Dynamics, conformational entropy, and frustration in protein-protein

Subscriber access provided by UNIV OF CAMBRIDGE

Dynamics, conformational entropy, and frustration in protein-protein interactions involving an intrinsically disordered protein domain Ida Lindström, and Jakob Dogan ACS Chem. Biol., Just Accepted Manuscript • DOI: 10.1021/acschembio.7b01105 • Publication Date (Web): 03 Apr 2018 Downloaded from http://pubs.acs.org on April 4, 2018

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

ACS Chemical Biology

Dynamics, conformational entropy, and frustration in protein-protein interactions involving an intrinsically disordered protein domain Ida Lindström1 and Jakob Dogan1* 1

Department of Biochemistry and Biophysics, Stockholm University, 10691 Stockholm, Sweden * Corresponding author: e-mail: [email protected] Telephone: 46-8-162470

1

ACS Paragon Plus Environment

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

Abstract Intrinsically disordered proteins (IDPs) are abundant in the eukaryotic proteome. However, little is known about the role of sub-nanosecond dynamics and the conformational entropy that it represents in protein-protein interactions involving IDPs. Using NMR side chain and backbone relaxation, stopped-flow kinetics, ITC, and computational studies, we have here characterized the interaction between the globular TAZ1 domain of CREB binding protein, and the intrinsically disordered transactivation domain of STAT2 (TAD-STAT2). We show that the TAZ1/TADSTAT2 complex retains considerable sub-nanosecond motions, with TAD-STAT2 undergoing only a partial disorder-to-order transition, We report here the first experimental determination of the conformational entropy change for both binding partners in an IDP binding interaction and find that the total change even exceeds in magnitude to the binding enthalpy and is comparable to the contribution from the hydrophobic effect, demonstrating its importance for the binding energetics. Furthermore, we show that the conformational entropy change for TAZ1 is also instrumental in maintaining a biologically meaningful binding affinity. Strikingly, a spatial clustering of very high amplitude motions and a cluster of more rigid sites in the complex exists, which we through computational studies found to overlap with regions that experience energetic frustration and are less frustrated, respectively. Thus, the residual dynamics in the bound state could be necessary for faster dissociation, which is important for proteins that interact with multiple binding partners.

2


Page 2 of 30

Page 3 of 30 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


Introduction Protein-protein interactions are instrumental for many life processes, and a better understanding of the structural basis for molecular recognition of such complexes has been obtained over the years. However, proteins are dynamic entities, which fluctuate over a wide range of amplitudes and time scales. On one end of this dynamic continuum are the so-called intrinsically disordered proteins (IDPs), which often undergo a disorder-to-order transition upon complex formation.1, 2 IDPs have been recognized for some time to be prevalent in the eukaryotic proteome, with about 30-40 % either being completely disordered or contain long disordered regions.2 IDPs are frequently involved in protein-protein interactions in processes such as signal transduction, transcription, cell-cycle control, or chaperoning.1 They have received a significant amount of attention the last decade due to their abundance and functional relevance as well as the fact that many are implicated in various kinds of diseases such as cancer and neurodegenerative disorders,3 which have made them attractive targets for drug development.4 However, the role of sub-nanosecond dynamics in protein-protein interactions involving IDPs, how much the conformational entropy change contributes to the binding energetics, its connection to energetic frustration, the dynamic nature of hot-spots, and to what extent IDP complexes retain dynamics compared to those formed by folded proteins remain unclear.

NMR relaxation is the most powerful experimental method to obtain site-resolved information on the dynamical behavior of proteins. Studies in recent years have emerged in which nuclear spin relaxation has been used to elucidate various roles for dynamics in protein function.5-14 When addressing the role of protein internal motions it is important to include the investigation of side chain picosecond-nanosecond (ps-ns) dynamics, motions that have been shown to correspond to

3



significant conformational entropy, are heterogeneously dynamic and can have a large dynamic response upon moving from one state to another, such as binding to a ligand.5, 15 Unfortunately such studies are much less prevalent than backbone dynamics investigations. We have here performed NMR relaxation experiments in order to investigate the dynamics of both partners in the bound state in a binary protein-protein interaction involving the globular transcriptional adaptor zinc-binding 1 (TAZ1) domain of CREB binding protein (CBP) and the intrinsically disordered transactivation domain of STAT2 (TAD-STAT2) (Figure 1A), as well as TAZ1 in the free state. To the best of our knowledge this is the first study in which both backbone amide and side chain methyl group ps-ns dynamics have been investigated for both binding partners in a protein-protein complex involving an IDP. We found that bound TAD-STAT2 still experiences extensive picosecond-nanosecond dynamics in the bound state. The conformational entropy change for both TAZ1 and TAD-STAT2 contributes significantly to the binding energetics. Strikingly, on one side of the complex there is a cluster of rigid sites whereas on the opposite side there is a clustering of sites that are highly dynamic, a region which we through computational studies also found to experience energetic frustration, with implications for the binding mechanisms of functionally promiscuous proteins being discussed.

Results and Discussion We have performed both side chain and backbone NMR relaxation experiments for bound TAZ1, bound TAD-STAT2, and free TAZ1, and complemented with ITC, stopped-flow kinetics, and computational studies, with the aim to investigate the role of sub-nanosecond internal motions in molecular recognition involving a disordered protein. The NMR solution 3D structures of the protein-protein complex and free TAZ1 have been previously determined.16, 17 TAD-STAT2 is

4


Page 4 of 30

Page 5 of 30 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


completely disordered in the free state, which is demonstrated by the negative {1H}-15N NOEs (Supporting Information Figure S1), as well as its CD properties,18 and the low dispersion in the 15

N-HSQC.16 TAZ1 is a well-folded protein as described previously.17 The dissociation constant,

Kd, was determined to be 45 ± 9 nM by ITC (Figure 1B). The binding reaction is enthalpically driven (∆H=−15.1 kcal/mol) with an unfavorable total entropy change (−T∆Stot=4.9 kcal/mol) at 303 K (Figure 1B). ITC measurements performed in HEPES or MES buffer gave the same results, thus there are no protonation/deprotonation effects associated with the binding reaction.

Figure 1. A) 3D structure of TAZ1/TAD-STAT216 (pdb code 2KA4), the backbone of TAZ1 is shown in gray, and TAD-STAT2 in blue. The helices of TAZ1 are labeled in the figure. B) Global binding thermodynamics was characterized by ITC. Shown here is ITC data in which TAD-STAT2 (170 µM) was titrated into TAZ1 (13 µM).

5



Backbone dynamics The model-free squared generalized order parameter, O2NH, reports on the degree of motional restriction for the backbone amide group, ranging from 0 (completely unrestricted) to 1 (complete restriction). The O2NH parameters for free TAZ1 are generally high at helix regions, with lower O2NH values in loops, at the initiation and termination of helices, and in the terminal regions (Supporting Information Figure S2). The average value of O2NH at the helix regions is 0.85 ± 0.12. In particular, the loop between helix 1 and helix 2 exhibits high ps-ns motional amplitudes with O2NH values as low as 0.3. The O2NH values for bound TAZ1 follows the same trend as in free TAZ1, i.e. helical regions are rigid, and with loops and termini being exceptions (Supporting Information Figure S2). Although, the pairwise average difference of O2NH for free and bound TAZ1 is zero (0.005 ± 0.013) (Figure 2A), there is to some extent a redistribution of backbone dynamics upon binding. For example, the N-terminal end of helix 1 and the third loop in TAZ1 rigidifies slightly upon binding whereas the C-terminal part of helix 3 become slightly more dynamic upon complex formation. The backbone dynamics for free and bound TAZ1 is quite typical for well-defined proteins, in that the motions at secondary structures are restricted whereas more mobility is seen at loops, at the initiation and termination of helices, and at the Nand C-terminals.

However, the analysis of

15

N backbone relaxation for bound TAD-STAT2 reveals a distinct

feature; although the N-terminal region of TAD-STAT2 becomes significantly rigidified (residues 791-815), with an average O2NH=0.86 ± 0.06, the C-terminal region (residues 816-838) is still highly dynamic on the ps-ns time scale, with an average O2NH=0.33 ± 0.16 (Figure 2B). Thus, TAD-STAT2 undergoes only a partial disorder-to-order transition on the backbone level upon binding. However, the C-terminal region does make interactions with TAZ1 as shown by 6


Page 6 of 30

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


the 3D structure of the complex, and comparison of the steady-state NOE values for free and bound TAD-STAT2 (Supporting Information, Figure S1), suggests that the ps-ns dynamics is attenuated to some degree even for the C-terminal part upon binding to TAZ1.

Figure 2. A) Difference between the backbone O2NH parameter of bound and free TAZ1 vs residue number. B) The backbone O2NH parameter vs residue number for bound TAD-STAT2.

Side chain dynamics for bound and free TAZ1 and bound TAD-STAT2 We investigated side chain ps-ns motions in methyl-containing residues by deuterium relaxation, from which the squared generalized methyl axis order parameter, O2axis, and the effective internal

7



correlation time, τ, were obtained. We were able to determine the O2axis parameter for 40 methyl groups out of a total of 51 in free TAZ1 (Figure 3A and Supporting Information Table S3). The remaining methyls were either overlapped or had too weak intensities, preventing accurate determination of 2H relaxation rates. The determined O2axis values show that free TAZ1 exhibit a wide range of motional amplitudes (Figure 3A), from highly unrestrained motions (O2axis=0.18 for Met-387ε) to very restricted ps-ns dynamics (O2axis=0.87 for Ala-363β). It does not appear to exist a clear correlation between methyl motions and solvent accessible surface area (Supporting Information Figure S3A). There are examples of methyl groups that are almost completely solvent exposed but have high O2axis values, such as Ala-406β with a O2axis value of 0.77, whereas some methyl groups have significant motion but are completely buried, as in the case of Leu-432δ1 with O2axis=0.23.

Overall, side chain methyl group dynamics displays more

heterogeneity compared to backbone dynamics.

8


Page 8 of 30

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


Figure 3. Motional character of methyl-containing side chains in TAZ1. A) Side chain methyl axis order parameters for free (red) and bound (black) TAZ1. B) Difference between the side chain methyl axis order parameter of bound and free TAZ1 vs residue number.

When bound to TAD-STAT2, O2axis for 41 methyl groups in TAZ1 could be determined, with values ranging from 0.06 (Thr-341γ) to 1.0 (Ala-409β) (Figure 3A and Supporting Information Table S4). For bound TAD-STAT2, O2axis parameters could be obtained for 24 out of 31 methyl probes, with values between 0.11 (Val-786γ1) to 0.86 (Leu-795δ1) (Figure 4 and Supporting Information Table S5). The O2axis values for bound TAD-STAT2 follows the same trend as in the

9



backbone order parameters, i.e. the N-terminal region (residues 786-815) have higher O2axis values than the C-terminal part (residues 816-838).

Figure 4. Side chain methyl axis order parameters for bound TAD-STAT2, plotted against the methyl name. Leu and Val methyls are not stereospecifically assigned.16

Only three methyl groups have O2axis values that are equal or higher than 0.7 (0.86 for Leu795δ1, 0.70 for Leu-795δ2 and 0.74 for Thr-800γ2). There are no correlations between O2axis and solvent accessible surface area for neither partner in the bound state (Supporting Information Figure S3A). Although the largest changes in O2axis are observed for methyl groups that are in direct contact with the binding partner, no clear correlation between O2axis and the distance to the binding site exists either (Supporting Information Figure S3B).

It has previously been observed in some systems that the distribution of methyl order parameters is multimodal, and often (but not always) three classes (J, α, and ω class) of motions are observed.19, 20 The J-class involves low order parameters and reflects motions between rotameric wells whereas the α-class, centered at ~0.6, involves to a large degree high amplitude motions within a rotameric well and also to some extent motions leading to rotameric interconversion, and finally the ω-class which contains high order parameters, which reflects restricted motions within

10


Page 10 of 30

Page 11 of 30 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


a rotameric well. The ω-class is more populated in the bound state of TAZ1 compared to free TAZ1 (Supporting Information Figure S4), which reflects a rigidification of TAZ1 upon binding to TAD-STAT2. Bound TAD-STAT2 is almost completely devoid of the ω-class (Supporting Information Figure S4), which signifies the highly dynamic nature of TAD-STAT2, especially the C-terminal region (Supporting Information Figure S5).

Dynamic character of the binding interface and energetic hot-spots The binding interface is a mixture of electrostatic and hydrophobic interactions, and contains several methyl-bearing side chains from both TAZ1 and TAD-STAT2.16 There is a hydrophobic groove in TAZ1, which contains Leu-352, Ile-353, Ile-414, and Ile-428 that favorably interacts with Leu-791, Leu-795, and Leu-798, and Met-803 at the N-terminal part of TAD-STAT2. Methyl groups at the binding interface are heterogeneously dynamic, with O2axis values ranging from 0.2 to 1.0. Comparison of O2axis values for bound TAD-STAT2 to a library of O2axis residuespecific values for 18 well-folded proteins21 shows that, although there are three methyl groups (L795-δ1, L795-δ2, and M803-ε) in bound TAD-STAT2 that are more rigid, and one methyl group (Thr-800γ2) in which no difference is observed to take place, it appears that the bound IDP is generally more dynamic on the ps-ns time scale than folded proteins, which is particularly the case for the 805-830 region of TAD-STAT2, whereas the earlier part of TAD-STAT2 contains a mixture of methyl groups that are either more mobile, with no difference in dynamics, or are more rigid compared to well-folded proteins (Supporting Information Figure S5). Binding hotspots are usually classified as residues that upon mutation to Ala reduce the binding affinity significantly (≥ 2 kcal/mol),22 and the hot-spot nature has often been interpreted in terms of enthalpic interactions. In a previous kinetic and mutational analysis of TAZ1/TAD-STAT2,18 it

11



was observed that the point-mutations L791A and L798A weakened the interaction to such an extent that these residues could be classified as hot-spots. The O2axis values for Leu-798 and Leu791 are between 0.36-0.43 (Figure 4 and Supporting Information Table S5). It is tempting to speculate that the conformational entropy at these specific positions could be favorable for the binding affinity (Supporting Information Figure S5) and thus an important reason for the hot-spot nature, but since the dynamics of the L798A, and the L791A mutations have not been investigated, it is at present difficult to assess if it is a different conformational entropy behavior at these position that is the reason for the reduced affinity for the mutations, or if the enthalpic interactions have been severely disturbed relative to the wild-type, or if a mixture of the two scenarios is at play. Nevertheless, the present study show that binding does not necessarily confine the motion of hot-spots, that these hot-spots are dynamic, which could potentially be an important element for their hot-spot properties.

A connection between spatial clustering of rigid and dynamic regions and energetic frustration Remarkably, on one side of the complex, which also contains the binding site for the highly dynamic C-terminal part of TAD-STAT2 (residues 816-838), there is a clustering of mobile methyl groups and backbone amide groups, contributed by both TAZ1 and TAD-STAT2, whereas on the opposite side of the complex, there is instead a cluster of sites from both binding partners that are more rigid, a region that also includes the interaction site for the N-terminal region of TAD-STAT2 (Figure 5B). Helix 2 in TAZ1, which is positioned between these regions, contains methyl-bearing side chains of which most of them have high amplitude motions (average O2axis in helix 2 is 0.50). Those that have O2axis < 0.65 have their side chains directed towards the more dynamic region of the complex whereas the side chains of Val-390, and Thr-

12


Page 12 of 30

Page 13 of 30 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


386 which contain the three most rigid methyl groups in helix 2 are pointing towards the more rigid region of the complex. Strikingly, the presence of energetic frustration in the complex as determined by the frustratometer23,

24

coincides with these different regions, i.e. the more

dynamic part of the complex displays more frustration than the region with less ps-ns motions (Figure 5A and Supporting Information Figure S6). Indeed, the 3D structure of the complex shows that the highly dynamic C-terminal region of TAD-STAT2, which is part of this frustrated region, is not as well defined as the N-terminal region, and contains side chains that are less protruding.

Figure 5. Energetic frustration and clustering of dynamic and rigid sites in the TAZ1/TADSTAT2 complex. A) Local frustration pattern mapped onto the 3D structure of the protein complex, with energetic frustration determined by the frustratometer23 using pdb code 2KA4. The backbones of TAZ1 and TAD-STAT2 are shown in gray. Minimally frustrated interactions are depicted with green lines, while red lines reflect highly frustrated interactions. B) Shown here as

13



spheres are all backbone amides and side chain methyl groups for which order parameters could be determined for both bound TAZ1 and TAD-STAT2. The spheres are color-coded according to the order parameter value from 0 (red) to 1 (yellow). The backbone of TAZ1 is shown in gray, while TAD-STAT2 is shown in blue. Helices in TAZ1 are labeled in the figure as well as the Nterminal and C-terminal regions of TAD-STAT2.

Protein folding is usually seen through the lens of a funneled energy landscape, in which a natively folded protein is at the bottom of this landscape with a well-defined global minimum, and thus are minimally frustrated,24 because inter-residue interactions are mostly not in conflict with each other, leading to a low-energy state. The concept of minimal frustration can also be applied to protein-protein interactions, and it has been previously argued that the binding sites in unbound proteins may experience some frustration, which is then minimized upon binding to their target.24 Wolynes and co-workers have developed the so-called frustratometer,23, 24 which determines the local frustration by evaluating how favorable a specific contact is relative to the set of all possible contacts in that location. The set of contacts are generated by substitution to other residues at these positions, and the energy for each replacement is calculated. The stabilization energy for the native contact is then compared to the statistics of energies for the different substitutions in that location as described in more detail elsewhere,23, 24 including the nature of the energy function that is employed.23, 25, 26 A so-called frustration index for the contact is calculated, and the frustratometer classifies the contact into one of three classes, depending on the value of the frustration index; (1) minimally frustrated, which means that replacement with other residues is not as favorable compared to the native residue; (2) highly frustrated, that is, substitutions to other residues is more tolerated compared to the native residue; (3) neutral, meaning that the interactions are not particularly favorable nor unfavorable.

14


Page 14 of 30

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


The role of this clustered dynamical behavior in TAZ1/TAD-STAT2 and its potential connection with frustration could be a result of the promiscuous binding properties of TAZ1, which has the ability of recognizing many different IDPs.1 Thus, such highly dynamic, frustrated regions in IDP complexes are likely to have an important functional role. There are limiting amounts of CBP in the cell, and many different targets compete for binding to CBP. Wright and colleagues27 showed in a recent study that the disordered CITED2 competes for binding to TAZ1 with the disordered Hif1α CAD in a highly effective manner. It was demonstrated that CITED2 out-competes Hif1α CAD by forming a transient ternary complex with TAZ1 and Hif1α CAD, which in turn promotes a faster release of Hif1α CAD, and it was proposed that formation of the ternary complex is facilitated by the initial binding of CITED2 to the TAZ1/Hif1α CAD complex in a region where Hif1α CAD interacts with its N-terminal region which contains a short dynamic helix.27 Indeed, here we used the frustratometer on the TAZ1/Hif1α CAD complex (Supporting Information Figure S6, panels C-E), which shows that this helix and in particular the following helix contains local frustration (Supporting Information Figure S6D and E). Thus, dynamic and frustrated regions in the complex, as seen for the TAZ1/TAD-STAT2 complex, could provide a contact point at which the competing target initially binds to, facilitating the release of the bound IDP. The promiscuous nature of TAZ1 recognition is also supported by the fact that different secondary structures are formed in the same binding sites in TAZ1 for the different IDP targets, and that a consensus binding sequence is lacking. Since TAZ1 is not optimized to bind only a single partner, then it is almost by necessity that TAZ1/IDP complexes experience some frustration.28 Further dynamic studies for both binding partners on other IDP complexes will reveal if this is a common feature for such associations.

15



It has been found that complexes involving IDPs have, on average, higher koff values than complexes formed between two well-folded proteins,29 which is an advantage when a fast release is desired, for instance, in signaling.1 We measured the dissociation rate constant, koff, for TAZ1/TAD-STAT2 (Supporting Information Figure S7) under the same conditions as those for NMR relaxation measurements, and determined it to be 2.33 ± 0.01 s-1, which is about 1000-fold higher than the average koff for complexes formed between well-defined proteins.29 Thus, the retained dynamics in the bound state could be a major reason for the higher koff values compared with those of folded proteins. An indication that this could be the case also comes from the interaction between the intrinsically disordered transactivation domain of p53 (TAD-p53) and the molten globular NCBD of CBP, where the measured backbone steady-state NOEs suggested that the main-chain in some areas of bound TAD-p53 retained considerable ps-ns dynamics,30 while the koff was determined to be as high as ~500 s-1 by stopped-flow fluorescence.31 Another example which may point to this trend is the interaction between the intrinsically disordered Cterminal domain of the measles virus nucleoprotein (NTAIL) and the globular X domain (XD) of the viral phosphoprotein,32, 33 in which a partial disorder-to-order transition takes place for NTAIL, and with a koff for NTAIL/XD that was reported to be about 230 s-1.33 Other IDP complexes have been reported as well that undergo a partial disorder-to-order transition.1, 2, 34 However, it is at present moment difficult to draw general conclusions about a potential correlation between residual dynamics and dissociation rate constants due to a lack of detailed site-resolved backbone and side-chain dynamics studies and kinetics measurements on other protein complexes involving IDPs, and must therefore await further experimental and theoretical studies.

Significant changes in internal motion upon binding for both TAZ1 and TAD-STAT2

16


Page 16 of 30

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


The majority of the methyl groups in TAZ1 that experience the largest changes in motion upon complex formation make significant contributions to the hydrophobic core of the binding interface to the N-terminal region of TAD-STAT2. These methyls are Leu-352δ1, Leu-352δ2, Ile-353δ1, Ile-353γ2, Leu-359δ1, Ile-414δ1, and Val-428γ1, all of which become more rigid upon binding (Figures 3B and 6). However, changes in internal motions were also observed for some methyl-bearing side chains that are not part of the binding interface, including sites with very little chemical shift changes (Leu-360δ1, Ala-378β Ala-399β, Val-405γ2, and Ala-409β) (Supporting Information Figure S8), which indicates a dynamic and not structural mediated response upon binding at these sites. The average change in O2axis for pairwise comparisons in TAZ1 is equal to 0.155 ± 0.046, i.e. binding results in an overall rigidification of TAZ1 upon binding. NMR relaxation provides a way of determining conformational entropy changes upon binding,35 with recent studies showing how one can determine the total conformational entropy change using derived relationships that take into account the dynamic response of the amino acid side chains.10, 11, 36 By using the so-called “entropy meter”, and also a relationship that has been derived for the determination of backbone conformational entropy changes, developed by Wand and co-workers10, 11, 37, 38, we calculated the change in conformational entropy based on changes in order parameters. Thus, the overall rigidification of TAZ1 corresponds to an unfavorable conformational entropy change of 11.9 kcal/mol. The use of the model free approach is problematic for free TAD-STAT2, due to the underlying assumption in the Lipari-Szabo modelfree analysis that global and internal motions are statistically independent, with a single global correlation time describing the IDP, which is suspect for IDPs.39 There are certain residues in both TAZ1 and TAD-STAT2 that have O2 values between 0.05-0.1, and are therefore extremely dynamic, with very low steady-state NOEs and some of them have similar values as those for free

17



Page 18 of 30

TAD-STAT2, which have {1H}-15N NOE values that are uniformly negative (Supporting Information Figure S1A). Assuming that the order parameters for unbound TAD-STAT2 are uniform and are equal to 0.1, the average difference of the order parameters ( and ) then corresponds to an unfavorable conformational entropy change11,

38

of 35.3

kcal/mol for TAD-STAT2, whereas an assumed value of 0.05 for free TAD-STAT2 returns a value of 38.8 kcal/mol. Therefore, the total change in conformational entropy is 47.2-50.8 kcal/mol for the binding reaction between TAZ1 and TAD-STAT2. There are empirical relationships that have been derived to get an estimation of the change in the heat capacity (∆Cp), and subsequently the change in solvent entropy, ∆Ssolv, based on changes in accessible surface areas (ASA).40-42 By calculating changes in polar and non-polar ASA for TAD-STAT2 (assuming a random coil for free TAD-STAT2, generated using pdb utility servers, spin.niddk.nih.gov), and for TAZ1 with binding in order to determine ∆Cp by using the parameterization as reported previously,40, 42 the change in solvent entropy,41 -T∆Ssolv, is then estimated to be between -61 kcal/mol40 and -68 kcal/mol.42 Different values have been proposed for the change in translational and rotational entropy (-T∆SRT), from ~3 to about 18 kcal/mol.43-45 Assuming a value of 10 kcal/mol for -T∆SRT and that the total entropy change, -T∆Stot, contains three terms (T∆Ssolv, -T∆Sconf, and -T∆SRT), the conformational entropy change, -T∆Sconf, is then estimated to be about 56-63 kcal/mol since -T∆Stot (4.9 kcal/mol) is known from ITC. Although, the structure of free TAD-STAT2 is unknown and there are uncertainties associated with such calculations and should be interpreted with caution, it agrees reasonably well with the NMR-determined value of T∆Sconf. The conformational entropy change for both TAZ1 and TAD-STAT2 contributes significantly to the binding energetics. The largest changes in TAZ1 are observed for residues that are in direct contact with TAD-STAT2 and form the hydrophobic groove (Figures 3B and 6),

18


Page 19 of 30 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


which is an important binding site for several other IDPs as well, such as Hif1α CAD, CITED2, and TAD-RelA. The unfavorable total change in conformational entropy is comparable and exceeds in magnitude to the change in solvent entropy, and the binding enthalpy, respectively. The largest contribution to the change in conformational entropy comes from the rigidification of TAD-STAT2. However, the analysis shows that the dynamic response of TAZ1 is also instrumental in maintaining a functionally desired binding affinity since an absence of the contribution of conformational entropy from TAZ1 to the binding energetics would result in an affinity of about 1 femtomolar, which is one million times stronger than any other reported binding affinities that involves the CBP. Overall, this clearly shows the potential of conformational entropy to modulate protein-target interactions, with important contributions from both the globular and the disordered protein, respectively.

Figure 6. Methyl axis order parameter changes for TAZ1 upon binding to TAD-STAT2. A) Methyl probes with relaxation data available in both states of TAZ1 are shown as spheres.

19



Changes in the O2axis parameter (bound minus free) are mapped according to a color gradient ranging from -0.1 (red) to 0.4 (yellow). B) Potential dynamic long-range transmission pathway in TAZ1. Methyl groups in TAZ1 that experience the largest changes in the O2axis parameter upon binding (>0.2) are highlighted as green spheres. The backbone of TAD-STAT2 is colored blue, whereas TAZ1 is in gray.

The methyl-bearing residues that experience large dynamic response upon binding (Figure 6A) are among the most conserved residues in TAZ1. Side chain methyl groups with significant differences in O2axis are spatially close to each other, with the potential of forming a contiguous long-range signal transmission pathway (Figure 6B). The largest changes in ps-ns dynamics are not only observed for the methyl-bearing residues that are part of the binding interface but also for residues that are distal to the binding site and in which chemical shift perturbations are absent or are very small (Supporting Information Figure S8). This shows that the interaction also involves long-range effects that are dynamical and not structural in nature and suggests a potential role for conformational entropy changes being utilized by nature to transmit energy through the protein.

Conclusions We have here demonstrated the importance of the conformational entropy change in modulating the binding thermodynamics of protein-protein interactions. However, TAD-STAT2 undergoes only a partial disorder-to-order. The methyl-bearing residues that experience the largest changes in methyl order axis parameters form a potential contiguous long-range transmission pathway, which also includes residues that are distal to the binding site and in which structural changes

20


Page 20 of 30

Page 21 of 30 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


upon binding are absent. The highly dynamic region of the complex is more energetically frustrated than the less mobile region of the complex. This is likely to be functionally important in an environment where different targets compete for binding to TAZ1. Also, the residual ps-ns dynamics that is present in the complex could be a major reason why complexes involving IDPs tend to have higher dissociation rate constants, which is believed to be one of the key advantages of being disordered. However, further studies on other IDP complexes needs to be carried out in order to see if there is a general correlation between residual dynamics and dissociation kinetics. Overall, this study shows that information on conformational entropy is crucial in order to get an in-depth understanding of molecular recognition processes involving IDPs.

Methods Protein expression and purification: Human TAZ1 (residues 340-439) and TAD-STAT2 (residues 786-838, C793H/C809S)16 were expressed and purified as described previously.18 15Nlabelled proteins were expressed in M9 minimal media supplemented with relaxation experiments were performed on

15

15

NH4Cl. Deuterium

N/13C/2H-labelled proteins that were expressed in

M9 minimal media46 containing 60 % D2O and supplemented with 15NH4Cl and 13C-glucose (U13

C6-99 %). TAZ1 was also obtained in a separate expression in which the minimal medium

contained 10%

13

C-glucose/90% unlabeled glucose, in order to stereospecifically assign the

valine and leucine methyls.47

Samples: All NMR samples were prepared in a final buffer containing 15 mM HEPES, 50 mM NaCl, 3 mM TCEP, pH=6.8, 0.01 % NaN3, and 7 % D2O. NMR samples of 15N- and 13C/15N/2Hlabelled free TAZ1 had both a protein concentration of 1.1 mM. For the protein complex, NMR

21



Page 22 of 30

samples containing labeled TAZ1 (15N, or 13C/15N/2H) / unlabeled TAD-STAT2 (1 mM/1.4 mM) or TAD-STAT2 (15N, or 13C/15N/2H) / unlabeled TAZ1 (1 mM/1.4 mM) were prepared. 10 %-13C labeled TAZ1 either free (1 mM) or bound to unlabeled TAD-STAT2 (1 mM/1.4 mM) was also prepared. Constant time 13C-HSQCs were recorded on 10 %-13C labeled free and bound TAZ1 at 700 MHz for stereospecific assignments of methyls in valines and leucines.

Relaxation experiments: Experiments were performed at 303 K and on Bruker NMR spectrometers operating at a Larmor (1H) frequency of 500 MHz, 600 MHz, and 700 MHz.

15

N

relaxation T1, T2, and steady state {1H}-15N NOEs for bound TAZ1 and bound TAD-STAT2 were measured at 500 MHz and 700 MHz. For free TAZ1,

15

N relaxation T1, T2, and {1H}-15N

NOEs were measured at 600 MHz whereas at 700 MHz

15

N T2, and {1H}-15N NOEs were

measured. 9-12 time points, of which three were duplicates to estimate peak intensity uncertainties, were collected. Time points for 15N T1 ranged from 0.02 s to 1.4 s at 700 MHz and 0.02 s to 1.0 s at 500 MHz for bound TAZ1 and TAD-STAT2. Times points for

15

N T2 was

between 16.5 ms and 115.5 ms at 700 MHz, and 17.0-135.7 ms at 500 MHz for bound TAZ and bound TAD-STAT2. For free TAZ1, time points at 600 MHz varied from 0.02 s to 0.94 s for 15N T1 and from 17.0 ms to 152.6 ms for 15N T2, whereas at 700 MHz, they ranged from 0.02 to 1.1 s (15N T1), and 16.5-148.9 ms (15N T2).

Deuterium relaxation experiments were carried out at both 700 MHz and 500 MHz for bound TAZ1, bound TAD-STAT2, and free TAZ1 in order to investigate side chain methyl dynamics. Relaxation rates of the spin coherences, IzCzDz, IzCz, and IzCzDy were determined.48 12 time points, of which three were duplicates, were collected. Time points ranged from 1.6 to 59.6 ms (IzCzDz), 2.2-61.7 ms (IzCz), and 0.4-15.5 ms (IzCzDy) for the bound proteins, whereas for free

22


Page 23 of 30 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


TAZ1 they varied from 1.6 to 59.2 ms for IzCzDz, 2.2 to 59.8 ms for IzCz, and 0.4-15.5 ms for IzCzDy.

Relaxation analysis: NMR data were processed using nmrPipe.49 Model free analysis50 of

15

N

backbone relaxation (R1, R2, and {1H}-15N NOE) was performed using the software relax.51, 52 The diffusion tensors sphere, oblate, prolate, and ellipsoid were tested using relaxation data obtained at two fields for residues that had a {1H}-15N NOE > 0.62 (at 500 MHz for bound TAZ1) or {1H}-15N NOE > 0.65 (at 600 MHz for free TAZ1). Each diffusion tensor as well as model-free models were optimized with model-free model selection being performed using the Akaike information criterion (AIC) as described previously.51, 52 Five different models of the local motion for each residue were considered53: (1), order parameter (O2); (2), O2 and an internal correlation time (τe); (3), O2 and a conformational exchange term, Rex; (4), O2, τe and Rex; (5), O2 for two time scales (Os2 and Of2) and the internal correlation time for the slower time scale, τe. A chemical shift anisotropy (CSA) value of -172 ppm, and a N-H bond length of 1.02 Å were used for these calculations. The diffusion tensor with the lowest AIC was selected (Supporting Information Tables S1 and S2), and subsequently local model-free models were minimized for all residues for which relaxation data were available using this diffusion tensor, with model-free model selection using AIC. For the optimization of the diffusion tensor for the protein complex we used relaxation data for the bound TAZ1, and these optimized tumbling parameters were then used for model-free analysis for bound TAD-STAT2. Uncertainties of the model-free parameters were determined by 500 Monte Carlo simulations. The best fit for unbound TAZ1 yielded an isotropic diffusion tensor with a rotational correlation time, τm, of 7.76 ns, which is in very good agreement with the correlation time obtained from HYDRONMR54 (7.70 ns), whereas for bound

23



Page 24 of 30

protein an ellipsoid diffusion tensor, with τm=11.1 ns and an anisotropy of 1.27 was selected as the best fit. The determined τm is in good agreement with the value obtained from HYDRONMR54 (10.9 ns). The average error for the O2NH values is 2.4 % for the complex, and 4.5 % for free TAZ1. Comparison of the order parameters from the ellipsoidal diffusion tensor with the best fit for the isotropic diffusion tensor for bound TAZ1 shows that rotational diffusion anisotropy has little effect on the order parameters since the average difference is 0.004 ± 0.029. For free TAZ1, the O2NH parameters obtained with the prolate diffusion tensor were virtually the same as those obtained with the spherical tensor, with an average difference of −0.001 ± 0.022. Thus, the O2 parameters are in this case rather insensitive to rotational diffusion anisotropy effects.

The methyl Dz and Dy rates were obtained by subtracting the IzCz rate from the IzCzDz and IzCzDy rates, respectively. The Dz and Dy rates were then used to determine model-free squared generalized order parameters and effective internal correlation times,19, correlation time obtained from

15

48

using the global

N backbone relaxation. A 2H quadrupolar constant of 167 kHz

was used in the calculations.55 The squared generalized methyl axis order parameters, O2axis, were determined by taking the motion about the methyl symmetry axis into consideration. Errors were determined by 500 Monte Carlo simulations.

Stopped-flow fluorimetry: Displacement experiments were performed using a SX-18MV stopped-flow spectrometer (Applied Photophysics), in order to determine the dissociation rate constant. Measurements were taken in 15 mM HEPES (pH=6.8), 50 mM NaCl, and 1 mM TCEP at 303 K. A preformed complex solution of TAZ1/TAD-STAT2 (1.2 µM/0.9 µM) was mixed

24


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


with varying concentrations of an excess of TAZ1W418Y, in which TAZ1 that is bound to TADSTAT2, is displaced by TAZ1W418Y. The displacement traces were fitted to a single-exponential function, and at high TAZ1W418Y concentrations, kobs is equal to koff.

Isothermal titration calorimetry: Measurements were taken using an iTC200-calorimeter (Malvern Instruments), at a temperature of 303 K. TAZ1 and TAD-STAT2 were dialyzed against 15 mM HEPES or MES, 50 mM NaCl, and 1 mM TCEP pH=6.8, before measurements were performed. TAD-STAT2 was loaded onto the syringe with a concentration of 170 µM, and a solution containing 13 µM TAZ1 was added to the sample cell. A preliminary 0.5 to 1.3 µL injection was followed by 19 subsequent 1.8 µL injections. Software provided by the manufacturer was used to fit the data to a one-to-one binding model.

Supporting information Steady-state {1H}-15N NOE values for free (Figure S1A) and bound (Figure S1B) TAD-STAT2; O2NH vs residue number for free and bound TAZ1 (Figure S2); dependence of the O2axis parameter on solvent exposed methyl area (Figure S3A), and distance to the binding site (Figure S3B); distribution of O2axis values for free and bound TAZ1, and bound TAD-STAT2 (Figure S4); comparison of O2axis values for bound TAD-STAT2 against the average O2axis residuespecific values from a library of 18 well-folded proteins (Figure S5); output figures from the frustratometer for TAZ1/TAD-STAT2 and TAZ1/Hif1α CAD (Figure S6); an example of a stopped-flow fluorescence kinetic trace in a displacement experiment, for the determination of koff; comparison of changes in O2axis with changes in backbone chemical shifts for TAZ1 upon binding (Figure S8); Summary of diffusion tensor optimization for the complex (Table S1) and

25



free TAZ1 (Table S2); O2axis parameters and effective internal correlation times for free TAZ1 (Table S3), bound TAZ1 (Table S4), and bound TAD-STAT2 (Table S5).

Acknowledgments This work was supported by the Swedish Research Council, the Magnus Bergvall Foundation, the Carl Trygger Foundations, the Ollie & Elof Ericsson Foundation, the Lars Hierta Memorial Foundation, and the Åke Wiberg Foundation.

References (1) Wright, P. E., and Dyson, H. J. (2015) Intrinsically disordered proteins in cellular signalling and regulation, Nat. Rev. Mol. Cell. Biol. 16, 18-29. (2) van der Lee, R., Buljan, M., Lang, B., Weatheritt, R. J., Daughdrill, G. W., Dunker, A. K., Fuxreiter, M., Gough, J., Gsponer, J., Jones, D. T., Kim, P. M., Kriwacki, R. W., Oldfield, C. J., Pappu, R. V., Tompa, P., Uversky, V. N., Wright, P. E., and Babu, M. M. (2014) Classification of intrinsically disordered regions and proteins, Chem. Rev. 114, 6589-6631. (3) Uversky, V. N., Oldfield, C. J., and Dunker, A. K. (2008) Intrinsically disordered proteins in human diseases: introducing the D2 concept, Annu. Rev. Biophys. 37, 215-246. (4) Babu, M. M., van der Lee, R., de Groot, N. S., and Gsponer, J. (2011) Intrinsically disordered proteins: regulation and disease, Curr. Opin. Struct. Biol. 21, 432-440. (5) Wand, A. J. (2013) The dark energy of proteins comes to light: conformational entropy and its role in protein function revealed by NMR relaxation, Curr Opin Struct Biol 23, 75-81. (6) Popovych, N., Sun, S., Ebright, R. H., and Kalodimos, C. G. (2006) Dynamically driven protein allostery, Nat. Struct. Mol. Biol. 13, 831-838. (7) Capdevila, D. A., Braymer, J. J., Edmonds, K. A., Wu, H., and Giedroc, D. P. (2017) Entropy redistribution controls allostery in a metalloregulatory protein, Proc. Natl. Acad. Sci. U. S. A. 114, 4424-4429. (8) Tzeng, S. R., and Kalodimos, C. G. (2009) Dynamic activation of an allosteric regulatory protein, Nature 462, 368-372. (9) Frederick, K. K., Marlow, M. S., Valentine, K. G., and Wand, A. J. (2007) Conformational entropy in molecular recognition by proteins, Nature 448, 325-329. (10) Marlow, M. S., Dogan, J., Frederick, K. K., Valentine, K. G., and Wand, A. J. (2010) The role of conformational entropy in molecular recognition by calmodulin, Nat. Chem. Biol. 6, 352-358. (11) Caro, J. A., Harpole, K. W., Kasinath, V., Lim, J., Granja, J., Valentine, K. G., Sharp, K. A., and Wand, A. J. (2017) Entropy in molecular recognition by proteins, Proc. Natl. Acad. Sci. U. S. A. 114, 6563-6568.

26


Page 26 of 30

Page 27 of 30 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


(12) Diehl, C., Engstrom, O., Delaine, T., Hakansson, M., Genheden, S., Modig, K., Leffler, H., Ryde, U., Nilsson, U. J., and Akke, M. (2010) Protein flexibility and conformational entropy in ligand design targeting the carbohydrate recognition domain of galectin-3, J. Am. Chem. Soc. 132, 14577-14589. (13) Tzeng, S. R., and Kalodimos, C. G. (2012) Protein activity regulation by conformational entropy, Nature 488, 236-240. (14) Eisenmesser, E. Z., Bosco, D. A., Akke, M., and Kern, D. (2002) Enzyme dynamics during catalysis, Science 295, 1520-1523. (15) Lee, A. L., Kinnear, S. A., and Wand, A. J. (2000) Redistribution and loss of side chain entropy upon formation of a calmodulin-peptide complex, Nat. Struct. Biol. 7, 72-77. (16) Wojciak, J. M., Martinez-Yamout, M. A., Dyson, H. J., and Wright, P. E. (2009) Structural basis for recruitment of CBP/p300 coactivators by STAT1 and STAT2 transactivation domains, EMBO J. 28, 948-958. (17) De Guzman, R. N., Wojciak, J. M., Martinez-Yamout, M. A., Dyson, H. J., and Wright, P. E. (2005) CBP/p300 TAZ1 domain forms a structured scaffold for ligand binding, Biochemistry 44, 490-497. (18) Lindström, I., and Dogan, J. (2017) Native Hydrophobic Binding Interactions at the Transition State for Association between the TAZ1 Domain of CBP and the Disordered TADSTAT2 Are Not a Requirement, Biochemistry 56, 4145-4153. (19) Igumenova, T. I., Frederick, K. K., and Wand, A. J. (2006) Characterization of the fast dynamics of protein amino acid side chains using NMR relaxation in solution, Chem. Rev. 106, 1672-1699. (20) Sharp, K. A., Kasinath, V., and Wand, A. J. (2014) Banding of NMR-derived methyl order parameters: Implications for protein dynamics, Proteins 82, 2106-2117. (21) Carbonell, P., and del Sol, A. (2009) Methyl side-chain dynamics prediction based on protein structure, Bioinformatics 25, 2552-2558. (22) Clackson, T., and Wells, J. A. (1995) A Hot-Spot of Binding-Energy in a HormoneReceptor Interface, Science 267, 383-386. (23) Parra, R. G., Schafer, N. P., Radusky, L. G., Tsai, M. Y., Guzovsky, A. B., Wolynes, P. G., and Ferreiro, D. U. (2016) Protein Frustratometer 2: a tool to localize energetic frustration in protein molecules, now with electrostatics, Nucl. Acids Res. 44, W356-W360. (24) Ferreiro, D. U., Hegler, J. A., Komives, E. A., and Wolynes, P. G. (2007) Localizing frustration in native proteins and protein assemblies, Proc. Natl. Acad. Sci. U. S. A. 104, 19819-19824. (25) Papoian, G. A., Ulander, J., and Wolynes, P. G. (2003) Role of water mediated interactions in protein-protein recognition landscapes, J. Am. Chem. Soc. 125, 9170-9178. (26) Papoian, G. A., Ulander, J., Eastwood, M. P., Luthey-Schulten, Z., and Wolynes, P. G. (2004) Water in protein structure prediction, Proc. Natl. Acad. Sci. U. S. A. 101, 3352-3357. (27) Berlow, R. B., Dyson, H. J., and Wright, P. E. (2017) Hypersensitive termination of the hypoxic response by a disordered protein switch, Nature 543, 447-+. (28) Jemth, P., Mu, X., Engstrom, A., and Dogan, J. (2014) A Frustrated Binding Interface for Intrinsically Disordered Proteins, J. Biol. Chem. 289, 5528-5533. (29) Shammas, S. L., Rogers, J. M., Hill, S. A., and Clarke, J. (2012) Slow, reversible, coupled folding and binding of the spectrin tetramerization domain, Biophys. J. 103, 2203-2214.

27



(30) Lee, C. W., Martinez-Yamout, M. A., Dyson, H. J., and Wright, P. E. (2010) Structure of the p53 transactivation domain in complex with the nuclear receptor coactivator binding domain of CREB binding protein, Biochemistry 49, 9964-9971. (31) Dogan, J., Jonasson, J., Andersson, E., and Jemth, P. (2015) Binding Rate Constants Reveal Distinct Features of Disordered Protein Domains, Biochemistry 54, 4741-4750. (32) Dosnon, M., Bonetti, D., Morrone, A., Erales, J., di Silvio, E., Longhi, S., and Gianni, S. (2015) Demonstration of a folding after binding mechanism in the recognition between the measles virus NTAIL and X domains, ACS Chem. Biol. 10, 795-802. (33) Gruet, A., Dosnon, M., Blocquel, D., Brunel, J., Gerlier, D., Das, R. K., Bonetti, D., Gianni, S., Fuxreiter, M., Longhi, S., and Bignon, C. (2016) Fuzzy regions in an intrinsically disordered protein impair protein-protein interactions, FEBS J. 283, 576-594. (34) Tompa, P., and Fuxreiter, M. (2008) Fuzzy complexes: polymorphism and structural disorder in protein-protein interactions, Trends Biochem. Sci. 33, 2-8. (35) Akke, M., Bruschweiler, R., and Palmer, A. G. (1993) NMR order parameters and freeenergy – an analytical approach and its application to cooperative Ca2+ binding by calbindin-D(9 k), J. Am. Chem. Soc. 115, 9832–9833. (36) Li, D. W., and Bruschweiler, R. (2009) A dictionary for protein side-chain entropies from NMR order parameters, J. Am. Chem. Soc. 131, 7226-7227. (37) Kasinath, V., Sharp, K. A., and Wand, A. J. (2013) Microscopic Insights into the NMR Relaxation-Based Protein Conformational Entropy Meter, J Am Chem Soc 135, 15092-15100. (38) Sharp, K. A., O'Brien, E., Kasinath, V., and Wand, A. J. (2015) On the relationship between NMR-derived amide order parameters and protein backbone entropy changes, Proteins 83, 922-930. (39) Peng, J. W. (2012) Exposing the Moving Parts of Proteins with NMR Spectroscopy, J. Phys. Chem. Lett. 3, 1039-1051. (40) Myers, J. K., Pace, C. N., and Scholtz, J. M. (1995) Denaturant m values and heat capacity changes: relation to changes in accessible surface areas of protein unfolding, Protein Sci. 4, 2138-2148. (41) DAquino, J. A., Gomez, J., Hilser, V. J., Lee, K. H., Amzel, L. M., and Freire, E. (1996) The magnitude of the backbone conformational entropy change in protein folding, Proteins 25, 143-156. (42) Spolar, R. S., Livingstone, J. R., and Record, M. T., Jr. (1992) Use of liquid hydrocarbon and amide transfer data to estimate contributions to thermodynamic functions of protein folding from the removal of nonpolar and polar surface from water, Biochemistry 31, 39473955. (43) Karplus, M., and Janin, J. (1999) Comment on: 'The entropy cost of protein association', Protein Eng. 12, 185-186; discussion 187. (44) Gomez, J., and Freire, E. (1995) Thermodynamic mapping of the inhibitor site of the aspartic protease endothiapepsin, J. Mol. Biol. 252, 337-350. (45) Luo, H., and Sharp, K. (2002) On the calculation of absolute macromolecular binding free energies, Proc. Natl. Acad. Sci. U. S. A. 99, 10399-10404. (46) Marley, J., Lu, M., and Bracken, C. (2001) A method for efficient isotopic labeling of recombinant proteins, J. Biomol NMR 20, 71-75. (47) Neri, D., Szyperski, T., Otting, G., Senn, H., and Wuthrich, K. (1989) Stereospecific nuclear magnetic resonance assignments of the methyl groups of valine and leucine in the

28


Page 28 of 30

Page 29 of 30 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


DNA-binding domain of the 434 repressor by biosynthetically directed fractional 13C labeling, Biochemistry 28, 7510-7516. (48) Muhandiram, D. R., Yamazaki, T., Sykes, B. D., and Kay, L. E. (1995) Measurement of H-2 T-1 and T-1p Relaxation-Times in Uniformly C-13-Labeled and Fractionally H-2-Labeled Proteins in Solution, J. Am. Chem. Soc. 117, 11536-11544. (49) Delaglio, F., Grzesiek, S., Vuister, G. W., Zhu, G., Pfeifer, J., and Bax, A. (1995) Nmrpipe - a Multidimensional Spectral Processing System Based on Unix Pipes, J Biomol. NMR 6, 277293. (50) Lipari, G., and Szabo, A. (1982) Model-free approach to the interpretation of nuclear magnetic resonance relaxation in macromolecules. 1. Theory and range of validity, J. Am. Chem. Soc. 104, 4546-4559. (51) d'Auvergne, E. J., and Gooley, P. R. (2008) Optimisation of NMR dynamic models II. A new methodology for the dual optimisation of the model-free parameters and the Brownian rotational diffusion tensor, J Biomol. NMR 40, 121-133. (52) d'Auvergne, E. J., and Gooley, P. R. (2008) Optimisation of NMR dynamic models I. Minimisation algorithms and their performance within the model-free and Brownian rotational diffusion spaces, J Biomol. NMR 40, 107-119. (53) Mandel, A. M., Akke, M., and Palmer, A. G., 3rd. (1995) Backbone dynamics of Escherichia coli ribonuclease HI: correlations with structure and function in an active enzyme, J. Mol. Biol. 246, 144-163. (54) de la Torre, J. G., Huertas, M. L., and Carrasco, B. (2000) HYDRONMR: Prediction of NMR relaxation of globular proteins from atomic-level structures and hydrodynamic calculations, J. Magn. Reson. 147, 138-146. (55) Mittermaier, A., and Kay, L. E. (1999) Measurement of methyl H-2 quadrupolar couplings in oriented proteins. How uniform is the quadrupolar coupling constant?, J. Am. Chem. Soc. 121, 10608-10613.

29



53x30mm (300 x 300 DPI)


Page 30 of 30

Dynamics, conformational entropy, and frustration in protein-protein

Recommend Documents