or SAXS Data for

Subscriber access provided by WEBSTER UNIV

Review

MD Simulations Combined with NMR and/or SAXS Data for Characterizing IDP Conformational Ensembles Maud Chan-Yao-Chong, Dominique Durand, and Tâp Ha-Duong J. Chem. Inf. Model., Just Accepted Manuscript • DOI: 10.1021/acs.jcim.8b00928 • Publication Date (Web): 06 Mar 2019 Downloaded from http://pubs.acs.org on March 7, 2019

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

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

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

Journal of Chemical Information and Modeling

MD Simulations Combined with NMR and/or SAXS Data for Characterizing IDP Conformational Ensembles Maud Chan-Yao-Chong1,2 , Dominique Durand2 and Tap Ha-Duong1,∗ 1 BioCIS,

Université Paris-Sud, CNRS, Université Paris-Saclay, 92290

Chˆ atenay-Malabry, France 2 Institut

de Biologie Intégrative de la Cellule, Université Paris-Sud, CNRS, Université

Paris-Saclay, 91400 Orsay, France. ∗ Corresponding

author: [email protected]

Abstract The concept of intrinsically disordered proteins (IDPs) has emerged relatively slowly, but over the past twenty years it has become an intense research area in structural biology. Indeed, due to their considerable flexibility and structural heterogeneity, the determination of IDP conformational ensemble is particularly challenging and often requires a combination of experimental measurements and computational approaches. With the improved accuracy of all-atom force fields and the increasing computing performances, molecular dynamics (MD) simulations have become more and more reliable to generate realistic conformational ensembles. And the combination of MD simulations with experimental approaches, such as nuclear magnetic resonance (NMR) and/or small-angle X-ray scattering (SAXS) allows to converge toward a more accurate and exhaustive description of IDP structures. In this review, we discuss the state of the art of MD simulations of IDP conformational ensembles, with a special focus on studies which back-calculated and directly compared theoretical and experimental NMR or SAXS observables, such as chemical shifts (CS), 3 Jcouplings (3 Jc), residual dipolar couplings (RDC), or SAXS intensities. We organize the review in three parts. In the first section, we discuss the studies which used NMR and/or SAXS data to test and validate the development of force fields or enhanced sampling techniques. In the second part, we explore different methods for the refinement of MDderived structural ensembles, such as NMR or SAXS data-restrained MD simulations or ensemble reweighting to better fit experiments. Finally, we survey some recent studies combining MD simulations with NMR and/or SAXS measurements to investigate the

1

ACS Paragon Plus Environment

Journal of Chemical Information and Modeling 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

relationship between IDP conformational ensemble and biological activity as well as their implication in human diseases. From this review, we noticed that quite few studies compared MD-generated conformational ensembles with both NMR and SAXS measurements to validate IDP structures at both local and global levels. Yet, beside the IDP propensity to form local secondary structures, their dynamic extension or compactness appears also important for their activity. Thus, we believe that a close synergy between MD simulations, NMR, and SAXS experiments would be greatly appropriate to address the challenges of characterizing the disordered structures of proteins and their complexes in relation to their biological functions.

Introduction In contrast to folded proteins, intrinsically disordered proteins (IDPs) sample very heterogeneous conformations and cannot be described by one or a few three-dimensional structures but rather by an ensemble of many diverse ones. Several biophysical experiments can be employed to structurally characterize IDP conformational ensembles [1–6], among which nuclear magnetic resonance (NMR) spectroscopy and small-angle X-ray scattering (SAXS) appear simple-to-apply and very complementary. Indeed, NMR measurements of residue-specific chemical shifts (CS) or residual dipolar couplings (RDC) provide information on IDP propensity to form transient local secondary structures [7, 8], while SAXS intensities can inform about IDP bound/unbound state, degree of disorder, as well as their global shape and size [9, 10]. However, a problem inherent to disordered proteins is to infer ensembles of three-dimensional conformations from sparse or low-resolution data. This degeneracy issue (different ensembles can fit the same experimental data) makes the determination of IDP conformational ensembles particularly challenging. There are mainly three different strategies to derive IDP conformational ensembles using experimental information [3, 11–14] (Fig. 1). The first one consists in using experimental data as restraints to guide the generation of an ensemble that agree with experiments (Restrained sampling). The second one consists in generating a pool of conformations and then optimizing the statistical weight of each structure to minimize the deviation between theoretical and experimental quantities (Ensemble reweighting).

2


Page 2 of 49

Page 3 of 49 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


Finally, the third strategy consists in generating a conformational ensemble as realistic as possible and assessing its quality against experimental measurements (Ensemble validation).

Figure 1: Strategies to generate IDP conformational ensemble that fit experimental data. NMR data, such as chemical shifts (CS), 3 J-couplings (3 Jc), or residual dipolar couplings (RDCs) provide information on IDP propensity to form transient local secondary structures. Small-angle X-ray scattered intensities I(q) as a function of the scattering momentum transfer q provide information regarding the shape of the scattering object and its size through the radius of gyration Rg , estimated from the Guinier approximation: I(q) = I(0)exp(−q 2 Rg2 /3) for sufficiently small values of q [15]. In these strategies, ensembles of IDP conformations can be generated using several computational approaches [11, 12, 16], among which statistical coil generators and molecular dynamics (MD) simulations are the most popular. The general principle of statistical coil generators, such as TraDES [17], Flexible-Meccano (FM) [18], or the more recent database-Markov method [19] and three-residue-based sampling [20], is to construct multiple conformations of a polypeptide main chain by assigning, to each amino acid, some φ/ψ values randomly taken in residue-specific data sets derived from experimentally determined structures of proteins. These very fast methods are in essence knowledge-based approaches which generally require, to be accurate, the input of amino acid secondary

3



structure propensities, either determined by experiments or predicted by computational techniques. An alternative to statistical coil generators is the MD simulation approach which can potentially generate realistic IDP conformational ensembles. Indeed, based on physics principles, MD simulations naturally provide the Boltzmann probability or statistical weight of each structure of a conformational ensemble, avoiding the problem of degenerate solutions. However, this approach has to overcome two limitations to provide relevant results: first, the force fields have to describe the interatomic interactions as accurately as possible, particularly, those between water molecules and apolar amino acids which are frequently exposed to solvent in disordered polypeptide chains. Recent improvements in this direction were made for the classical force fields AMBER and CHARMM [21– 23]. Secondly, explorations of IDP conformational ensembles have to be as exhaustive as possible, which generally requires computationally expensive calculations which can last up to several months. Several excellent review articles discussed the developments and applications of enhanced sampling techniques to generate statistically correct IDP conformational ensembles [16, 24, 25]. Given the better accuracy of all-atom force fields and the more and more powerful high performance computers, MD simulations have become a major approach to correctly characterize protein conformational ensembles. In the present review, we provide a survey of the recent achievements of MD simulations for generating realistic conformational ensembles of various IDP that agree with experimental data. We will particularly emphasize the MD studies that were able to fit both NMR and SAXS measurements, indicating both local and global correct structures. We hope that this review will encourage the molecular modeling community to validate IDP ensembles with both NMR and SAXS data and possibly deposit their structures in the Protein Ensemble Database (pE-DB) [26]. In the first part of this review, we will recall the studies that use experimental data to assess the MD simulation approach, especially those which develop force fields and enhanced sampling techniques for IDP. In the second part, we will shortly survey the methods to refine IDP conformational ensembles generated by MD simulations. Lastly, we will discuss the contribution of recent MD studies combining NMR and SAXS measurements to our understanding of IDP biological functions, notably the role of the transient secondary structures and compactness in their binding mechanism to protein partners.

4


Page 4 of 49

Page 5 of 49 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


It should be noted that studies using paramagnetic relaxation enhancement (PRE) experiments will not be considered in this review. The results provided by this NMR technique are more difficult to process than standard NMR experiments, as shown by Sasmal et al. [27], because it requires to introduce a reporter tag which can induce some perturbations in structural ensembles. Similarly, this review will not include many studies of IDPs that use fluorescence resonance energy transfer (FRET) measurements which also provides useful information on IDP overall size [28, 29].

1

NMR and SAXS data for validating MD simulations

1.1

Toward force fields for IPDs

In the late 20th century, the most popular all-atom force fields for proteins were GROMOS96 [30], OPLS-AA [31], CHARMM22 [32], or AMBER99 [33] (Fig. 2). These force fields were optimized to well reproduce the structure and dynamics of folded proteins, such as enzyme and receptors, in explicit solvent, using generally a three-site water model, such as the SPC/E [34] or TIP3P [35]. However, with the emergence of IDPs in structural biology, limitations in accuracy of modeling protein disordered regions appeared at the beginning of the 2000s. More specifically, it was pointed out that MD simulations with traditional force fields often generated conformational ensembles of disordered segments with over-estimated propensities to form secondary structures and overly compact shapes when compared to experimental data. Since then, great efforts were made to develop new force fields optimized for both structured and disordered parts of proteins. In this section, we survey these developments which have been reported in a about fifteen publications since 2001. Early efforts to optimize force fields were focused on improving the backbone φ and ψ dihedral potentials to better reproduce the two-dimensional Ramachandran probability distributions, notably resulting in the refined force fields OPLS-AA/L [36] and AMBER03 [37]. More explicitly, MacKerell et al. introduced in CHARMM22 cross term energy functions of φ and ψ angles to better account for their correlations. These grid-based energy correction maps (CMAP) gave the name CHARMM22-CMAP to this new force field, also referred to as CHARMM27 [38, 39]. Since, many improvements of the backbone φ and ψ dihedral energy potentials were continuously made, leading to a serie of force fields more or less able to well reproduce polypeptide propensities to form local secondary

5



Page 6 of 49

structures. This includes the AMBER99SB [40], AMBER99SB∗ , AMBER03∗ [41], AMBER03w [42], CHARMM22∗ [43], CHARMM36 [44], and AMBER14SB [45] force fields (Fig. 2). It could be noted that some of these models were subsequently completed with improved sides chain torsion potentials to refine side chain rotamers, resulting in the ”-ILDN” extension of some force field names, such as AMBER99SB-ILDN [46]. More recently, updated CMAP potentials were implemented in several force fields to specifically correct the φ and ψ probability distributions of intrinsically disordered proteins, leading to the new force fields AMBER99IDPs [47], AMBER14IDPSFF [48], CHARMM36m [23], and CHARMM36IDPSFF [49]. From a methodological viewpoint, one should recall that all these developments were mainly validated with comparisons between back-calculated and experimentally measured NMR observables, such as scalar 3 J-couplings or residual dipolar couplings (RDCs), on both folded and unstructured polypeptides. Most recent ones, such as CHARMM36m, also used SAXS data for additional validations.

Figure 2: Developments of force fields and water models for IDPs. SPC/E [50].

(b) CHARMM-TIP3P:

(c) TIP4P-Ew:

(a) SPC:

initial model of

specific TIP3P model for CHARMM force fields [32].

reparameterization of the TIP4P model for use with Ewald techniques [51].

(d) AMBER12SB:

preliminary version of AMBER14SB [52].

In spite of these improvements in reproducing local structures of folded and disordered polypeptides, several studies pointed out another limitation of current force fields, that is they generally produce overly compact conformations of disordered proteins [22, 53– 55]. The main reason of this shortcoming is now understood and originates from the fact that current water models underestimate their dispersion interactions with amino acids, resulting in protein structures that are too poorly solvated [21, 22]. In the recent

6


Page 7 of 49 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


literature, we found three different approaches to remedy this situation. First, Best et al. proposed to slightly modify their previous combination of AMBER03w force field [42] with the highly optimized TIP4P/2005 water model [56]: they simply rescaled by a factor γ = 1.10 the Lennard-Jones well depth parameter Oi between the water oxygen and all protein atoms, resulting in the force field name AMBER03ws [21] (in which the water model is thus slightly different from TIP4P/2005 and can be renamed TIP4P/2005s). In a more force field agnostic approach, Piana et al. introduced a new water model called TIP4P-D derived from the model TIP4P/2005 [56] but with a dispersion parameter C6 = 900 (kcal/mol).˚ A6 significantly larger (by about 50%) than in other popular water models [22]. This water model was showed to significantly improve the average radius of gyration of several IDPs simulated with several force fields. Notably, a combination of TIP4P-D with a derived version of AMBER99SB-ILDN with optimized torsion potentials, resulting in the force field AMBER99SB-disp, allowed to obtain very good accuracy in simulations of several IDPs [57]. Lastly, a third approach was reported in the communication of CHARMM36m force field [23], in which the authors proposed to accentuate the Lennard-Jones well depth parameter H of the water hydrogen of the specific CHARMM-TIP3P model from -0.046 to -0.10 kcal/mol (a model that could be named CHARMM-TIP3Pm to be distinguished from the original one). This change improved the computed average radius of gyration of some proteins but also deteriorated results for other IDPs, suggesting that this H parameter could be further optimized [23]. Regarding the methodology of these developments, it should be emphasized that all these improvements were validated by comparing back-calculated SAXS profiles against experimental intensities. It should be stressed that comparisons of only theoretical and experimental radii of gyration are partial assessments and do not represent complete validations of IDP conformational ensembles.

1.2

Force field benchmark using NMR and/or SAXS data

Over the past four years, several benchmarks of previously mentioned force fields were performed on various disordered peptides in order to guide the choice of the most appropriate ones for IDP simulations. The IDPs which frequently served as test cases are amyloidogenic peptides, particularly Aβ40 and Aβ42 related to the Alzheimer’s disease, or the 37-residue IAPP involved in the type II diabetes. These disordered peptides are known to have a significant propensity to form transient β-strands accountable for their

7



aggregation and fibrillization. When the force field capability to reproduce their secondary structure propensities is tested against NMR measurements of chemical shifts, 3 J-couplings,

or residual dipolar couplings, then a general trend seems to emerge: AM-

BER03 and CHARMM22 are not appropriate for this kind of β-peptides since they over-stabilize α-helices, while OPLS-AA and GROMOS96 similarly produces unstructured ensembles with low secondary structures propensities [58–60]. In contrast, AMBER99SB, AMBER99SB-ILDN, AMBER99SB∗ -ILDN, and AMBER03w provides similar results in good agreement with NMR data, and CHARMM22∗ slightly performs better than all the previously mentioned force fields [59–61]. This hierarchy is overall confirmed by other studies on shorter peptides of about 10 to 20 residues [62, 63]. Interestingly, AMBER99SB-ILDN and AMBER99SB∗ -ILDN were also shown to produce secondary structure propensities consistent with NMR data for disordered peptides with α-helix tendencies, such as the central region (380-490) of axin-1 protein [64] or the TAD2 peptide (41-62) from the human p53 tumor suppressor [65]. Lastly, it could be noted that, when the CMAP energy correction term for backbone (φ, ψ) dihedrals is implemented in a AMBER99SD-ILDN, AMBER14SB, or CHARMM36m, then the resulting IDP-specific force fields AMBER99IDPs, AMBER14IDPSFF, and CHARMM36IDPSFF, respectively, seem to systematically improve the agreement between the back-calculated chemical shifts and NMR measurements for dozens of known disordered peptides and proteins [48, 49, 66, 67]. However, although the tested force fields yield local structures consistent with NMR chemical shifts, 3 J-couplings, or residual dipolar couplings, it is not clear whether they can well reproduce the IDP global shape and size. The assessment of this capacity can be performed using comparisons between back-calculated and experimental SAXS intensities, or at least between theoretical and experimental radii of gyration. From studies which reported such comparisons, it appears that the model of water is a key determinant for accurate MD simulations of IDPs. First, it seems that three-site water models TIP3P and SPC/E nearly always induce overly collapsed conformations, this general trend being reported by several studies on various disordered peptides or proteins [22, 55, 68, 69, 71]. Regarding the four-site water models, some MD studies on several IDPs indicate that TIP4P and TIP4P-Ew models are also unable to yield average radii of gyration consistent with experimental measurements (Table 1) [22, 69]. Actually, it appears that only the two water models TIP4P/2005s, in association with AMBER03ws [21, 68], and TIP4P-

8


Page 8 of 49

Page 9 of 49 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


IDP

Force field

Water model

< Rg > (nm)

ACTR

AMBER03w

TIP4P/2005

1.5

(71 residues)

AMBER03ws

TIP4P/2005s

2.1

AMBER03w

TIP4P/2005

1.3

AMBER03ws

TIP4P/2005s

1.5

CHARMM22∗

TIP4P-D

1.4

AMBER99SB-ILDN

TIP4P-Ew

1.7

α-Synuclein

AMBER12

TIP4P-D

3.0

(140 residues)

AMBER99SB-ILDN

TIP4P-D

2.5

CHARMM22∗

TIP4P-D

2.3

α-Synuclein

AMBER03ws

TIP4P/2005s

3.0

(140 residues)

CHARMM22∗

TIP4P-D

2.2

AMBER12

TIP4P-Ew

1.9

Protymosin α

AMBER12

TIP4P-D

4.0

(111 residues)

AMBER99SB-ILDN

TIP4P-D

4.3

CHARMM22∗

TIP4P-D

4.6

AMBER99SB∗ -ILDN

TIP4P

1.3

AMBER99SB∗ -ILDN

TIP4P-D

2.1

OPLS-AA/L

TIP4P

1.3

Histatin 5

AMBER99SB-ILDN

TIP4P-D

1.3

(24 residues)

AMBER03ws

TIP4P/2005s

1.3

SAXS Rg (nm)

Ref.

2.5

[21]

1.3

[55]

3.5

[22]

3.6

[68]

3.8

[22]

2.4

[69]

1.4

[70]

RS peptide (24 residues)

Nucleoporin 153 (82 residues)

Table 1: Comparison of average radius of gyration from MD simulations using a four-site water model and SAXS-derived measurement for several IDPs.

9



D, which can be combined with different force fields [22], can generate ensemble-average radii of gyration close to experimental values derived from SAXS intensities (Table 1). All together, these benchmarks emphasize the importance to verify the agreement between the models and the experiments at both local (NMR) and global (SAXS) levels when developing or applying new force fields for IDPs.

1.3

Enhanced sampling techniques assessed using NMR/SAXS data

Comparisons between MD simulations and experimental measurements were also employed to develop and test enhanced sampling techniques in order to improve the exhaustiveness of the conformational exploration. For instance, Karp et al. showed on intein proteins which are partially disordered that calculated chemical shifts, using SHIFTX2 or SPARTA+, are closer to experimental values when they are averaged over ensembles generated by accelerated molecular dynamics (aMD) rather than by standard MD simulations (both approaches using AMBER99SB/TIP3P force field) [72]. Interestingly, improvements were observed for residues found in spatial clusters but not for residues in flexible terminal tails. The authors suggested that boost potentials used in aMD are suited for escaping deep energy wells, but in the case of shallow energy surfaces, they might lead to incorrect distributions of IDP conformations [72]. In contrast, Sz¨ ollösi et al. showed that replica exchange discrete molecular dynamics (RX-DMD), a collision driven algorithm, using CHARMM force field and an implicit solvent model, can retrieve in 25 IDP conformational ensembles their known α-pre-structural motifs (α-PreSMo) determined by NMR [73]. However, in this study, the authors did not report back-calculation and comparison of simulated chemical shifts against NMR measurements, which does not allow to assess the conformation distributions produced by this method. Enhanced sampling methods were also tested using SAXS data. Such a study was carried out, for example, by Wen et al. with their technique called amplified collective motion (ACM) which uses a few collective normal modes to guide and accelerate the conformational ensemble exploration [74]. For two highly flexible multidomain proteins (the bacteriophage T4 lysozyme and the tandem WW domain of the formin-binding protein 21), they demonstrated that ACM-guided MD simulations explore conformational ensembles more extensively than standard MD (both using CHARMM27/TIP3P force field). They also showed that subsequent selections (using EOM [75]) of a small subset of conformations from the ACM ensemble allow to better fit the SAXS intensities than

10


Page 10 of 49

Page 11 of 49 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


when selected from the standard MD trajectory [74]. This emphasizes the importance of generating a conformational ensemble as exhaustive as possible before selection to accurately fit SAXS data. However, the authors recognized that the ACM technique does not perform equilibrium simulations and thus does not generate proper Boltzmann distributions [74]. Another approach to enhance sampling of IDP conformational ensemble is to use coarse-grained (CG) models which reduce the number of simulated particles to typically one bead per residue. Terakawa and Takada developed such a model from a combination of all-atom MD simulations of few fragments of a target IDP and a structural dababase of loop segments [76]. They reported that their CG approach generated a conformational ensemble of the N-terminal transcription domain of protein p53, which yielded backcalculated SAXS intensity (using CRYSOL [77]) and NMR residual dipolar couplings (using PALES [78]) very close to experimental measurements [76]. In a later work, the same group used a similar bottom-up approach to derive CG long-range contacts from allatom simulations of the p53 linker region (between the core and tetramerization (TET) domain). The authors validated the obtained CG model by generating a conformational ensemble of a tetramer of p53 core-linker-TET domains which back-calculated SAXS intensity (using CRYSOL) agrees very well with experimental data [79].

2

Methods for refinement of MD-derived ensembles

With the increasing power of high performance computers, MD simulations can now generate wide IDP conformational ensembles. However, even with enhanced techniques, sampled ensembles do not always completely account for experimental data. Thus, several methods were developed to refine ensembles from MD trajectories in complete agreement with experimental data, reducing the imperfections of force fields and samplings. In this section we briefly survey these developments to which MD simulations have also contributed. Before this, it should be emphasized that comparisons between theoretical and experimental NMR or SAXS data require to back-calculate or predict these observables from ensembles of atomic structures. Regarding NMR chemical shift calculation, various approaches (sequence-based, structure-based, or quantum mechanics-based) were implemented in many software, including SHIFTS [80], SPARTA+ [81], CAMSHIFT [82], or SHIFTX2 [83]. Although it is still delicate to calculate IDP chemical shifts with high

11



accuracy, several studies showed that all these methods yielded comparable accurate and sensitive chemical shift predictions [84–86]. In contrast, back-calculations of protein SAXS intensities, obtained as the scattering curve of the protein solution minus the scattering of the buffer, seem to be more tricky. Most popular software for SAXS intensity calculation, including CRYSOL [77, 87], FoXS [88, 89], and Pepsi-SAXS [90], use implicit models of solvation. These approaches generally have reduced computational costs, but require to fix a few free parameters, particularly the solvent excluded volume and excess density of the hydration layer, which is not a trivial task. It should be emphasized here that, unlike for chemical shift calculations, very few studies benchmarked these implicit solvent software and investigated how these free parameters impact the SAXS intensities of disordered proteins calculated over conformational ensembles generated by MD simulations. It is notably found that small variations of the excess density of the hydration layer strongly affect the protein scattering profiles [91, 92]. This suggests that users of these implicit solvent approaches should carefully assess these free parameters before calculating and interpreting SAXS intensity profiles [14, 93]. Another type of SAXS intensity calculations explicitly takes into account solvent molecules around protein to perform this prediction [14]. These approaches, such as implemented in WAXSiS [94, 95], are more time-consuming than implicit solvent programs, but naturally provide an atomic description of protein hydration layers without requiring to adjust free parameters. They proved to be efficient and useful until hundreds of protein conformations [96] with computational costs comparable to the unavoidable time costs for SAXS or NMR experiments. Assuming that software for back-calculation of NMR and SAXS observables are sufficiently reliable, they can be employed to derive IDP conformational sub-ensembles that agree well with experiments. This can be performed following two approaches, either data-restrained MD simulations (restrained sampling) or reweighting the conformations of MD trajectories (ensemble reweighting) (Fig. 1). To help the reader in identifying the various techniques and software discussed in this text, we listed their names and purposes in Table 2.

2.1

Experimental data-restrained MD simulations

MD simulations restrained by NMR-derived atomic distances have been developed for several decades to derive ensemble of folded protein structures [106–108]. This method,

12


Page 12 of 49

Page 13 of 49 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


Purpose

Name

Ref.

TraDES

[17]

Flexible-Meccano

[18]

AMBER

[97]

CHARMM

[98]

GROMACS

[99]

NAMD

[100]

SHIFTS

[80]

SPARTA+

[81]

CAMSHIFT

[82]

SHIFTX2

[83]

PALES

[78]

CRYSOL

[77]

FoXS

[89]

Pepsi-SAXS

[90]

WAXSiS

[95]

EOM

[75]

ASTEROIDS

[101]

SES

[102]

ENSEMBLE

[103]

EROS

[104]

COPER

[105]

Statistical coil generator

Molecular dynamics simulations

Chemical shift back-calculation

RDC back-calculation

SAXS intensity back-calculation

Sub-ensemble selection (maximum parsimony)

Ensemble reweighting (maximum entropy)

Table 2: Name and purpose of main software employed for conformational ensemble validation, restrained sampling, or refinement using NMR or SAXS data. A more exhaustive list can be found in Ref. [12].

13



which basically consists in adding penalty energy functions of differences between backcalculated and experimental data to physics-based hamiltonians, was thereafter extended for characterizing IDP conformations using chemical shifts, 3 J-couplings, or RDCs [109– 111]. It could be noted that a slightly different type of NMR data-restrained MD simulations was developed by Granata et al., consisting in performing bias-exchange metadynamics with chemical shifts as collective variables [112, 113]. Intuitively, one could think that restrained MD simulations sample a limited region of protein conformational space, depending on the penalty energy functions. Nevertheless, it was interestingly shown that, when simulations are performed on several replicas and when restraints are imposed on back-calculated observables averaged over all the replicas, then data-restrained MD approaches are equivalent to ensemble reweighting using the maximum entropy principle (see below) [114–117]. Compared to NMR data-restrained simulations, very few SAXS-restrained MD studies were found in the literature. As far as we know, only Chen and Hub reported the development of a SAXS/WAXS intensity-restrained MD algorithm to refine conformational ensembles against scattering measurements [118]. It is worthy to note that their restrained MD simulations in explicit solvent use a penalty energy function of differences between experimental and theoretical intensities back-calculated on-the-fly by their explicit solvent approach WAXSiS [94]. Otherwise, in a very similar spirit as the previously mentioned bias-exchange metadynamics using chemical shifts as collective variables [112], Kimanius et al. developed a SAXS-guided bias-exchange metadynamics using scattering intensities I(q) at four small angles q = 0.08, 0.14, 0.20, and 0.28 ˚ A−1 as collective variables [119]. In their case, theoretical SAXS intensities were back-calculated using the implicit solvent software FoXS [88].

2.2

Ensemble reweighting of MD trajectories

These computational approaches, also called sample-and-select, are not specific to MD simulations, but can be applied to conformational ensembles generated with any methods, notably statistical coil generators. Several review articles [12, 13] already surveyed the various software implementing these approaches which can be divided into two types: maximum parsimony methods, such as EOM [75], ASTEROIDS [101] or SES [102], which aim at finding a minimal subset of protein conformations that satisfactorily fits the experimental data, or maximum entropy methods, such as ENSEMBLE [103], EROS [104]

14


Page 14 of 49

Page 15 of 49 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


or COPER [105], which seek to determine a conformation probability distribution consistent with experiments but keeping a maximum of structures from the pre-sampled ensembles. Since disordered proteins generally have many various conformations, and since MD simulations are supposed to generate physics-based ensembles close to real ones, maximum entropy methods appear more appropriate than maximum parsimony approaches for refining IDP ensembles from MD trajectories. This was illustrated, for instance, in the study by Ball et al. who comparatively refined statistical coil and MD ensembles of peptides Aβ40 and Aβ42 with the programs ENSEMBLE and ASTEROIDS using several NMR measurements, including 3 J-couplings, RDCs, and NOEs. They concluded that MD simulations generate IDP conformational ensembles which constitute excellent starting pools for subsequent refinements [120]. To close this section, we should mention that significant improvements of conformational ensemble reweighting can be obtained by incorporating errors on experimental measurements into refinement processes. This is generally performed using Bayesian inference approaches which allows not only to take into account the quality of various heterogeneous experimental data, but also to estimate the uncertainty and accuracy of the output conformation probability distributions [121–123]. Interestingly, the Bayesian inference formalism can be implemented in data-restrained replica MD simulations in which the force constants of the penalty energy functions are treated as variables depending on the noise in experimental data and inferred like the conformation probability uncertainties [123, 124].

3

IDP studies using MD and NMR/SAXS data

After the two previous sections mainly dedicated to methodological aspects of the modeling of IDP conformational ensembles, we now review some recent studies combining MD simulations and NMR and/or SAXS measurements to gain insight into the molecular mechanism of IDP biological activity [125–127] but also their implication in human diseases [128]. We will first look over studies of amyloid proteins related to degenerative diseases, because MD simulations can be very helpful in characterizing their structures, which is challenging by only experimental means due to their strong aggregation propensity. Then, we will survey recent MD applications on some IDPs involved in cellular regulation or signaling mechanisms, and lastly on some IDPs from viruses.

15



3.1

Amyloid proteins

There are about thirty human amyloid proteins which form insoluble fibrils mainly deposited in the extracellular spaces of tissues resulting in a family of diseases called amyloidosis [129]. Most of the structural investigations of these proteins are focused on three of them, the human islet amyloid polypeptide (hIAPP), amyloid-β (Aβ), and α-synuclein (αSyn), involved in the type II diabetes, the Alzheimer’s, and the Parkinson’s disease, respectively (Table 3). A fourth frequently studied protein is the microtubule-stabilizing tau protein which abnormal phosphorylation leads to aggregation into intracellular neurofibrillary tangles also involved in the Alzheimer’s disease. Cumulative evidence indicates that these amyloid proteins are fully or partially disordered in their monomeric state, but their mechanisms of aggregation still remain unclear. To shed light on the early steps of their oligomerization, several studies, discussed below, combine MD simulations and NMR data to accurately characterize their conformational ensemble and to identify specific structures, generally rich in β-stands, that are prone to aggregate. Regarding hIAPP, its monomeric structure bound to micelles is rather well characterized and known to be mostly α-helical [153, 154]. In solution, experimental studies indicate that hIAPP is predominantly in random coil but transiently samples helical conformations [155–157]. Strikingly, although many MD simulations studies were reported in the literature [158], very few of them compared their results to available experimental measurements (Table 3). Laghaei et al. used a coarse-grained model with implicit solvent to sample hIAPP conformational ensemble and found that its region 5-16 transiently adopts α-helical structures [130] consistently with previously mentioned NMR data. The all-atom REMD study by Qi et al. confirmed that hIAPP region 11-25 transiently visits α-helical conformations in agreement with chemical shift data. These transient α-helices would be intermediate elements of recognition in hIAPP oligomer formation [131]. For the monomeric Aβ peptide in water, we found much more structural studies that combined simulations and experimental data than for the three other amyloid proteins. Most of them used replica-exchange molecular dynamics (REMD) simulations and NMR chemical shifts, 3 J-couplings, and/or RDCs to assess the sampled ensembles (Table 3). Whereas most of these studies just validated their MD ensembles against NMR data, only Ball et al. further reweighted their MD-generated conformations [141] (using the ENSEMBLE software [103]), and Granata et al. performed bias-exchange metadynamics with chemical shifts as collective variables [113]. It is worthy to note that, in all studies

16


Page 16 of 49

Page 17 of 49 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


IDP

Disease

Method

Force field | Water model

Exp. data

Ref.

hIAPP

Type II

HT-REMD

OPEP(a) | OPEP

3 Jc

[130]

(37 residues)

Diabetes

REMD

OPLS-AA/L | TIP4P

CS

[131]

REMD

OPLS-AA/L | TIP3P

3 Jc

[132]

REMD

AMBER99SB | GB(b)

CS

[133]

HT-REMD

OPEP | OPEP

3 Jc

[134]

H-REMD

AMBER03 | TIP3P

CS

[135]

REMD

AMBER99SB | TIP4P-Ew

REMD


MD

GROMOS96 | SPC

CS

[138]

MD

GROMOS96 | SPC

CS

[139]

REMD

OPLS-AA/L | TIP3P

3 Jc

[140]

Amyloid-β Alzheimer (40/42 residues)

REMD +


3 Jc, 3 Jc,

3 Jc,

RDC

[136]

CS, RDC

[137]

CS, RDC

[141]

Reweighting Restrained

CHARMM22∗ | TIP3P

CS

[113]

Metadynamics REMD

OPLS-AA/L | SPC/E

3 Jc,

CS

[142]

MD

AMBER14SB | TIP3P

3 Jc,

CS

[143]

REMD

CHARMM27 | TIP3P

CS

[144]

AMBER99 | SPC/E

SAXS

[145]

OPLS-AA/L | AGBNP(c)

RDC

[146]

Tau Alzheimer (441 residues)

Metadynamics + Reweighting REMD +

α-Synuclein Parkinson (140 residues)

Reweighting Restrained

CHARMM19 | EEF1(d)

3 Jc,

RDC

[147]

REMD AMBER03ws | TIP4P/2005s

REMD

3 Jc,

CS

[68]

SAXS

Table 3: MD studies of monomeric amyloid proteins in solution combined with NMR measurements of chemical shifts (CS), 3 J-couplings (3 Jc), residual dipolar couplings (RDC), or SAXS experiments.

(a) OPEP

an implicit solvent model [148, 149]. (c) AGBNP

is a coarse-grained protein force field using

(b) Generalized-Born

is another implicit solvent model [151].

implicit solvation model [150].

(d) CHARMM19

energy function (EEF1) for implicit protein solvation [152].

17


associated effective


that compared chemical shifts, agreements between back-calculated and experimental values are always very good, with Pearson correlation coefficients around 0.9. This indicates that chemical shifts are not sufficient metrics to validate an conformational ensemble or to discriminate between two different ones [135, 137]. Various sampling approaches, force fields, and water models used in these studies led to different structural ensembles with various level of agreement with NMR data, ranging from moderate to very good. But, overall, most of these studies seem to converge toward a similar description of Aβ conformations: the peptide is mostly in statistical coil but transiently samples, at various extents, structures with α-helices and, most importantly, β-strands at consensus regions (16-21, 30-35, and 39-41). The latter can be arranged in β-hairpins or β-sheets of three β-strands which probably seed the Aβ self-recognition and the early steps of the oligomerization [142]. The isoform hTau40 of human protein Tau is a 441-residue microtubule-associated protein. Its microtubule-binding domain (197-441) includes a proline-rich domain (197243) and a repeat domain (244-372) which can form amyloid fibrils with cross-β structures. Tau is natively unfolded and highly dynamics in solution, but some transient secondary structures were identified in several regions [159]. For example, MD simulations detected a transient β-strand and a polyproline II (PPII) helix in segments 223-247 and 225-250 [143], in good agreement with NMR measurements of chemical shifts and RDCs. Extensive REMD simulations of the long repeat domain 244-372 generated a conformational ensemble in quantitative correlation with experimental chemical shifts with transient α-helices and notably high β-strand propensities at the segments 275-280 and 306-311, both crucial for Tau aggregation [144]. Actually, Luo et al. found that the repeat domain conformational ensemble is a mixture of disordered and ordered structure which is rather compact with long-range contacts between the repeats [144]. The compactness of the whole protein was quantified by Battisti et al. in a rare study combining Metadynamics simulations (using the protein radius of gyration as collective variable) and ensemble reweighting (using EOM) based on SAXS intensities. At 293 K, the authors found an average radius of gyration of 6.6 nm, before selection by EOM. After ensemble reweighting, they reported a value of 6.9 nm [145], indicating that the hTau40 conformational ensemble sampled by Metadynamics was quite close to the one revealed by SAXS experiments, at least at the global level. Like protein Tau, α-Synuclein has a conformational ensemble more compact than

18


Page 18 of 49

Page 19 of 49 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


expected for a random coil protein of this size [160, 161]. However, characterizing the transient long-range contacts between the transient secondary structures of this ”premolten” globular state still remains challenging. To this end, paramagnetic relaxation enhancement (PRE) and fluorescence resonance energy transfer (FRET) experiments are particularly well suited for probing long-range contacts in proteins. Although these techniques were considered out of scope of this review, we would like to highlight some investigations which combined MD simulations with PRE or FRET and NMR data (Table 3). These studies described conformational ensembles of α-Synuclein, generated by REMD simulations, in good agreement with PRE or FRET data and other NMR measurements [68, 146, 147]. These ensembles notably comprise heterogeneous structures, entirely collapsed, fully extended, or semi-compact in which the N-terminal domain makes contacts to the central non amyloid-β component (NAC), the C-terminal domain remaining mainly extended. It could be noted that former similar PRE studies reported that α-Synuclein NAC domain is rather contacted by the C-terminal region [161, 162], highlighting the difficulties to provide accurate conformational ensembles from sparse experimental data. Indeed, different ensembles can quantitatively fit PRE data, but could be divergent about local secondary structures and disagree with 3 J-coupling or RDC measurements. The latter are therefore recommended to be used as additional restraints or in reweighting refinements to obtain a more accurate description of IDP conformational ensembles [146, 147].

3.2

Regulation and signaling proteins

IDPs are now recognized as key players in regulation of cellular processes as well as in cell signal transduction. These functions are generally performed through cascades of high-specificity and low-affinity association/dissociation events of intrinsically disordered regions (IDRs) with various protein partners. In many IDR binding process, a disorderto-order transition is observed in which a random coil segment in free state adopts a secondary structure when bound to a partner. Since their uncovering by bioinformatics studies [163, 164], these segments, called molecular recognition features (MoRFs), have been the subject of many MD studies to characterize their disorder-to-order transition and better understand their mechanism of binding. We survey here recent ones which also combined NMR or SAXS experiments (Table 4). A first interesting and representative study was performed by Yoon et al. on the C-

19



Type

MoRF

IDR

Partners

Methods MD +

Reweighting(a)

Ref. 3 Jc

C-ter domain of p21

CaM

(residues 145-164)

PCNA

(CHARMM19 | EEF1)

TAD domain of p53

MDM2

REMD + CS

(residues 15-29)

(N-ter domain)

(AMBER03w | TIP4P/2005)

TAD domain of p53

MDM2

RE-GA(b) + CS, RDC

(residues 1-61)

(N-ter domain)

(CHARMM27 | GBSW/SA(c) )

bZip domain of GNC4

+ CS,

MD + CS

DNA

Gab2

Grb2

Restrained REMD + CS

(residues 503-524)

(C-ter SH3 domain)


FBP21 linker

between WW1 and

MD + Reweighting + SAXS

(residues 33-46)

WW2 domains

(CHARMM27 | TIP3P)

Disordered loop and linker

[165]

[166]

[167]

[168]

(AMBER99SB | TIP3P)

(residues 225-280)

Disordered domain

Page 20 of 49

RfaH linker

between NTD and

MD + CS

(residues 101-114)

CTD domains

(RSFF1(d) | TIP4P-Ew)

TCTP long loop

TCTP

MD + CS, 3 Jc

(residues 37-68)

core domain

(AMBER99SB-ILDN | TIP3P)

NCBD domain of CBP

ACTR

REMD + CS + SAXS

(residues 2059-2117)

IRF-3


P18-S65 region of 4E-BP2

REMD + CS

eIF4E

(RSFF2(d) | TIP3P)

(residues 18-65) C-ter domain of 4.1G

NuMA

REMD + CS

(residues 939-1005)

(residues 1800-1825)

(CHARMM27 | TIP3P)

[169]

[170]

[171]

[172]

[173]

[174]

[175]

Table 4: MD studies of IDR in solution combined with NMR or SAXS measurements. (a) Energy-Minima

Mapping and Weighting (EMW) method.

Guided Annealing [176].

(c) Generalized-Born

/ Surface Area [177–179].

(d) Residue-Specific

(b) Replica

Exchange with

model with simple SWitching function

Force Field 1 and Residue-Specific Force

Field 2 are modifications of OPLS-AA/L and AMBER99SB based on protein coil library, respectively [180–182].

20


Page 21 of 49 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


terminal domain of protein p21 (residues 145-164) which binds several proteins including calmodulin (CaM), the proliferating cell nuclear antigen (PCNA), or the human papilloma virus E7 oncoprotein [165]. Using MD simulations and a reweighting optimization based on Cα chemical shifts, the authors generated a conformational ensemble of p21 segment 145-164 in excellent agreement with several other NMR measurements. They showed that the N-terminal half of the peptide transiently adopt an α-helical structure which is similar to its CaM-bound conformation, while its C-terminal half is preferentially in an extended strand similar to the one adopted in the complex with PCNA. These results suggest that proteins CaM and PCNA bind the p21 peptide α-MoRF and β-MoRF, respectively, by a conformational selection of these secondary structures preexisting in its unbound state [165]. Regarding PCNA in association with the intrinsically disordered PCNA associated factor p15, it should be mentioned that Cordeiro et al. recently characterized different conformations of the complex by combining MD simulations of PCNA, conformational ensembles of p15 generated with Flexible-Meccano [18], SAXS, and NMR experiments [96]. Bioinformatics analyses of residues in MoRFs estimated that, upon binding, 27% of them adopt an α-helical conformation (α-MoRF), 12% a β-strand (β-MoRF), and about 48% an irregular structure (i-MoRF). The remaining 13% are residues with missing coordinates in complex structures [164]. α-MoRFs are frequently observed in IDP-protein interactions, such as in the case of the widely studied p53 TAD domain which conformational ensemble was simulated in agreement with NMR data by Mittal et al. [166] or Ganguly and Chen [167], for example. α-MoRFs are also found in IDP-DNA associations, such as in the DNA binding of the basic leucine-zipper (bZip) domain of the yeast transcription factor GCN4 which MD-generated conformational ensemble was validated against NMR chemical shifts by Robustelli et al. [168]. Among the irregular i-MoRFs, it should be pointed out that some of them could be polyproline II (PPII) left-handed helical structures which were often misplaced among unstructured, extended, or random coil conformations. The study of such PPII-MoRF was reported, for instance, by Krieger et al. who used chemical shift measurements in restrained MD simulations of the fragment 503-524 of protein Gab2 to characterize its conformational properties. The authors showed that the free disordered peptide had significant amounts of residual extended (up to 20%) and PPII (25-30%) structures and transiently visited preorganized conformations highly similar to the one found in its complex with the C-terminal SH3 domain of

21



the growth factor receptor-bound protein 2 (Grb2) [169]. The detection of preexisting preorganized bound structures in the IDR free state conformational ensemble does not prove that the binding to partners proceeds through a conformational selection mechanism. Nevertheless, the two previously mentioned studies on p53 TAD domain [166, 167] and the one about Gab2 fragment 503-524 [169] interestingly reported that mutations in these disordered peptides modify their propensity to form α-helical or PPII structures in qualitative correlation with the binding affinity to their partners. This indicates that, although the bound structure of MoRFs is not necessarily the one which is selected by protein partners, at least, a pathway from disordered state to these conformations should exist and not be energetically unfavorable, so that the MoRF secondary structures can easily be formed to fit the binding sites. Other functions of intrinsically disordered segments are to serve as flexible linker between two functional folded domains. Structure and dynamics of these linkers generally modulate the relative orientation of the two linked domains and their possible binding for cooperative function (Table 4). Conformational ensembles of these highly flexible proteins can be efficiently characterized using MD simulations and sub-ensemble selection based on SAXS intensities, as Zhang et al., for instance, have done on the FBP21 protein [170]. In another study using MD simulations validated against chemical shift measurements, Xun et al. showed that the disordered linker of RfaH protein, which is probably extended when the two linked N-terminal and C-terminal domains are separated, become a long loop that participates in stabilizing the close state in which the two domains are in contact [171]. The role of another disordered long loop, the TCTP protein one, was also investigated by Malard et al. using MD simulations validated against NMR chemical shifts and 3 J-couplings. This study showed that TCTP disordered loop can be detached from or bound to its core domain in response to various cellular conditions, indicating that the loop modulates the accessibility of the protein core domain to its partners [172]. Above tails, linkers, and loops, a higher level of disorder can be found in molten or pre-molten globules which are partially or fully unfolded domains or proteins but still collapsed [183]. Neverteless, these compact unfolded domains still have a role in specific binding other proteins. For example, the nuclear coactivator binding domain (NCBD) of the transcriptional coactivator CBP is a molten globule with some helical segments but with heterogeneous tertiary conformations. The REMD study of unbound NCBD, car-

22


Page 22 of 49

Page 23 of 49 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


ried out by Knott and Best and validated against both NMR measurements and SAXS intensities, identified in its conformational ensemble some tertiary structures which have preformed interface competent for binding ACTR protein [173]. A similar observation was made by Zeng et al. on the P18-S65 region of protein 4E-BP2, a repressor of the eukaryotic initiation factor eIF4E. The conformational ensemble of the unbound domain, generated by REMD simulations in agreement with chemical shift data, exhibits collapsed conformations including some folded α-helical motifs which can be superimposed on the crystal structure of 4E-BP2 bound to eIF4E [174]. Interestingly, these helical structures are less sampled in the conformational ensemble of the phosphorylated protein which is shifted toward β-folds, consistently with the fact that 4E-BP2 phosphorylations suppress its binding to eIF4E. A last example of such collapsed and disordered domain is provided by the C-terminal domain (CTD) of protein 4.1G which conformational ensemble was recently investigated using REMD simulations and NMR spectroscopy by Wu et al. [175]. This domain was shown to sample heterogeneous globular structures organized alongside a stable α-helix, in good agreement with chemical shift measurements. However, this stable helix is not the binding site of the protein partner NuMA (nuclear mitotic apparatus) which rather interacts with multiple sites on the globular disordered part of 4.1G, forming a dynamic fuzzy complex in which both proteins remain disordered [175].

3.3

Disordered proteins from viruses

Bioinformatical disorder predictions unveiled that the prevalence of IDPs in proteomes increases from prokaryotes (10-30%) to more complex eukariotic organisms (40-50%) [184]. Interestingly, these studies also revealed that, compared to eukariotes, viruses have similar amounts of short intrinsically disordered segments (30-60 residues) but contains fewer long disordered regions. An analysis of these viral IDP biological functions indicates that they are mainly related to protein-protein interactions, regulation, and signaling [184]. All together, these observations suggest that IDPs from viruses seem to mimic the host disordered proteins in order to hijack the host cellular machineries for their benefit. As such, characterizing viral IDP conformational ensemble helps not only to develop new antiviral compounds but also to better understand IDP mechanism (Table 5). As for other organism IDPs, NMR secondary chemical shifts can inform about the propensity of viral disordered segments to form transient local secondary structures which population and extension can be quantified in detail with MD simulations. For example,

23



Virus

Disease

Human Papillomavirus

Genital infection,

(HPV)

wart, cancer

IDR

Page 24 of 49

Methods

Ref.

N-ter domain of REMD + CS oncoprotein E7

[185]

(AMBER99SB | GB/SA)

(residues 1-46) Measles Virus

C-ter domain (NT AIL ) Measles infection

of nucleoprotein N

(MeV)

REMD + CS (CHARMM22∗ | SPC

[186]

(residues 484-504) Human Metapneumo-

Core region (PCED ) Respiratory infection

of phosphoprotein P

virus (HMPV)

MD + Reweighting + SAXS

[187]

(AMBER-99SB-ILDN | SPC/E)

(residues 158-237) N-ter tail of Dengue Virus

Dengue fever

MD + Reweighting + SAXS protein C

and infection

[188]

(AMBER03ws | TIP4P)

(residues 1-100)

Table 5: MD studies of IDRs involved in infection diseases combined with NMR or SAXS measurements the intrinsically disordered N-terminal region (residues 1-46) of the oncoprotein E7 from human papillomavirus (HPV) was shown by several NMR experiments to adopt transient α-helical structures at segments 7-13 and 20-26 [185, 189], the latter containing the LxCxE motif which allows the binding of E7 to the host retinoblastoma tumor suppressor protein pRB. The relative population of these helical structures was further estimated to 10-20% by REMD simulations [185] (Table 5). Disordered regions are also frequently found in virus capsid structural proteins which often form complexes with the viral genome. For instance, the nucleoprotein N of the measles virus (MeV) has a C-terminal domain (NT AIL ) which is essential for the virus replication by interacting with the phosphoprotein P of the viral polymerase complex. NT AIL was proven to be intrinsically disordered by various biophysical methods, including NMR and SAXS experiments [190, 191]. Subsequent measurements of chemical shifts and RDCs revealed that NT AIL transiently adopts conformations having an α-helix of various length (from 6 to 18 residues) in the interacting region 484-504 [192]. Again, REMD simulations allowed to generate a conformational ensemble of this region in agreement with NMR secondary chemical shifts, and to quantify the residue-specific helical propensity between 10 and 60% [186] (Table 5). Furthermore, this theoretical study indicated that the unbound NT AIL samples the long α-helix conformation observed in the bound form, but, interestingly, this helical structure is not selected by its protein partner. The latter seems to rather preliminary binds one of the NT AIL conformations with a shorter or partial α-helix which subsequently elongates following an induced fold-

24


Page 25 of 49 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


ing mechanism [186]. Another surprising finding of this study is that the disordered conformations of the unbound NT AIL segment 484-504 is on average more compact than in the folded bound state as indicated by its radius of gyration distributions. However, these theoretical conclusions are based on protein and water models that might not be optimized for IDPs and were not experimentally confirmed using for instance SAXS experiments (Table 5). This emphasizes the necessity to accurately characterize IDP conformational ensembles at both local and global levels in order to deeply understand their mechanism of binding. The human metapneumovirus (HMPV) is another virus from the Paramyxoviridae family to which the previously mentioned measles virus belongs. Similarly to all members of this family, its phosphoprotein P forms homotetramers through a core region, named PCED (residues 158-237), which is structured in α-helical coiled coil. However, the length of this structural motif can widely vary among the Paramyxoviridae, and the one from HMPV was not clearly determined. From an MD-derived conformational ensemble of the PCED tetramer, Leyrat et al. selected (using EOM) a subset of conformations that fits very well the SAXS data obtained on this sytem [187] (Table 5). The authors thus revealed that the HMPV PCED tetrameric coiled coil is the shortest one among the Paramyxoviridae, spanning residues 171-194. The protein C-terminal region (195-237) was found disordered but with a significant propensity to form α-helices which may be involved in binding nucleoprotein N [187]. A similar approach, combining MD simulations and sub-ensemble selection (with EOM) based on SAXS intensities, was used by Boon et al. to prove the intrinsically disordered state of the N-terminal tail of the capsid protein C of the dengue virus [188]. The core domain of protein C is composed of four helices and form homodimers characterized by an important hydrophobic patch on their surface that binds lipid droplets for the viral particle formation [193]. Although disordered, the C protein N-terminal tails can sample α-helical structures at region 9-18, as revealed by MD simulations. The dynamics of these transient α-helices would allow to occlude the hydrophobic patch to auto-inhibit the protein C [188]. From a methodological viewpoint, this study interestingly showed that a good fit to SAXS data could only be obtained by reweighting the ensemble generated with the AMBER-03ws force field. The other tested models (AMBER-14sb, CHARMM-36, CHARMM-36m, and GROMOS-96-54a7) generated overly compact conformations with radii of gyration too far from the average value determined from SAXS data [188] (Ta-

25



ble 5). This emphasizes again the importance to generate IDP conformational ensembles as realistic as possible, even if refinements are performed after.

Conclusions Since the beginning of the century, an increasing number of theoretical and computational studies have been reported on the conformational ensemble of intrinsically disordered proteins, domains, or segments. Among them, MD simulations have become a more and more reliable approach to generate realistic conformational ensembles, thanks to the better accuracy of all-atom force fields (section 1.2), the more efficient enhanced sampling techniques (section 1.3), as well as the more powerful high performance computers. However, not all of them validated the simulated ensembles against available experimental data, such as NMR or SAXS measurements. And very few investigations combined MD simulations with both local and global information from NMR and SAXS experiments, respectively. We encourage the biomolecular simulation community to validate their IDP ensembles with NMR and SAXS data to gain a correct understanding of their structures and dynamics, their mechanisms of action, and their implication in human diseases. Additionally, IDP conformational ensembles can be accurately generated using NMR/SAXS data-restrained simulations (section 2.1) or reweighting MD ensembles (section 2.2). Many studies combined MD simulations and NMR measurements to investigate the role of amyloid protein conformational ensembles in the associated human diseases. However, few of them used SAXS data to provide subsequent validations of these conformational ensembles (section 3.1). Regarding the relationship between IDP conformational ensembles and biological functions, a frequent conclusion emerges from the reviewed MD studies combined with NMR/SAXS experiments: in their free state, IDPs are characterized by highly heterogeneous conformations, some of them having secondary or tertiary structures similar to those found in complexes with their partners, suggesting a conformational selection mechanism of binding (section 3.2). These preformed bindingcompetent structures were also identified and characterized using combined approaches in the conformational ensembles of viral IDPs, likely enabling them to bind and hijack host proteins (section 3.3). However, it should be reminded that binding of IDPs is not necessarily coupled to a disorder-to-order transition as emphasized, for instance, by

26


Page 26 of 49

Page 27 of 49 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


Sigalov et al. [194]. Many IDPs remain disordered when bound to their targets, forming so-called ”fuzzy” complexes which the highly dynamic structure might allow fast conformational responses to various modifications of the cellular conditions [195–197]. These various levels of disorder in IDP unbound and bound states represent a great challenge for molecular modeling approaches to exhaustively and accurately characterize their structural ensembles. Besides, this also may call into question our unconscious search for order.

Acknowledgements This work is supported by the ”IDI 2016” project funded by the IDEX Paris-Saclay (ANR-11-IDEX-0003-02).

Conflict of interests The authors declare no conflict of interests regarding the publication of this article.

References [1] Bracken, C.; Iakoucheva, L.; Romero, P.; Dunker, A. Combining Prediction, Computation and Experiment for the Characterization of Protein Disorder. Curr. Opin. Struct. Biol. 2004, 14, 570–576. [2] Receveur-Bréchot, V.; Bourhis, J.; Uversky, V.; Canard, B.; Longhi, S. Assessing Protein Disorder and Induced Folding. Proteins 2005, 62, 24–45. [3] Mittag, T.; Forman-Kay, J. Atomic-level Characterization of Disordered Protein Ensembles. Curr. Opin. Struct. Biol. 2007, 17, 3–14. [4] Eliezer, D. Biophysical Characterization of Intrinsically Disordered Proteins. Curr. Opin. Struct. Biol. 2009, 19, 23–30. [5] Uversky, V. Biophysical Methods to Investigate Intrinsically Disordered Proteins: Avoiding an Elephant and Blind Men Situation. In Intrinsically Disordered Proteins Studied by NMR Spectroscopy; Felli, I.; Pierattelli, R., Eds.; Advances in Experimental Medicine and Biology, Springer International Publishing, 2015; pp. 215–260.

27



[6] Wei, G.; Xi, W.; Nussinov, R.; Ma, B. Protein Ensembles: How Does Nature Harness Thermodynamic Fluctuations for Life? The Diverse Functional Roles of Conformational Ensembles in the Cell. Chem. Rev. 2016, 116, 6516–6551. [7] Kragelj, J.; Ozenne, V.; Blackledge, M.; Jensen, M. Conformational Propensities of Intrinsically Disordered Proteins from NMR Chemical Shifts. ChemPhysChem 2013, 14, 3034–3045. [8] Kurzbach, D.; Kontaxis, G.; Coudevylle, N.; Konrat, R. NMR Spectroscopic Studies of the Conformational Ensembles of Intrinsically Disordered Proteins. In Intrinsically Disordered Proteins Studied by NMR Spectroscopy; Felli, I.C.; Pierattelli, R., Eds.; Advances in Experimental Medicine and Biology, Springer International Publishing, 2015; pp. 149–185. [9] Receveur-Bréchot, V.; Durand, D. How Random are Intrinsically Disordered Proteins? A Small Angle Scattering Perspective. Current Protein & Peptide Science 2012, 13, 55–75. [10] Kikhney, A.; Svergun, D. A Practical Guide to Small Angle X-ray Scattering (SAXS) of Flexible and Intrinsically Disordered Proteins. FEBS Letters 2015, 589, 2570–2577. [11] Fisher, C.; Stultz, C. Constructing Ensembles for Intrinsically Disordered Proteins. Curr. Opin. Struct. Biol. 2011, 21, 426–431. [12] Ravera, E.; Sgheri, L.; Parigi, G.; Luchinat, C. A Critical Assessment of Methods to Recover Information from Averaged Data. Phys. Chem. Chem. Phys. 2016, 18, 5686–5701. [13] Bonomi, M.; Heller, G.; Camilloni, C.; Vendruscolo, M. Principles of Protein Structural Ensemble Determination. Curr. Opin. Struct. Biol. 2017, 42, 106– 116. [14] Hub, J. Interpreting Solution X-ray Scattering Data using Molecular Simulations. Curr. Opin. Struct. Biol. 2018, 49, 18–26. [15] Jacques, D.; Trewhella, J. Small-Angle Scattering for Structura BiologyExpanding the Frontier while Avoiding the Pitfalls. Protein Sci. 2010, 19, 642–657.

28


Page 28 of 49

Page 29 of 49 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


[16] Rauscher, S.; Pomès, R. Molecular Simulations of Protein Disorder. Biochem. Cell Biol. 2010, 88, 269–290. [17] Feldman, H.; Hogue, C. Probabilistic Sampling of Protein Conformations: New Hope for Brute Force? Proteins 2002, 46, 8–23. [18] Ozenne, V.; Bauer, F.; Salmon, L.; Huang, J.; Jensen, M.; Segard, S.; Bernadó, P.; Charavay, C.; Blackledge, M. Flexible-meccano: A Tool for the Generation of Explicit Ensemble Descriptions of Intrinsically Disordered Proteins and their Associated Experimental Observables. Bioinformatics 2012, 28, 1463–1470. [19] Cukier, R. Generating Intrinsically Disordered Protein Conformational Ensembles from a Database of Ramachandran Space Pair Residue Probabilities using a Markov Chain. J. Phys. Chem. B 2018, 122, 9087–9101. [20] Estana, A.; Sibille, N.; Delaforge, E.; Vaisset, M.; Cortés, J.; Bernad´ o, P. Realistic Ensemble Models of Intrinsically Disordered Proteins using a StructureEncoding Coil Database. Structure 2019, 27, 1–11. [21] Best, R.; Zheng, W.; Mittal, J. Balanced Protein-Water Interactions Improve Properties of Disordered Proteins and Non-Specific Protein Association. J. Chem. Theory Comput. 2014, 10, 5113–5124. [22] Piana, S.; Donchev, A.; Robustelli, P.; Shaw, D. Water Dispersion Interactions Strongly Influence Simulated Structural Properties of Disordered Protein States. J. Phys. Chem. B 2015, 119, 5113–5123. [23] Huang, J.; Rauscher, S.; Nawrocki, G.; Ran, T.; Feig, M.; de Groot, B.; Grubm¨ uller, H.; MacKerell, A. CHARMM36m: An Improved Force Field for Folded and Intrinsically Disordered Proteins. Nat. Methods 2017, 14, 71–73. [24] Mollica, L.; Bessa, L.; Hanoulle, X.; Jensen, M.; Blackledge, M.; Schneider, R. Binding Mechanisms of Intrinsically Disordered Proteins: Theory, Simulation, and Experiment. Front. Mol. Biosci. 2016, 3, 1–18. [25] Schor, M.; Mey, A.; MacPhee, C. Analytical Methods for Structural Ensembles and Dynamics of Intrinsically Disordered Proteins. Biophys. Rev. 2016, 8, 429– 439.

29



Page 30 of 49

[26] Varadi, M.; Kosol, S.; Lebrun, P.; Valentini, E.; Blackledge, M.; Dunker, A.; Felli, I.; Forman-Kay, J.; Kriwacki, R.; Pierattelli, R.; Sussman, J.; Svergun, D.; Uversky, V.; Vendruscolo, M.; Wishart, D.; Wright, P.; Tompa, P.

pE- DB:

A Database of Structural Ensembles of Intrinsically Disordered and of Unfolded Proteins. Nucleic Acids Res. 2014, 42, D326–D335. [27] Sasmal, S.; Lincoff, J.; Head-Gordon, T. Effect of a Paramagnetic Spin Label on the Intrinsically Disordered Peptide Ensemble of Amyloid-β. Biophys. J. 2017, 113, 1002–1011. [28] Merchant, K.; Best, R.; Louis, J.; Gopich, I.; Eaton, W. Characterizing the Unfolded States of Proteins Using Single-molecule FRET Spectroscopy and Molecular Simulations. Proceedings of the National Academy of Sciences 2007, 104, 1528–1533. [29] Fuertes, G.; Banterle, N.; Ruff, K.; Chowdhury, A.; Mercadante, D.; Koehler, C.; Kachala, M.; Estrada-Girona, G.; Milles, S.; Mishra, A.; Onck, P.; Gr¨ ater, F.; Esteban-Mart´ın, S.; Pappu, R.; Svergun, D.; Lemke, E. Decoupling of Size and Shape Fluctuations in Heteropolymeric Sequences Reconciles Discrepancies in SAXS vs. FRET Measurements. Proceedings of the National Academy of Sciences 2017, 114, E6342–E6351. [30] van Gunsteren, W.; Billeter, S.; Eising, A.; H¨ unenberger, P.; Kr¨ uger, P.; Mark, A.; Scott, W.; Tironi, I. Force Fields and Topology Data Set. In Biomolecular Simulation: The GROMOS96 Manual and User Guide; Vdf Hochschulverlag AG an der ETH Z¨ urich: Switzerland, 1996; Vol. 3. [31] Jorgensen, W.; Maxwell, D.; Tirado-Rives, J. Development and Testing of the OPLS All-Atom Force Field on Conformational Energetics and Properties of Organic Liquids. J. Am. Chem. Soc. 1996, 118, 11225–11236. [32] MacKerell, A.; Bashford, D.; Bellott, M.; Dunbrack, R.; Evanseck, J.; Field, M.; Fischer, S.; Gao, J.; Guo, H.; Ha, S.; Joseph-McCarthy, D.; Kuchnir, L.; Kuczera, K.; Lau, F.; Mattos, C.; Michnick, S.; Ngo, T.; Nguyen, D.; Prodhom, B.; Reiher, W.; Roux, B.; Schlenkrich, M.; Smith, J.; Stote, R.; Straub, J.; Watanabe, M.; Wiórkiewicz-Kuczera, J.; Yin, D.; Karplus, M. All-Atom Empirical Potential for

30


Page 31 of 49 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


Molecular Modeling and Dynamics Studies of Proteins. J. Phys. Chem. B 1998, 102, 3586–3616. [33] Wang, J.; Cieplak, P.; Kollman, P. How Well Does a Restrained Electrostatic Potential (RESP) Model Perform in Calculating Conformational Energies of Organic and Biological Molecules? J. Comput. Chem. 2000, 21, 1049–1074. [34] Berendsen, H.; Grigera, J.; Straatsma, T. The Missing Term in Effective Pair Potentials. J. Phys. Chem. 1987, 91, 6269–6271. [35] Jorgensen, W.; Chandrasekhar, J.; Madura, J.; Impey, R.; Klein, M. Comparison of Simple Potential Functions for Simulating Liquid Water. J. Chem. Phys. 1983, 79, 926–935. [36] Kaminski, G.; Friesner, R.; Tirado-Rives, J.; Jorgensen, W. Evaluation and Reparametrization of the OPLS-AA Force Field for Proteins via Comparison with Accurate Quantum Chemical Calculations on Peptides. J. Phys. Chem. B 2001, 105, 6474–6487. [37] Duan, Y.; Wu, C.; Chowdhury, S.; Lee, M.; Xiong, G.; Zhang, W.; Yang, R.; Cieplak, P.; Luo, R.; Lee, T.; Caldwell, J.; Wang, J.; Kollman, P. A Pointcharge Force Field for Molecular Mechanics Simulations of Proteins Based on Condensed-phase Quantum Mechanical Calculations. J. Comput. Chem. 2003, 24, 1999–2012. [38] MacKerell, A.; Feig, M.; Brooks, C. Improved Treatment of the Protein Backbone in Empirical Force Fields. J. Am. Chem. Soc. 2004, 126, 698–699. [39] Mackerell, A.; Feig, M.; Brooks, C. Extending the Treatment of Backbone Energetics in Protein Force Fields: Limitations of Gasphase Quantum Mechanics in Reproducing Protein Conformational Distributions in Molecular Dynamics Simulations. J. Comput. Chem. 2004, 25, 1400–1415. [40] Hornak, V.; Abel, R.; Okur, A.; Strockbine, B.; Roitberg, A.; Simmerling, C. Comparison of Multiple Amber Force Fields and Development of Improved Protein Backbone Parameters. Proteins: Structure, Function, and Bioinformatics 2006, 65, 712–725.

31



[41] Best, R.; Hummer, G. Optimized Molecular Dynamics Force Fields Applied to the Helix-Coil Transition of Polypeptides. J. Phys. Chem. B 2009, 113, 9004– 9015. [42] Best, R.; Mittal, J. Protein Simulations with an Optimized Water Model: Cooperative Helix Formation and Temperature-Induced Unfolded State Collapse. J. Phys. Chem. B 2010, 114, 14916–14923. [43] Piana, S.; Lindorff-Larsen, K.; Shaw, D. How Robust are Protein Folding Simulations with Respect to Force Field Parameterization? Biophys. J. 2011, 100, L47– L49. [44] Huang, J.; MacKerell, A. CHARMM36 All-atom Additive Protein Force Field: Validation Based on Comparison to NMR Data.

J. Comput. Chem. 2013,

34, 2135–2145. [45] Maier, J.; Martinez, C.; Kasavajhala, K.; Wickstrom, L.; Hauser, K.; Simmerling, C. FF14SB: Improving the Accuracy of Protein Side Chain and Backbone Parameters from FF99SB. J. Chem. Theory Comput. 2015, 11, 3696–3713. [46] Lindorff-Larsen, K.; Piana, S.; Palmo, K.; Maragakis, P.; Klepeis, J.; Dror, R.; Shaw, D. Improved Side-chain Torsion Potentials for the Amber FF99SB Protein Force Field. Proteins 2010, 78, 1950–1958. [47] Wang, W.; Ye, W.; Jiang, C.; Luo, R.; Chen, H. New Force Field on Modeling Intrinsically Disordered Proteins. Chem. Biol. Drug. Des. 2014, 84, 253–269. [48] Song, D.; Luo, R.; Chen, H. The IDP-Specific Force Field FF14IDPSFF Improves the Conformer Sampling of Intrinsically Disordered Proteins. J. Chem. Inf. Model. 2017, 57, 1166–1178. [49] Liu, H.; Song, D.; Lu, H.; Luo, R.; Chen, H.

Intrinsically Disordered

Protein-specific Force Field CHARMM36IDPSFF. Chem. Biol. Drug Des. 2018, 92, 1722–1735. [50] Berendsen, H.; Postma, J.; Van Gunsteren, W.; Hermans, J.; Pullman, B. Interaction Models for Water in Relation to Protein Hydration. Nature 1981, 11, 331–342.

32


Page 32 of 49

Page 33 of 49 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] Horn, H.; Swope, W.; Pitera, J.; Madura, J.; Dick, T.; Hura, G.; Head-Gordon, T. Development of an Improved Four-site Water Model for Biomolecular Simulations: TIP4P-Ew. J. Chem. Phys. 2004, 120, 9665–9678. [52] Case, D.; Darden, T.; Cheatham, T.; Simmerling, C.; Wang, J.; Duke, R.; Luo, R.; Walker, R.; Zhang, W.; Merz, K.; Roberts, B.; Hayik, S.; Roitberg, A.; Seabra, G.; Swails, J.; G¨ otz, A.; Kolossváry, I.; Wong, K.; Paesani, F.; Vanicek, J.; Wolf, R.; Liu, J.; Wu, X.; Brozell, S.; Steinbrecher, T.; Gohlke, H.; Cai, Q.; Ye, X.; Wang, J.; Hsieh, M.; Cui, G.; Roe, D.; Mathews, D.; Seetin, M.; Salomon-Ferrer, R.; Sagui, C.; Babin, V.; Luchko, T.; Gusarov, S.; Kovalenko, A.; Kollman, P. Amber12: Reference Manuel. Technical report, University of California, San Francisco, 2012. [53] Nettels, D.; Muller-Spath, S.; Kuster, F.; Hofmann, H.; Haenni, D.; Ruegger, S.; Reymond, L.; Hoffmann, A.; Kubelka, J.; Heinz, B.; Gast, K.; Best, R.; Schuler, B. Single-molecule Spectroscopy of the Temperature-induced Collapse of Unfolded Proteins. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 20740–20745. [54] Piana, S.; Klepeis, J.; Shaw, D. Assessing the Accuracy of Physical Models used in Protein-folding Simulations: Quantitative Evidence from Long Molecular Dynamics Simulations. Curr. Opin. Struct. Biol. 2014, 24, 98–105. [55] Rauscher, S.; Gapsys, V.; Gajda, M.; Zweckstetter, M.; de Groot, B.L.; Grubm¨ uller, H. Structural Ensembles of Intrinsically Disordered Proteins Depend Strongly on Force Field: A Comparison to Experiment. J. Chem. Theory Comput. 2015, 11, 5513–5524. [56] Abascal, J.; Vega, C. A General Purpose Model for the Condensed Phases of Water: TIP4P/2005. J. Chem. Phys. 2005, 123, 234505. [57] Robustelli, P.; Piana, S.; Shaw, D. Developing a Molecular Dynamics Force Field for Both Folded and Disordered Protein States. Proc. Natl. Acad. Sci. U.S.A. 2018, 115, E4758–E4766. [58] Gerben, S.; Lemkul, J.; Brown, A.; Bevan, D. Comparing Atomistic Molecular Mechanics Force Fields for a Difficult Target: A Case Study on the Alzheimer’s Amyloid β-peptide. J. Biomol. Struct. Dyn. 2014, 32, 1817–1832.

33



[59] Somavarapu, A.; Kepp, K. The Dependence of Amyloid-β Dynamics on Protein Force Fields and Water Models. ChemPhysChem 2015, 16, 3278–3289. [60] Hoffmann, K.; McGovern, M.; Chiu, C.; Pablo, J. Secondary Structure of Rat and Human Amylin across Force Fields. PLOS One 2015, 10, e0134091. [61] Carballo-Pacheco, M.; Strodel, B. Comparison of Force Fields for Alzheimer’s Aβ42: A Case Study for Intrinsically Disordered Proteins. Protein Science 2016, 26, 174–185. [62] Palazzesi, F.; Prakash, M.; Bonomi, M.; Barducci, A. Accuracy of Current AllAtom Force-Fields in Modeling Protein Disordered States. J. Chem. Theory Comput. 2015, 11, 2–7. [63] Chen, W.; Shi, C.; MacKerell, A.; Shen, J. Conformational Dynamics of Two Natively Unfolded Fragment Peptides: Comparison of the AMBER and CHARMM Force Fields. J. Phys. Chem. B 2015, 119, 7902–7910. [64] Bomblies, R.; Luitz, M.; Scanu, S.; Madl, T.; Zacharias, M. Transient Helicity in Intrinsically Disordered AXIN-1 Studied by NMR Spectroscopy and Molecular Dynamics Simulations. PLoS One 2017, 12, e0174337. [65] Ouyang, Y.; Zhao, L.; Zhang, Z. Characterization of the Structural Ensembles of p53 TAD2 by Molecular Dynamics Simulations with Different Force Fields. Phys. Chem. Chem. Phys. 2018, 20, 8676–8684. [66] Ye, W.; Ji, D.; Wang, W.; Luo, R.; Chen, H. Test and Evaluation of FF99IDPs Force Field for Intrinsically Disordered Proteins. J. Chem. Inf. Model. 2015, 55, 1021–1029. [67] Duong, V.; Chen, Z.; Thapa, M.; Luo, R. Computational Studies of Intrinsically Disordered Proteins. J. Phys. Chem. B 2018, 122, 10455–10469. [68] Graen, T.; Klement, R.; Grupi, A.; Haas, E.; Grubm¨ uller, H. Transient Secondary and Tertiary Structure Formation Kinetics in the Intrinsically Disordered State of α-Synuclein from Atomistic Simulations. ChemPhysChem 2018, 19, 2507–2511.

34


Page 34 of 49

Page 35 of 49 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


[69] Mercadante, D.; Milles, S.; Fuertes, G.; Svergun, D.; Lemke, E.; Gr¨ ater, F. Kirkwood-Buff Approach Rescues Overcollapse of a Disordered Protein in Canonical Protein Force Fields. J. Phys. Chem. B 2015, 119, 7975–7984. [70] Henriques, J.; Skep¨ o, M. Molecular Dynamics Simulations of Intrinsically Disordered Proteins: On the Accuracy of the TIP4P-D Water Model and the Representativeness of Protein Disorder Models. J. Chem. Theory Comput. 2016, 12, 3407–3415. [71] Henriques, J.; Cragnell, C.; Skep¨ o, M. Molecular Dynamics Simulations of Intrinsically Disordered Proteins: Force Field Evaluation and Comparison with Experiment. J. Chem. Theory Comput. 2015, 11, 3420–3431. [72] Karp, J.; Erylimaz, E.; Cowburn, D. Correlation of Chemical Shifts Predicted by Molecular Dynamics Simulations for Partially Disordered Proteins. J. Biomol. NMR 2015, 61, 35–45. [73] Sz¨ ollösi, D.; Horv´ ath, T.; Han, K.; Dokholyan, N.; Tompa, P.; Kalmár, L.; Heged¨ us, T. Discrete Molecular Dynamics can Predict Helical Prestructured Motifs in Disordered Proteins. PLoS One 2014, 9, e95795. [74] Wen, B.; Peng, J.; Zuo, X.; Gong, Q.; Zhang, Z. Characterization of Protein Flexibility using Small-Angle X-Ray Scattering and Amplified Collective Motion Simulations. Biophys. J. 2014, 107, 956–964. [75] Bernad´ o, P.; Mylonas, E.; Petoukhov, M.; Blackledge, M.; Svergun, D. Structural Characterization of Flexible Proteins using Small-Angle X-ray Scattering. J. Am. Chem. Soc. 2007, 129, 5656–5664. [76] Terakawa, T.; Takada, S. Multiscale Ensemble Modeling of Intrinsically Disordered Proteins: p53 N-Terminal Domain. Biophys. J. 2011, 101, 1450–1458. [77] Franke, D.; Petoukhov, M.; Konarev, P.; Panjkovich, A.; Tuukkanen, A.; Mertens, H.; Kikhney, A.; Hajizadeh, N.; Franklin, J.; Jeffries, C.; Svergun, D. ATSAS 2.8: A Comprehensive Data Analysis Suite for Small-Angle Scattering from Macromolecular Solutions. J. Appl. Crystallogr. 2017, 50, 1212–1225. [78] Zweckstetter, M.; Bax, A. Prediction of Sterically Induced Alignment in a Dilute

35



Liquid Crystalline Phase: Aid to Protein Structure Determination by NMR. J. Am. Chem. Soc. 2000, 122, 3791–3792. [79] Terakawa, T.; Higo, J.; Takada, S. Multi-scale Ensemble Modeling of Modular Proteins with Intrinsically Disordered Linker Regions: Application to p53. Biophys. J. 2014, 107, 721–729. [80] Xu X.P..; Case D.A.. Probing Multiple Effects on 15N, 13Cα, 13Cβ, and 13C’ Chemical Shifts in Peptides using Density Functional Theory. Biopolymers 2002, 65, 408–423. [81] Shen, Y.; Bax, A. SPARTA+: A Modest Improvement in Empirical NMR Chemical Shift Prediction by Means of an Artificial Neural Network. J. Biomol. NMR 2010, 48, 13–22. [82] Kohlhoff, K.; Robustelli, P.; Cavalli, A.; Salvatella, X.; Vendruscolo, M. Fast and Accurate Predictions of Protein NMR Chemical Shifts from Interatomic Distances. J. Am. Chem. Soc. 2009, 131, 13894–13895. [83] Han, B.; Liu, Y.; Ginzinger, S.; Wishart, D. SHIFTX2: Significantly Improved Protein Chemical Shift Prediction. J. Biomol. NMR 2011, 50, 43. [84] Seidel, K.; Etzkorn, M.; Schneider, R.; Ader, C.; Baldus, M. Comparative Analysis of NMR Chemical Shift Predictions for Proteins in the Solid Phase. Solid State Nucl. Magn. Reson. 2009, 35, 235–242. [85] Christensen, A.; Sauer, S.; Jensen, J. Definitive Benchmark Study of Ring Current Effects on Amide Proton Chemical Shifts. J. Chem. Theory Comput. 2011, 7, 2078–2084. [86] Christensen, A.; Linnet, T.; Borg, M.; Boomsma, W.; Lindorff-Larsen, K.; Hamelryck, T.; Jensen, J. Protein Structure Validation and Refinement using Amide Proton Chemical Shifts Derived from Quantum Mechanics. PLOS One 2013, 8, e84123. [87] Svergun, D.; Barberato, C.; Koch, M. CRYSOL - a Program to Evaluate X-ray Solution Scattering of Biological Macromolecules from Atomic Coordinates. J. Appl. Crystallogr. 1995, 28, 768–773.

36


Page 36 of 49

Page 37 of 49 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


[88] Schneidman-Duhovny, D.; Hammel, M.; Tainer, J.; Sali, A. Accurate SAXS Profile Computation and its Assessment by Contrast Variation Experiments. Biophys. J. 2013, 105, 962–974. [89] Schneidman-Duhovny, D.; Hammel, M.; Tainer, J.; Sali, A. FoXS, FoXSDock and MultiFoXS: Single-state and Multi-state Structural Modeling of Proteins and their Complexes Based on SAXS Profiles. Nucl. Acids Res. 2016, 44, W424– W429. [90] Grudinin, S.; Garkavenko, M.; Kazennov, A. Pepsi-SAXS: An Adaptive Method for Rapid and Accurate Computation of Small-Angle X-ray Scattering Profiles. Acta Crystallogr. D Struct. Biol. 2017, 73, 449–464. [91] Henriques, J.; Arleth, L.; Lindorff-Larsen, K.; Skepö, M. On the Calculation of SAXS Profiles of Folded and Intrinsically Disordered Proteins from Computer Simulations. J. Mol. Biol. 2018, 430, 2521–2539. [92] Chan-Yao-Chong, M.; Deville, C.; Pinet, L.; van Heijenoort, C.; Durand, D.; Ha-Duong, T. Structural Characterization of N-WASP Domain V Combining NMR and SAXS Data with MD Simulations. Accepted 2018. [93] Trewhella, J.; Duff, A.; Durand, D.; Gabel, F.; Guss, J.; Hendrickson, W.; Hura, G.; Jacques, D.; Kirby, N.; Kwan, A.; Pérez, J.; Pollack, L.; Ryan, T.; Sali, A.; Schneidman-Duhovny, D.; Schwede, T.; Svergun, D.; Sugiyama, M.; Tainer, J.; Vachette, P.; Westbrook, J.; Whitten, A. 2017 Publication Guidelines for Structural Modelling of Small-Angle Scattering Data from Biomolecules in Solution: An Update. Acta Crystallogr. D Struct. Biol. 2017, 73, 710–728. [94] Chen, P.; Hub, J. Validating Solution Ensembles from Molecular Dynamics Simulation by Wide-Angle X-ray Scattering Data. Biophys. J. 2014, 107, 435– 447. [95] Knight, C.; Hub, J. WAXSiS: A Web Server for the Calculation of SAXS/WAXS Curves Based on Explicit-solvent Molecular Dynamics. Nucl. Acids Res. 2015, 43, W225–W230. [96] Cordeiro, T.; Chen, P.; De Biasio, A.; Sibille, N.; Blanco, F.; Hub, J.; Crehuet, R.; Bernadó, P. Disentangling Polydispersity in the PCNA–p15PAF Complex, a

37



Disordered, Transient and Multivalent Macromolecular Assembly. Nucleic Acids Res. 2017, 45, 1501. [97] Salomon-Ferrer, R.; Case, D.; Walker, R. An Overview of the Amber Biomolecular Simulation Package: Amber Biomolecular Simulation Package. WIREs Comput. Mol. Sci. 2013, 3, 198–210. [98] Brooks, B.; Brooks, C.; Mackerell, A.; Nilsson, L.; Petrella, R.; Roux, B.; Won, Y.; Archontis, G.; Bartels, C.; Boresch, S.; Caflisch, A.; Caves, L.; Cui, Q.; Dinner, A.; Feig, M.; Fischer, S.; Gao, J.; Hodoscek, M.; Im, W.; Kuczera, K.; Lazaridis, T.; Ma, J.; Ovchinnikov, V.; Paci, E.; Pastor, R.; Post, C.; Pu, J.; Schaefer, M.; Tidor, B.; Venable, R.; Woodcock, H.; Wu, X.; Yang, W.; York, D.; Karplus, M. CHARMM: The Biomolecular Simulation Program. J. Comput. Chem. 2009, 30, 1545–1614. [99] Abraham, M.; Murtola, T.; Schulz, R.; Pall, S.; Smith, J.; Hess, B.; Lindahl, E. GROMACS: High Performance Molecular Simulations through Multi-level Parallelism from Laptops to Supercomputers. SoftwareX 2015, 1-2, 19–25. [100] Phillips, J.; Braun, R.; Wang, W.; Gumbart, J.; Tajkhorshid, E.; Villa, E.; Chipot, C.; Skeel, R.; Kalé, L.; Schulten, K. Scalable Molecular Dynamics with NAMD. J. Comput. Chem. 2005, 26, 1781–1802. [101] Nodet, G.; Salmon, L.; Ozenne, V.; Meier, S.; Jensen, M.; Blackledge, M. Quantitative Description of Backbone Conformational Sampling of Unfolded Proteins at Amino Acid Resolution from NMR Residual Dipolar Couplings. J. Am. Chem. Soc. 2009, 131, 17908–17918. [102] Berlin, K.; Castaneda, C.; Schneidman-Duhovny, D.; Sali, A.; Nava-Tudela, A.; Fushman, D. Recovering a Representative Conformational Ensemble from Underdetermined Macromolecular Structural Data. J. Am. Chem. Soc. 2013, 135, 16595–16609. [103] Choy, W.; Forman-Kay, J. Calculation of Ensembles of Structures Representing the Unfolded State of an SH3 Domain. J. Mol. Biol. 2001, 308, 1011–1032. [104] Rozycki, B.; Kim, Y.; Hummer, G. SAXS Ensemble Refinement of ESCRT-III CHMP3 Conformational Transitions. Structure 2011, 19, 109–116.

38


Page 38 of 49

Page 39 of 49 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


[105] Leung, H.; Bignucolo, O.; Aregger, R.; Dames, S.; Mazur, A.; Bernèche, S.; Grzesiek, S. A Rigorous and Efficient Method to Reweight very Large Conformational Ensembles using Average Experimental Data and to Determine their Relative Information Content. J. Chem. Theory Comput. 2016, 12, 383–394. [106] Lautz, J.; Kessler, H.; Boelens, R.; Kaptein, R.; van Gunsteren, W. Conformational Analysis of a Cyclic Thymopoietin-analogue by 1H N.M.R. Spectroscopy and Restrained Molecular Dynamics Simulations. Int. J. Pept. Protein Res. 1987, 30, 404–414. [107] Han, K.; Syi, J.; Brooks, B.; Ferretti, J. Solution Conformations of the B-loop Fragments of Human Transforming Growth Factor Alpha and Epidermal Growth Factor by 1H Nuclear Magnetic Resonance and Restrained Molecular Dynamics. Proc. Natl. Acad. Sci. U.S.A. 1990, 87, 2818–2822. [108] Kemmink, J.; Scheek, R. Dynamic Modelling of a Helical Peptide in Solution using NMR Data: Multiple Conformations and Multi-spin Effects. J. Biomol. NMR 1995, 6, 33–40. [109] Lindorff-Larsen, K.; Best, R.; DePristo, M.; Dobson, C.; Vendruscolo, M. Simultaneous Determination of Protein Structure and Dynamics. Nature 2005, 433, 5. [110] Fuentes, G.; Nederveen, A.; Kaptein, R.; Boelens, R.; Bonvin, A. Describing Partially Unfolded States of Proteins from Sparse NMR Data. J. Biomol. NMR 2005, 33, 175–186. [111] Camilloni, C.; Vendruscolo, M. Statistical Mechanics of the Denatured State of a Protein using Replica-Averaged Metadynamics. J. Am. Chem. Soc. 2014, 136, 8982–8991. [112] Granata, D.; Camilloni, C.; Vendruscolo, M.; Laio, A. Characterization of the Free-energy Landscapes of Proteins by NMR-guided Metadynamics. Proc. Natl. Acad. Sci. U.S.A. 2013, 110, 6817–6822. [113] Granata, D.; Baftizadeh, F.; Habchi, J.; Galvagnion, C.; De Simone, A.; Camilloni, C.; Laio, A.; Vendruscolo, M. The Inverted Free Energy Landscape of

39



Page 40 of 49

an Intrinsically Disordered Peptide by Simulations and Experiments. Sci. Rep. 2015, 5. [114] Pitera, J.; Chodera, J. On the Use of Experimental Observations to Bias Simulated Ensembles. J. Chem. Theory Comput. 2012, 8, 3445–3451. [115] Roux, B.; Weare, J. On the Statistical Equivalence of Restrained-ensemble Simulations with the Maximum Entropy Method. J. Chem. Phys. 2013, 138, 084107. [116] Cavalli, A.; Camilloni, C.; Vendruscolo, M. Molecular Dynamics Simulations with Replica-averaged Structural Restraints Generate Structural Ensembles According to the Maximum Entropy Principle. J. Chem. Phys. 2013, 138, 094112. [117] Boomsma, W.; Ferkinghoff-Borg, J.; Lindorff-Larsen, K. Combining Experiments and Simulations using the Maximum Entropy Principle. PLOS Comput. Biol. 2014, 10, e1003406. [118] Chen, P.; Hub, J. Interpretation of Solution X-Ray Scattering by Explicit-Solvent Molecular Dynamics. Biophys. J. 2015, 108, 2573–2584. [119] Kimanius, D.; Pettersson, I.; Schluckebier, G.; Lindahl, E.; Andersson, M. SAXSGuided Metadynamics. J. Chem. Theory Comput. 2015, 11, 3491–3498. [120] Ball, K.; Wemmer, D.; Head-Gordon, T. Comparison of Structure Determination Methods for Intrinsically Disordered Amyloid-β Peptides. J. Phys. Chem. B 2014, 118, 6405–6416. [121] Rieping, W.; Habeck, M.; Nilges, M. Inferential Structure Determination. Science 2005, 309, 303–306. [122] Fisher, C.; Huang, A.; Stultz, C. Modeling Intrinsically Disordered Proteins with Bayesian Statistics. J. Am. Chem. Soc. 2010, 132, 14919–14927. [123] Hummer, G.; K¨ ofinger, J. Bayesian Ensemble Refinement by Replica Simulations and Reweighting. J. Chem. Phys. 2015, 143, 243150. [124] Bonomi, M.; Camilloni, C.; Cavalli, A.; Vendruscolo, M. A Bayesian Inference Method for Heterogeneous Systems. 2, e1501177.

40


Metainference: Sci. Adv. 2016,

Page 41 of 49 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


[125] Dunker, A.; Brown, C.; Obradović, Z. Identification And Functions of Usefully Disordered Proteins. In Unfolded Proteins; Academic Press, 2002; Vol. 62, Advances in Protein Chemistry, pp. 25–49. [126] Dunker, K.; Brown, C.; Lawson, D.; Iakoucheva, L.; Obradović, Z. Intrinsic Disorder And Protein Function. Biochemistry 2002, 41, 6573–6582. [127] Tompa, P. The Interplay between Structure and Function in Intrinsically Unstructured Proteins. FEBS Letters 2005, 579, 3346–3354. [128] Uversky, V.; Oldfield, C.; Dunker, K. Intrinsically Disordered Proteins in Human Diseases: Introducing the D2 Concept. Annu. Rev. Biophys. 2008, 37, 215–246. [129] Sipe, J.; Benson, M.; Buxbaum, J.; Ikeda, S.; Merlini, G.; Saraiva, M.; Westermark, P. Amyloid Fibril Protein Nomenclature: 2010 Recommendations from the Nomenclature Committee of the International Society of Amyloidosis. Amyloid 2010, 17, 101–104. [130] Laghaei, R.; Mousseau, N.; Wei, G.

Effect of the Disulfide Bond on the

Monomeric Structure of Human Amylin Studied by Combined Hamiltonian and Temperature Replica Exchange Molecular Dynamics Simulations. J. Phys. Chem. B 2010, 114, 7071–7077. [131] Qi, R.; Luo, Y.; Ma, B.; Nussinov, R.; Wei, G. Conformational Distribution and α-Helix to β-Sheet Transition of Human Amylin Fragment Dimer. Biomacromolecules 2014, 15, 122–131. [132] Sgourakis, N.; Yan, Y.; McCallum, S.; Wang, C.; Garcia, A. The Alzheimers Peptides Aβ40 and 42 Adopt Distinct Conformations in Water: A Combined MD / NMR Study. J. Mol. Biol. 2007, 368, 1448–1457. [133] Yang, M.; Teplow, D. Amyloid β-Protein Monomer Folding: Free-Energy Surfaces Reveal Alloform-Specific Differences. J. Mol. Biol. 2008, 384, 450–464. [134] Cˆ oté, S.; Derreumaux, P.; Mousseau, N. Distinct Morphologies for Amyloid Beta Protein Monomer: Aβ 1-40 , Aβ 1-42 , And Aβ 1-40 (D23N). J. Chem. Theory Comput. 2011, 7, 2584–2592. [135] Wood, G.; Rothlisberger, U. Secondary Structure Assignment of Amyloid-β Peptide using Chemical Shifts. J. Chem. Theory Comput. 2011, 7, 1552–1563.

41



[136] Sgourakis, N.; Merced-Serrano, M.; Boutsidis, C.; Drineas, P.; Du, Z.; Wang, C.; Garcia, A. Atomic-level Characterization of the Ensemble of the Aβ(1-42) Monomer in Water using Unbiased Molecular Dynamics Simulations and Spectral Algorithms. J. Mol. Biol. 2011, 405, 570–583. [137] Ball, A.; Phillips, A.; Nerenberg, P.; Fawzi, N.; Wemmer, D.; Head-Gordon, T. Homogeneous and Heterogeneous Tertiary Structure Ensembles of Amyloid-β Peptides. Biochemistry 2011, 50, 7612–7628. [138] Olubiyi, O.; Strodel, B. Structures of the Amyloid β-Peptides Aβ1-40 and Aβ1-42 as Influenced by pH and a d-Peptide. J. Phys. Chem. B 2012, 116, 3280–3291. [139] Viet, M.; Li, M. Amyloid Peptide Aβ40 Inhibits Aggregation of Aβ42: Evidence from Molecular Dynamics Simulations. J. Chem. Phys. 2012, 136, 245105. [140] Rosenman, D.; Connors, C.; Chen, W.; Wang, C.; Garc´ıa, A. Aβ Monomers Transiently Sample Oligomer and Fibril-like Configurations: Ensemble Characterization using a Combined MD/NMR Approach. J. Mol. Biol. 2013, 425, 3338– 3359. [141] Ball, K.; Phillips, A.; Wemmer, D.; Head-Gordon, T. Differences in β-strand Populations of Monomeric Aβ40 and Aβ42. Biophys. J. 2013, 104, 2714–2724. [142] Tran, L.; Basdevant, N.; Prévost, C.; Ha-Duong, T. Structure of Ring-shaped Aβ42 Oligomers Determined by Conformational Selection. Sci. Rep. 2016, 6. [143] Lyons, A.; Gandhi, N.; Mancera, R. Molecular Dynamics Simulation of the Phosphorylation-induced Conformational Changes of a Tau Peptide Fragment: Tau Peptide Conformation Changes: An MD Study. Proteins 2014, 82, 1907– 1923. [144] Luo, Y.; Ma, B.; Nussinov, R.; Wei, G. Structural Insight into Tau Proteins Paradox of Intrinsically Disordered Behavior, Self-Acetylation Activity, and Aggregation. J. Phys. Chem. Lett. 2014, 5, 3026–3031. [145] Battisti, A.; Ciasca, G.; Grottesi, A.; Tenenbaum, A. Thermal Compaction of the Intrinsically Disordered Protein Tau: Entropic, Structural, and Hydrophobic Factors. Phys. Chem. Chem. Phys. 2017, 19, 8435–8446.

42


Page 42 of 49

Page 43 of 49 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


[146] Narayanan, C.; Weinstock, D.; Wu, K.; Baum, J.; Levy, R. Investigation of the Polymeric Properties of α-Synuclein and Comparison with NMR Experiments: A Replica Exchange Molecular Dynamics Study. J. Chem. Theory Comput. 2012, 8, 3929–3942. [147] Allison, J.; Rivers, R.; Christodoulou, J.; Vendruscolo, M.; Dobson, C. A Relationship between the Transient Structure in the Monomeric State and the Aggregation Propensities of α-Synuclein and β-Synuclein. Biochemistry 2014, 53, 7170–7183. [148] Maupetit, J.; Tuffery, P.; Derreumaux, P. A Coarse-grained Protein Force Field for Folding and Structure Prediction. Proteins: Struct., Funct., Bioinf. 2007, 69, 394–408. [149] Chebaro, Y.; Pasquali, S.; Derreumaux, P. The Coarse-Grained OPEP Force Field for Non-Amyloid and Amyloid Proteins. J. Phys. Chem. B 2012, 116, 8741– 8752. [150] Mongan, J.; Simmerling, C.; McCammon, J.; Case, D.; Onufriev, A. Generalized Born Model with a Simple, Robust Molecular Volume Correction. J. Chem. Theory Comput. 2007, 3, 156–169. [151] Gallicchio, E.; Levy, R. AGBNP: An Analytic Implicit Solvent Model Suitable for Molecular Dynamics Simulations and High-resolution Modeling. J. Comput. Chem. 2004, 25, 479–499. [152] Lazaridis, T.; Karplus, M. Effective Energy Function for Proteins in Solution. Proteins: Struct., Funct., Bioinf. 1999, 35, 133–152. [153] Patil, S.; Xu, S.; Sheftic, S.; Alexandrescu, A. Dynamic α-Helix Structure of Micelle-bound Human Amylin. J. Biol. Chem. 2009, 284, 11982–11991. [154] Nanga, R.; Brender, J.; Vivekanandan, S.; Ramamoorthy, A. Structure and Membrane Orientation of IAPP in its Natively Amidated Form at Physiological pH in a Membrane Environment. Biochim. Biophys. Acta - Biomembranes 2011, 1808, 2337–2342. [155] Yonemoto, I.; Kroon, G.; Dyson, H.; Balch, W.; Kelly, J. Amylin Proprotein

43



Processing Generates Progressively More Amyloidogenic Peptides that Initially Sample the Helical State. Biochemistry 2008, 47, 9900–9910. [156] Williamson, J.; Loria, J.; Miranker, A. Helix Stabilization Precedes Aqueous and Bilayer-Catalyzed Fiber Formation in Islet Amyloid Polypeptide. J. Mol. Biol. 2009, 393, 383–396. [157] Liu, G.; Prabhakar, A.; Aucoin, D.; Simon, M.; Sparks, S.; Robbins, K.; Sheen, A.; Petty, S.; Lazo, N. Mechanistic Studies of Peptide Self-Assembly: Transient α-Helices to Stable β-Sheets. J. Am. Chem. Soc. 2010, 132, 18223–18232. [158] Tran, L.; Ha-Duong, T. Insights into the Conformational Ensemble of Human Islet Amyloid Polypeptide from Molecular Simulations. Current Pharmaceutical Design 2016, 22, 3601–3607. [159] Mukrasch, M.; Bibow, S.; Korukottu, J.; Jeganathan, S.; Biernat, J.; Griesinger, C.; Mandelkow, E.; Zweckstetter, M. Structural Polymorphism of 441-Residue Tau at Single Residue Resolution. PLOS Biol. 2009, 7, e1000034. [160] Uversky, V. Natively Unfolded Proteins: A Point where Biology Waits for Physics. Protein Sci. 2002, 11, 739–756. [161] Dedmon, M.; Lindorff-Larsen, K.; Christodoulou, J.; Vendruscolo, M.; Dobson, C. Mapping Long-Range Interactions in α-Synuclein using Spin-Label NMR and Ensemble Molecular Dynamics Simulations. J. Am. Chem. Soc. 2005, 127, 476– 477. [162] Bertoncini, C.; Jung, Y.; Fernandez, C.; Hoyer, W.; Griesinger, C.; Jovin, T.; Zweckstetter, M. Release of Long-range Tertiary Interactions Potentiates Aggregation of Natively Unstructured α-synuclein. Proceedings of the National Academy of Sciences 2005, 102, 1430–1435. [163] Fuxreiter, M.; Simon, I.; Friedrich, P.; Tompa, P. Preformed Structural Elements Feature in Partner Recognition by Intrinsically Unstructured Proteins. J. Mol. Biol. 2004, 338, 1015–1026. [164] Mohan, A.; Oldfield, C.; Radivojac, P.; Vacic, V.; Cortese, M.; Dunker, A.; Uversky, V. Analysis of Molecular Recognition Features (MoRFs). J. Mol. Biol. 2006, 362, 1043–1059.

44


Page 44 of 49

Page 45 of 49 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


[165] Yoon, M.; Venkatachalam, V.; Huang, A.; Choi, B.; Stultz, C.; Chou, J. Residual Structure Within the Disordered C-terminal Segment of P21Waf1/Cip1/Sdi1 and its Implications for Molecular Recognition. Protein Sci. 2009, 18, 337–347. [166] Mittal, J.; Yoo, T.; Georgiou, G.; Truskett, T. Structural Ensemble of an Intrinsically Disordered Polypeptide. J. Phys. Chem. B 2013, 117, 118–124. [167] Ganguly, D.; Chen, J. Modulation of the Disordered Conformational Ensembles of the p53 Transactivation Domain by Cancer-Associated Mutations. PLoS Comput. Biol. 2015, 11, e1004247. [168] Robustelli, P.; Trbovic, N.; Friesner, R.; Palmer, A. Conformational Dynamics of the Partially Disordered Yeast Transcription Factor GCN4. J. Chem. Theory Comput. 2013, 9, 5190–5200. [169] Krieger, J.; Fusco, G.; Lewitzky, M.; Simister, P.; Marchant, J.; Camilloni, C.; Feller, S.; De Simone, A. Conformational Recognition of an Intrinsically Disordered Protein. Biophys. J. 2014, 106, 1771–1779. [170] Zhang, Y.; Wen, B.; Peng, J.; Zuo, X.; Gong, Q.; Zhang, Z. Determining Structural Ensembles of Flexible Multi-domain Proteins using Small-Angle X-ray Scattering and Molecular Dynamics Simulations. Protein Cell 2015, 6, 619–623. [171] Xun, S.; Jiang, F.; Wu, Y. Intrinsically Disordered Regions Stabilize the Helical Form of the C-terminal Domain of RFAH: A Molecular Dynamics Study. Bioorg. Med. Chem. 2016, 24, 4970–4977. [172] Malard, F.; Assrir, N.; Alami, M.; Messaoudi, S.; Lescop, E.; Ha-Duong, T. Conformational Ensemble and Biological Role of the TCTP Intrinsically Disordered Region: Influence of Calcium and Phosphorylation. J. Mol. Biol. 2018, 430, 1621–1639. [173] Knott, M.; Best, R. A Preformed Binding Interface in the Unbound Ensemble of an Intrinsically Disordered Protein: Evidence from Molecular Simulations. PLoS Comput. Biol. 2012, 8, e1002605. [174] Zeng, J.; Jiang, F.; Wu, Y. Mechanism of Phosphorylation-Induced Folding of 4E-BP2 Revealed by Molecular Dynamics Simulations. J. Chem. Theory Comput. 2017, 13, 320–328.

45



[175] Wu, S.; Wang, D.; Liu, J.; Feng, Y.; Weng, J.; Li, Y.; Gao, X.; Liu, J.; Wang, W. The Dynamic Multisite Interactions between Two Intrinsically Disordered Proteins. Angew. Chem. Int. Ed. 2017, 56, 7515–7519. [176] Zhang, W.; Chen, J. Replica Exchange with Guided Annealing for Accelerated Sampling of Disordered Protein Conformations. J. Comput. Chem. 2014, 35, 1682–1689. [177] Im, W.; Lee, M.; Brooks, C. Generalized Born Model with a Simple Smoothing Function. J. Comput. Chem. 2003, 24, 1691–1702. [178] Im, W.; Feig, M.; Brooks, C. An Implicit Membrane Generalized Born Theory for the Study of Structure, Stability, And Interactions of Membrane Proteins. Biophys. J. 2003, 85, 2900–2918. [179] Chen, J.; Im, W.; Brooks, C. Balancing Solvation and Intramolecular Interactions. J. Am. Chem. Soc. 2006, 128, 3728–3736. [180] Jiang, F.; Han, W.; Wu, Y. The Intrinsic Conformational Features of Amino Acids from a Protein Coil Library and their Applications in Force Field Development. Phys. Chem. Chem. Phys. 2013, 15, 3413–3428. [181] Jiang, F.; Zhou, C.; Wu, Y. Residue-Specific Force Field Based on the Protein Coil Library. RSFF1: Modification of OPLS-AA/L. J. Phys. Chem. B 2014, 118, 6983–6998. [182] Zhou, C.; Jiang, F.; Wu, Y. Residue-Specific Force Field Based on Protein Coil Library. RSFF2: Modification of AMBER FF99SB. J. Phys. Chem. B 2015, 119, 1035–1047. [183] Uversky, V.; Oldfield, C.; Dunker, A. Showing your ID: Intrinsic Disorder as an ID for Recognition, Regulation and Cell Signaling. J. Mol. Recognit. 2005, 18, 343–384. [184] Xue, B.; Blocquel, D.; Habchi, J.; Uversky, A.; Kurgan, L.; Longhi, S. Structural Disorder in Viral Proteins, 2014. [185] Lee, C.; Kim, D.; Lee, S.; Su, J.; Han, K. Structural Investigation on the Intrinsically Disordered N-terminal Region of HPV16 E7 Protein. BMB Rep. 2016, 49, 431–436.

46


Page 46 of 49

Page 47 of 49 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


[186] Wang, Y.; Chu, X.; Longhi, S.; Roche, P.; Han, W.; Wang, E.; Wang, J. Multiscaled Exploration of Coupled Folding and Binding of an Intrinsically Disordered Molecular Recognition Element in Measles Virus Nucleoprotein. Proc. Natl. Acad. Sci. U.S.A. 2013, 110, E3743–E3752. [187] Leyrat, C.; Renner, M.; Harlos, K.; Grimes, J. Solution and Crystallographic Structures of the Central Region of the Phosphoprotein from Human Metapneumovirus. PLOS One 2013, 8, e80371. [188] Boon, P.; Saw, W.; Lim, X.; Raghuvamsi, P.; Huber, R.; Marzinek, J.; Holdbrook, D.; Anand, G.; Gr¨ uber, G.; Bond, P. Partial Intrinsic Disorder Governs the Dengue Capsid Protein Conformational Ensemble. ACS Chem. Biol. 2018, 13, 1621–1630. [189] Noval, M.; Gallo, M.; Perrone, S.; Salvay, A.; Chemes, L.; de Prat-Gay, G. Conformational Dissection of a Viral Intrinsically Disordered Domain Involved in Cellular Transformation. PLOS One 2013, 8, e72760. [190] Milles, S.; Jensen, M.; Lazert, C.; Guseva, S.; Ivashchenko, S.; Communie, G.; Maurin, D.; Gerlier, D.; Ruigrok, R.; Blackledge, M. An Ultraweak Interaction in the Intrinsically Disordered Replication Machinery is Essential for Measles Virus Function. Science Advances 2018, 4, eaat7778. [191] Troilo, F.; Bignon, C.; Gianni, S.; Fuxreiter, M.; Longhi, S. Chapter Seven - Experimental Characterization of Fuzzy Protein Assemblies: Interactions of Paramyxoviral NTAIL Domains with their Functional Partners. In Intrinsically Disordered Proteins; Rhoades, E., Ed.; Academic Press, 2018; Vol. 611, Methods in Enzymology, pp. 137 – 192. [192] Jensen, M.; Communie, G.; Ribeiro, E.; Martinez, N.; Desfosses, A.; Salmon, L.; Mollica, L.; Gabel, F.; Jamin, M.; Longhi, S.; Ruigrok, R.; Blackledge, M. Intrinsic Disorder in Measles Virus Nucleocapsids. Proc. Natl. Acad. Sci. U.S.A. 2011, 108, 9839–9844. [193] Samsa, M.; Mondotte, J.; Iglesias, N.; Assuncao-Miranda, I.; Barbosa-Lima, G.; Da Poian, A.; Bozza, P.; Gamarnik, A. Dengue Virus Capsid Protein Usurps Lipid Droplets for Viral Particle Formation. PLOS Pathogens 2009, 5, 1–14.

47



[194] Sigalov, A.; Zhuravleva, A.; Vladislav Yu. Orekhov, V. Binding of Intrinsically Disordered Proteins is not Necessarily Accompanied by a Structural Transition to a Folded Form. Biochimie 2007, 89, 419–421. [195] Tompa, P.; Fuxreiter, M. Fuzzy Complexes: Polymorphism and Structural Disorder in Protein-protein Interactions. Trends Biochem. Sci. 2008, 33, 2–8. [196] Sharma, R.; Raduly, Z.; Miskei, M.; Fuxreiter, M. Fuzzy Complexes: Specific Binding without Complete Folding. FEBS Lett. 2015, 589, 2533–2542. [197] Olsen, J.; Teilum, K.; Kragelund, B. Behaviour of Intrinsically Disordered Proteins in Protein-protein Complexes with an Emphasis on Fuzziness. Cell. Mol. Life Sci. 2017, 74, 3175–3183.

48


Page 48 of 49

Page 49 of 49 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


NMR data MD ensemble

SAXS intensity


or SAXS Data for

Recommend Documents