Research ArticleBIOCHEMISTRY

The force-dependent mechanism of DnaK-mediated mechanical folding

See allHide authors and affiliations

Science Advances  09 Feb 2018:
Vol. 4, no. 2, eaaq0243
DOI: 10.1126/sciadv.aaq0243
  • Fig. 1 DnaJ binds to ubiquitin and blocks refolding.

    (A) Scheme of the distinct protein conformations visited by ubiquitin [Protein Data Bank (PDB): 1UBQ] when exposed to mechanical force, namely, native, collapsed, and extended (22). (B) Schematics of the single-molecule nanomechanical experiment, whereby a ubiquitin polyprotein (Ubi)9 is tethered between a gold substrate and an atomic force microscopy cantilever tip. (C) A force-quench protocol demonstrates that ubiquitin can quantitatively refold in tq = 5 s, marked by the presence of 20-nm steps in the test pulse. (D) By contrast, upon addition of 5 μM DnaJ, the refolding efficiency is significantly decreased. (E) The kinetics of ubiquitin refolding is measured by changing tq, spanning the range 0.5 to 15 s (blue symbols; tq = 0.5 s, n = 37 individual trajectories; tq = 1 s, n = 56; tq = 2 s, n = 36; tq = 5s, n = 46; tq = 10 s, n = 44; tq = 15 s, n = 26). A single exponential fit to the refolding kinetics of wild-type ubiquitin displays a folding rate kf = 0.52 ± 0.05 s−1. In the presence of DnaJ (green symbols), the refolding rate is decreased down to kf = 0.29 ± 0.07 s−1 (tq = 0.5 s, n = 36 trajectories; tq = 1 s, n = 30; tq = 2 s, n = 39; tq = 5 s, n = 37; tq = 10 s, n = 35; tq = 15 s, n = 52). Such a reduced refolding rate can be explained in terms of the delayed ubiquitin collapse in the presence of DnaJ (fig. S1). (F) Whereas ubiquitin reaches a high refolding yield (~65%), the addition of DnaJ (green symbols) significantly (P < 0.0001) reduces the yield down to ~30% at a quench time of tq = 5 s. (G) The unfolding kinetics of ubiquitin at 120 pN in the absence and presence of 5 μM DnaJ does not result in a change in the unfolding rate [ku = 0.93 ± 0.06 s−1 (n = 195 unfolding trajectories) and ku = 0.94 ± 0.07 s−1 (n = 138 unfolding trajectories), respectively], suggesting that DnaJ does not bind to the native state of ubiquitin. (H) These results are confirmed with pull-down assays, demonstrating that DnaJ and the folded ubiquitin do not coprecipitate (P, pellet; S, supernatant).

  • Fig. 2 DnaJ binds to the unfolded and extended state of ubiquitin.

    (A) Changing the time of the initial pulse in the presence of DnaJ enables binding to the unfolded ubiquitin. After a short extended time text = 1 s, ubiquitin shows a high refolding efficiency, fingerprinted by a large number of 20-nm steps in the test pulse. (B) Upon increasing text = 15 s, the refolding efficiency is markedly decreased. (C) Evolution of the refolding efficiency as a function of text for a constant tq = 5 s. In the presence of 5 μM DnaJ, the folding efficiency decreases exponentially with the text with an associated rate k = 0.27 ± 0.05 s−1 (green symbols). The number of individual trajectories used was text = 1 s, n = 33; text = 2 s, n = 43; text = 3 s, n = 57; text = 5 s, n = 84; text = 10 s, n = 52; text = 15 s, n = 45; text = 30 s, n = 21. Decreasing the concentration of DnaJ down to 1 μM increases the refolding percentage for a constant text = 15 s (n = 63, light green symbol), demonstrating that DnaJ binding to ubiquitin is a concentration-dependent mechanism with an associated association constant, kon. In the absence of DnaJ, the refolding of ubiquitin is largely independent of text (blue symbols; text = 1 s, n = 26; text = 5 s, n = 33; text = 15 s, n = 17).

  • Fig. 3 DnaJ binding to the unfolded ubiquitin is force-dependent.

    (A) Changing the force at which ubiquitin is left stretched after mechanical unfolding results in a marked change in the refolding yield. In these experiments, a high force of 170 pN for a short time t = 1 s triggers ubiquitin unfolding. The force is then varied within the range X = 50 to 300 pN for a period t = 15 s before a tq = 5 s pulse is applied to trigger protein refolding. The refolding efficiency is tested in the test pulse, when the protein is stretched back at 120 pN. (B) Dependency of the refolding percentage with the pulling force (F = 30 pN, n = 33; F = 50 pN, n = 30; F = 90 pN, n = 25; F = 120 pN, n = 36; F = 150 pN, n = 28; F = 170 pN, n = 45; F = 200 pN, n = 43; F = 300 pN, n = 22). (C) Snapshot of the interaction of DnaJ (PDB: 1NLT) with the unfolded and mechanically stretched ubiquitin (orange ribbon): The interaction surface involves seven residues (red), exhibiting the consensus sequence GKQLEDG. Inset: Zoom of the reconstructed (see Materials and Methods) DnaJ-ubiquitin fragment system. The bonds of the amino acids from the consensus sequence are shown as red tubes, and the corresponding Cα’s are shown as red balls. (D) Five couples (ϕ, ψ) of backbone dihedral angles can be defined for the consensus sequence to estimate the free-energy contribution corresponding to the overall energetic cost underlying the force-induced remodeling of the dihedral angles of the ubiquitin interaction fragment upon DnaJ binding (see Materials and Methods and the Supplementary Materials for details about the calculation). Error bars correspond to the SDs observed among the five independent runs at each force, and the data are normalized with respect to the lowest free-energy value (observed at 150 pN).

  • Fig. 4 DnaK(ADP) hinders protein folding by binding to the intermediate collapsed states.

    (A) Refolding of ubiquitin is severely compromised in the presence of 5 μM DnaK(ADP). (B) The associated refolding kinetics (orange symbols; tq = 0.5 s, n = 30 individual trajectories; tq = 1 s, n = 46; tq = 2 s, n = 43; tq = 5 s, n = 51; tq = 10 s, n = 27; tq = 15 s, n = 31) plateaus at a low refolding yield of ~30% and cannot be captured by a single exponential. Instead, a calculated function obtained from a multiparametric fit (see Materials and Methods and figs. S11 and S12) reproduces the experimental data with high accuracy (orange fit). (C) The kinetics of ubiquitin unfolding at 120 pN (ku = 0.93 ± 0.06 s−1, n = 195 unfolding trajectories) is not affected upon addition of DnaK (ku = 0.92 ± 0.05 s−1, n = 194 unfolding trajectories), hence suggesting that DnaK does not bind the native state of ubiquitin. (D) Similarly, the refolding yield is not significantly modified with text (text = 2 s, n = 49; text = 5 s, n = 21; text = 15 s, n = 22; text = 30 s, n = 26). Combined, these observations suggest that DnaK binds the intermediate, collapsed conformations.

  • Fig. 5 The complete DnaKJE system enhances ubiquitin refolding.

    (A) Individual folding trajectories corresponding to low tq = 0.5 s reveal that the addition of the complete DnaKJE system improves ubiquitin refolding. (B) Comparison of the refolding kinetics of ubiquitin in the absence (blue symbols) and the presence (violet symbols; tq = 0.5 s, n = 69 individual trajectories; tq = 1 s, n = 32; tq = 2 s, n = 40; tq = 5 s, n = 46; tq = 10 s, n = 17; tq = 15 s, n = 28) of the DnaKJE system. Single exponential fit to the data demonstrates that the rate of refolding kf = 0.52 ± 0.05 s−1 is also increased in the presence of the DnaK system (kf = 0.92 ± 0.1 s−1). (C) As a consequence of this faster folding rate, at short quench times tq = 0.5 to 2 s, the folding increase is more pronounced [P = 0.064 (tq = 0.5 s); P = 0.01163 (tq = 1 s); P = 0.0346 (tq = 2 s)] when compared to longer tq values. (D) Comparison of the refolding percentage for a given tq = 5 s highlights the additive role of each of the molecular components of the DnaKJE system [DnaK(ADP) in orange, n = 51 trajectories; DnaK(ATP) in light brown, n = 40; DnaK(ATP) + GrpE in dark brown, n = 36; and DnaK(ATP) + GrpE + DnaJ in violet, n = 46].

  • Fig. 6 Schematic representation of the effect of the DnaK system on the one-dimensional projection of the folding free-energy landscape under force and its associated folding kinetics.

    (A) One-dimensional representation of the force modulation of the ubiquitin free-energy landscape, highlighting the different conformations (native, collapsed, and extended) (31). While DnaJ binds the unfolded and extended state, preventing protein refolding, DnaK stabilizes the collapsed conformations instead. (B) By contrast, the whole DnaKJE complex catalyzes the collapsed–to–native state transition. (C) Kinetic scheme highlighting the dynamic interaction—hallmarked by the related binding (kon) and dissociation constants (koff)—between each chaperone and the distinct conformation of the ubiquitin substrate (at a force of 170 pN for the extended conformation). The protein-chaperone interaction has direct implications for protein elasticity.

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/4/2/eaaq0243/DC1

    Supplementary Information Methods

    fig. S1. The collapse dynamics of the extended polyubiquitin chain is slowed down upon DnaJ binding.

    fig. S2. Specifically designed mutations in ubiquitin impair protein folding.

    fig. S3. DnaJ blocks I27 refolding by binding to the collapsed (and not the native or extended) states.

    fig. S4. Energetic cost of changing the dihedral angles upon DnaJ binding to the stretched ubiquitin.

    fig. S5. Ramachandran plots for each ubiquitin fragment residue that interacts with DnaJ as a function of the pulling force.

    fig. S6. The refolding kinetics of ubiquitin in the presence of DnaK cannot be captured by a single exponential.

    fig. S7. The ATPase activity of DnaK is not increased upon incubation with ubiquitin.

    fig. S8. DnaK recognizes the collapsed states of I27, blocking refolding.

    fig. S9. Calculation of the binding (kon) and unbinding (koff) rate constants reveals that DnaJ and DnaK associate to different conformations of ubiquitin and I27.

    fig. S10. Calculation of the binding Formula and unbinding Formula constants of DnaJ to the extended conformations of ubiquitin.

    fig. S11. BSA does not affect the kinetics of ubiquitin refolding.

    fig. S12. Independent addition of the different components of the KJE system improves the DnaK-mediated refolding of ubiquitin.

    fig. S13. The complete DnaKJE system significantly enhances the rate and the extent of I27 refolding.

    fig. S14. The DnaK system refolds the folding-inefficient titin (Z1)8 polyprotein.

    fig. S15. Proposed sequences can be recognized by DnaK.

    table S1. Summary of the kinetic parameters obtained after fitting for ubiquitin and I27.

    Reference (49)

  • Supplementary Materials

    This PDF file includes:

    • Supplementary Information Methods
    • fig. S1. The collapse dynamics of the extended polyubiquitin chain is slowed down upon DnaJ binding.
    • fig. S2. Specifically designed mutations in ubiquitin impair protein folding.
    • fig. S3. DnaJ blocks I27 refolding by binding to the collapsed (and not the native or extended) states.
    • fig. S4. Energetic cost of changing the dihedral angles upon DnaJ binding to the stretched ubiquitin.
    • fig. S5. Ramachandran plots for each ubiquitin fragment residue that interacts with DnaJ as a function of the pulling force.
    • fig. S6. The refolding kinetics of ubiquitin in the presence of DnaK cannot be captured by a single exponential.
    • fig. S7. The ATPase activity of DnaK is not increased upon incubation with ubiquitin.
    • fig. S8. DnaK recognizes the collapsed states of I27, blocking refolding.
    • fig. S9. Calculation of the binding (kon) and unbinding (koff) rate constants reveals that DnaJ and DnaK associate to different conformations of ubiquitin and I27.
    • fig. S10. Calculation of the binding (kextendedon) and unbinding (kextendedoff ) constants of DnaJ to the extended conformations of ubiquitin.
    • fig. S11. BSA does not affect the kinetics of ubiquitin refolding.
    • fig. S12. Independent addition of the different components of the KJE system improves the DnaK-mediated refolding of ubiquitin.
    • fig. S13. The complete DnaKJE system significantly enhances the rate and the extent of I27 refolding.
    • fig. S14. The DnaK system refolds the folding-inefficient titin (Z1)8 polyprotein.
    • fig. S15. Proposed sequences can be recognized by DnaK.
    • table S1. Summary of the kinetic parameters obtained after fitting for ubiquitin and I27.
    • Reference (49)

    Download PDF

    Files in this Data Supplement:

Navigate This Article