Spontaneous binding of single-stranded RNAs to RRM proteins visualized by unbiased atomistic simulations with a rescaled RNA force field.

Recognition of single-stranded RNA (ssRNA) by RNA recognition motif (RRM) domains is an important class of protein-RNA interactions. Many such complexes were characterized using nuclear magnetic resonance (NMR) and/or X-ray crystallography techniques, revealing ensemble-averaged pictures of the bound states. However, it is becoming widely accepted that better understanding of protein-RNA interactions would be obtained from ensemble descriptions. Indeed, earlier molecular dynamics simulations of bound states indicated visible dynamics at the RNA-RRM interfaces. Here, we report the first atomistic simulation study of spontaneous binding of short RNA sequences to RRM domains of HuR and SRSF1 proteins. Using a millisecond-scale aggregate ensemble of unbiased simulations, we were able to observe a few dozen binding events. HuR RRM3 utilizes a pre-binding state to navigate the RNA sequence to its partially disordered bound state and then to dynamically scan its different binding registers. SRSF1 RRM2 binding is more straightforward but still multiple-pathway. The present study necessitated development of a goal-specific force field modification, scaling down the intramolecular van der Waals interactions of the RNA which also improves description of the RNA-RRM bound state. Our study opens up a new avenue for large-scale atomistic investigations of binding landscapes of protein-RNA complexes, and future perspectives of such research are discussed.

Molecular dynamics simulations reveal the parallel stranded d(GGGA)3GGG DNA quadruplex folds via multiple paths from a coil-like ensemble.

G-quadruplexes (G4s) are non-canonical nucleic acid structures that fold through complex processes. Characterization of the G4 folding landscape may help to elucidate biological roles of G4s but is challenging both experimentally and computationally. Here, we achieved complete folding of a three-quartet parallel DNA G4 with (GGGA)3GGG sequence using all-atom explicit-solvent enhanced-sampling molecular dynamics (MD) simulations. The simulations suggested early formation of guanine stacks in the G-tracts, which behave as semi-rigid blocks in the folding process. The folding continues via the formation of a collapsed compact coil-like ensemble. Structuring of the G4 from the coil then proceeds via various cross-like, hairpin, slip-stranded and two-quartet ensembles and can bypass the G-triplex structure. Folding of the parallel G4 does not appear to involve any salient intermediates and is a multi-pathway process. We also carried out an extended set of simulations of parallel G-hairpins. While parallel G-hairpins are extremely unstable when isolated, they are more stable inside the coil structure. On the methodology side, we show that the AMBER DNA force field predicts the folded G4 to be less stable than the unfolded ensemble, uncovering substantial force-field issues. Overall, we provide unique atomistic insights into the folding landscape of parallel-stranded G4 but also reveal limitations of current state-of-the-art MD techniques.

Conformational Heterogeneity of RNA Stem-Loop Hairpins Bound to FUS-RNA Recognition Motif with Disordered RGG Tail Revealed by Unbiased Molecular Dynamics Simulations.

RNA-protein complexes use diverse binding strategies, ranging from structurally well-defined interfaces to completely disordered regions. Experimental characterization of flexible segments is challenging and can be aided by atomistic molecular dynamics (MD) simulations. Here, we used an extended set of microsecond-scale MD trajectories (400 µs in total) to study two FUS-RNA constructs previously characterized by nuclear magnetic resonance (NMR) spectroscopy. The FUS protein contains a well-structured RNA recognition motif domain followed by a presumably disordered RGG tail that binds RNA stem-loop hairpins. Our simulations not only provide several suggestions complementing the experiments but also reveal major methodological difficulties in studies of such complex RNA-protein interfaces. Despite efforts to stabilize the binding via system-specific force-field adjustments, we have observed progressive distortions of the RNA-protein interface inconsistent with experimental data. We propose that the dynamics is so rich that its converged description is not achievable even upon stabilizing the system. Still, after careful analysis of the trajectories, we have made several suggestions regarding the binding. We identify substates in the RNA loops, which can explain the NMR data. The RGG tail localized in the minor groove remains disordered, sampling countless transient interactions with the RNA. There are long-range couplings among the different elements contributing to the recognition, which can lead to allosteric communication throughout the system. Overall, the RNA-FUS systems form dynamical ensembles that cannot be fully represented by single static structures. Thus, albeit imperfect, MD simulations represent a viable tool to investigate dynamic RNA-protein complexes.

Residues flanking the ARKme3T/S motif allow binding of diverse targets to the HP1 chromodomain: Insights from molecular dynamics simulations.

BACKGROUND: The chromodomain (CD) of HP1 proteins is an established H3K9me3 reader that also binds H1, EHMT2 and H3K23 lysine-methylated targets. Structural experiments have provided atomistic pictures of its recognition of the conserved ARKme3S/T motif, but structural dynamics' contribution to the recognition may have been masked by ensemble averaging. METHODS: We acquired ~350 µs of explicit solvent molecular dynamics (MD) simulations of the CD domain interacting with several peptides using the latest AMBER force fields. RESULTS: The simulations reproduced the experimentally observed static binding patterns well but also revealed visible structural dynamics at the interfaces. While the buried K0me3 and A-2 target residues are tightly bound, several flanking sidechains sample diverse sites on the CD surface. Different amino acid positions of the targets can substitute for each other by forming mutually replaceable interactions with CD, thereby explaining the lack of strict requirement for cationic H3 target residues at the -3 position. The Q-4 residue of H3 targets further stabilizes the binding. The recognition pattern of the H3K23 ATKme3A motif, for which no structure is available, is predicted. CONCLUSIONS: The CD reads a longer target segment than previously thought, ranging from positions -7 to +3. The CD anionic clamp can be neutralized not only by the -3 and -1 residues, but also by -7, -6, -5 and +3 residues. GENERAL SIGNIFICANCE: Structural dynamics, not immediately apparent from the structural data, contribute to molecular recognition between the HP1 CD domain and its targets. Mutual replaceability of target residues increases target sequence flexibility.

Recognition of N6-Methyladenosine by the YTHDC1 YTH Domain Studied by Molecular Dynamics and NMR Spectroscopy: The Role of Hydration.

The YTH domain of YTHDC1 belongs to a class of protein "readers", recognizing the N6-methyladenosine (m6A) chemical modification in mRNA. Static ensemble-averaged structures revealed details of N6-methyl recognition via a conserved aromatic cage. Here, we performed molecular dynamics (MD) simulations along with nuclear magnetic resonance (NMR) and isothermal titration calorimetry (ITC) to examine how dynamics and solvent interactions contribute to the m6A recognition and negative selectivity toward an unmethylated substrate. The structured water molecules surrounding the bound RNA and the methylated substrate's ability to exclude bulk water molecules contribute to the YTH domain's preference for m6A. Intrusions of bulk water deep into the binding pocket disrupt binding of unmethylated adenosine. The YTHDC1's preference for the 5'-Gm6A-3' motif is partially facilitated by a network of water-mediated interactions between the 2-amino group of the guanosine and residues in the m6A binding pocket. The 5'-Im6A-3' (where I is inosine) motif can be recognized too, but disruption of the water network lowers affinity. The D479A mutant also disrupts the water network and destabilizes m6A binding. Our interdisciplinary study of the YTHDC1 protein-RNA complex reveals an unusual physical mechanism by which solvent interactions contribute toward m6A recognition.

UUCG RNA Tetraloop as a Formidable Force-Field Challenge for MD Simulations.

Explicit solvent atomistic molecular dynamics (MD) simulations represent an established technique to study structural dynamics of RNA molecules and an important complement for diverse experimental methods. However, performance of molecular mechanical (MM) force fields (ff's) remains far from satisfactory even after decades of development, as apparent from a problematic structural description of some important RNA motifs. Actually, some of the smallest RNA molecules belong to the most challenging systems for MD simulations and, among them, the UUCG tetraloop is saliently difficult. We report a detailed analysis of UUCG MD simulations, depicting the sequence of events leading to the loss of the UUCG native state during MD simulations. The total amount of MD simulation data analyzed in this work is close to 1.3 ms. We identify molecular interactions, backbone conformations, and substates that are involved in the process. Then, we unravel specific ff deficiencies using diverse quantum mechanical/molecular mechanical (QM/MM) and QM calculations. Comparison between the MM and QM methods shows discrepancies in the description of the 5'-flanking phosphate moiety and both signature sugar-base interactions. Our work indicates that poor behavior of the UUCG tetraloop in simulations is a complex issue that cannot be attributed to one dominant and straightforwardly correctable factor. Instead, there is a concerted effect of multiple ff inaccuracies that are coupled and amplifying each other. We attempted to improve the simulation behavior by some carefully tailored interventions, but the results were still far from satisfactory, underlying the difficulties in development of accurate nucleic acid ff's.

Role of Fine Structural Dynamics in Recognition of Histone H3 by HP1γ(CSD) Dimer and Ability of Force Fields to Describe Their Interaction Network.

Human heterochromatin protein 1 (HP1) is a key factor in heterochromatin formation and maintenance. Its chromo-shadow domain (CSD) homodimerizes, and the HP1 dimer acts as a hub, transiently interacting with diverse binding partner (BP) proteins. We analyze atomistic details of interactions of the HP1γ(CSD) dimer with one of its targets, the histone H3 N-terminal tail, using molecular dynamics (MD) simulations. The goal is to complement the available X-ray crystallography data and unravel potential dynamic effects in the molecular recognition. Our results suggest that HP1(CSD)-BP recognition involves structural dynamics of both partners, including structural communication between adjacent binding pockets that may fine-tune the sequence recognition. For example, HP1 Trp174 sidechain substates may help in distinguishing residues bound in the conserved HP1(CSD) ±2 hydrophobic pockets. Further, there is intricate competition between the binding of negatively charged HP1 C-terminal extension and solvent anions near the ±2 hydrophobic pockets, which is also influenced by the BP sequence. Phosphorylated H3 Y41 can interact with the same site. We also analyze the ability of several pair-additive force fields to describe the protein-protein interface. ff14SB and ff99SB-ILDN* provide the closest correspondence with the crystallographic model. The ff15ipq local dynamics are somewhat less consistent with details of the experimental structure, while larger perturbations of the interface commonly occur in CHARMM36m simulations. The balance of some interactions, mainly around the anion binding site, also depends on the ion parameters. Some differences between the simulated and experimental structures are attributable to crystal packing.

DNA Damage Changes Distribution Pattern and Levels of HP1 Protein Isoforms in the Nucleolus and Increases Phosphorylation of HP1ß-Ser88.

The family of heterochromatin protein 1 (HP1) isoforms is essential for chromatin packaging, regulation of gene expression, and repair of damaged DNA. Here we document that γ-radiation reduced the number of HP1α-positive foci, but not HP1ß and HP1γ foci, located in the vicinity of the fibrillarin-positive region of the nucleolus. The additional analysis confirmed that γ-radiation has the ability to significantly decrease the level of HP1α in rDNA promoter and rDNA encoding 28S rRNA. By mass spectrometry, we showed that treatment by γ-rays enhanced the HP1ß serine 88 phosphorylation (S88ph), but other analyzed modifications of HP1ß, including S161ph/Y163ph, S171ph, and S174ph, were not changed in cells exposed to γ-rays or treated by the HDAC inhibitor (HDACi). Interestingly, a combination of HDACi and γ-radiation increased the level of HP1α and HP1γ. The level of HP1ß remained identical before and after the HDACi/γ-rays treatment, but HDACi strengthened HP1ß interaction with the KRAB-associated protein 1 (KAP1) protein. Conversely, HP1γ did not interact with KAP1, although approximately 40% of HP1γ foci co-localized with accumulated KAP1. Especially HP1γ foci at the periphery of nucleoli were mostly absent of KAP1. Together, DNA damage changed the morphology, levels, and interaction properties of HP1 isoforms. Also, γ-irradiation-induced hyperphosphorylation of the HP1ß protein; thus, HP1ß-S88ph could be considered as an important marker of DNA damage.

QM/MM Calculations on Protein-RNA Complexes: Understanding Limitations of Classical MD Simulations and Search for Reliable Cost-Effective QM Methods.

Although atomistic explicit-solvent Molecular Dynamics (MD) is a popular tool to study protein-RNA recognition, satisfactory MD description of protein-RNA complexes is not always achieved. Unfortunately, it is often difficult to separate MD simulation instabilities primarily caused by the simple point-charge molecular mechanics (MM) force fields from problems related to the notorious uncertainties in the starting structures. Herein, we report a series of large-scale QM/MM calculations on the U1A protein-RNA complex. This experimentally well-characterized system has an intricate protein-RNA interface, which is very unstable in MD simulations. The QM/MM calculations identify several H-bonds poorly described by the MM method and thus indicate the sources of instabilities of the U1A interface in MD simulations. The results suggest that advanced QM/MM computations could be used to indirectly rationalize problems seen in MM-based MD simulations of protein-RNA complexes. As the most accurate QM method, we employ the computationally demanding meta-GGA density functional TPSS-D3(BJ)/def2-TZVP level of theory. Because considerably faster methods would be needed to extend sampling and to study even larger protein-RNA interfaces, a set of low-cost QM/MM methods is compared to the TPSS-D3(BJ)/def2-TZVP data. The PBEh-3c and B97-3c density functional composite methods appear to be suitable for protein-RNA interfaces. In contrast, HF-3c and the tight-binding Hamiltonians DFTB3-D3 and GFN-xTB perform unsatisfactorily and do not provide any advantage over the MM description. These conclusions are supported also by similar analysis of a simple HutP protein-RNA interface, which is well-described by MD with the exception of just one H-bond. Some other methodological aspects of QM/MM calculations on protein-RNA interfaces are discussed.

MD and QM/MM Study of the Quaternary HutP Homohexamer Complex with mRNA, l-Histidine Ligand, and Mg2.

The HutP protein from B. subtilis regulates histidine metabolism by interacting with an antiterminator mRNA hairpin in response to the binding of l-histidine and Mg2+. We studied the functional ligand-bound HutP hexamer complexed with two mRNAs using all-atom microsecond-scale explicit-solvent MD simulations performed with the Amber force fields. The experimentally observed protein-RNA interface exhibited good structural stability in the simulations with the exception of some fluctuations in an unusual adenine-threonine interaction involving two closely spaced H-bonds. We further investigated this interaction by comparing QM/MM and MM optimizations, using the QM region comprising almost 350 atoms described at the DFT-D3 level. The QM/MM method clearly improved the adenine-threonine interaction compared to MM, especially when the X-H bond lengths were frozen during the MM optimization to mimic the use of SHAKE in the MD simulations. Thus, both the MM approximation and the use of SHAKE can compromise the description of H-bonds at protein-RNA interfaces. The simulations also revealed a notable Mg2+-parameter dependence in the behavior of the ligand-binding pocket (LBP). With the SPC/E water model, the 12-6 Åqvist and Li&Merz parameters provided an entirely stable LBP structure, but the 12-6 Allnér and 12-6-4 Li&Merz parametrizations resulted in a progressive loss of direct nitrogen-Mg2+ LBP coordination. The Allnér and Li&Merz 12-6 parametrizations were also tested with the TIP3P water model; the LBP was destabilized in both cases. This illustrates the difficulty of consistently describing different Mg2+ interactions using nonpolarizable force fields. Overall, the simulations support the hypothesis that HutP protein becomes fully structured upon ligand binding. Subsequent RNA binding does not affect the protein structure, in keeping with the mechanism inferred from experimental structures.