Stability Effect of Quinary Interactions Reversed by Single Point

Subscriber access provided by Macquarie University

Article

Stability Effect of Quinary Interactions Reversed by Single Point Mutations David Gnutt, Stepan Timr, Jonas Ahlers, Benedikt König, Emily Manderfeld, Matthias Heyden, Fabio Sterpone, and Simon Ebbinghaus J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.8b13025 • Publication Date (Web): 11 Feb 2019 Downloaded from http://pubs.acs.org on February 11, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society

Stability Effect of Quinary Interactions Reversed by Single Point Mutations David Gnutt†,‡, Stepan Timr#, Jonas Ahlers‡, Benedikt König‡, Emily Manderfeld‡, Matthias Heyden§, Fabio Sterpone#, and Simon Ebbinghaus†,‡* † Institute

of Physical and Theoretical Chemistry, Technical University Braunschweig, Rebenring 56,

38106 Braunschweig, Germany ‡

Department of Physical Chemistry II, Ruhr University Bochum, Universitätsstrasse 150, 44801 Bochum,

Germany §

School of Molecular Sciences, Arizona State University, 551 E. University Dr., Tempe, AZ 85281, USA

#

CNRS Laboratoire de Biochimie Théorique, Institut de Biologie Physico-Chimique, Université Paris

Denis Diderot, Sorbonne Paris Cité, PSL Research University, 13 rue Pierre et Marie Curie, 75005, Paris, France *E-mail: [email protected] Abstract In cells, proteins are embedded in a crowded environment that controls their properties via manifold avenues including weak protein-macromolecule interactions. A molecular level understanding of these quinary interactions and their contribution to protein stability, function and localization in the cell is central to modern structural biology. Using a mutational analysis to quantify the energetic contributions of single amino acids to the stability of the ALS related protein superoxide dismutase I (SOD1) in mammalian cells, we show that quinary interactions destabilize SOD1 by a similar energetic offset for most of the mutants, but there are notable exceptions: Mutants that alter its surface properties can even lead to a stabilization of the protein in the cell compared to the test tube. In conclusion, quinary interactions can amplify and even reverse the mutational response of proteins, being a key aspect in pathogenic protein misfolding and aggregation. Introduction The ability of cells to maintain their macromolecules folded and functional under different conditions is a remarkable aspect of biology. However, most macromolecules only show marginal stability inside cells to allow for conformational transitions that often accompany function such as enzyme domain movement upon catalysis,1 RNA hairpins as transcription initiators2,3 or chromatin packing upon mitosis.4 All these processes occur in the densely crowded and highly organized environment of living cells.5–7 Most of our

1 ACS Paragon Plus Environment

Journal of the American Chemical Society 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 22

knowledge of these processes is derived from in vitro studies (e.g. NMR, crystallography, spectroscopy) or simulations that are conducted in aqueous solution.8 However, in certain cases the local cellular environment of a biomolecule changes its structure and function.7 For proteins, such interactions are often named quinary interactions,9 referring to the fifth level of organization of a protein embedded in a protein-protein network. They play an important role in forming various multi-enzyme complexes for example in metabolic pathways9–11 or macromolecular assemblies such as membraneless organelles.12 Experimental techniques that are applicable to conduct in-cell measurements include fluorescence based techniques, nuclear magnetic resonance (NMR) spectroscopy and mass spectrometry based methods.13–15 Such experiments showed that quinary interactions, compared to tertiary and quaternary levels of structural organization, are transient and weak (~1 kcal mol-1).11 Here, we study how quinary interactions alter the in-cell folding stability of mutated proteins. We measure the stability (in reference to dilute solution) of a superoxide dismutase I (SOD1) folding reporter by Fast Relaxation Imaging (FReI)16 and complement these measurements with microscopic insights from coarse-grained and atomistic molecular dynamics (MD) simulations.17,18 The folding reporter is based on a truncated variant of superoxide dismutase I (SOD1),19 a key protein in familial amyotrophic lateral sclerosis (fALS), a severe motor neuron disease. Many disease-related mutations affect either the stability of the wildtype homodimeric enzyme or the stability of the apo state fold.20 In this study, we use SOD1bar, a monomeric variant that maintains the folding characteristics of the apo state, featuring a reversible two-state unfolding transition.19 We show that most disease-related mutations are destabilized by quinary interactions, but distinct point mutations can reverse or significantly enhance this effect. Materials and Methods Materials. All reagents and chemicals were purchased from Sigma-Aldrich if not otherwise specified. Cell culture and transfection. HeLa cells were grown as adherent culture in T25 flasks (Sarstedt) in DMEM supplemented with 10% FBS, 100 U/mL penicillin and 0.1 mg/mL streptomycin (all Sigma) in a humidified atmosphere (37°C, 5% CO2). They were passaged every 2-3 d and subcultured in a 1:4 to 1:6 ratio using trypsin digestion. Prior to the experiment, cells were seeded on 6 well plates (Sarstedt). They were transfected at 80-90 % confluence using Lipofectamine 3000 (Thermo Fisher) reagent according the manufacturer’s protocol. 2 µg plasmid DNA were mixed with 4 µL P3000 reagent in 125 µL Opti-MEM (Thermo Fisher). The mixture was transferred to and mixed with 4 µL Lipofectamine 3000 reagent in 125 µL Opti-MEM. Cell culture medium was supplemented to medium with antibiotics before transfection. The full transfection mixture was incubated with the cells for 6 hours. Cells were then passaged using trypsin


Page 3 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


digestion and seeded on 35 mm glass bottom dishes (WPI), where they were incubated for 2 d using the DMEM/10%FBS/1%P/S at 37°C and 5% CO2 before the experiment. Plasmid preparation. SOD1 barrel DNA in pET3a was a kind gift from Mikael Oliveberg (Stockholm University). It was subcloned into a modified pDream2.1 vector with a C-terminal mCherry and an N-terminal AcGFP1, used previously. For cloning, a PCR based approach (In-Fusion, Clontech) was used. Point mutations were introduced using site-directed mutagenesis (Quik Change Lightning, Agilent). Sequence identity was verified via an in-house DNA sequencing service (Ruhr-Universität Bochum, Department of Biochemistry 1). Plasmid DNA was amplified via transformation into XL10 gold (Agilent), Stellar (Clontech) or NEB5a (New England Biolabs) and purified using the Zyppy Miniprep for small scale plasmid preparations (Zymo) or Zymo Pure Midiprep for large scale plasmid preparation (Zymo). DNA concentration was determined using UV/Vis spectroscopy (NanoDrop 2000, Thermo). Protein purification. For protein purification, plasmid DNA was transformed into NiCo21 bacteria. A single colony was grown to OD 0.6 at 37°C and 220 rpm before induction with 500 mM IPTG. Expression was allowed for 18 hours at 18-19°C. Cells were harvested using centrifugation and lysed using Clontech Xtractor buffer. Proteins were affinity purified using gravity flow His60 columns (Clontech) according to the manufacturer’s protocol. After purification, buffer was exchanged using Amicon Ultra 30 kDa (Merck) to 100 mM NaCl and 50 mM phosphate adjusted to pH 7.4 (PBS). Proteins were aliquoted and used directly or shock-frozen in liquid nitrogen for prolonged storage. Sample preparation for FReI. For measurements in vitro, purified protein was diluted to 10 µM with PBS and cosolutes (e.g. BSA or ATP (complexed with equimolar amount of MgCl2)) were added at the desired concentrations. 20 µL of each sample were placed between a clean glass cover slip (Menzel No 1.0) and a glass bottom dish (FluoroDish, WPI) separated by a 120 µm thick imaging spacer (Secure Seal, Sigma). For in-cell measurements, cell growth medium was aseptically removed and the cell coated glass bottom dishes were placed on a glass cover slip with a 120 µm thick imaging spacer covered with 30 µL Leibovitz’s L15 medium supplemented with 30% FBS. FReI measurements and data evaluation. Fast Relaxation Imaging (FReI) is a technique combining rapid temperature jumps with fluorescence microscopy, as described previously.16 The utilized experimental set-up was described in detail before.2 Briefly, an IR diode laser is used to incrementally heat the sample by 2.3 °C each 25 s from 23 °C to ~60 °C (16 to 18 temperature jumps for each measurement) to record unfolding kinetics and equilibrium data at each temperature. Temperature jumps were calibrated using the temperature sensitive dye rhodamine B as described earlier.2,21 A sample rhodamine B intensity profile is shown in Figure S1. The temperature jumps did not affect cell viability which was assessed based on morphological criteria (intact cell membrane, nuclear and cytoplasm integrity, full attachment to the dish) as used elsewhere22–24 (Fig. S2). Furthermore, no leakage of the protein from the cell or nucleus was



Page 4 of 22

observed (Fig. S2C,D). AcGFP1 fluorescence was excited by using 470 nm LED light at constant exposure for each sample. Emission of the AcGFP1-donor (D) (497 - 527 nm) and mCherry-acceptor (A) (581 - 679 nm) was separated using a dichroic mirror und imaged simultaneously by two CCD cameras at a frame rate of 5 frames per second. The data was evaluated using ImageJ (US NIH) and self-written Matlab code. The intensities were extracted for every cell separately by first using intensity thresholding of D intensities to separate the cytosol from the nuclear region (Fig S2C). For each ROI (region of interest) and at each time point, donor and acceptor intensities were then averaged throughout the ROI and background subtracted for each channel separately. Only cytosolic intensities were used for further analysis to prevent any skewing of the data by nuclear regions, e.g. diffusion of proteins into and out of the nucleus.

For data analysis, the thermodynamics from kinetics method was used as originally developed by the Gruebele and Chemla labs25. The method improves the measurement of protein stability from denaturation experiments by analyzing the corresponding folding kinetic amplitudes. The idea is to use kinetic information for the separation of timescales that give rise to either the population change between the native and unfolded states (desired) or population changes within either of the states (giving rise to folded and unfolded state baselines). The kinetic separation of the timescales allows to measure the population shift essentially without baselines. Compared to fitting of a two-state model to a sigmoidal melting curve, the described method yields more reliable results especially if the unfolded or folded baselines are not fully resolved25. Briefly, for each temperature jump, the 𝐷 ― 𝛼𝐴 values were calculated26 and fitted using a single exponential function. The corresponding amplitudes 𝐷 ― 𝛼𝐴 (T) were plotted against the temperature.25 The resulting data was fitted to the equation derived by Gruebele and Chemla25 using a two-state approximation (see Results and Discussion for justification) yielding g1 and Tm,25

―𝛿𝑔1Δ𝑇 ⋅ 𝑇𝑚

𝑒

𝐷 ― 𝛼𝐴 (𝑇) = 𝑅(𝑇 ― Δ𝑇/2)2 ⋅ (𝐴0 + 𝑚𝑎 ⋅ (𝑇 ― 𝑇𝑚)) ⋅

(

)((

(

Δ𝑇 Δ𝑇 ―𝛿𝑔1 𝑇 ― ― 𝑇𝑚 ⋅ 𝑅 𝑇 ― 2 2

1+𝑒

(

)((

))

Δ𝑇 Δ𝑇 ―𝛿𝑔1 𝑇 ― ― 𝑇𝑚 ⋅ 𝑅 𝑇 ― 2 2

―1

))

―1 2

(1)

)

where T is the amplitude of the temperature jump and A0 and mA fitting parameters for the linear baseline (mA was set to 0). The standard free energy of folding at 37°C (Gf0’) was calculated by a linear Taylor approximation of a two-state population change as described earlier26. To test for statistical difference between conditions, Gf0’ of individual measurements was averaged and analysis performed using Graphpad Prism 6. Details regarding samples sizes and statistical test performed are denoted in each figure legend.


Page 5 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


LBMD simulation. The Lattice-Boltzmann Molecular Dynamics (LBMD) simulation17 was performed using MUPHY,27 a code for multiphysics simulation. We considered a system of 5 SOD1 molecules and 15 BSA molecules placed in a cubic box of dimensions 20.1 x 20.1 x 20.1 nm3, corresponding to a BSA concentration of 200 g/L. Proteins structures obtained from the PDB database28 (PDB ID 4BCZ29 for SOD1 and PDB ID 4F5S30 for BSA) were converted into a coarse-grain resolution and parameterized with the OPEP force field31 combined with an elastic network model (distance cutoff 0.6 nm, force constant 2092 kJ·mol-1·nm-2). The friction coefficient, controlling the coupling between the protein particles and the fluid, was set to  = 0.004 fs-1 and  = 0.001 fs-1 for SOD1 and BSA, respectively, i.e., to values reproducing the respective HYDROPRO32 diffusion coefficients in dilute conditions. The resolution of the lattice was 0.3 nm, and the kinematic viscosity of the fluid was set to 0 = 1.66·10-3 nm2·fs-1, a value corresponding to liquid water. The system was propagated for 1 s with a time step of 10 fs at T = 300 K. Back-mapping to the all-atom resolution. Atomistic protein structures were aligned to coarse-grain geometries from the LBMD simulation by minimizing the root-mean square deviation (RMSD) between the corresponding C atoms. In the case of BSA, a crystal structure (PDB ID 4F5S30) was used for the alignment. The atomistic geometry of SOD1 was selected from an ensemble of relaxed conformations as the one exhibiting the fewest clashes with BSA after the alignment. The relaxed SOD1 conformations were obtained from a 200 ns MD simulation performed in dilute conditions using the ff99sb/TIP3P force fields33,34 in the NAMD 2.10 software.35 A hydration layer of water molecules with distances below 0.35 nm from SOD1 was also transferred to the back-mapped system unless an overlap with BSA atoms occurred. Hydrogen atoms were added to the BSA structures using the Gromacs pdb2gmx tool,36 and the simulation box was filled with water molecules, including potassium cations to neutralize the net negative charge of the system. In total, the simulation box contained 26748 water molecules and 19 K+ cations for the state 1 and 31002 water molecules and 35 K+ cations for the state 5. All-atom MD and REST2 simulations of crowded systems. All-atomistic MD and REST218,37 simulations were performed in the GROMACS 5.1.4 software36 patched with the Plumed 2.3.1 package.38 A cutoff of 1.0 nm was applied to short-range electrostatic interactions while long-range electrostatics was calculated with the use of the particle mesh Ewald method.39 Van der Waals interactions were truncated at 1.0 nm. All protein bonds were constrained by the LINCS algorithm,40 and water molecules were kept rigid by the SETTLE algorithm.41 The proteins were parameterized with the ff99SB*-ILDN force field,33,42,43 the water molecules were described with the TIP3P model,34 and the K+ cations were parameterized with the default parameters available in the ff99SB*-ILDN force field. All simulations were run at a constant temperature of 300 K, maintained by the velocity rescaling thermostat with a stochastic term44 using a time constant of 1 ps. The thermostat was coupled separately to the proteins and to the rest of the system. Unless stated otherwise, the pressure was kept constant at 1.01 bar by employing the Parrinello-Rahman barostat45



Page 6 of 22

with a time constant of 1 ps, and Newton’s equations of motion were integrated using the leap-frog algorithm46 with a time step of 2 fs. The atomistic systems obtained by the back-mapping procedure described above were energy minimized and subsequently equilibrated by performing a series of six equilibration steps (see Table S1 for more details on the equilibration procedure). The first two short simulations were run in the NVT ensemble, while the following four steps were NpT trajectories simulated at a pressure of 1.01 bar. Harmonic restraints were imposed on all heavy atoms of the proteins. Different force constants, gradually decreasing to zero, were applied to restrain the backbone and the side chains of SOD1 and BSA (Table S1). The time constants of the velocity rescaling thermostat44 and the Berendsen barostat47 were set to 1 ps. The six equilibration steps were followed by a 300 ns NpT unrestrained MD simulation. The final structures obtained from the unrestrained MD simulations each served as a starting point for a 500 ns REST2 simulation18,37 with 24 replicas of the system simulated at 300 K. The REST2 scheme was previously used with success to investigate protein conformations and thermal stabilities in both dilute48 and crowded49 conditions. In the present simulations, the SOD1 barrel formed the solute, i.e., its potential energy terms were rescaled in line with the REST2 approach to enhance the sampling of its conformations and its interactions with BSA. The solute temperatures ranged from 289 K to 652 K (Table S2), and exchanges of replicas were attempted every 5 ps. The temperature spacing was chosen so as to ensure sufficiently high average transition probabilities (> 0.2) for neighboring pairs of solute temperatures. Results were analyzed for the solute temperature T = 300 K, sampling the statistical ensemble of interest. In addition, a 500 ns REST2 simulation was performed for SOD1 in dilute conditions; this simulation was started from the final geometry of the 200 ns MD trajectory mentioned in the paragraph “Back-mapping to the all-atom resolution”. Two representative snapshots of the system were selected from the REST2 simulation of the state 5, and for each of these two geometries, the H46S and H46R mutations were introduced into the SOD1 structure using the PyMOL software.50 After an energy minimization, 500 ns direct MD trajectories were run for each these mutated geometries. The trajectories were compared with an MD trajectory which was started from the same respective geometry before the mutation. All-atom MD simulations of individual protein mutants. Based on the crystal structure of the engineered SOD1bar protein (PDB ID 4BCZ), we generated the point mutations H46S, H46R, A4V and G41D via sidechain replacement. The wt and mutant structures were solvated by ~13200 water molecules with sodium ions to neutralize the total charge. Protein and water force fields and constraint algorithms were employed as described for the all-atom simulations of crowded systems; minor differences of other simulation parameters are not expected to impact the reported observables. Simulations were performed with the gromacs-4.6.1 software package.51 The short-range cutoff for electrostatic and van-der-Waals


Page 7 of 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


interactions was set to 0.9 nm, which was combined with a corresponding shift of the potential and a correction for the system’s energy and pressure. The particle-mesh Ewald method39 was used to describe long-range electrostatics using a 0.12 nm grid spacing and a 4th order interpolation. After an initial energy minimization all systems were equilibrated in the NPT ensemble for 1 ns with harmonic position restraints acting on all non-hydrogen atoms of the protein with a force constant of 1000 kJ/(mol nm2). This was followed by an additional 1 ns equilibration without position restraints. During equilibration, the temperature and pressure were weakly coupled to an external bath using a Berendsen thermostat and barostat47 with a 1.0 ps time constant set to 300 K and 1 bar, respectively. All equilibrations were performed with a 1 fs timestep. Following equilibration, production simulations were performed for each system for a total duration of 500 ns using a time step of 2 fs. During the production simulations, the temperature and pressure of the system were controlled by a Nose-Hoover52,53 thermostat and Parrinello-Rahman45 barostat with time constants of 1.0 ps. Solvent accessible surface areas were monitored with the g_sas tool of the gromacs software package using a solvent probe with a 0.14 nm radius. Protein atoms with a point charge |q|>0.2 were considered polar; atoms with a point charge |q| 4 for in vitro and n > 18 for in-cell measurements (exception, I35A n = 11). (D) Comparison of different histidine substitutions based on A4V as template. HallS is a mutant in which all histidines were replaced by serines. Statistical significance was tested using one-way ANOVA analysis with a post-hoc Tukey test to correct for multiple comparisons. If not otherwise specified, asterisks indicate the comparison between cell and buffer for a single mutant. *** p < 0.001 Error bars represent mean and s.d. (E) Gf0’ of A4V and G41D in BSA (200 mg mL-1) and PBS shown as mean and s.d. Statistical significance was tested by unpaired non-parametric t-tests (Mann-Whitney) comparing each BSA condition to its PBS control. ** p

Stability Effect of Quinary Interactions Reversed by Single Point

Recommend Documents