The Role of the Interdomain Interactions on RfaH Dynamics and

Subscriber access provided by CMU Libraries - http://library.cmich.edu

Article

The Role of the Interdomain Interactions on RfaH Dynamics and Conformational Transformation Jeevan B. GC, Bernard S Gerstman, and Prem P. Chapagain J. Phys. Chem. B, Just Accepted Manuscript • DOI: 10.1021/acs.jpcb.5b05681 • Publication Date (Web): 15 Sep 2015 Downloaded from http://pubs.acs.org on September 19, 2015

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.

The Journal of Physical Chemistry B 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 26

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

The Journal of Physical Chemistry

The Role of the Interdomain Interactions on RfaH Dynamics and Conformational Transformation Jeevan B. GC, Bernard S. Gerstman, and Prem P. Chapagain* Department of Physics, Florida International University, Miami, FL 33199 Abstract The transcription anti-terminator RfaH has been shown to undergo major structural rearrangements to perform multiple functions. Structural determination of the C-terminal domain (CTD) of RfaH showed that it can exist as either an α-helix bundle when interfacing with the N-terminal domain (NTD), or as a β-barrel conformation when it is not interfacing with the NTD. In this paper, we investigate the full RfaH with both CTD and NTD using a variety of all-atom Molecular Dynamics (MD) simulation techniques, including Targeted Molecular Dynamics, Steered Molecular Dynamics, and adaptive biasing force, and calculate Potentials of Mean Force. We also use network analysis to determine communities of amino acids that are important in transferring information about structural changes. We find that the CTD-NTD interdomain interactions constitute the main barrier in the CTD α-helix to β-barrel structural conversion. Once the interfacial interactions are broken, the structural conversion of the CTD is relatively easy. We determined which amino acids play especially important roles in controlling the interdomain motions and also describe subtle structural changes that may be important in the functioning of RfaH.

* Corresponding Author, Email: [email protected]; phone: +1 305-348-6266

1

ACS Paragon Plus Environment


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

I. INTRODUCTION Proteins perform a myriad of tasks in the cell such as molecular transport, enzymatic activities, energy transfer, and signaling. In order to perform their physiological functions, most proteins must fold to their functional, native state structure, which are highly specific three-dimensional conformations1. Depending on the biochemical environment, some proteins are observed to have multiple functions. A protein that exhibits multiple functions is known as a moonlighting protein2. In performing alternate functions, such a moonlighting protein can utilize separate protein interfaces of the same configuration 3, or take advantage of small-scale protein structural flexibility as well as larger-scale domain separation or oligomerization4. Intrinsically unstructured proteins may have sufficient structural flexibility to interact with multiple partners, whereas globular proteins can undergo structural changes that allow them to perform different functions 5. Globular proteins such as lymphotactin6 and Mad27 that undergo small structural rearrangements that lead to different functions are referred to as metamorphic proteins. Recently, the transcription anti-terminator RfaH has been shown to undergo major structural rearrangements to perform multiple functions8,9. When interfacing with its Nterminal domain (NTD), the C-terminal domain (CTD) is in an α-helix bundle, but transforms into a β-barrel conformation when not interfaced with its NTD. The α-form of the RfaH-CTD masks the RNAP binding site and regulates transcription, whereas the βform of the CTD enhances translation by recruiting a ribosome to an mRNA lacking a ribosome-binding site. NusG10 is a transcription factor that also has two domains, and whose NTD has a similar structure to the NTD of RfaH. The NTDs of RfaH and NusG have the same functional roles in RNA Polymerase (RNAP) binding. Though the NTD of RfaH and NusG are similar, the CTD conformations are different, with the native state of the NusG CTD not interfaced with the NTD and having an all-β fold. As in NusG, RfaH can bind to RNAP and protein S10 only when the CTD is in the β-form, which is not in contact with the NTD and thus unmasks the binding site. This separation of the NTDCTD domains in RfaH may be allosterically activated by the interactions with an ops element11,12 followed by spontaneous transformation of the CTD into the β-structure conformation. In an earlier work13, molecular dynamics (MD) simulations were used to investigate the folding/unfolding dynamics of the isolated CTD of RfaH without the NTD and explored the α-helix to β-structural conversions using the EEF1 force field with CHARMM19. Simulations of the isolated CTD showed that without the NTD interactions, the CTD Helix I loses its structural integrity and unfolds faster than Helix II. A recent computational study highlighted the differences in the structural propensities of the CTD helices and showed that leucine-rich regions in Helix II provide structural stability in both α-helix and β-structure14. Similarly, Li et. al used Markov State models to investigate the structural transformation and found various parallel folding pathways

2


Page 2 of 26

Page 3 of 26

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


with different mechanisms15. Consistent with earlier observations, Helix I was found to be unstable in the absence of the NTD interactions. These studies on the isolated CTD provide information on the α-helix to βstructure conversion in the absence of the interdomain interactions with the NTD. However, in the functioning protein, interdomain interactions are critical in determining the conformational preference of the RfaH CTD between α-helix versus β-barrel structure. In this paper, we investigate the full RfaH with both CTD and NTD using a variety of all-atom Molecular Dynamics (MD) simulation techniques. Since the interdomain contacts stabilize the CTD α-helical state, and the conversion into the βbarrel structure requires these interdomain interactions to weaken or break, here we focus our attention on the role of the NTD-CTD interfacial contacts in the structural conversion of the RfaH CTD. Recently, Xiong et al.16 used a coarse grained off-lattice model to investigate the folding and allostery in the full RfaH as well as structural conversion in the isolated CTD. Thermodynamic analysis showed that the heat capacity curve in the full RfaH displays two peaks, corresponding to the CTD and NTD folding. Most recently, Ramirez-Sarmiento et al.17 used a dual basin model that biases the protein to fold into two different forms and found that tuning the interdomain interaction bias strengths can cause the protein to switch the structure between the α- and β-forms in the full RfaH. Their results also suggested that F130 is an important residue for CTD stability in both alpha and beta forms. In order to further understand the all-α to all-β transformation of the CTD in the full RfaH, we performed Targeted Molecular Dynamics (TMD) simulations and Steered Molecular Dynamics (SMD) simulations18,19 in explicit water. Targeted MD is especially useful for investigating conformational changes or large molecular motions that are otherwise unfeasible to access in reasonable computational times. This method has been extensively used in various molecular systems that undergo large-scale conformational changes that are relevant to protein function. For example, Targeted MD was used to elucidate the ATP binding mechanism, and to determine the transition pathways between the open and closed states of GroEL20. Similarly, Targeted MD has been used to investigate transitions between open and closed states of ion channels and their gating mechanisms in various systems such as a nicotinic receptor21, KcsA potassium channels22,23, mechanosensitive channels24, and the AMPA receptor 25. More relevant to the transitions in RfaH, the dynamical transition between two distinct states in the Mad2 metamorphic protein was recently studied using a combination of conventional and Targeted MD and the structural characteristics as well as dynamic transition mechanisms were explored26. To understand the role of the interfacial interactions in the structural rearrangements of the RfaH-CTD, we also used SMD to calculate the Potential of Mean Force (PMF) from the ensemble average of the SMD trajectories using the Jarzynski equality 27. SMD simulations have been valuable in understanding many biophysical

3



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

processes such as mechanical responses to biomolecules28,29,30,31,32, protein dissociation and conformational changes33,34,35, stability of protein-protein or domain interfaces36,37,38, and ligand binding 18, 39. Through the TMD and SMD simulations, we found that the interdomain interactions constitute the main barrier in the α-helix to β-barrel structural conversion. Once the interfacial interactions are broken, the structural conversion of the CTD occurs relatively easily. We identified key residues that are involved in the interdomain motion and structural transformation. II. METHOD The x-ray crystal structure of RfaH with α-CTD (PDB ID 2OUG) was obtained from the Protein Data Bank. Missing residues, including those in the 14-residue NTDCTD linker segment were modelled using Modeller40,41. The lowest energy structure was chosen based upon the discrete optimized protein energy (DOPE) score. The linker was modeled without a template and thus had no secondary structure, which is consistent with the flexibility that made the linker difficult to be resolved by x-ray crystallography. The modeled configuration had no van der Waals clash and placed the linker away from the domain interface. Although the modeled linker segment was flexible, the protein structure remained stable during a 100-ns simulation. Several different MD simulation techniques were performed for the full RfaH as discussed below. MD setup for equilibration and targeted MD: The full RfaH with the helical CTD was solvated with TIP3P water in a box with 10Å padding around the protein. The system was neutralized by adding two Cl- ions using VMD 42. The dimensions of the simulation box were 56×54×78 Å3 and the resulting system contained a total of 24,082 atoms. MD simulations were run with the CHARMM36 43 force-field using NAMD 19. The particle mesh Ewald method 44 was used to treat the long-range interactions with a 12 Å nonbonded cut-off. Using the conjugate gradient and line search algorithm, the energy was minimized for 5 ps. Heating was performed with a linear gradient of 20 K/ps from 20 to 300 K. Each system was then equilibrated with 2-ns NPT run using a 1-fs integration time step, followed by 5-ns NVT run also using a 1 fs integration time step. Temperature was maintained at 300 K using Langevin dynamics with a damping constant of 1 ps−1 was used. The RATTLE algorithm was used to constrain protein bonds involving hydrogen and SETTLE was used to maintain water geometry. Multiple time-stepping algorithms (1-2-4) were used to compute interactions between atoms. Interactions through covalent bonds were calculated at every time step whereas the calculations of short-range non-bonded interactions and long-range electrostatic forces were performed every other step and every fourth step respectively. A 10 ns equilibration run was performed before commencing each targeted MD simulation. The target RfaH structure with the beta conformation of CTD was created using Modeller by connecting the RfaH-NTD and the β-CTD (PDB ID: 2LCL) with a

4


Page 4 of 26

Page 5 of 26

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


linker. For α β conversion, force was applied only on the N=777 heavy (non hydrogen) atoms of the helical CTD (residues 115-162). For β α conversion, force was applied on N=1322 heavy atoms which include all of the CTD (residues 115-162) as well as an additional 30 residues of NTD that participate in interdomain interaction. The 2 targeted MD calculation uses the potential energy function = [ − ] . 2

Here, rmsd(t) is the root mean squared deviation (rmsd) of the simulated structure with respect to the target structure and RMSD(t) is the assigned, target rmsd, which decrease linearly over time. A force constant per atom of katom=k/N=0.4 kcal/mol/Å2 was used for all TMD simulations. Potential of Mean Force (PMF) calculations were performed using Adaptive Biasing Forces (ABF)45,46. The PMF was calculated using 20 ns ABF simulations with a time step of 0.5 fs using rmsd as the reaction coordinate in the NAMD colvar module. The upper and lower boundary wall constant was 310 kcal/mol and a bin width of 0.001 Å. Force samples were accrued for 5000 steps until the adaptive biasing force was applied. The Langevin damping constant was 1 ps-1. MD setup for SMD: The full RfaH was solvated with TIP3P water in rectangular boxes. Pulling occurred along the x-axis. A cut-off of 10 Å was used for the y- and z-directions, whereas a large padding was added in the x-direction to allow enough space for protein 3

extension due to pulling. The resulting box dimensions were 198×56×81 Å with a total of 85,567 atoms, which includes two Cl- ions to neutralize the system. The particle mesh Ewald method was used to treat the long-range interactions with a 12 Å non-bonded cutoff. Energy minimization, heating, and equilibration were performed as described earlier. Simulations were run with the CHARMM27 force field using NAMD. For SMD, the dummy atom attached to the C-terminal Cα SMD atom (residue L162) via a virtual spring (k=3 kcal/mol/Å2) was pulled in the negative x-direction with a constant speed of 2 Å/ns while keeping the Cα atom of the N-terminal residue (M1) fixed. SMD was performed with a 1 fs time step and the steering forces were saved every 1 ps. Trajectories were obtained by alternating pulling and relaxation, i.e. 10 ns of SMD followed by 10 ns of relaxation. During relaxation, the extension was preserved by keeping the first Cα in the NTD and the final SMD Cα in the CTD fixed (i.e. setting the velocity of each to zero), but the protein’s internal structure was allowed to relax after the forced stretching 30, 47. The average work done due to pulling was calculated using nine independent simulations, and the potential of mean force (PMF) was calculated from the Jarzynski equality using an ensemble average of the nine SMD trajectories. Although PMF calculations with ABF have been shown to represent the reaction more accurately than with constant velocity SMD 48, the SMD simulations give important insights into the barriers along the pulling reaction coordinates.

5



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

III. RESULTS AND DISCUSSION A. Structural analysis of the RfaH α-CTD The RfaH-CTD consists of the protein segment from residues 115 to 156. These residues form a two-helix bundle when interfaced with the RfaH-NTD, but prefer a βbarrel conformation when not interfaced with the NTD. Its NusG paralog has a similar NTD structure but the CTD is never interfaced with the NTD, and the CTD is found only in the β-barrel structure. In both NusG and RfaH, RNAP and protein S10 bind with the NTD when the CTD is in its β-form. Since the α-form of the RfaH-CTD is stabilized by interdomain interactions (Table S1), it is important to examine these interactions for understanding the structural transformation of the CTD. We performed a 100-ns NVT simulation on the native state configuration of the full RfaH with the CTD in the α-helix form. Figure 1a shows overlays of 100 sampled conformations taken from the 100 ns trajectory. These, together with the root mean squared fluctuations (rmsf) for all Cα plotted in Fig. S1, provide insight on the conformational fluctuations during the simulation. The rmsf figure (Fig. S1 a) shows that both the NTD and the CTD are stable but large fluctuations occur in the linker and loop segments, as expected. Specifically, the segment composed of residues 35-45, corresponding to the loop region of the β-hairpin motif in the NTD shows particularly large fluctuations. Also, the long NTD-CTD linker segment (residues 101-114) shows large flexibility. Interestingly, the NTD helix (90-100) at the domain interface is slightly more flexible than the CTD helices (Helix I and Helix II). Dynamic network analysis can be used to understand the correlated molecular motions49,50,51. Dynamic analysis from the 100 ns NVT simulation of the full RfaH resulted in 11 communities that are highlighted in color in Fig. 1a and the constituent amino acids for each community are listed in Table S2. A community is composed of amino acids that are in close proximity and have highly correlated motions. The size of the circle for each node is proportional to the number of residues in the community, and the line thickness signifies the inter-community edge weights. As shown in Fig. 1b, community H (lime green) in the NTD β-sheet core region seems to have the most correlated motion with other communities. We found three communities (C,D,G) that each are composed of residues from both the NTD and CTD. The amino acids that comprise these three communities are highlighted in Fig. 1c. By examining the composition of each community, we were able to locate the amino acids pairs that most strongly bridge the NTD to the CTD within each community. These pairs are shown in Fig. 1c: L96-F126 (community C), F33-F130 (community D), and F81-I118 (community E). In all of these cases, phenylalanine residues seem to play important roles in modulating the interdomain motions. F126 plays a major role in interdomain communication and therefore it is a functionally important residue for modulating the domain separation. Alanine substitution for F126 using FoldX 52 revealed that this mutation results in an increase in ∆∆G by ~3 kcal/mol (Fig. S1 (b)), indicating that the 6


Page 6 of 26

Page 7 of 26

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


F126A substitution destabilizes the domain interface, and is one of the largest ∆∆G among all residue positions upon alanine mutation. Figure S1 (b) also shows that more than 75% of the interface residues do not favor alanine substitution. We performed a 100 ns MD simulation on the RfaH-F126A mutant and investigated its kinetic/dynamic behavior. Figure S2 shows that the Cα rmsf for the WT and the RfaH-F126A are comparable. The F126A mutation caused slightly larger fluctuations in the loop/turn region (residues 35-50) in the β-turn-β motif of the NTD, but the linker segment connecting the two domains, as well as the some other loop segments are more restricted in the mutant. This appears to be due to the interactions with the CTD tail segment (residues 158-162), which became flexible in the WT after ~50 ns, whereas it remained intact for the entire 100 ns in F126A. A significantly longer simulation time may be required for a proper equilibrium sampling of the tail opening and closing for both the WT and the mutant. The network analysis displayed in Fig. S2 shows that the F126A mutation causes changes in the community structure. Specifically, the interdomain correlated motion through F126 disappears in the mutant.

Figure 1. RfaH structural fluctuations and correlated motions. a) Structural fluctuations of amino acids in the full RfaH determined from MD simulations. b) Community map of RfaH showing 11 different communities. The size of the circle for each node represents the number of residues in each community, and the line thickness represents the intercommunity edge weights. The communities are colored as in Fig. 1a. c) Three communities especially relevant for the interdomain interactions are highlighted.

7



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

B. Structural transformation with targeted MD The α-helical conformation of the CTD (α-CTD) in the full RfaH was targeted to map onto the β-barrel CTD (β-CTD) and vice versa using targeted MD. Several targeted MD simulations of length 10 ns were performed. Each simulation resulted in conversion to the desired target structure. A typical α-helix to β-barrel structural transformation process is shown in Fig. 2, which shows snapshots of the protein conformation at various stages of the structural transformation. In general, the α-helix secondary structure is lost by the mid-part of the simulation where the target RMSD has decreased by 50%. For the CTD, once the tertiary contacts with the NTD are lost as well as the CTD helicity, folding into the CTD β-structure begins to occur. By the end of the simulation, CTD β-sheets are fully formed and arranged into the final β-barrel structure. The final structure was within 0.94 Å rmsd of the experimental structure 9. To confirm the stability of the transformed structure, we performed a 100-ns conventional MD and found that the resulting βstructure was very stable at 300K.

Figure 2. Snapshots of RfaH conformations during TMD simulation for the α-CTD to βCTD conversion in the full RfaH. The N-terminal domain is shown as gray and the transforming C-terminal domain is highlighted in green. We plot the time evolution of the secondary structure of the full RfaH in Fig. 3a from the TMD. The structural transformation of the CTD (residues 115-162) can be divided into two parts: helix-to-coil transition (approximately 2–5 ns), and coil-to-beta transition (approximately 6-9 ns). The total number of intradomain CTD hydrogen bonds also follows this pattern, as shown in Fig. 3b. Hydrogen bonds are calculated with a distance cut-off of 3.5 Å and an angle cut-off of 30°. As discussed later, the coil structure that is intermediate between the α-helix and β-barrel structure is in fact a collapsed globule state of the protein. 8


Page 8 of 26

Page 9 of 26

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


Figure 3. Results from RfaH TMD simulations for the α-CTD β-CTD structural conversion. The CTD is composed of residues 115-162. a) Evolution of the secondary structure b) Evolution of the intradomain hydrogen bonds in the CTD as a function of time. The number of hydrogen bonds decreases as the CTD helices unfold (2-5 ns) but increases with the formation of the CTD β-structure (6-9 ns). Unfolding of the helix to the collapsed coil state: As shown in Fig. 3a, the CTD begins to lose its Helix II helicity around 2 ns. Interestingly, Helix I (residues 116-131) retains its structural integrity better than Helix II (residues 135-156). This result using the full RfaH is in contrast to previous computational results 13-14 obtained for the isolated CTD. Simulations for the isolated CTD showed faster unfolding of Helix I due to the absence of the interdomain tertiary interactions. Based on amino acid secondary structural propensity calculations, Balasco et al. 14 showed that Helix I has weaker stability as compared to Helix II, whereas the interdomain tertiary interactions with the NTD for both helices are comparable. Targeted MD results for the full RfaH show that in the presence of the NTD interactions, the structural transformation proceeds through the pathway that rearranges the amino acids in Helix II first. We further explored this process with steered MD simulations, as discussed later. As seen in Fig. 3b, during the helix unfolding the number of hydrogen bonds decreases by ~70%. Although some residual hydrogen bonds are still present in the coil state (5-6 ns), the CTD helix secondary structure is completely lost by 5 ns (Fig. 3a).

9



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

Initially, Helix II is stabilized by both CTD Helix I-Helix II interactions as well as the NTD-CTD inter-domain interactions during the structural conversion. Part of Helix II unravels to form a loop/turn at 2 ns, and then a transient 310-helix at 3 ns. The CTD interhelical contact is lost once Helix II is sufficiently destabilized. Key stages during the helix to coil conversion of the CTD are shown in Fig. 4.

Figure 4. CTD conformations at various stages of structural transformation during helix unfolding (top, left to right) and beta-folding (bottom, right to left). A closer examination of the CTD coil structure (5-6 ns) reveals that it has significant internal hydrophobic interactions, making it more like a molten-globule state than a random coil structure. This is also evidenced by only a small difference in the solvent accessible surface area (SASA) calculated for the hydrophobic amino acids in the CTD (Fig. 5, top). This hydrophobic collapsed state constitutes the intermediate structure that connects the CTD α-helix and β-barrel basins in the energy landscape. This is different from the observed mechanism for the structural transformation in the isolated CTD 13 in which the helices first unfolded to a random coil structure, followed by the folding into the β-barrel.

10


Page 10 of 26

Page 11 of 26

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


Figure 5. Solvent accessible surface area (SASA) calculated for the hydrophobic amino acids in the CTD (top-green) and at the CTD-NTD interface (bottom-red). Folding into the β-barrel: The structural rearrangement of the collapsed globule state of the CTD around 6 ns in Fig. 3a allows the formation of the first anti-parallel β-strands. At this point, the SASA for the CTD hydrophobic residues (Fig. 5, top) increases slightly and then steadily decreases as the hydrophobic residues bury due to reorganization. This also causes the SASA for the NTD-CTD interface residues (Fig. 5, bottom) to increase due to increased exposure of the hydrophobic residues. During 7-8 ns, the middle two βstrands of the CTD, β2 (residues 127-133) and β3 (residues 138-143) form first. This is followed by the alignment of β1 (residues 115-117) with β2 as well as the alignment of β4 (residues 150-155) with β3. Finally, β5 (residues 160-161) forms and aligns with β1 to complete the β-barrel structure. This sequence of β-structure formation is consistent with observations by Li et al.15 and Xiong et al.16. To investigate the dynamics of the resulting RfaH β-CTD, we performed a 100-ns conventional MD simulation. Figure S3 (inset) shows that the rmsd calculated only for the CTD is well below 5 Å, which indicates that the β-structure of the CTD is very stable at 300K. However, the absence of the interdomain contacts, combined with the flexible nature of the linker, allows large-scale diffusive motions of the β-CTD relative to the NTD, which can be important for its function. This is displayed in Fig. S3 (a) which shows an increasing rmsd for the full RfaH. In Figure S3 (b), we also plot the time evolution of the CTD-NTD center of mass separation distance (dcm), as well as the rootmean-squared distance (rmsd) from the starting structure during the 100 ns simulation and compared these with the results for the RfaH α-CTD. The rmsd and dcm for the RfaH β-CTD shows significant fluctuations as compared to those of the RfaH α-CTD. As the linker segment undergoes large fluctuations, dcm tends to increase in the β-CTD whereas the dcm for the closed form α-CTD remains constant due to CTD-NTD interactions. In addition, the rmsf for the NTD show that the residue fluctuations are comparable in the

11



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

beta and alpha forms, suggesting that the NTD integrity is not significantly perturbed by domain dissociation. However, the intradomain network connectivity in the NTD changes significantly upon domain separation (Fig. S2).We also observed that when the E48R138 interdomain salt-bridge interaction breaks, E48 makes a different ionic interaction with R11 which is within the NTD, giving additional structural stability. This R11-E48 interaction was also observed in the 100-ns simulation of the RfaH in the α-CTD form when the E48-R138 interaction is intermittently broken. Dynamics of the interdomain contacts: The RfaH-CTD in the α-helical state is stabilized by interdomain contacts. We calculated the time-evolution of the fraction of native contacts (Q) of the NTD-CTD interfacial residues and CTD intradomain contacts and plotted the results in Fig. 6a. Native contacts were defined as contacts between any heavy-atoms within 4 Å of distance in the native state configuration. NTD-CTD interfacial native contacts are almost completely lost by 4 ns. Intradomain alpha CTD contacts are lost by 6 ns, at which time the native contacts for the CTD β-structure begin to form. Interestingly, some residue pairs that were part of the native intradomian contacts in the α-CTD were also part of the native intradomian contacts in β-CTD. Specifically, residue pairs G135-K139 and A128-L141 were within a cutoff distance of ~4 Å in both the α-helix and β-structure of the CTD. Persistence of the hydrophobic contact between residues A128 and L141 is an interesting feature of the structure conversion. One of the major interdomain interactions in the RfaH with α-CTD is the saltbridge interaction between E48 and R138. Experimental observations9 show that E48S mutation favors the dissociated state leading to the β-barrel structure of the CTD. As shown in Fig. S1, alanine substitution at either E48 or R138 destabilizes this interdomain interaction and the E48S mutation increases the ∆∆G by 0.82 kcal/mol. Figure 6b shows the distance (∆r) between the anionic carboxylate oxygen of the E48 side chain and the cationic ammonium nitrogen of the R138 side chain. The distance ∆r remains around 5Å until 3 ns but increases steadily once the salt-bridge is broken. The time evolution of the interdomain contacts in Fig 6a shows a sharp decrease in Q around 4 ns, which is when the salt-bridge interaction is completely broken. Therefore, the E48-R138 salt-bridge is one of the last interdomain contacts that are still present during the structural transformation. This is explored further with SMD in the next section.

12


Page 12 of 26

Page 13 of 26

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


Figure 6. a) Fraction of the native contacts Q at the NTD-CTD the interface (red), and intradomain for the CTD for α-CTD (green) and β-CTD (blue). b) Distance (∆r) between the E48 oxygen and the R138 nitrogen that are involved in the interdomain salt-bridge interaction. C. Potential of mean force using adaptive biasing forces To further investigate the αβ structural transformation of the CTD in the full RfaH, we calculated the Potential of Mean Force (PMF) using the Adaptive Biasing Force (ABF) method implemented in NAMD, and display the results in Fig. 7. In this method, a free-energy profile is computed using a thermodynamic integration scheme in which the biasing force is adapted continuously in the Hamiltonian until the system overcomes an energy barrier. This has proven to be a very useful tool to investigate and explain protein interactions as well as dynamic pathways in a complex energy landscape53,54.

13



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

Figure 7. PMF as a function of rmsd for the structural transformation of the RfaH-CTD. Figure 7 shows the calculated PMF with the root-mean-squared-deviation (rmsd) as the reaction co-ordinate. The rmsd is calculated based on an initial reference structure that is obtained after a 100 ns equilibration simulation of the full RfaH with the alpha form of the CTD. Figure 7 shows that readjustment (AC) of the residues at the CTDNTD interface requires overcoming a large barrier B and ultimately leads the system to an especially low free-energy configuration at C. The deep basin at C corresponds to a structure that has subtle, but important changes in amino acid positions and interactions compared to the initial structure in A, which represents the native state basin. Specifically, a slight bending in the CTD Helix I allows its amino acid I118 to make a hydrophobic contact with the CTD tail residue V154 as shown in Fig 8a. Similarly, the NTD residue A91 makes a hydrophobic contact with the CTD tail residue F159. This interaction also allows the CTD tail segment (residues 158-161) to align with the NTDCTD linker segment (residues 110-113) in an antiparallel fashion. Both the linker segment and the CTD tail segment were parts of the segments that were added using Modeller and their interactions in this region seem to yield a more compact and stable structure.

14


Page 14 of 26

Page 15 of 26

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


Figure 8. a) Snapshot of the structures corresponding to state A (left) and state C (right) showing the rearrangement of the hydrophobic residues in the CTD. b) Displacement of I93 of the NTD helix in state A (left) slightly exposes the hydrophobic core (right, state C). The changes in structure between Fig. 7, Structure A and the structure corresponding to the minimum at C due to small rearrangements of the CTD may be functionally important. For RNAP binding to RfaH, the clamp helix known as β’ CH of the RNAP β-subunit binds to the RfaH NTD hydrophobic region composed of residues W4, Y54 and F56 11 (Fig. 8b). This binding of β’ CH is important for reducing the transcriptional pausing and excluding the RNAP binding of NusG-NTD 9. In the native state (A of Fig. 7 and the left side of Fig. 8b), these hydrophobic residues interact with I93 of the NTD helix and form a completely buried hydrophobic core, which may not allow the β-subunit binding. However, in structure C (right side of Fig. 8b), this NTD hydrophobic core is slightly exposed as the Y54-I93 interdomain hydrophobic contact is broken as I93 is pulled away by the CTD Helix II through a hydrophobic interaction with F126. Figure 7 shows a large barrier separating states A and C and overcoming the large barrier at B is not possible merely due to thermal fluctuations, but requires assistance. It was suggested11 that the ops element binding at another binding site (residues R16 and R73) in RfaH initiates the domain dissociation and subsequent binding of the RNAP. Consistent with this suggested mechanism, the minimum at C might constitute the structure that is accessible for RNAP binding. These structural changes present correlations with the functionally relevant structures, though we note that the allosteric changes caused by the ops binding may be different from the structural changes observed here under biasing forces. The transition from C to D in Fig. 7 involves another large barrier in the freeenergy profile. The structure corresponding to the state D exhibits significant solvent exposure of the hydrophobic core residues at the RNAP binding site. This suggests that 15



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

the RNAP β-subunit binding may facilitate the CD transition and allow further domain separation. The complete domain separation occurs when the rmsd exceeds 6 Å, with a corresponding domain-domain distance of ~16 Å. Once the domains are separated, significant structural changes (Fig. 7, Structures E, F, G) take place during the helix unfolding and beta-structure formation. The free-energy plot shows a rugged landscape, as expected for the calculations with rmsd as a reaction coordinate54. The β-structure formation occurs between rmsd 9 to 11 Å. The state E in Fig. 7 corresponds to a structure that shows complete loss of the CTD helicity as well as the initiation of the antiparallel βsheet. After overcoming a small free-energy barrier, β-structure formation is completed at state G. Most structures beyond state G are part of the native state basin of the β-CTD. Folding of the β-CTD into the α-CTD in the full RfaH: Through the assessment of RfaH transcription activities as well as NMR spectroscopy, it was shown experimentally55 that the denatured CTD spontaneously folds into the α-form in the presence of the NTD. Folding of the denatured CTD into the alpha form in the full RfaH even when the domains were swapped, suggested that the β-CTD could refold into the closed α-CTD form of RfaH. To investigate the β α refolding dynamics, we performed a 10 ns TMD for the RfaH starting in the β-CTD with the α-CTD as the target structure. All heavy-atoms in the CTD were targeted, and to promote correct interfacial association of the CTD with the NTD, several interfacial residues in the NTD were also included as targeted residues. With this, the resulting final structure was very close to the original RfaH with α-CTD. The observed folding and unfolding pathways are different for the forward and back conversions. The time evolution of the secondary structure for the β α conversion is displayed in Fig. S4, which shows that the unfolding of the β-barrel takes longer than the unfolding of the helix bundle during the α β conversion in Fig. 3. Figure S4 also shows the changes in the number of CTD hydrogen bonds as well as the changes of the inter-domain hydrogen bonds during the structural conversion. The unfolding of the β-barrel proceeds by first breaking the β1-β5 contacts, followed by the collapse of the hydrophobic core of the β-barrel. A significant rearrangement of the hydrophobic residues takes place before the alpha structure begins to form. This rearrangement involves formation of multiple non-native intermediary contacts. Immediately after the β-structure is unfolded around 6 ns, a transient, non-native saltbridge interaction forms between residues R11 and D134 and remains intact for more than 2 ns. During this time, α-helix secondary structure begins to form. Folding of the αhelices, followed by the formation of the E48- R138 native salt-bridge contact completes the refolding of the RfaH to its α-CTD closed form. D. Interdomain and CTD inter-helical interactions using SMD To further investigate the NTD-CTD interdomain, as well as the CTD intradomain interactions, we performed Steered Molecular Dynamics (SMD) simulations

16


Page 16 of 26

Page 17 of 26

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


on the full RfaH. Before applying the pulling force, the system was first equilibrated for 5 ns in an NPT simulation in a long rectangular box with enough size to accommodate the pulling. The N-terminal Cα atom was held fixed and a dummy atom pulls the C-terminal atom Cα as the SMD atom for 10 ns, followed by another 10 ns without pulling to allow the molecular system to relax from mechanical stretching. This process was repeated until the SMD atom was close to the box edge, with a final end-to-end distance (dee) of ~130 Å. Thus, SMD simulations were performed for intervals (0-10, 20−30, 40−50, 60−70, 80−90, 100−110, 120−130) ns with relaxation simulations in-between. The pulling/relaxation simulation is displayed in Movie S1. Figure 9 shows representative snapshots of various intermediate stages during the mechanical stretching. After the initial rearrangement of the CTD loop segment, the positions of the CTD helices as well as the NTD helix at the interface are affected around 40 ns. Consistent with our description given above with respect to Figs. 7 and 8 and the ABF simulations, Fig. 9 shows (40 ns) that the NTD helix initially moves together with the CTD helices, opening up the NTD hydrophobic pocket at the RNAP binding site (residues W4, Y54 and F56). After 45 ns of pulling (80 ns frame in Fig. 9), contacts between the CTD helices and the NTD helix weaken, which ultimately lead to the separation of the CTD from the NTD helix. As shown in Fig. S5, the SASA for the hydrophobic core opens up around 100 ns due to a slight movement of I93.

Figure 9. Snapshots of protein conformations at 1 ns, 40 ns, 80 ns, and 120 ns from the SMD pulling and relaxation trajectory. 17



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

Figure 10. Various dynamic and structural parameters as a function of time during SMD: a) force profiles during pulling intervals, b) number of hydrogen bonds during SMD (red) and relaxation (blue) c) secondary structural conversion, and d) Distance (∆r) between the E48 oxygen and the R138 nitrogen. The SMD simulation is set up so that the dummy atom (attached to the CTD Cα SMD atom) travels at constant velocity while being pulled. This requires a pulling force that varies in strength in response to the protein’s resistance. The helical form of the CTD is stabilized by interdomain interactions with the NTD, which include the E48-R138 saltbridge (involving the CTD Helix II) and other hydrophobic and hydrogen bond (involving both CTD helices). The force profile displayed in Fig. 10a shows a general increase in the pulling force with time until a large drop at 80 ns. The force rises again, and another significant drop in force is observed around 90 ns that coincides with breaking of major interdomain interactions, including the E48-R138 salt-bridge and some hydrogen bonds. Hydrogen bond analysis (Fig. 10b) shows an initial increase in hydrogen bonds due to rearrangement of the CTD-tail segment (residues 158-162). Figure 10c displays the changes in secondary structure of all residues during the pulling and relaxation. At around 60 ns, a significant reduction in helicity is observed for the CTD Helix II, whereas Helix I is relatively unperturbed and retains most of its secondary structure. This is consistent with our results presented above in Fig. 3. As shown in Fig. S6, the interdomain interacting hydrophobic residues F33-F130 separate in two stages. 18


Page 18 of 26

Page 19 of 26

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


They begin to separate and become more exposed to the solvent (increasing solvent accessible surface area - SASA) around 80ns, which allows solvent molecules to penetrate the interfacial region, weakening the interdomain interactions 56. This facilitates breaking of the E48-R138 salt-bridge interaction around 90 ns (increasing ∆r in Fig. 10d) and increases the SASA of E48-R138 (Fig. S6). As the E48-R138 salt-bridge breaks, the CTD moves far away from the NTD and the F33-F130 separate further, resulting in a larger SASA. E. Potential of mean force calculation We performed nine independent 60 ns SMD simulations (pulling only) to calculate the potential of mean force (PMF) using the Jarzynski equality, which relates the non-equilibrium work done (W) with the PMF as ∆ = 〈 〉 where, 〈 〉 refers to the ensemble average. The free-energy profile can be obtained from ∆ = 〈〉 −

〈 〉〈〉 !"

(1)

where k is the Boltzmann constant. The harmonic force applied to the virtual spring attached to the C-terminal Cα was integrated over the pulled distance to calculate the "

work done over the simulation time using = #& $ % , where v is the pulling speed, f(t) is the force as a function of time, and v dt=dx is a small increment in the pulling distance. All nine simulations were used in eq. 1 to calculate the average and the standard deviation for calculating the free-energy profile. Figure 11 shows the resulting free-energy profile as a function of M1-CαL162-Cα end-to-end distance, dee. State B at 40 Å (~15 ns) represents the native state minimum after the rearrangement of the 158-162 CTD tail segment residues from structure A, and before any structural changes occurred at the NTD-CTD interface. From B to C, the initial transient interactions of the tail segment with the NTD are broken, followed by the rugged energy landscape from state C to D representing multiple structural transitions. The minimum at D around 110 Å (~ 56 ns) corresponds to the structure in which the CTD is mostly detached but the E48-R138 salt-bridge is still intact. The final, major barrier between D and E corresponds to breaking of the salt-bridge interaction. This is followed by the complete detachment of the CTD at E, where all interdomain contacts are lost. At this stage, the CTD-helices still retain their secondary structure as a helix-bundle. The PMF in Fig. 11 calculated from SMD describes the results of the pulling simulations that are relevant for AFM experiments.

19



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

Figure 11. PMF as a function of end-to-end (Cα-Cα) distance between M1 and L162. IV. CONCLUSION We investigated the mechanisms of the structural transformation of the CTD in the full RfaH using various molecular dynamics techniques. Through dynamic network analysis, we determined the communities of amino acids that are involved in interdomain and intradomain communications. Examination of these communities showed that amino acid F126 in the CTD shows correlated motions with the NTD amino acids I93, S97, and L96. It also affects the RNAP binding site through interaction with I93, making it a possible functionally relevant residue. Using targeted MD, we investigated the αβ as well as the βα structural transformation of the CTD in the full RfaH. The observed folding and unfolding pathways are different for the forward and back conversions. The observed mechanism for the folding of the CTD into the β-barrel structure is consistent with those from previous computational studies of the isolated CTD. In the presence of the NTD, however, the CTD helix unfolding process leading to the initiation of the βstructure is different. Instead of going through a completely unfolded random coil state, the α-helix to β-structure transition occurs through a collapsed globule state in which significant residual hydrophobic interactions are present. Also, the order of the unfolding of the CTD helices is reversed due to interactions with the NTD. We used the Adaptive Biasing Force method to calculate the free-energy profile for the structural transformation. We find that the CTD-NTD interdomain interactions constitute the main barrier in the CTD α-helix to β-barrel structural conversion. Once the interfacial interactions are broken, the structural conversion of the CTD is relatively easy. SMD calculations elucidated specific amino acids at the interdomain coupling between

20


Page 20 of 26

Page 21 of 26

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


movement of the CTD helices and the NTD helix, and show how solvent accessible surface area is affected at the RNAP binding site. Supporting Information Movie S1 displays the 130-ns pulling and relaxation during SMD simulation. Additional figures showing the time-evolution of the structural parameters as well as the rmsf are also presented. This material is available free of charge via the Internet at http://pubs.acs.org. ACKNOWLEDGMENTS We thank the Instructional & Research Computing Center (IRCC) at Florida International University for providing computing resources.

21



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

References: (1) Anfinsen, C. B. Principles that govern the folding of protein chains. Science 1973, 181, 223-230. (2) Jeffery, C. J. Moonlighting proteins--an update. Mol. Biosyst. 2009, 5, 345-350. (3) Jeffery, C. J. Molecular mechanisms for multitasking: recent crystal structures of moonlighting proteins. Curr. Opin. Struct. Biol. 2004, 14, 663-668. (4) Tokuriki, N.; Tawfik, D. S. Protein dynamism and evolvability. Science 2009, 324, 203-207. (5) Tompa, P.; Szasz, C.; Buday, L. Structural disorder throws new light on moonlighting. Trends. Biochem. Sci. 2005, 30, 484-489. (6) Tuinstra, R. L.; Peterson, F. C.; Kutlesa, S.; Elgin, E. S.; Kron, M. A.; Volkman, B. F. Interconversion between two unrelated protein folds in the lymphotactin native state. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 5057-5062. (7) Mapelli, M.; Massimiliano, L.; Santaguida, S.; Musacchio, A. The Mad2 conformational dimer: structure and implications for the spindle assembly checkpoint. Cell 2007, 131, 730-743. (8) Knauer, S. H.; Artsimovitch, I.; Rösch, P. Transformer proteins. Cell Cycle 2012, 11, 4289-4290. (9) Burmann, B. M.; Knauer, S. H.; Sevostyanova, A.; Schweimer, K.; Mooney, R. A.; Landick, R.; Artsimovitch, I.; Rosch, P. An alpha helix to beta barrel domain switch transforms the transcription factor RfaH into a translation factor. Cell 2012, 150, 291-303. (10) Mooney, R. A.; Schweimer, K.; Rosch, P.; Gottesman, M.; Landick, R. Two structurally independent domains of E. coli NusG create regulatory plasticity via distinct interactions with RNA polymerase and regulators. J. Mol. Biol. 2009, 391, 341-358. (11) Belogurov, G. A.; Sevostyanova, A.; Svetlov, V.; Artsimovitch, I. Functional regions of the N-terminal domain of the antiterminator RfaH. Mol. Microbiol. 2010, 76, 286-301. (12) Belogurov, G. A.; Mooney, R. A.; Svetlov, V.; Landick, R.; Artsimovitch, I. Functional specialization of transcription elongation factors. EMBO J. 2008, 28, 112-122. (13) GC, J. B.; Bhandari, Y. R.; Gerstman, B. S.; Chapagain, P. P. Molecular dynamics investigations of the alpha-helix to beta-barrel conformational transformation in the RfaH transcription factor. J. Phys. Chem. B 2014, 118, 5101-5108. (14) Balasco, N.; Barone, D.; Vitagliano, L. Structural conversion of the transformer protein RfaH: new insights derived from protein structure prediction and molecular dynamics simulations. J. Biomol. Struct. Dyn. 2015, 33, 1-7. (15) Li, S.; Xiong, B.; Xu, Y.; Lu, T.; Luo, X.; Luo, C.; Shen, J.; Chen, K.; Zheng, M.; Jiang, H. Mechanism of the all-α to all-β conformational transition of RfaH-CTD: Molecular dynamics simulation and Markov state model. J. Chem. Theory Comp. 2014, 10, 2255-2264. (16) Xiong, L.; Liu, Z. Molecular dynamics study on folding and allostery in RfaH. Proteins 2015, 83, 1582-1593.

22


Page 22 of 26

Page 23 of 26

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


(17) Ramirez-Sarmiento, C. A.; Noel, J. K.; Valenzuela, S. L.; Artsimovitch, I. Interdomain contacts control native state switching of RfaH on a dual-funneled landscape. PLoS Comput. Biol. 2015, 11, e1004379. (18) Isralewitz, B.; Gao, M.; Schulten, K. Steered molecular dynamics and mechanical functions of proteins. Curr. Opin. Struct. Biol. 2001, 11, 224-230. (19) Phillips, J. C.; Braun, R.; Wang, W.; Gumbart, J.; Tajkhorshid, E.; Villa, E.; Chipot, C.; Skeel, R. D.; Kale, L.; Schulten, K., Scalable molecular dynamics with NAMD. J. Comput. Chem. 2005, 26, 1781-1802. (20) Ma, J.; Sigler, P. B.; Xu, Z.; Karplus, M. A dynamic model for the allosteric mechanism of GroEL. J. Mol. Biol. 2000, 302, 303-313. (21) Cheng, X.; Wang, H.; Grant, B.; Sine, S. M.; McCammon, J. A. Targeted molecular dynamics study of C-loop closure and channel gating in nicotinic receptors. PLoS Comput. Biol. 2006, 2, e134. (22) Mashl, R. J.; Jakobsson, E. End-point targeted molecular dynamics: large-scale conformational changes in potassium channels. Biophys. J. 2008, 94, 4307-4319. (23) Compoint, M.; Picaud, F.; Ramseyer, C.; Girardet, C. Targeted molecular dynamics of an open-state KcsA channel. J. Chem. Phys. 2005, 122, 134707. (24) Kong, Y.; Shen, Y.; Warth, T. E.; Ma, J. Conformational pathways in the gating of Escherichia coli mechanosensitive channel. Proc. Natl. Acad. Sci. U. S. A. 2002, 99, 5999-6004. (25) Mendieta, J.; Gago, F.; Ramirez, G. Binding of 5'-GMP to the GluR2 AMPA receptor: insight from targeted molecular dynamics simulations. Biochemistry 2005, 44, 14470-14476. (26) Li, C.; Zhu, Y.; Wang, Y.; Chen, G. Molecular dynamics simulation on the conformational transition of the mad2 protein from the open to the closed state. Int. J. Mol. sciences 2014, 15, 5553-5569. (27) Park, S.; Schulten, K. Calculating potentials of mean force from steered molecular dynamics simulations. J. Chem. Phys. 2004, 120, 5946-5961. (28) Hamed, E.; Keten, S. Hierarchical cascades of instability govern the mechanics of coiled coils: helix unfolding precedes coil unzipping. Biophys. J. 2014, 107, 477484. (29) Kwansa, A. L.; De Vita, R.; Freeman, J. W. Mechanical recruitment of N- and Ccrosslinks in collagen type I. Matrix biol. 2014, 34, 161-169. (30) Liu, Y.; Hsin, J.; Kim, H.; Selvin, P. R.; Schulten, K. Extension of a three-helix bundle domain of myosin VI and key role of calmodulins. Biophys. J. 2011, 100, 2964-2973. (31) Sotomayor, M.; Schulten, K. Single-molecule experiments in vitro and in silico. Science 2007, 316, 1144-1148. (32) Zheng, W. All-atom and coarse-grained simulations of the forced unfolding pathways of the SNARE complex. Proteins 2014, 82, 1376-1386. (33) He, J.; Xu, L.; Zhang, S.; Guan, J.; Shen, M.; Li, H.; Song, Y. Steered molecular dynamics simulation of the binding of the beta2 and beta3 regions in domainswapped human cystatin C dimer. J. Mol. Model. 2013, 19, 825-832. (34) Cuendet, M. A.; Michielin, O. Protein-protein interaction investigated by steered molecular dynamics: the TCR-pMHC complex. Biophys. J. 2008, 95, 35753590. 23



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

(35) Giorgino, T.; De Fabritiis, G. A High-throughput steered molecular dynamics study on the free energy profile of ion permeation through Gramicidin A. J. Chem. Theory and Comput. 2011, 7, 1943-1950. (36) Cheung, L. S.; Shea, D. J.; Nicholes, N.; Date, A.; Ostermeier, M.; Konstantopoulos, K. Characterization of monobody scaffold interactions with ligand via force spectroscopy and steered molecular dynamics. Scientific reports 2015, 5, 8247. (37) Xu, L.; Hasin, N.; Shen, M.; He, J.; Xue, Y.; Zhou, X.; Perrett, S.; Song, Y.; Jones, G. W. Using steered molecular dynamics to predict and assess Hsp70 substrate-binding domain mutants that alter prion propagation. PLoS Comput. Biol. 2013, 9, e1002896. (38) Guzman, D. L.; Randall, A.; Baldi, P.; Guan, Z. Computational and singlemolecule force studies of a macro domain protein reveal a key molecular determinant for mechanical stability. Proc. Natl. Acad. Sci. U. S. A. 2010, 107, 19891994. (39) Patel, J. S.; Berteotti, A.; Ronsisvalle, S.; Rocchia, W.; Cavalli, A. Steered molecular dynamics simulations for studying protein-ligand interaction in cyclindependent kinase 5. J. Chem. Info. Model. 2014, 54, 470-480. (40) Eswar, N.; Webb, B.; Marti-Renom, M. A.; Madhusudhan, M. S.; Eramian, D.; Shen, M. Y.; Pieper, U.; Sali, A. Comparative protein structure modeling using Modeller. Current Protocols in Bioinformatics; John Wiley & Sons, Inc., Supplement 15, 2006. (41) Sanchez, R.; Sali, A. Comparative protein structure modeling. Introduction and practical examples with modeller. Methods Mol. Biol. 2000, 143, 97-129. (42) Humphrey, W.; Dalke, A.; Schulten, K. VMD: visual molecular dynamics. J. Mol. Graph 1996, 14, 33-38. (43) Vanommeslaeghe, K.; Hatcher, E.; Acharya, C.; Kundu, S.; Zhong, S.; Shim, J.; Darian, E.; Guvench, O.; Lopes, P.; Vorobyov, I.; Mackerell, A. D., Jr. CHARMM general force field: A force field for drug-like molecules compatible with the CHARMM allatom additive biological force fields. J. Comput. Chem. 2010, 31, 671-690. (44) Ulrich, E.; Lalith, P.; Max, L. B.; Tom, D.; Hsing, L.; Lee, G. P. A smooth particle mesh Ewald method. J. Chem. Phys. 1995, 103, 8577-8593. (45) Darve, E.; Rodriguez-Gomez, D.; Pohorille, A. Adaptive biasing force method for scalar and vector free energy calculations. J. Chem. Phys. 2008, 128, 144120. (46) Hénin, J.; Fiorin, G.; Chipot, C.; Klein, M. L. Exploring multidimensional free energy landscapes using time-dependent biases on collective variables. J. Chem. Theory Comput. 2010, 6, 35-47. (47) Zheng, W. Anharmonic normal mode analysis of elastic network model improves the modeling of atomic fluctuations in protein crystal structures. Biophys. J. 2010, 98, 3025-3034. (48) Martin, H. S.; Jha, S.; Coveney, P. V. Comparative analysis of nucleotide translocation through protein nanopores using steered molecular dynamics and an adaptive biasing force. J. Comput. Chem. 2014, 35, 692-702. (49) Girvan, M.; Newman, M. E. Community structure in social and biological networks. Proc. Natl. Acad. Sci. U. S. A. 2002, 99, 7821-7826. (50) Sethi, A.; Eargle, J.; Black, A. A.; Luthey-Schulten, Z. Dynamical networks in tRNA:protein complexes. Proc. Natl. Acad. Sci. U. S. A. 2009, 106, 6620-6625. 24


Page 24 of 26

Page 25 of 26

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


(51) McClendon, C. L.; Kornev, A. P.; Gilson, M. K.; Taylor, S. S. Dynamic architecture of a protein kinase. Proc. Natl. Acad. Sci. U. S. A. 2014, 111, E4623E4631. (52) Schymkowitz, J.; Borg, J.; Stricher, F.; Nys, R.; Rousseau, F.; Serrano, L. The FoldX web server: an online force field. Nucleic acids Res. 2005, 33, W382-W388. (53) Bai, Q.; Zhang, Y.; Li, X.; Chen, W.; Liu, H.; Yao, X. Computational study on the interaction between CCR5 and HIV-1 entry inhibitor maraviroc: insight from accelerated molecular dynamics simulation and free energy calculation. Phys. Chem. Chem. Phys. 2014, 16, 24332-24338. (54) Gumbart, J. C.; Roux, B.; Chipot, C. Standard binding free energies from computer simulations: What is the best strategy? J. Chem. Theory Comput. 2013, 9, 794-802. (55) Tomar, S. K.; Knauer, S. H.; NandyMazumdar, M.; Rösch, P.; Artsimovitch, I. Interdomain contacts control folding of transcription factor RfaH. Nucleic acids Res. 2013, 41, 10077-10085. (56) Lemkul, J. A.; Bevan, D. R. A comparative molecular dynamics analysis of the amyloid beta-peptide in a lipid bilayer. Arch. Biochem. Biophys. 2008, 470, 54-63.

25



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

Table of Content Graphic

26


Page 26 of 26

The Role of the Interdomain Interactions on RfaH Dynamics and

Recommend Documents