Nature Structural Biology
9, 406 - 408 (2002)
doi:10.1038/nsb0602-406
A new twist for an Hsp70 chaperoneJoanna F. Swain1
& Lila M. Gierasch21 Joanna F. Swain is in the Department of
Biochemistry Molecular Biology at the University of Massachusetts,
Amherst, Massachusetts 01003, USA.
feltham@nsm.umass.edu 2 Lila M. Gierasch is in the Department of
Biochemistry & Molecular Biology and the Department of Chemistry at the
University of Massachusetts, Amherst, Massachusetts
01003, USA. A new study has demonstrated that the E. coli Hsp70, DnaK,
can catalyze cis-trans isomerization of non-prolyl peptide
bonds.It has long been recognized that isomerization of Xaa-Pro peptide bonds
can cause slow phases in protein folding1,
2, and many helper
proteins that catalyze the isomerization, called peptidyl prolyl isomerases,
have been characterized3. Recent studies have shown that
isomerization of non-prolyl (secondary) peptide bonds can also significantly
slow the refolding of certain proteins4,
5, but relevant
isomerases have not been identified until now. On page
419 of this issue of Nature Structural Biology, Schiene-Fischer
et al.6 provide unexpected evidence that the
Escherichia coli Hsp70 molecular chaperone, DnaK, is capable of
catalyzing secondary peptide bond isomerization, albeit modestly. This is
indeed an intriguing observation; whether DnaK has evolved to perform this
function in the cell remains an open question.
Secondary amide isomerization during folding
Due to the double bond character of the peptide bond, a significant
energy barrier prevents free rotation7,
8,
9 (Fig.
1). Of the two available conformations, the trans state is
energetically favored by  15 kJ mol-1 over the
cis state for secondary peptide bonds because of the close approach of
the two C atoms in the cis form9. For Xaa-Pro
peptide bonds, however, the two conformers are nearly identical in energy8. As a result, in native protein structures, cis Xaa-Pro bonds
are more commonly found than are cis non-prolyl peptide bonds (5.2%
versus 0.03%, respectively)7.
 | | |
Recently, the -amylase inhibitor tendamistat was observed to
display a slow folding phase that was proposed to arise from isomerization of
non-prolyl peptide bonds5. In this case, the slow phase appears
to be contributed by a very small proportion of non-prolyl peptide bonds that
equilibrate to the cis form in the unfolded state and must isomerize to
trans in the native state. Since all peptide bonds are synthesized in
the trans configuration on the ribosome10, one might argue
whether proteins have sufficient time between synthesis and folding (or binding
to chaperones) for significant equilibration to cis. On the other hand,
native protein structures containing a cis peptide bond must isomerize
from the trans form, and catalysis of this process would be
advantageous.
The study by Schiene-Fischer et al.6 demonstrates
that, in addition to its established role in chaperoning nascent protein
chains, the E. coli Hsp70 DnaK has the potential to accelerate
isomerization of non-prolyl peptide bonds. These authors first used an Ala-Ala
dipeptide in an absorbance-based isomerization assay11 to purify
isomerase activity from E. coli extracts, and DnaK was identified by
N-terminal sequencing as the active component. The sequence specificity for the
isomerase activity of DnaK, determined using a panel of Ala-Xaa dipeptides, is
surprisingly different from its binding preferences for unfolded polypeptides.
Whereas DnaK prefers to bind hydrophobic and positively charged residues, such
as Leu, Ile, Val, Tyr, Phe, Arg and Lys12,
13,
14, the isomerase
activity was highest for Met, Ala and Ser, and unobservable for many residues,
including Val, Tyr, Phe and Arg.
Advantageous signal dispersion for the cis isomer allowed the use
of NMR to quantify isomerization rates9 for the Ala-Tyr and
Tyr-Ala peptide bonds in an Ala-Ala-Tyr-Ala-Ala peptide (Box
1). Interestingly, DnaK only catalyzes isomerization of the Ala 2-Tyr 3
peptide bond, and the assisted isomerization is blocked by addition of
substance P, a neurotransmitter peptide that is known to bind in the
peptide-binding pocket of DnaK14. This suggests that the
catalysis site and the peptide-binding pocket are one and the same. Since ATP
binding to DnaK decreases peptide-binding affinity 100-fold, with increases in
both on and off rates15, it is perplexing that DnaK-catalyzed
isomerization is ATP-independent. In order to show that the isomerase activity
of DnaK has relevance for protein folding, the authors followed refolding of an
RNase T1 mutant that contains a native state cis non-prolyl peptide
bond4. DnaK addition caused a dose-dependent increase in the
first order rate constant of folding, which could be reversed by addition of a
peptide (NR) that binds to the peptide-binding site of DnaK16.
Complementary roles for trigger factor and DnaK?
In many ways, it makes physiological sense for DnaK to be a secondary
amide isomerase, based on its functional similarity to trigger factor, a known
peptidyl prolyl isomerase17. Both are present at the ribosome
during protein synthesis and cooperate in the handling of nascent protein
chains18,
19. E. coli can survive loss of either the gene
encoding DnaK or that encoding trigger factor, but deletion of both at once
results in synthetic lethality. Whereas trigger factor tends to bind smaller
proteins and can catalyze isomerization at Xaa-Pro bonds, DnaK preferentially
binds to proteins >30 kDa in size18, and we now learn that it
can speed isomerization of the other peptide bonds. Because large proteins have
many more peptide bonds, one might imagine that they have even more need for a
secondary amide isomerase. However, trigger factor is a much better catalyst.
In the native state of RNase T1, the Tyr 38-Pro 39 peptide bond is cis;
trigger factor (0.5  M) enhances the folding rate of RNase T1 45-fold17. Mutation of Pro 39 to Ala results in retention of the cis
peptide bond (now Tyr-Ala) in the native state4, but an even
higher amount of DnaK (2 M) can only accelerate folding of this mutant by a
factor of 2.5.
Possible mechanisms of catalysis
Catalysis of peptide bond rotation could exploit transition state
stabilization by physical distortion of the peptide group or by selection for
the uncharged resonance form that rotates more freely (Fig.
1). The latter effect could arise either by destabilization of the
charge-separated state via appropriate placement of charged side chains
or by stabilization of the non-charged state with a hydrophobic binding pocket.
Inspection of the crystal structure of the peptide-binding domain of DnaK bound
to a seven-residue peptide16 (NR, Asn-Arg-Leu-Leu-Leu-Thr-Gly,
Fig. 2) demonstrates that only the middle five residues
of the peptide interact with the chaperone, which limits the site of catalysis
to one of four possible peptide bonds. All peptide bonds of NR are
trans, and all are within 2° of 180° except for that between Leu
3 and Leu 4, which is 175°. While this is not a large deviation, refinement
protocols used to generate protein structures typically favor a planar peptide
bond, so a larger deviation may have been obscured during refinement. If
catalysis occurs by distortion, it would likely be a small effect and confined
to the peptide bond between protein sites -1 and 0, which in the case of
the NR peptide corresponds to the Leu 3-Leu 4 bond (Fig.
2). While there are no charges close to the bound peptide, the central
peptide-binding pocket is highly hydrophobic, especially sites -1 and 0.
Thus, by these two criteria, the peptide bond between protein sites -1
and 0 would be the most likely site of catalysis.
| | |
Remaining conundrums
The physiological significance of the reported secondary amide isomerase
activity of DnaK remains to be established. The fact that DnaK binds to nascent
chains immediately after synthesis, when most peptide bonds have not had a
chance to assume the cis isomer, as well as the fact that DnaK binds to
discrete sites in each protein sequence13, suggests that DnaK
probably does not facilitate isomerization of the small population of
cis at many sites. Thus the most likely use of this isomerase activity
would be to assist the folding of proteins that have one or more cis
peptide bonds in the native state. Of the 34 proteins containing non-prolyl
cis peptide bonds in a non-redundant set of protein structures from the
Protein Data Bank, 20 are of bacterial origin7 and could be
substrates for the isomerase activity of DnaK. The high homology of
substrate-binding pockets in different Hsp70 molecules20 begs the
question of whether this activity is conserved throughout the Hsp70 family and
might play a role in eukaryotic protein folding as well. Nonetheless, the
modest degree of catalysis by DnaK leaves open the possibility that its
isomerase activity is simply an adventitious byproduct of its principal
function: binding and stabilizing unfolded polypeptide chains.
REFERENCES
- Brandts, J.F., Halvorson, H.R. & Brennan, M. Biochemistry 14, 4953−4963 (1975). | Article | PubMed | ISI | ChemPort |
- Schmid, F.X. & Baldwin, R.L. Proc. Natl. Acad. Sci. USA 75, 4764−4768 (1978). | PubMed | ChemPort |
- Schiene, C. & Fischer, G. Curr. Opin. Struct. Biol. 10, 40−45 (2000). | Article | PubMed | ISI | ChemPort |
- Odefey, C., Mayr, L.M. & Schmid, F.X. J. Mol. Biol. 245, 69−78 (1995). | Article | PubMed | ISI | ChemPort |
- Pappenberger, G. et al. Nature Struct. Biol. 8, 452−458 (2001). | Article | PubMed | ISI | ChemPort |
- Schiene-Fischer, C., Habazettl, J., Schmid, F.X. & Fischer, G. Nature Struct. Biol. 9, 419−424 (2002). | Article | PubMed | ISI | ChemPort |
- Jabs, A., Weiss, M.S. & Hilgenfeld, R. J. Mol. Biol. 286, 291−304 (1999). | Article | PubMed | ISI | ChemPort |
- Stein, R.L. Adv. Protein Chem. 44, 1−24 (1993). | PubMed | ISI | ChemPort |
- Scherer, G., Kramer, M.L., Schutkowski, M., Reimer, U. & Fischer, G. J. Am. Chem. Soc. 120, 5568−5574 (1998). | Article | ISI | ChemPort |
- Lim, V.I. & Spirin, A.S. J. Mol. Biol. 188, 565−574 (1986). | PubMed | ISI | ChemPort |
- Schiene-Fischer, C. & Fischer, G. J. Am. Chem. Soc. 123, 6227−6231 (2001). | Article | PubMed | ISI | ChemPort |
- Gragerov, A. & Gottesman, M.E. J. Mol. Biol. 241, 133−135 (1994). | Article | PubMed | ISI | ChemPort |
- Rüdiger, S., Germeroth, L., Schneider-Mergener, J. & Bukau, B. EMBO J. 16, 1501−1507 (1997). | PubMed |
- de Crouy-Chanel, A., Kohiyama, M. & Richarme, G. J. Biol. Chem. 271, 15486−15490 (1996). | Article | PubMed | ChemPort |
- Pierpaoli, E.V., Gisler, S.M. & Christen, P. Biochemistry 37, 16741−16748 (1998). | Article | PubMed | ISI | ChemPort |
- Zhu, X. et al. Science 272, 1606−1614 (1996). | PubMed | ISI | ChemPort |
- Stoller, G. et al. EMBO J. 14, 4939−4948 (1995). | PubMed | ISI | ChemPort |
- Teter, S.A. et al. Cell 97, 755−765 (1999). | Article | PubMed | ISI | ChemPort |
- Deuerling, E., Schulze-Specking, A., Tomoyasu, T., Mogk, A. & Bukau, B. Nature 400, 693−696 (1999). | Article | PubMed | ISI | ChemPort |
- Rüdiger, S., Buchberger, A. & Bukau, B. Nature Struct. Biol. 4, 342−349 (1997). | Article | PubMed | ISI | ChemPort |
- Kraulis, P. J. Appl. Crystallogr. 24, 946−950 (1991). | Article | ISI |
- Merritt, E.A. & Bacon, D.J. Methods Enzymol. 277, 505−524 (1997). | Article | ISI | ChemPort |
|
