Molecular chaperones

Plugging the transport gap

Molecular chaperones are generally thought to protect newly synthesized proteins and ensure that they fold into the correct shape. But it seems that two chaperones also help to target certain proteins to mitochondria.

Mitochondria are the major energy-producing organelles of animal and fungal cells, and contain more than 700 different proteins. Fewer than 5% of these proteins are encoded by mitochondrial genes and produced inside mitochondria. Instead, most are encoded by genes in the cell nucleus. They are synthesized as preproteins in the cytoplasm by complex molecular machines called ribosomes, before being transported into developing mitochondria. Writing in Cell, Young et al.1 report that the transport of certain preproteins into mitochondria requires the chaperone proteins Hsp70 and Hsp90. The chaperones act by binding preproteins to a specific structure in a receptor protein called Tom70, which lies in the outer mitochondrial membrane (Fig. 1). This paper helps to plug the gap left by two reports2,3 from the early years of 'chaperonology', which indicated that Hsp70 is involved in transporting proteins into mitochondria, although the transport mechanism was unknown. It also provides the first evidence that Hsp90 is involved in targeting proteins to a specific organelle.

Figure 1: Targeting proteins to mitochondria, as proposed by Young et al.1.

Newly synthesized preproteins that are destined to function in mitochondrial membranes are imported via the Tom70 receptor. From left to right: in mammals, the chaperones Hsp90 and Hsp70 bind to the preprotein in the cytosol (in yeast, just Hsp70 is required). The chaperones dock onto Tom70's clamp domain, allowing internal targeting sequences in the preprotein to be recognized by Tom70's core domain. Further Tom70 proteins are then recruited. ATP hydrolysis, probably by the chaperones, is needed to transfer the preproteins from Tom70, through the import pore in the outer mitochondrial membrane, to import machinery in the inner membrane. (Modified from ref. 1.)

Mitochondrial preproteins face several problems. After they have been released from ribosomes, they must be targeted to mitochondria but not to other organelles in the cell. They must also adopt conformations that can unfold easily, because they move as extended chains through the two membranes (outer and inner) that surround mitochondria. Such incompletely folded proteins often display hydrophobic regions on their surface, especially if they are membrane proteins, creating the danger of self-aggregation. These problems are combated by molecular chaperones, a large class of proteins that assist the correct folding and assembly of other proteins without remaining bound to them once they are performing their normal functions4.

The targeting problem is solved in part by the fact that preproteins contain one of two types of amino-acid sequence that enable them to be recognized by one of two types of receptor in the outer mitochondrial membrane5. These are called Tom receptors — short for 'translocase of the outer mitochondrial membrane'. The Tom20 receptor binds preferentially to 'presequences' found at the amino terminus of some preproteins; presequences form positively charged helices that are removed after transport through the mitochondrial membrane, and so do not appear in the mature protein. In contrast, the Tom70 receptor binds to internal sequences found in membrane proteins such as the phosphate and adenine-nucleotide carrier proteins of the inner mitochondrial membrane. Although such sequences are not removed after transport, these carriers are also called preproteins at the stage before transport.

It now seems, however, that this sequence information in preproteins is not the only means of targeting them to mitochondria; their chaperones also have a part to play, at least in the case of membrane preproteins1. Previous work had shown that a portion of Tom70 located just outside the mitochondrion (that is, in the cell cytosol) recognizes the internal targeting sequences of membrane preproteins6. This cytosolic portion also contains several so-called tetratricopeptide repeat (TPR) sequences, which form a characteristic structure within a protein that consists of a helix, then a turn, then a helix. Three of Tom70's TPR sequences are rather similar to sequences in a class of proteins that do not bind preproteins themselves, but instead act as cochaperones by binding to the chaperones Hsp70 or Hsp90. The TPR sequences in the cochaperones form what is called a dicarboxylate clamp domain, which interacts with the aspartate amino acid at the carboxy-terminal end of Hsp70 and Hsp90. Specificity for either Hsp70 or Hsp90 is determined by hydrophobic contacts with neighbouring amino acids in the chaperones.

Young et al.1 now build on these findings. They first suggest that, given the similarity between the TPR motifs of Tom70 and those of the cochaperones, these sequences in Tom70 also form a clamp domain. They further propose that this domain contributes to preprotein targeting by binding the chaperones.

The authors find that the cytosolic segment of human Tom70 binds both Hsp90 and Hsp70 in crude cell extracts, but cannot do so if a key arginine amino acid in Tom70's presumed clamp domain is mutated to alanine. Interestingly, yeast Tom70 binds to Hsp70 only, but this interaction is also abolished by the clamp mutation. This difference between yeast and mammals is due to the receptor, not the chaperones. As an indication of the importance of Tom70's clamp domain to mitochondrial protein import, yeast cells die if the normal Tom70 receptor is replaced by one with the clamp mutation. Moreover, import of the phosphate-carrier protein by isolated rat liver mitochondria is inhibited by adding surplus Hsp90 carboxy-terminal domain, which competes with the complete Hsp90 for binding to Tom70's clamp domain. The import of preproteins that use the Tom20 receptor is not affected. Similar experiments with yeast mitochondria indicate that Hsp70–Tom interactions are also essential for preprotein import via Tom70.

Young et al. also start to disentangle the respective roles of Hsp90 and Hsp70. Both chaperones have the ability to hydrolyse ATP molecules, and the turnover of ATP causes release of bound preprotein. The ATP-hydrolysing activity of Hsp90, unlike that of Hsp70, is inhibited by the antibiotic geldanamycin, providing a means of testing the specific involvement of Hsp90 in protein transport. Young et al. find that geldanamycin inhibits the import of the phosphate-carrier protein both into isolated rat liver mitochondria and into mitochondria of a mammalian cell line. As might be expected, it does not affect the import of the adenine-nucleotide carrier into isolated yeast mitochondria. Moreover, the carboxy-terminal domain of another cochaperone, Bag-1, binds to the ATP-hydrolysing site of Hsp70 but not to that of Hsp90. Surplus Bag-1 carboxy-terminal domain inhibits the import of the two carrier preproteins into mitochondria from either rat liver or yeast by interfering with the turnover of ATP by Hsp70.

These observations suggest that Tom70 is not only a preprotein receptor, but also a cochaperone that aids chaperones in the targeting of carrier preproteins to mitochondria (Fig. 1). This dual function of Tom70 can be rationalized in terms of the need to protect hydrophobic membrane proteins from aggregation as they pass from ribosomes to mitochondria — a hazardous journey for which such proteins need the continued protection of chaperones. Cytosolic Hsp70 was known to chaperone the folding of many newly synthesized proteins7, whether they are destined for mitochondria or remain in the cytosol, but Hsp90 was thought to be largely restricted to the folding and assembly of signal-transduction proteins in the cytosol8. Why mammalian mitochondria require Hsp90 for carrier-protein import whereas yeast mitochondria do not is just one of the questions raised by these observations1. The study of molecular chaperones continues to spring surprises.


  1. 1

    Young, J. C., Hoogenraad, N. J. & Hartl, F. U. Cell 112, 41–50 (2003).

    CAS  Article  Google Scholar 

  2. 2

    Deshaies, R. J. et al. Nature 332, 800–805 (1988).

    ADS  CAS  Article  Google Scholar 

  3. 3

    Murakami, H., Pain, D. & Blobel, G. J. Cell Biol. 107, 2051–2057 (1988).

    CAS  Article  Google Scholar 

  4. 4

    Ellis, R. J. Biochem. Biophys. Res. Commun. 238, 687–692 (1997).

    CAS  Article  Google Scholar 

  5. 5

    Pfanner, N. & Geissler, A. Nature Rev. Mol. Cell Biol. 2, 339–349 (2001).

    CAS  Article  Google Scholar 

  6. 6

    Brix, J., Dietmeier, K. & Pfanner, N. J. Biol. Chem. 272, 20730–20735 (1997).

    CAS  Article  Google Scholar 

  7. 7

    Frydman, J. Annu. Rev. Biochem. 70, 603–649 (2001).

    CAS  Article  Google Scholar 

  8. 8

    Young, J. C., Moarefi, I. & Hartl, F. U. J. Cell Biol. 154, 267–273 (2001).

    CAS  Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to R. John Ellis.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Ellis, R. Plugging the transport gap. Nature 421, 801–802 (2003).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.