Mechanism‐Dependent Modulation of Ultrafast Interfacial Water Dynamics in Intrinsically Disordered Protein Complexes

Abstract The recognition of intrinsically disordered proteins (IDPs) is highly dependent on dynamics owing to the lack of structure. Here we studied the interplay between dynamics and molecular recognition in IDPs with a combination of time‐resolving tools on timescales ranging from femtoseconds to nanoseconds. We interrogated conformational dynamics and surface water dynamics and its attenuation upon partner binding using two IDPs, IBB and Nup153FG, both of central relevance to the nucleocytoplasmic transport machinery. These proteins bind the same nuclear transport receptor (Importinβ) with drastically different binding mechanisms, coupled folding–binding and fuzzy complex formation, respectively. Solvent fluctuations in the dynamic interface of the Nup153FG‐Importinβ fuzzy complex were largely unperturbed and slightly accelerated relative to the unbound state. In the IBB‐Importinβ complex, on the other hand, substantial relative slowdown of water dynamics was seen in a more rigid interface. These results show a correlation between interfacial water dynamics and the plasticity of IDP complexes, implicating functional relevance for such differential modulation in cellular processes, including nuclear transport.

Abstract: The recognition of intrinsically disordered proteins (IDPs) is highly dependent on dynamics owing to the lacko f structure.Here we studied the interplay between dynamics and molecular recognition in IDPs with ac ombination of timeresolving tools on timescales ranging from femtoseconds to nanoseconds.W ei nterrogated conformational dynamics and surface water dynamics and its attenuation upon partner binding using two IDPs,IBB and Nup153FG,b oth of central relevance to the nucleocytoplasmic transport machinery.These proteins bind the same nuclear transport receptor (Importinb) with drastically different binding mechanisms,c oupled folding-binding and fuzzy complex formation, respectively.S olvent fluctuations in the dynamic interface of the Nup153FG-Importinb fuzzy complex were largely unperturbed and slightly accelerated relative to the unbound state.I nt he IBB-Importinb complex, on the other hand, substantial relative slowdown of water dynamics was seen in amore rigid interface. These results show ac orrelation between interfacial water dynamics and the plasticity of IDP complexes,i mplicating functional relevance for such differential modulation in cellular processes,including nuclear transport.
Hydration is crucial to the expression of bio-molecular functionality. [1] Consequently,h ydration dynamics are intimately related to the dynamics and function of several biomolecules,a ss een in numerous examples where ac orre-lation of hydration dynamics and functionality has been found;f or example,i nterfacial hydration dynamics has been related to the speed of polymerases and specificity of enzymes. [2] Ther ole of hydration dynamics,a lbeit poorly understood, and dynamics in general is central to understanding intrinsically disordered proteins (IDPs), which unlike folded proteins populate as tructural ensemble in their native state instead of aunique structure. [3] As IDPs lack the luxury of rich structural elements present in folded proteins to guide molecular recognition, the molecular recognition of IDPs is largely governed by dynamical features.T his makes the question of the role of hydration dynamics in IDPs ap ertinent, albeit hugely neglected one. Mechanistically,I DP recognition comes in at least two core characters:c oupled folding-binding and fuzzy complex formation. [4] Thef ormer involves ap rocess where an IDP assumes af olded structure upon binding the partner,a nd in the latter,I DPs retain their disorder even after partner binding.Considering the amplified importance of dynamics in IDP recognition, af airly obvious and crucial, but still unanswered question is whether hydration dynamics play ar ole in governing IDP binding mechanisms.W ith an emphasis on solvation dynamics,here we directly interrogate the differential dynamics in two distinct IDP complexes formed via coupled folding-binding and fuzzy mechanisms using two disordered proteins,I BB and Nup153FG,b oth crucial players in nucleocytoplasmic transport. [5] Our results reveal differential dynamics in these systems across timescales ranging from femtoseconds to nanoseconds,a nd elucidate possible molecular underpinnings and functional relevance of mechanism-dependent differential IDP recognition.
Nup153FG is a600aa long IDP domain that belongs to the class of FGNups,which are F-and G-rich IDPs that constitute the permeability barrier of the nuclear pore complex (NPC) and facilitate transport across the NPC by binding nuclear transport receptors (NTRs). [5] We and others have previously shown that FGNups,v ia FG motif facilitated ultrafast multivalent interactions,engage NTRs,such as Importinb,without substantial conformational change and with diffusion-limited kinetics forming fuzzy complexes. [6] IBB (Importinb-binding domain of Importina)i sp art of the N-terminal disordered region of Importina which also binds to Importinb,b ut with acoupled folding-binding mechanism. While unbound IBB is an IDP,i ta ssumes af olded helical structure in the bound state. [7] Recognition of Importinb by IBB is ac rucial step in the formation of an import cargo complex that can traverse the NPC. [8] At the very outset we see ac lear difference in conformational changes in these two proteins upon partner binding using multiparameter single-molecule fluorescence resonance energy transfer (smFRET) spectroscopy with pulsed interleaved excitation (PIE), which allows us to probe FRET specifically in molecules that bear both donor and acceptor dye labels and thus the distance between two labeling sites in ap rotein chain. [6a,c, 9] IBB (S24C/S55C) labeled site-specifically with the FRET pair Alexa488 (donor dye) and Alexa594 (acceptor dye) showed E FRET % 0.8 which changed to % 0.3 upon Importinb binding (Figure 1A,B and Figure S1). This suggested an increased distance between the labeling sites and reduced dynamics which is commensurate with the disorder-to-helix transition (Figure S2). Besides confirming the conformational change associated with IBB-Importinb interaction, the experiment also largely exonerates dye labeling at these sites in IBB from interfering with binding;t his is crucial as we will use S24C below to characterize the IBB-Importinb interface. Nup153FG shows av ery different behavior from IBB when binding to Importinb.Nup153FG has anonrandom sequence distribution of FG motifs,w ith ac oncentration of FxFG repeats in the N-terminal region. Nup153FG labeled (see the Supporting Information for labeling details) in this region at positions S883 and S938 with the same FRET pair (Alexa 488/ 594) showed within experimental precision unchanged conformational dynamics and E FRET of % 0.7 for both bound (confirmed via fluorescence correlation spectroscopy,F igure S1) and unbound states in line with our previous studies. [6a,c] To characterize the interface of Nup153FG and IBB in complex with Importinb,w en ext employed steady-state and picosecond-resolved fluorescence encoded by site-specifically labeled acrylodan, athiol-selective fluorophore that is highly sensitive to microenvironment polarity. [10] ForNup153FG we engineered multiple single-cysteine mutants spanning the length of the two different sequence stretches in the protein which are rich in FxFG motifs (883C and 990C) and PxFG motifs (1330C,1 355C and 1391C), to check for the effect of sequence propensity (Figure 2A). Forthe much smaller IBB, we used the site S24C,which is at the base of the helix formed upon Importinb binding,d irectly probing the interface pointing towards Importinb ( Figure 2B). We saw large blueshifted acrylodan emission and increased fluorescence lifetime in both the Nup153FG mutants and IBB upon Importinb binding ( Figures S3 and S4). Thea crylodan emission spectrum is highly sensitive to the local polarity with more redshifted emission indicating more polar environment and vice versa. [10] Thus the blue-shifted emission upon Importinb binding indicated al ess polar environment at the interface compared to the surfaces of free IBB or Nup153FG.This also manifested in ad rop in the calculated apparent relative permittivity by % 12 units in the bound states from the free ones. ( Figure 2C and Figure S5). As relative permittivity is ad irect comparative measure of environment polarity,t his likely resulted from desolvation, that is water release,f rom the IDP surface when the interface formed upon partner binding.Interestingly enough, the extent of inferred desolvation, and thus the associated entropic gain, upon partner binding was similar for IBB and Nup153FG,d espite the fact that the binding mechanism and the structure of the IDP in the bound state are very different.
We next employed picosecond-resolved anisotropy experiments,w hich probe the depolarization kinetics of af luorophore on ap icosecond-nanosecond timescale.F or IDPs this provides information about segmental motion, that is,t he conformational motion of as egment of the chain (several amino acids), and information about very local dynamics in the immediate vicinity of the labeling site (very few amino acids), as we have shown before. [11] Here we used  acrylodan, which due to its very small size combined with avery short linker enables highly sensitive anisotropy studies. Both acrylodan-labeled IBB and Nup153FG in the free state showed bi-exponential relaxation behavior ( Figure 2D,E) with an anosecond and sub-nanosecond component and some residual anisotropy.L ikely the nanosecond component related to segmental motion, while the sub-nanosecond component related to local dynamics and the much slower hydrodynamic rotation of the entire molecule manifested itself in the residual anisotropy.For Nup153FG upon addition of Importinb the fast sub-nanosecond component remained and the nanosecond component became slower together with an increase in residual anisotropy ( Figure 2D,F igure S6, and Table S1). Aglobal fit of the anisotropy decays at the tested 5 different sites of Nup153FG with two rotational correlation times showed that the longer time constant from segmental motion slowed down from % 5nst o% 25 ns.Intriguingly,the faster component appeared to become slightly faster (from % 0.52 ns to 0.37 ns) ( Figure 2D,F igure S6, and Table S1). Thes lowing down of the segmental motion at different Nup153FG sites and increase in residual anisotropy by % 0.1 for 883C ( Figure 2D,T able S1) provided direct evidence of Importinb binding.T he tentative acceleration of the fast picosecond component indicated increased picosecond motion at multiple sites upon Importinb binding.W es peculate that such behavior originated from many FG motifs anchoring on the multivalent Importinb sites and this might prevent self-interaction of the IDP chain yielding greater flexibility around the labeling site (very short length scale). Such behavior could also compensate for the entropic penalty of binding caused by the slowed segmental motion of the chain at comparatively longer length scales.F or IBB,i nt he presence of Importinb,t he sub-nanosecond component completely vanished and the entire decay was described by asingle component of % 30 ns that likely originated from the hydrodynamic tumbling motion of the entire complex (Figure 2E and Figure S7). This indicated avery rigid interface in the IBB-Importinb complex where all picosecond motions were frozen. Due to the limited time resolution of % 100 ps, we likely underestimated the initial anisotropy whenever ps dynamics were present (i.e.u nbound state), explaining the increased initial anisotropy in the bound state that lack such dynamics ( Figure 3E). These experiments highlight as tark difference between alabile interface with picosecond motion in the Nup153FG-Importinb complex and am ore rigid and dynamically frozen interface in the IBB-Importinb complex.
We next set out to probe the modulation of surface water dynamics at the interface of IBB and Nup153FG complexed with Importinb,c ompared to the unbound proteins.W e employed dipolar relaxation to probe water dynamics from femtoseconds to picoseconds by monitoring the time-dependent Stokes shift (TDSS) of the site-specifically labeled acrylodan. [12] To probe 5o rders of magnitude of solvation dynamics,w ec ombined broadband femtosecond transient absorption (fsTA, 200 fs À2ns) with time-correlated singlephoton counting (TCSPC) to probe TDSS from 200 fs to 20 ns ( Figures S8 and S9 and the Supporting Information). Broadband spectroscopy allowed us to obtain model-free TDSS data bypassing the need for convolution-based fitting of the kinetics at multiple wavelengths,which is typically used in solvation studies. [2a,b,13] Our analysis avoided an artificial smoothing from model-based fitting and retained the experimental noise in the data. Ther elaxation process in all cases was dominated by sub-20 ps processes.I na ll experiments, ! 75 %o ft he Stokes shift had occurred under 20 ps (Figure 3A-D). TheS tokes shift in all our TDSS data in the bound state (for IBB and Nup153FG) was always at higher frequencies compared to the free state and the total Stokes shift was greater in the free state;this corresponds to the redshifted emission and the more solvated environment in the free state,r espectively ( Figure 3A,B). This was in line with the conclusion from above that polarity decreased at the interface compared to the IDP surface on its own. Therelaxation functions shown in Figure 3could not be fit with simple bi-exponential or tri-exponential functions, indicating rather ac ontinuum of timescales in the solvent relaxation process.T his led us to fit the relaxation functions with as imple power law type function ( Figure S10) that described well the entire range of the Stokes shift and gave us as imple minimalistic empirical descriptor of the solvation process,w ithout the need for extrapolative assumptions of acertain number of exponential terms apriori. Thepower law exponent changed from 0.34 in free IBB to 0.26 in the bound state,suggesting aslowdown of solvent relaxation in the IBB-Importinb interface compared to the free IBB surface.F or Nup153FG we saw ad iametrically opposite behavior where the exponent changed from 0.32 to 0.37 indicating the overall relaxation to be faster in the bound state.
Then ature of IDP solvation has been proposed to be distinctly different from that of folded proteins in their native states. [14] While we cannot ascertain what molecular attributes give rise to such power law type relaxation in the IDPs that we have measured, unlike exponential relaxation behavior typically seen on folded protein surfaces,i ti sd efinitely an interesting question that warrants further systematic investigation. Power law type solvation dynamics have,h owever, been seen in several DNAs tructures and its origin has been am atter of debate. [15] Thep ower law exponent is ad irect indicator of the timescale of the solvent relaxation process, with ah igher exponent indicating af aster relaxation process and vice versa. It has to be noted that for power law type relaxation processes even am odest change in the exponent implies at remendous change in the timescale of the whole relaxation process.Acloser inspection of our TDSS indicates that up to 20 ps the dynamics in the bound and unbound states of Nup153FG 883C remained constant, in contrast to alarge slowdown observed in IBB as also supported by the scaling exponents obtained from power law fits up to 20 ps (Figure 3F). Since ! 75 %( dashed line Figure 3C,D) of the relaxation process in all cases was completed within 20 ps,we can conclude that alarge part of the water dynamics occurred on the sub-20 ps timescale;this part remained unperturbed in Nup153FG while it was substantially slowed down in the IBB case.T his suggested that overall acceleration of the Nup153FG solvation as seen from power law exponents occurred primarily from water relaxation at longer timescales. This was also evidenced by comparing the ratio S(t)(free)/ S(t)(bound) of normalized relaxation functions S(t)f or the bound and unbound scenarios for Nup153FG and IBB,where the ratio quickly approached values larger than one for IBB while for Nup153FG the values stayed close to one and started decreasing slowly at longer times ( Figure 3G). We also measured solvation dynamics at ad ifferent site in Nup153FG,n amely 1391C,w hich showed aq ualitatively similar trend to 883C manifested by an increase in the power law exponent in the bound state from 0.25 to 0.32 (Figure S11). While seeing differences,interms of the dynamical attributes,a tm ultiple sites in Nup153FG in different sequence contexts would have also been an interesting outcome,t he qualitative similarity we see at the different sites was also reassuring and validates the robustness of the findings.
Lacking experimental evidence of discrete relaxation processes in the solvation dynamics,wecannot assign distinct molecular species of water molecules such as free water, bound water, coupled protein-water dynamics,e tc.w idely used in literature for less "fuzzy" systems. [16] However,t he results strongly suggest that while the interfacial water dynamics in IBB-Importinb were slowed down substantially compared to the free state,for Nup153FG in the bound state most of the water dynamics remain unperturbed with some acceleration at longer times.Our experiments provide crucial insight into the very different nature of interfacial dynamics in IDP complexes depending on the mechanisms that drive their formation. Af olded structure is associated with dynamic stability and consequently in the IBB-Importinb complex formed by coupled folding-binding,t he interface is dynamically rigid and the interfacial solvation dynamics is substantially retarded. In the fuzzy Nup153FG-Importinb complex the accelerated picosecond motions seen in the time-resolved anisotropy might well be driven by the accelerated solvation dynamics in the longer (> 20 ps) timescales,c onsidering the recently established paradigm where solvation dynamics beget conformational fluctuations of the protein. [16] While interfacial water dynamics in any protein complex and hence ipso facto any IDP complex is hitherto unmeasured, it is still interesting to speculate to what extent this could relate to general differences in IDP binding mechanisms.I nb iology, coupled folding-binding complexes are often associated with kinetic stability, [3] which can be attributed to arigid interface and slowed interfacial solvation dynamics.F uzzy complexes, on the other hand, are typically associated with more transient interactions,w hich can be more advantageous for several biological functions. [3, 4b] Ad ynamic interface associated with unperturbed or somewhat accelerated solvation dynamics might facilitate fuzziness.Asimple activation barrier based argument can explain how slowed down interfacial solvation dynamics vs.l argely unperturbed interfacial solvation dynamics might facilitate kinetic stability and kinetic lability,r espectively ( Figure S12). This is aq uestion relevant to cargo transport through the nuclear pore complex, which is known to be fast, yet specific. [6] In the case of Nup153FG and IBB,s uch behavior can be directly linked to their functions as well. Kinetic stability in the IBB complex is crucial to maintain the integrity of the import cargo complex during transport across the NPC,w hile kinetic lability is important for FGNups to facilitate fast transport. Minimal perturbation of most (! 75 %) of the solvation dynamics in FGNups poses minimal energy barriers from solvation towards unbinding events.T his might be the crucial barrierreducing mechanism that allows FGNups to remain mobile on NTR surfaces. [6,17] We hypothesize that such drastically different interfacial solvation dynamics might be ag eneral mechanism for IDPs to tune the plasticity of complexes ranging from kinetically stable to fuzzy,a nd thus encode various functionality.I ns ummary our work underscores the supreme importance of ultrafast dynamics,e specially that of the solvent milieu.