Fig. 4.

The ¹H_N-Γ₂ PRE profiles recorded on nitroxide spin-labeled ΔST-DNAJB6b. PREs recorded on [²H/¹⁵N]-labeled ΔST-DNAJB6b with the nitroxide spin-label (R1) located at (A) N96C-R1 (60 μM) and (B) G106C-R1 (100 μM). PREs for the JD and CTD domains are shown in red and light blue, respectively. Data for the 2 flexible linkers are not shown for clarity. The ¹H-¹⁵N cross-peaks that are broadened beyond detection are shown as red crosses. The black circles show a control experiment recorded on a sample of [²H/¹⁵N]-labeled ΔST-DNAJB6b in the presence of an equimolar concentration of the nitroxide free radical TEMPO (120 μM) each). No PREs are observed from TEMPO indicating that the interdomain PREs observed for N96C-R1 and G106-R1 are due to transient interdomain interaction and not from hydrophobic interactions between the nitroxide R1 label and the CTD domain. The solid lines are the calculated PREs obtained from Xplor-NIH PRE-driven simulated annealing calculations (see SI Appendix for full details) (36). The nitroxide spin labels are represented by 5-membered ensembles, and their positions and correlation times were optimized by simulated annealing to fit the intradomain PRE data from N96C-R1 and G106C-R1 to helices 1 to 4 of the JD domain. The calculated interdomain PRE data from N96C-R1 and G106-R1 to the CTD domain were obtained by ensemble PRE-driven simulated annealing in which the JD domain (including helix 5 and the nitroxide labels) were held fixed, the CTD domain was allowed to move as a rigid body, and linker 2 (residues 103 to 137) was given torsional degrees of freedom. The CTD and linker 2 were represented by an ensemble size of N = 6 (Fig. 6). The open blue circles for the PREs from N96C-R1 and involving strand β5 were not included in the simulated annealing calculations but back-calculated from the ensembles; the excellent agreement between observed and calculated values for these PREs serves to cross-validate the results.