The Conformational Change in Elongation Factor Tu Involves

Subscriber access provided by Universitaetsbibliothek | Johann Christian Senckenberg

Article

Conformational change in Elongation factorTu involves separation of its domains Jonathan Lai, Zhaleh Ghaemi, and Zaida Luthey-Schulten Biochemistry, Just Accepted Manuscript • DOI: 10.1021/acs.biochem.7b00591 • Publication Date (Web): 18 Oct 2017 Downloaded from http://pubs.acs.org on October 19, 2017

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 free 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 accessible to all readers and 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.

Biochemistry 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 28

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

Biochemistry

Conformational change in Elongation factor-Tu involves separation of its domains Jonathan Lai,‡,¶ Zhaleh Ghaemi,∗,‡,¶ and Zaida Luthey-Schulten∗,‡,§,k,⊥ ‡Department of Chemistry, University of Illinois at Urbana-Champaign, Urbana, IL, USA §Center for the Physics of Living Cells, University of Illinois, Urbana, IL USA kBeckman Institute, University of Illinois, Urbana, IL, USA ⊥Carl Woese Institute for Genomic Biology, University of Illinois, Urbana, IL, USA ¶Contributed equally to this work E-mail: [email protected]; [email protected] Phone: (217) 333-3518. Fax: (217) 244-3186

1 ACS Paragon Plus Environment

Biochemistry

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

Abstract Elongation factor Tu (EF-Tu) is a highly conserved GTPase responsible for supplying the aminoacylated tRNA to the ribosome. Upon binding to the ribosome, EF-Tu undergoes GTP hydrolysis which drives a major conformational change, triggering the release of aa-tRNA to the ribosome. Using a combination of molecular simulation techniques, we studied the transition between the pre- and post-hydrolysis structures through two distinct pathways. We show that the transition free energy is minimum along a non-intuitive pathway, that involves "separation” of the GTP binding domain (Domain 1) from the OB folds (Domains 2 and 3), followed by Domain 1 rotation, and, eventually, locking EF-Tu conformation in the post-hydrolysis state. The domain separation also leads to a slight extension of the linker connecting Domain 1 to Domain 2. Using docking tools and correlation-based analysis, we identified and characterized the EF-Tu conformations that release the tRNA. These calculations suggest that the EF-Tu can release the tRNA before the domains separate and after Domain 1 rotates by 25◦ . We also examined the EF-Tu conformations in the context of the ribosome. Given the high sequence similarities with other translational GTPases, we predict a similar separation mechanism is followed.

Abbreviations Elongation factor Thermo unstable (EF-Tu), transfer RNA (tRNA), molecular dynamics (MD)

Introduction Cells assemble proteins by linking monomeric amino acids together in the ribosome. Crucial to supplying amino acids to the ribosome is a heterotrimeric GTPase, Elongation factor Tu 2 ACS Paragon Plus Environment

Page 2 of 28

Page 3 of 28

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

Biochemistry

(EF-Tu) in Bacteria—EF-1α in Eukaryota and Archaea. In the cytoplasm, EF-Tu forms a ternary complex with both GTP and aminoacylated tRNA (aa-tRNA). Association of the ternary complex to the ribosome and the anticodon of the aa-tRNA to the codon of the mRNA triggers GTP hydrolysis (1 ). The release of the GTP γ-phosphate in EF-Tu induces a conformational change (2 ) which increases the dissociation constant of the ternary complex by three orders of magnitude—from the nanomolar (3 ) to micromolar range (4 ). Because of the speed of translation (5 ), EF-Tu needs to rapidly switch from its pre- and post-hydrolysis conformations (6 –9 ) Once the aa-tRNA is released, a cognate or near-cognate aa-tRNA can rapidly move into the ribosome and participate in peptide bond formation (10 ). X-ray crystallography and cryo-electron microscopy have captured structures of EF-Tu in complex with GTP or GDP ligands bound to tRNA and ribosome (11 –14 ). Structures of Thermus aquaticus (T. aq.) and Escherichia coli (E. coli) bacterial EF-Tu in both the pre- and post-hydrolysis states have been also resolved using X-ray crystallography (6 , 7 ). Comparison of the pre- and post-hydrolysis structures reveal three major structural differences (Fig 1): 1) using the OB folds (Domains 2 and 3) for alignment, the GTP binding domain (Domain 1) of the post-hydrolysis EF-Tu rotates by 90◦ with respect to the Domain 1 of the pre-hydrolysis state; 2) switch I, residues 41 to 63 (40 to 62 in E. coli) changes from an α-helix and adopts a β-sheet conformation; 3) switch II, residues 81 to 101 (80 to 100 in E. coli), partially unwinds and rotates approximately a quarter of a turn (7 ). It is thought that the switch II favors the post-hydrolysis conformation which can only occur when the hydrogen bond between the γ-phosphate of the GTP and H85 (84) of switch II is broken (15 ). Experimental and computational methods have been extensively applied to study: 1) possible mechanisms of GTP hydrolysis in EF-Tu (16 –20 ); 2) accommodation of the tRNA into the A-site of the ribosome (21 ); and 3) potential transient interactions between EF-Tu and other components involved with protein synthesis (22 –25 ). Although the pre- and post3 ACS Paragon Plus Environment

Biochemistry

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 ACS Paragon Plus Environment

Page 4 of 28

Page 5 of 28

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

Biochemistry

ensemble of EF-Tu conformations along the minimum free energy path (MFP) that releases the aa-tRNA. We suggest that because the crucial residues for the EF-Tu conformational change are highly conserved among other translational GTPases, our proposed mechanism is also likely to be conserved.

Materials and methods Molecular dynamics Models of EF-Tu bound to GTP in the pre-hydrolysis (PDB: 1B23 (6 )) and with GDP in the post-hydrolysis conformation (PDB: 1TUI (7 )) are prepared using the CHARMM36 (30 , 31 ) force field for protein and nucleic acid. Systems were ionized and solvated following Eargle et al. (32 ). Briefly, the crystal structures were neutralized by placing K + ions at electrostatic minima near the protein (using the program Ionize (33 )). Afterwards, the system is solvated using Solvate1.0 (34 ) and VMD 1.9.3 (35 ) to ensure that all dications have a full solvation shell and to maximize the protein-solvent interactions. The models have a minimum water buffer of at least 15 Å in all directions. Random water molecules distant to EF-Tu are replaced with K + and Cl− to bring the total salt concentration to 0.150 M. Models contain approximately 95,000 atoms and have an approximate box size of 100×100×100Å3 . Simulations were performed in the NPT ensemble, with periodic boundary conditions, using either a Langevin (with a coupling constant of 1 ps) or Nose-Hoover (36 , 37 ) thermostat set at 300K and pressure at 1 atm Berendsen barostat. Long-range electrostatics are calculated using the Particle-Mesh Ewald algorithm (38 ) with a cutoff of 12Å. SETTLE (39 )/LINCS (40 ) is used and the timestep is set to 2 fs. All simulations are performed using either NAMD 2.10 (41 ) and Gromacs 5.1 (42 ). Models are minimized and equilibrated—using a combination of contraints and the conjugate gradient method—in a four step process. Initially, all non-hydrogen atoms are 5 ACS Paragon Plus Environment

Biochemistry

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

constrained and the hydrogen atoms are directly minimized for 2000 steps. Next, the constraints on the water oxygens are released and minimized for 5000 steps. Afterwards, constraints on the amino acid sidechains and nucleic acid bases and sugar moieties are released and the system is minimized for 10,000 steps. Finally, all constraints are released and the entire system is minimized for 20,000 steps. Likewise, the system is systematically heated (0K, 100K, 200K, 300K) using the same constraints and simulation protocol.

Preparing the transition pathways To model the transition of EF-Tu from state a to state f, we constructed a heavy-atom SMOG model of EF-Tu (43 ). The native potential was constructed using a cut-off based contact map of EF-Tu in state f. Default parameters were used for all other calculations. Simulations were performed with a 0.5 fs timestep and at a reduced temperature of 0.2 (25K) in the NVT ensemble. Frames were written every 5 fs. Twenty unique SMOG trajectories and velocity distributions were generated and the RMSD difference between pairs of trajectories were calculated. The SMOG trajectories then were clustered and the most “central pathway”, i.e. had the smallest RMSD to the other trajectories, was chosen as our initial guess; The RMSD of the path is 0 Å and 0.6 Å of the pre- and post-hydrolysis states, respectively. We then converted the coarse-grained SMOG (initial) trajectory into all-atom models. This all-atom path was then iteratively optimized following Lodola, et al. (44 ). Briefly, the EF-Tu conformational change is described in terms of path collective variables (S, Z) (45 ), as implemented in Plumed 2.2 (46 ). Path collective variables take the form: ΣN i · exp(−λ · RMSD[χ − χi ]) 1 S = i=1 and Z = − ln[ΣN i=1 exp(−λ · RMSD[χ − χi ])] N λ Σi=1 exp(−λ · RMSD[χ − χi ]) where i is an index that iterates over frames of the pathway (ranging from 1 to to 24 in this study), RMSD[χ − χi ] defines the RMS deviation of a current structure (χ) with respect to 6 ACS Paragon Plus Environment

Page 6 of 28

Page 7 of 28

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

Biochemistry

the frames of the pre-defined pathway, χi . Because the snapshots in the initial trajectory are separated by a RMSD of 1.8Å from each other, Branduardi et al. (45 ) recommends using a λ of around 13Å−1 . The following atoms were chosen for the RMSD calculation: Cα atoms of the Domain 1, heavy atoms of the switch I domain and amino acids at the interface of all three domains, protein backbone atoms everywhere else. S, so defined, weighs the various frames such that the one with the smallest RMSD to the initial pathway has the greatest weight. Z can be thought of as a measure of the fluctuations (entropy) of the current trajectory from the initial trajectory. Visually, one can think of a cylinder enclosing the initial pathway. If the initial pathway agrees well with the all-atom simulation, Z should only be a few Angstroms. When the trajectory deviates, Z can be much larger. To keep all of our plots consistent, we only used the optimized initial SMOG trajectory for our S and Z plots. In the separation pathway studied in Fig. S1, Fig. S2, Z is larger than normal because the domain separation is not predicted in the initial SMOG trajectories with default parameters. We then performed SMD on the initial pathway in the space of path collective variables, extracted new snapshots separated by a RMSD of 1.8Å, and updated the pathway. This optimization process was done repeatedly until no qualitative change has been observed when the paths are mapped in the space of intuitive collective variables—namely domain rotation, domain separation, and contacts between the domain. To generate the direct pathway, we pulled the pre-hydrolysis state towards the posthydrolysis conformation along the path prescribed by the SMOG trajectory. No prescribed SMOG trajectory existed for the separation pathway. It was first observed by a well-tempered metadynamics study which indicated Domain 1 could separate from Domain 2 and 3 starting at S=4.8 and ending at S=18. To generate the complete separation pathway, we used SMD to 1) pull Domain 1 from the other domains at S value of 4.8 (the separation of domains from state b to c) and 2) pull Domain 1 towards Domains 2, 3 at S value of 18 (the rejoining 7 ACS Paragon Plus Environment

Biochemistry

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

of domains from state d to e). We used targeted molecular dynamics (47 ) to connect the ends of the SMD runs. Figures Fig. S3 and Fig. S4, show the convergence of the optimization protocol for the separation and direct pathways, respectively. This separation path was then optimized with the aforementioned protocol. To generate conformations that were orthogonal to the initial pathway (i.e. along the Z), we performed either well-tempered metadynamics (48 ) (Z < 3Å) or extracted snapshots from the SMD (Z > 3Å).

Free energy calculations We used 2-dimensional umbrella sampling method along the S and Z coordinates. Umbrellas were placed on a regular grid along S (from 1 to 24 every 0.25) and Z (every 1 Å), totaling 306 umbrellas. Each umbrella was restrained with a harmonic potential with a force constant of 25

·

·

·

·

kcal mol−1 S−1 and 2500 kcal mol−1 Å−1 (along Z), and equilibrated for 2 ns. Production run for each umbrella was at least 8 ns. The free energy surface was calculated with weighted histogram analysis method (WHAM) (49 ) using the same grid spacing as the the umbrellas. To check the convergence of the calculation, the free energy was calculated from either the first 5 ns of the production run or from all 8 ns; the average difference between the two

·

free energy surface was 0.45 kcal mol−1 . Additionally, we checked the overlap of umbrellas and plotted the number of overlapping umbrellas in S and Z space (Fig. S5) as well as the density of states sampled for each S and Z bin (Fig. S6). To examine the obtained free energy we launched unbiased MD simulations from the two minima (Fig. S7) and from state c and state d (Fig. S8).


Page 8 of 28

Page 9 of 28

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

Biochemistry

Estimating the binding of EF-Tu to tRNA The EF-Tu bound to tRNA, with or without A76, was pulled along the minimum free energy path (MFP) using constant-velocity SMD restraining Z> 2 Å with force constants

·

·

·

·

of 1000 kcal mol−1 S−1 and 1000 kcal mol−1 Å−1 . The protein and tRNA conformations were extracted from the SMD and interaction energies between the two were scored using Autodock Vina (50 ).

Correlation calculations The SMD trajectory is divided into 1 ns segments (i.e. 500 frames saved every 2 ps). Within each segment, generalized correlations (51 ) are calculated between pairs of residues using the center of mass protocol from Van Wart, et. al. (52 ) Nodes were defined on the center of mass for amino acids and sugar and bases of the nucleic acids. Correlations are only calculated between pairs of nodes if any heavy atoms from the two corresponding residues are within 4.5Å of each other for at least 75% of the trajectory; in all other cases, the correlations are zero. The correlation coefficients of the interface residues between the EF-Tu and aa-tRNA have been averaged for different functional regions of aa-tRNA (acceptor-stem, T-stem and A76). The SMD trajectory is divided into 1 ns segments (i.e. 500 frames saved every 2 ps).

Generation of sequences and conservation GTP-binding proteins (GO:0005525) are extracted from the UniProt database and clustered based on their titles, discarding clusters with fewer than 20 entries. These clusters are as follows: EF-Tu/EF-1A: 925, EF-G: 30, RF2: 53, Eef2: 51, LepA: 3545, Guf1: 93, RRF3: 179, IF2: 1044, and Era: 2912. Curated sequences of Ras (mouse and human) and selenocysteinespecific elongation factor (SelB E. coli), from Swiss-Prot, are added to the sample dataset ex post facto. Clusters are aligned to the reference EF-Tu sequence in PDB: 1B23 using MAFFT9 ACS Paragon Plus Environment

Biochemistry

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 10 of 28

L-INS-i (53 ) (ver. 7) with default parameters. Mutual information was calculated using Weblogo (54 ). Structural alignments are performed using MultiSeq (55 ).

Results and Discussions Pre- to post-hydrolysis conformational change of EF-Tu involves separation of its domains To generate the initial path connecting the pre- to post-hydrolysis conformations, state a and state f respectively, twenty SMOG trajectories were launched—starting from the prehydrolysis crystal structure (PDB: 1B23). An exemplar path from the trajectories was selected and iteratively optimized (44 ) (see Methods for details) (Fig. S3, Fig. S4). We used path-collective variables (45 ) to describe the conformational change both along (S, a dimensionless quantity denoting conformational change progress) and orthogonal (Z, distance from a path) of the optimized pathway. We sampled two distinct pathways: the first one directly connects state a and f based on Domain 1 rotation (direct pathway) (Fig. 1); the second one includes the separation of Domain 1 from Domains 2 and 3 in addition to the domain rotation (separation pathway). The direct pathway can be divided into three segments: initial conformational change (a → b), large scale rotation (b → e), and final conformational change (e → f ) (Figure. 2A, red dashed line). The separation pathway, on the other hand, can be divided into five parts: (a →b), separation of the domains (b →c), free diffusion of the Domain 1 (c →d), rejoining of domains (d →e), and (e →f ) (Fig. 2A, gray dashed line). Both pathways approached within 1.08 Å backbone RMSD of the post-hydrolysis crystal structures (Fig. S9). The amino acids that deviated the most from the crystal structure are found either in the flexible region of the switch I or in loops connecting the secondary structure elements in Domain 1 and


Page 11 of 28

Biochemistry


Biochemistry

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 12 of 28

formed. Below, we describe the details of the conformational changes along the separation pathway. Amino acid numbering is based on T. aq. structure (PDB: 1B23) following by E. coli residue numbers in parenthesis. State a to b: Initial conformational changes from pre-hydrolysis state In the pre-hydrolysis crystal structure (PDB: 1B23), a key Mg2+ is coordinated by both the β and γ phosphates of the GTP ligand as well as T62 (61) of the switch I region. Upon GTP hydrolysis, T62 releases the Mg2+ , causing residues A53—I61 of switch I to partially open up and become more solvent exposed. Concomitantly, the Mg2+ moves in between the α and β phosphates of the GDP ligand (Fig. 2B), where other nearby oxygen atoms—namely the side chains of T25 (25) and D51 (50), and water molecules—can coordinate with Mg2+ and complete its solvation shell. Direct coordination of D51 to the Mg2+ pulls N41—I50 towards the GDP binding site, allowing A53—I61 to pivot out further into the solvent. From the free energy calculations (Fig. 2A), the conformational change from a to b leads to a 20◦ rotation of Domain 1 and a domain separation of about 3Å. The free energy difference

·

for the conformational change is 10 kcal mol−1 —arising from breakage of more than 80% of the native hydrogen bonds between Domains (Fig. S11). In addition, R241 (230) switches hydrogen bonding partner from D100 (99) to Q98 (97). To examine this part of the free energy surface, we launched short unbiased MD simulations from b: all simulations move towards the pre-hydrolysis state within 50 ns, confirming that the initial movement of EF-Tu is energetically expensive. We also performed a long-time scale simulation of EF-Tu with the CHARMM22* (58 ) force field, starting from a, to check if a is truly a minimum on the free energy surface; after 1µs, the system remained mainly in state a (Fig. S12).


Page 13 of 28

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

Biochemistry

State b to c: separation of the domains As the GTP binding Domain 1 separates from Domain 2, the remaining inter-domain hydrogen bonds continue to break. The final breaking of the hydrogen bonds occur between the side chain of Q98 (97) and R241 (230), and the backbone of G283 (271) (Fig. 2C). Calculations with Autodock Vina (50 ) show that once the the center of the Domain 1 moves beyond 38 Å from the center of Domain 3 (Z >4Å), the interaction energy between the domains trends towards zero (Fig. S2). State c to d: free diffusion of Domain 1 In this step, the rotation of Domain 1 towards the post-hydrolysis conformation completes. Because the domains are not interacting at state c, the conformational change resembles an iso-energetic diffusion (Fig. 2A, Fig. S1). Unbiased MD simulations launched from either c or d states freely moved towards the other endpoint (Fig. S8), confirming that the transition from c to d is unobstructed. Interestingly, this is where we predict the helix containing H85 (84) of switch II partially unwinds, adjusting to the post-hydrolysis crystal structure (Fig. 2D). Simultaneously, the linker connecting to Domain 1 to Domain 2 extends from 26Å to 30Å (Fig. S13). State d to e: rejoining of domains As Domain 1 rejoins Domain 3, switch I region adopts the extended β sheet conformation seen in the post-hydrolysis crystal structure, allowing the positively charged residues (e.g. R59 (58)) in the switch I to bind to a negative patch on Domain 3 (Fig. 2E). The rejoining causes the linker to contract to 25Å from 30Å Fig. S13. Simultaneously, R385 (373) of Domain 3 docks in-between the switch II and adjacent α-helix (I120—V128) of Domain 1 and hydrogen bonds to the backbone of the switch II (Fig. 2E). On the separation pathway, the docking of R385 can occur quite easily because the domain separation minimizes any potential non-native contacts 13 ACS Paragon Plus Environment

Biochemistry

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 14 of 28

that would interfere with R385 during the domain rotation. On the direct pathway, however, R385 tends to sterically clash with I120—V128 as the domains rotate; the entanglement of R385 contribute to the large transition barrier seen on the direct pathway. State e to f: final conformational rearrangements towards post-hydrolysis state As EF-Tu converges near the post-hydrolysis state, switch II can lock into its final position. The switch II region contains two highly conserved residues: P83 (82) and Y88 (87) (Fig. 3A). From state e to f, Y88 is initially solvent exposed (purple) (Fig. 3A). As Domain 1 adjusts its conformation, Y88 slips in between residues Y70 to K90 and binds to a hydrophobic pocket which acts as a lock for the post-hydrolysis conformation. This pocket can only form if P83 moves away from Domain 1 and occupies the same location of T62 (pink) as found in the GTP bound state. Unbiased MD simulations launched from the pre- and post-hydrolysis crystal structures (PDB: 1B23 and PDB: 1TUI, respectively) show that the Y88 exists either a solvent exposed or solvent shielded state—with minimal overlap between the states.

Effects of the inorganic phosphate retention on the MFP Experiments by Kothe, et al. (2 ) suggest that the timescale for inorganic phosphate (Pi) release is of the same order of magnitude as the timescale for the global conformational change; therefore, we investigated the effects of the presence of Pi on our MFP and conformational changes. To probe these effects, we built a model EF-Tu in complex with a GDP·Pi and used SMD to bias the system along our pre-described minimum free energy pathway. As shown in Fig. S14, there is no considerable difference between the pathways with and without the Pi along the main collective variables, namely, the domain rotation and separation. Instead, the largest effect of Pi retention appears to be in the placement of the SW1 region (Fig. S15). In the state a to b transition, the MFP predicted that the backbone of the SW1 region move closer to the GDP β phosphate; beyond state b, the SW1 region is 14 ACS Paragon Plus Environment

Page 15 of 28

Biochemistry


Biochemistry


Page 16 of 28

Page 17 of 28

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

Biochemistry

As the protein changes conformation from state a to b, the correlations between Domain 2 and Domain 3 of EF-Tu to the acceptor stem and the T-stem, respectively, decrease (Fig. 4C). This means that these parts of EF-Tu and tRNA can move more independently. Moreover, the correlation between the Domain 3 of EF-Tu and T-stem for the state b to c transition increases, namely movement of Domain 3 is more dependent on the T-stem movement (Fig. 4C). The correlation between Domain 2 and A76 mainly increases during both transitions. These results support the idea that 5′ end of aa-tRNA acceptor stem detaches from EF-Tu first followed by the T-stem and then A76 nucleotide. In all of the SMD simulations, A76 remains hydrogen bonds to E271 (259) while the backbone of the amino acid charged on the A76 can transiently bind to N285 (273). Moreover, the rotation of EF-Tu moves most of the interacting residues, as identified in Eargle, et al (22 ), away from the aa-tRNA (Fig. S16). After state b, no energetic contributions are seen from the amino acid side chain charged on the A76. Instead, intramolecular hydrogen bonding forces the ester-linked amino-acid to rotate towards C75, decreasing the overall solvent accessibility surface area for the entire aminoacylated-A76 residue. Because the acyl-linkage of the aa-tRNA is vulnerable to hydrolytic attack, Domain 2 probably protects the aminoacylated bond during the release of aa-tRNA into the ribosome. We also repeated the same SMD simulations without A76 to accelerate the aa-tRNA dissociation event. These simulations show that, the aa-tRNA easily dissociates from EF-Tu, suggesting that the rate-limiting step for aa-tRNA release is the interaction between Domain 2 and A76. We also docked EF-Tu into the ribosome and estimated its binding energy using Autodock Vina with default parameters at 25◦ c. Based on the docking study, EF-Tu has sufficient space to move away from the ribosomal large subunit during the conformational change (Fig. S16). Furthermore, EF-Tu remains bound to the ribosome for rotation angles less than 30◦ (Fig. S17 a,b) and only detaches from the ribosome after passing through state c (Fig. S17 C). 17 ACS Paragon Plus Environment

Biochemistry

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 18 of 28

Universality of the conformational change mechanism among all translational GTPases Given the sequence and structural conservation of Domain 1 and 2 amongst translational GTPases in all domains of life (61 , 62 ), the amino acids involved in the EF-Tu conformational change can also be involved in the mechanistic function of other translational GTPases. To test this possibility, we extracted and aligned more than 8835 translational GTPases sequences from all domains of life (Fig. 5). As expected, key residues in either switch I or switch II, such as D51 and T62 (state a to b transition), H85 (state d to e), P83 and Y88 (locking the post-hydrolysis state), are present in all GTPases (Fig. 5). Interestingly, residues Q98, D100 (or a negatively charged amino acid), and R241—all predicted to be involved in the a to b transition—and R385—d to e transition—are conserved amongst other GTPases (Fig. 5). Because the linker region connects the domains together, it ultimately limits the amount of separation between the domains. Therefore, to determine whether the separation mechanism is structurally feasible in the other GTPases, we estimated the end-to-end distance of the linker in several families (EF-1α (PDB: 3WXM), SelB (PDB: 4ZU9), EF-G (PDB: 1FNM), LepA (PDB: 3CB4), and RRF3 (PDB: 3VQT) and compared it to that of EF-Tu in the pre-hydrolysis conformation (Fig. S18 and Tab. S1). In the other GTPases, the primary sequences of the linkers are as long or longer than the linker in EF-Tu Tab. S1. Structurally, however, all of the linkers appear to be the same length Fig. S18, which suggests that the there is no structural hindrance for other GTPase to follow the separation mechanism.

Conclusion To study the conformational change of EF-Tu after GTP hydrolysis, we used path-based collective variables to sample several transition pathways connecting pre- and post-hydrolysis states. Interestingly, we have found that the pathway with lower free energy involves 18 ACS Paragon Plus Environment

Page 19 of 28

Biochemistry


Biochemistry

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 20 of 28

Supporting Information Supporting Information contains: the path optimization plots, convergence of the umbrella sampling calculations, results of the unbiased MD simulations on critical regions of the free energy surface, RMSD analysis of the crystalized and post-hydrolysis EF-Tu structures, free energy surfaces along path collective variables (S and Z) for both direct and separation pathways, variation of the number of hydrogen bonds as a function of conformational change, a µs long simulation result with Charmm22* force field, interaction energies between the Domain 1 and Domain 3 along the separation pathway, end-to-end distance of the linker in EF-Tu, comparison of EF-Tu trajectory with and without the Pi, docked EF-Tu conformations in the ribosome, binding energy of EF-Tu to the ribosome at different domain rotation and separation values, alignment of several GTPase crystal structures in pre-hydrolysis state. A table showing the comparison of linkers from other GTPases.


Page 21 of 28

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

Biochemistry

References 1. Rodnina, M. V., Pape, T., Fricke, R., and Wintermeyer, W. (1995) Elongation factor Tu, a GTPase triggered by codon recognition on the ribosome: mechanism and GTP consumption. Biochem. Cell Biol. 73, 1221–1227, PMID: 8722040. 2. Kothe, U., and Rodnina, M. V. (2006) Delayed release of inorganic phosphate from elongation factor Tu following GTP hydrolysis on the ribosome. Biochemistry 45, 12767– 12774, doi: 10.1021/bi061192z. 3. Sanderson, L. E., and Uhlenbeck, O. C. (2007) Exploring the specificity of bacterial elongation factor Tu for different tRNAs. Biochemistry 46, 6194–6200, PMID: 17489561. 4. Pingoud, A., Block, W., Wittinghofer, A., Wolf, H., and Fischer, E. (1982) The elongation factor Tu binds aminoacyl-tRNA in the presence of GDP. J. Biol. Chem. 257, 11261–7. 5. Bremer, H., and Dennis, P. P. (2008) Modulation of chemical composition and other parameters of the cell at different exponential growth rates. EcoSal Plus 3 . 6. Nissen, P., Thirup, S., Kjeldgaard, M., and Nyborg, J. (1999) The crystal structure of Cys-tRNACys–EF-Tu–GDPNP reveals general and specific features in the ternary complex and in tRNA. Structure 7, 143–156. 7. Polekhina, G., Thirup, S., Kjeldgaard, M., Nissen, P., Lippmann, C., and Nyborg, J. (1996) Helix unwinding in the effector region of elongation factor EF-Tu-GDP. Structure 4, 1141–1151. 8. Pape, T., Wintermeyer, W., and Rodnina, M. V. (1998) Complete kinetic mechanism of elongation factor Tu-dependent binding of aminoacyl-tRNA to the A site of the E.coli ribosome. EMBO J. 17, 7490–7497. 9. Maracci, C., and Rodnina, M. V. (2016) Review: translational GTPases. Biopolymers 105, 463–475. 10. Pape, T. (1999) Induced fit in initial selection and proofreading of aminoacyl-tRNA on


Biochemistry

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 22 of 28

the ribosome. The EMBO Journal 18, 3800–3807. 11. Schmeing, T. M., Voorhees, R. M., Kelley, A. C., Gao, Y.-G., Murphy, F. V., Weir, J. R., and Ramakrishnan, V. (2009) The crystal structure of the ribosome bound to EF-Tu and aminoacyl-tRNA. Science 326, 688–694. 12. Agirrezabala, X., Schreiner, E., Trabuco, L. G., Lei, J., Ortiz-Meoz, R. F., Schulten, K., Green, R., and Frank, J. (2011) Structural insights into cognate versus near-cognate discrimination during decoding. EMBO J 30, 1497–1507. 13. Behrmann, E., Loerke, J., Budkevich, T. V., Yamamoto, K., Schmidt, A., Penczek, P. A., Vos, M. R., Bürger, J., Mielke, T., Scheerer, P., and Spahn, C. M. (2015) Structural snapshots of actively translating human ribosomes. Cell 161, 845–857. 14. Loveland, A. B., Demo, G., Grigorieff, N., and Korostelev, A. A. (2017) Ensemble cryo-EM elucidates the mechanism of translation fidelity. Nature 546, 113–117. 15. Maracci, C., Peske, F., Dannies, E., Pohl, C., and Rodnina, M. V. (2014) Ribosomeinduced tuning of GTP hydrolysis by a translational GTPase. Proc. Natl. Acad. Sci. USA 111, 14418–14423. 16. Daviter, T., Wieden, H.-J., and Rodnina, M. V. (2003) Essential role of histidine 84 in elongation factor Tu for the chemical step of GTP hydrolysis on the ribosome. J. Mol. Biol. 332, 689–699. 17. Adamczyk, A. J., and Warshel, A. (2011) Converting structural information into an allosteric-energy-based picture for elongation factor Tu activation by the ribosome. Proc. Natl. Acad. Sci. USA 108, 9827–9832. 18. Aleksandrov, A., and Field, M. (2013) Mechanism of activation of elongation factor Tu by ribosome: catalytic histidine activates GTP by protonation. RNA 19, 1218–1225. 19. Åqvist, J., and Kamerlin, S. C. (2015) Exceptionally large entropy contributions enable the high rates of GTP hydrolysis on the ribosome. Sci. Rep. 5, 1–10. 20. Carvalho, A. T., Szeler, K., Vavitsas, K., Åqvist, J., and Kamerlin, S. C. (2015) Modeling 22 ACS Paragon Plus Environment

Page 23 of 28

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

Biochemistry

the mechanisms of biological GTP hydrolysis. Arch. Biochem. Biophys. 582, 80–90. 21. Noel, J. K., and Whitford, P. C. (2016) How EF-Tu can contribute to efficient proofreading of aa-tRNA by the ribosome. Nature Comm. 7, 13314. 22. Eargle, J., Black, A. A., Sethi, A., Trabuco, L. G., and Luthey-Schulten, Z. (2008) Dynamics of recognition between tRNA and elongation factor Tu. J. Mol. Biol. 377, 1382 – 1405. 23. Sethi, A., Eargle, J., Black, A. A., and Luthey-Schulten, Z. (2009) Dynamical networks in tRNA: protein complexes. Proc. Natl. Acad. Sci. USA 106, 6620. 24. Alexander, R., Eargle, J., and Luthey-Schulten, Z. (2010) Experimental and computational determination of tRNA dynamics. FEBS Lett. 584, 376–386. 25. Rodnina, M. V., Fischer, N., Maracci, C., and Stark, H. (2017) Ribosome dynamics during decoding. Phil. Trans. Royal Soc. B 372, 20160182. 26. Frank, J., Sengupta, J., Gao, H., Li, W., Valle, M., Zavialov, A., and Ehrenberg, M. (2004) The role of tRNA as a molecular spring in decoding, accommodation, and peptidyl transfer. FEBS Letters 579, 959–962. 27. Burnett, B. J., Altman, R. B., Ferrao, R., Alejo, J. L., Kaur, N., Kanji, J., and Blanchard, S. C. (2013) Elongation factor Ts directly facilitates the formation and disassembly of the Escherichia coli elongation factor Tu·GTP·aminoacyl-tRNA ternary complex. J. Biol. Chem. 288, 13917–13928. 28. Liu, W., Chen, C., Kavaliauskas, D., Knudsen, C. R., Goldman, Y. E., and Cooperman, B. S. (2015) EF-Tu dynamics during pre-translocation complex formation: EFTu·GDP exits the ribosome via two different pathways. Nucl. Acids Res. 43, 9519–9528. 29. Morris, G. M., Huey, R., Lindstrom, W., Sanner, M. F., Belew, R. K., Goodsell, D. S., and Olson, A. J. (2009) AutoDock4 and AutoDockTools4: automated docking with selective receptor flexibility. J. Comp. Chem. 30, 2785–2791. 30. Denning, E. J., Priyakumar, U. D., Nilsson, L., and Mackerell, A. D. (2011) Impact of 23 ACS Paragon Plus Environment

Biochemistry

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 24 of 28

2′ -hydroxyl sampling on the conformational properties of RNA: update of the CHARMM all-atom additive force field for RNA. J. Chem. Theory Comput. 32, 1929–1943. 31. Best, R. B., Zhu, X., Shim, J., Lopes, P. E. M., Mittal, J., Feig, M., and MacKerell, A. D. (2012) Optimization of the additive CHARMM all-atom protein force field targeting improved sampling of the backbone φ, ψ and side-chain χ1 and χ2 dihedral angles. J. Chem. Theory Comput. 8, 3257–3273. 32. Eargle, J., and Luthey-Schulten, Z. A. In RNA 3D Structure Analysis and Prediction; Leontis, N., and Westhof, E., Eds.; 2012; Chapter 12, pp 213–238. 33. Balaeff, A., Eargle, J., and Roberts, E. Ionize v 1.6 ; University of Illinois – UrbanaChampaign: Urbana, Illinois, 2005. 34. Grubmueller, H., and Groll, V. Solvate v 1.0 ; Max Planck Institute for Biophysical Chemistry: Göttingen, Germany, 1996. 35. Humphrey, W., Dalke, A., and Schulten, K. (1996) VMD–visual molecular dynamics. J. Mol. Graphics 14, 33–38. 36. Nose, S. (1984) A unified formulation of the constant temperature molecular dynamics methods. J. Chem. Phys. 81, 511–519. 37. Hoover, W. G. (1985) Canonical dynamics: equilibrium phase-space distributions. Phys. Rev. A 31, 1695–1697. 38. Ewald, P. P. (1921) Die berechnung optischer und elektrostatischer gitterpotentiale. Annalen der Physik 369, 253–287. 39. Miyamoto, S., and Kollman, P. A. (1992) SETTLE: an analytical version of the SHAKE and RATTLE algorithm for rigid water models. J. Comp. Chem. 13, 952–962. 40. Hess, B., Bekker, H., Berendsen, H. J. C., and Fraaije, J. G. E. M. (1997) LINCS: a linear constraint solver for molecular simulations. J. Comp. Chem. 18, 1463–1472. 41. Phillips, J. C., Braun, R., Wang, W., Gumbart, J., Tajkhorshid, E., Villa, E., Chipot, C., Skeel, R. D., Kale, L., and Schulten, K. (2005) Scalable molecular dynamics with NAMD. 24 ACS Paragon Plus Environment

Page 25 of 28

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

Biochemistry

J. Comp. Chem. 26, 1781–1802. 42. Pronk, S., Pall, S., Schulz, R., Larsson, P., Bjelkmar, P., Apostolov, R., Shirts, M. R., Smith, J. C., Kasson, P. M., van der Spoel, D., Hess, B., and Lindahl, E. (2013) GROMACS 4.5: a high-throughput and highly parallel open source molecular simulation toolkit. Bioinformatics 29, 845–854. 43. Noel, J. K., Whitford, P. C., Sanbonmatsu, K. Y., and Onuchic, J. N. (2010) SMOG@ctbp: simplified deployment of structure-based models in GROMACS. Nucl. Acids Res. W657– 661, PMID: 20525782. 44. Lodola, A., Branduardi, D., Vivo, M. D., Capoferri, L., Mor, M., Piomelli, D., and Cavalli, A. (2012) A catalytic mechanism for cysteine N-terminal nucleophile hydrolases, as revealed by free energy simulations. PLoS ONE 7, e32397. 45. Branduardi, D., Gervasio, F., and Parrinello, M. (2007) From A to B in free energy space. J. Chem. Phys. 126, 05103. 46. Bonomi, M., Branduardi, D., Bussi, G., Camilloni, C., Provasi, D., Raiteri, P., Donadio, D., Marinelli, F., Pietrucci, F., Broglia, R. A., and Parrinello, M. (2009) PLUMED: A portable plugin for free-energy calculations with molecular dynamics. Comput. Phys. Commun. 180, 1961 – 1972. 47. Schlitter, J., Engels, M., and Krüger, P. (1994) Targeted molecular dynamics: a new approach for searching pathways of conformational transitions. J. Mol. Graph 12, 84–89. 48. Barducci, A., Bussi, G., and Parrinello, M. (2008) Well-tempered metadynamics: a smoothly converging and tunable free-energy method. Phys. Rev. Lett. 100, 020603. 49. Grossfield, A. WHAM: the weighted histogram analysis method v2.0.9 (2013). 2013. 50. Trott, O., and Olson, A. J. (2010) AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. J. Comp. Chem. 31, 455–461. 51. Lange, O. F., and Grubmuller, H. (2006) Generalized correlation for biomolecular dy25 ACS Paragon Plus Environment

Biochemistry

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 26 of 28

namics. Proteins: Struct., Func., Bioinf. 62, 1053–1061. 52. VanWart, A., Eargle, J., Luthey-Schulten, Z., and Amaro, R. (2012) Exploring residue component contributions to dynamical network models of allostery. J. Chem. Theory Comp. 8, 2949–2961. 53. Katoh, K., and Standley, D. M. (2016) A simple method to control over-alignment in the MAFFT multiple sequence alignment program. Bioinformatics 32, 1933–1942. 54. Crooks, G. E., Hon, G., Chandonia, J. M., and Brenner, S. E. (2004) WebLogo: a sequence logo generator. Genome Research 14, 1188–1190. 55. Roberts, E., Eargle, J., Wright, D., and Luthey-Schulten, Z. (2006) MultiSeq: unifying sequence and structure data for evolutionary analysis. BMC Bioinf. 7, 382. 56. Jensen, M., Cool, R. H., Mortensen, K. K., Clark, B. F. C., and Parmeggiani, A. (1989) Structure-function relationships of elongation factor Tu. Eur. J. Biochem. 182, 247–255. 57. Gromadski, K. B., Wieden, H.-J., and Rodnina, M. V. (2002) Kinetic mechanism of elongation factor Ts-catalyzed nucleotide exchange in elongation factor Tu. Biochemistry 41, 162–169. 58. Piana, S., Lindorff-Larsen, K., and Shaw, D. E. (2011) How robust are protein folding simulations with respect to force field parameterization? Biophys. J. 100, L47–L49. 59. Schrader, J. M., Chapman, S. J., and Uhlenbeck, O. C. (2011) Tuning the affinity of aminoacyl-tRNA to elongation factor Tu for optimal decoding. Proc. Natl. Acad. Sci. USA 108, 5215–5220. 60. Mittelstaet, J., Konevega, A. L., and Rodnina, M. V. (2013) A kinetic safety gate controlling the delivery of unnatural amino acids to the ribosome. J. Am. Chem. Soc. 135, 17031–17038. 61. Atkinson, G. (2015) The evolutionary and functional diversity of classical and lesserknown cytoplasmic and organellar translational GTPases across the tree of life. BMC Genomics 16, 78–93. 26 ACS Paragon Plus Environment

Page 27 of 28

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

Biochemistry

62. Wittinghofer, A., and Vetter, I. R. (2011) Structure-function relationships of the G domain, a canonical switch motif. Annu. Rev. Biochem. 80, 943–971, doi: 10.1146/annurevbiochem-062708-134043.


Biochemistry

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

Graphical TOC Entry


Page 28 of 28

The Conformational Change in Elongation Factor Tu Involves

Recommend Documents