A.1 ATPase activity of Hsp90 point mutants
A detailed characterization is done for the studied protein variants depicted in Fig.A.1.
This makes clear that rather the labeled protein determines the ATPase activity of each studied Hsp90 variant than the unlabeled cysteine mutant.
Figure A.1: The ATPase activity of the depicted constructs compared to the wild-type (WT) activity. The point mutation 51C exhibits a similar activity, but loses more than 70% of activity upon dye attachment (Atto550). The position 431C gains activity due to labeling. The ATPase activity of 517C does not significantly change upon the addition of dyes (Atto550 or Atto647N).
The exchange (a 1:1 mix, which is heated 30 minutes at 47∘C) does not significantly affect the ATPase activity of the depicted variants. ATPase activity is measured with a regenerating assay with 1𝜇M protein under saturating ATP concentration (2 mM) at 37∘C. The error bars represent the standard deviations from three replicates.
A broad variation of activities was observed (summarized in Table A.1), but as in this project we study the structure of Hsp90 and not the kinetics of conformational transitions (which are suspected to be linked somehow to the ATPase activity, even though not directly coupled), we only excluded Hsp90 variants from our analysis that do not exhibit ATPase activity at all (which is e.g. the labeled S51C mutant). Other mutants that do not exhibit ATPase activity at all are labeled at the amino acid positions 141, 109, 179, 466.
Interestingly, we find a correlation of complete loss of ATPase activity and incomplete
107
closing after addition of AMP-PNP (as monitored by smFRET). Thus, this is checked for all constructs as well.
A.1 ATPase activity of Hsp90 point mutants 109
Table A.1: The ATPase activities of Hsp90 point mutants. The following single cysteine point mutations labeled with the maleiimide derivative of the depicted label are tested at 30 ∘C, 2 mM ATP and 1 𝜇M protein. Two numbers indicate double cysteine mutants, ex refers to exchanged samples that are mixed in a 1:1 ratio and incubated for 30 minutes at 47∘C.
Cysteine mutation Label ATPase activity (monomer−1min−1)
9 - 1.06
Ex 61 & 431 Atto550 & Atto647N 0.76
Ex 61/452 & WT Atto550/Atto647N & - 0.37
A.2 Combined anisotropy threshold for open Hsp90
The RMSD between the optimal structure arrangement for the open conformation of Hsp90 and the smFRET derived distances depends on the combined anisotropy 𝑟𝑐, similar to what we find for the closed state of yHsp90. We apply a threshold on 𝑟𝑐 at 0.21, because lower thresholds do not result in an improved RMSD. The remaining 81 distances have an RMSD of about 4 Å.
Figure A.2: RMSD between optimal structure arrangement for the open state of Hsp90 and the𝑟𝑐 threshold applied on the set of distances.
A.3 Molecular dynamics simulations
Molecular dynamic (MD) simulations on the structure of Hsp90 are conducted by Florian Kandzia and Martin Zacharias, TU München, Munich. Their methods and implementation are described in the following. The text is retrieved from the joint publication on the dynamic structure of open Hsp90 [3].
Restrained MD simulations
MD simulations and energy minimization to generate Hsp90 structural models are performed using the Sander module of the Amber14 molecular simulation package [184]. During structural modeling the core structures of the single domains are restrained to experimental reference structures by adding distance restraints between C𝛼 atoms within 15 Å(force constant 3 kcal mol−1Å−2). This allows translational and rotational mobility but limits internal flexibility in each domain. Additionally, the core structure of the C-domains is kept close to the geometry of the X-ray structure owing to the corresponding experimental distances being equal in the presence of ADP and AMP-PNP. For the structure generation steps, an implicit solvent model with a distance dependent dielectric constant (𝜀= 4r) is employed.
As start structure the -2𝜎 domain arrangement is used. The connection between the NTD and MD is modeled by a short Gly10 linker inserted between residues 216 and 261.
For inclusion of the FRET-derived distance data, pseudoatoms with a van-der-Waals radius of 6.5 Å, representing approximately the size of the dyes, are connected to the C𝛽 atoms of the dye-labeled residues using a pseudobond (linker) of 11 Å. The distance between each donor–acceptor pseudoatom pair is allowed to vary within an interval given
A.3 Molecular dynamics simulations 111
by the experimentally determined FRET-derived distance range. Beyond the allowed distance interval, further variation is penalized with a force constant of 3 kcal mol−1Å−2. In addition, the optimal spatial NTD/MD-arrangement in the dimer serves as additional positional restraints (on heavy atoms, with a force constant 1 kcal mol−1Å−2) during MD based structure generation. To keep the structure symmetric with respect to the spatial arrangement of each monomer, the option to penalize the difference between two distances implemented in the sander module of the Amber14 package is employed. The option is applied to a subset of equivalent backbone distances within each monomer and between monomers. MD simulations are performed for 0.5 ns at 450 K followed by 0.5 ns at 300 K using a time step of 0.001 ps. Structures are energy minimized within 7,500 steps of conjugate gradient minimization. Structural models of the open Hsp90 structure are re-evaluated by the FPS to check compatibility with the FRET data. The final mean structure yield an average deviation from the FRET distance data of 𝜒2 < 0.3.
Unrestrained MD simulations
In addition to restrained MD simulations in implicit solvent for structure generation, unrestrained simulations starting from the mean open ADP-bound Hsp90 structure are performed in the presence of explicit solvent and surrounding ions. The structure is solvated in a truncated octahedral box with explicit TIP3P water molecules [185] and neutralized with sodium and chloride ions by means of the leap module and employing the parm14SB force field [186]. Long-range electrostatic interactions are calculated with the particle mesh Ewald (PME) method and a real space cutoff radius of 9 Å[187]. During 0.5 ns equilibration the system is heated up to 300 K while heavy atoms of the protein are harmonically restrained (25 kcal mol−1Å−2) to positions in the starting structure.
Positional restraints are gradually removed during another 0.15 ns. Then the structures are equilibrated for 0.2 ns in an unrestrained simulation. The simulation is extended to 100 ns at 300 K and a pressure of 1 bar. For comparison, explicit solvent MD simulations are started from the closed crystal structure of Hsp90 (PDB 2CG9) after removing the co-chaperone Sba1 and supplementing with the same Gly10 linker between N- and M-domain as used for modeling the open Hsp90 structures. Equilibration follows the same protocol, and simulations are extended to 100 ns. Several thousand structures in the form of simulation snapshots are evaluated by the FPS, resulting in an average 𝜒2 of 1.5. The generated trajectories are evaluated using the cpptraj module of the Amber14 package.
Calculations of the RMSD of the backbone and RMSF are performed using the cpptraj module. The RMSF is calculated as the mean over all heavy atoms of each residue. For a subset of 200 snapshots of the open and closed simulations, the buried surface area (BSA) of the CTD/MD interface are calculated. Therefore, the solvent-accessible surface area of the CTD, MD and both domains is calculated atom-wise using the SHRAKE algorithm [188]. This is done for both monomers of the Hsp90 dimer, and the arithmetic mean is used in the following. For each residue the normalized total contribution as well as the number of contributions to the BSA is determined. The same procedure is repeated for the NM interface.