent with the alpha 2 beta 2 structure of E1. The present results demonstrate that chaperonins groEL and groES promote folding and assembly of heterotetrameric proteins of mammalian mitochondrial origin.