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

https://doi.org/10.1016/j.bbagen.2020.129771Get rights and content

Highlights

  • Atomistic simulations of HP1 chromodomain (CD) bound to diverse targets are reported.

  • The binding includes structural dynamics that is not visible from experimental data.

  • Residues as distant as −7 to +3 positions from Kme3 affect target binding to HP1 CD.

  • Amino acids in different positions can replace each other in the binding.

  • Recent AMBER force fields provide improved description of the CD-target binding.

Abstract

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.

Introduction

HP1 (Heterochromatin Protein 1) proteins have multiple functions, including participation in the formation and maintenance of heterochromatin, transcription regulation and RNA splicing, DNA repair, telomere maintenance and chromosome separation [[1], [2], [3], [4], [5]]. Their versatility can be attributed to their ability to bind numerous partner proteins as well as nucleic acids, and typically they have multiple isoforms. There are three isoforms in mammals – HP1α, β, and γ (also called CBX5, 1 and 3, respectively) – with differing chromosomal localizations and post-translational modification (PTM) patterns [[6], [7], [8]], which further tune their interactions.

HP1 proteins have two structured domains, which are connected and flanked by flexible, intrinsically disordered regions (IDRs, Fig. 1A). The chromodomain (CD, from CHROmatin MOdifier) is a well-known reader of histone protein H3 N-terminal tail lysine 9 methylation, a key heterochromatin-associated PTM marker [[9], [10], [11]]. CD has also been shown to bind other IDRs with methyllysine, namely N-terminal regions of histone H1 [12] and histone methyltransferase EHMT2 (also called G9a protein) [13], see Fig. 1C. There is also evidence of its binding to H3 methylated at K23 [14,15]. The chromo-shadow domain (CSD) is responsible for HP1 homodimerization. CSD homodimers mediate most HP1 interactions with partner proteins [5,[16], [17], [18], [19], [20]]. An IDR separating the two domains, the hinge region, mediates HP1s' interactions with nucleic acids. In this manner, HP1s not only associate with chromatin but also affect transcription and co-transcriptional splicing processes [2,[21], [22], [23]]. Flexible N- and C-terminal HP1 regions modulate binding affinities and specificities of the CD and CSD domains, respectively [18,24,25]. Broad ranges of interactions mediated by different regions of the HP1s and their ability to dynamically exchange binding partners allow them to serve as chromatin-associated hub proteins [26,27].

Structures of HP1(CD) with H3K9, H1K26 and EHMT2K185 lysine-methylated peptides have been experimentally determined [28,29]. All three of these target peptides share an ARKme3S/T motif, which is replaced by ATKme3A in the H3K23 target (Fig. 1C). CD recognizes the Kme3 (or Kme2), referred to as the 0 position of the bound peptide, via a hydrophobic cage (Fig. 1B). The backbone of the preceding segment is bound in a β-sheet style in the peptide-binding groove formed between the CD N-terminal tail and L3 loop. The −2 residue of the target, alanine, is anchored in a small hydrophobic pocket in the binding groove. –1 (arginine) and − 3 residue sidechains are oriented in the opposite direction, towards an anionic clamp formed by CD residues Glu29 (N-tail) and Asp68 (loop L3). H3 targets do not have clamp-neutralizing residues at the −3 position but the other two targets do (Fig. 1C). In the +1 position, serine or threonine form a sidechain H-bond to CD Glu62. Available experimental structures do not show any specific contacts of other residues of the bound peptides, and there is no structural information on CD binding to H3K23 peptide.

Experimental techniques are required to determine structures of proteins and their complexes, but they provide limited information on dynamics of the studied systems due to the ensemble-averaged nature of the collected data. Specific limitations of each technique inevitably restrict both the accuracy and detail of the resulting structures. Among others, potential roles of molecular motions and rare conformations involved in binding may be obscured. These limitations may be addressed by combining experimental structural data with computational approaches, particularly atomistic molecular dynamics (MD) simulations, which can portray detailed, atomic-level structural dynamics on pico- to millisecond time scales [[30], [31], [32]].

Some details of HP1 CD-H3K9 binding have already been characterized using MD simulations. Simulations of the CD-H3K9 complex with different analogues of Kme3 led to proposals that hydrophobic effects play more important roles in Kme3 readout than cation-π interactions [25]. It has also been suggested that the H3K9 peptide has higher observed affinity for the HP1β isoform than the HP1γ isoform mainly due to enhanced electrostatic interactions of R8 (R−1) and K14 (K+5) with the CD N-terminal stretch, which is more acidic in the β isoform [33]. In the same study, attempts were made to explain experimentally observed [34,35] weakening of the CD-H3K9 binding by the H3S10phos PTM. An increase in flexibility following this PTM was observed in 100 ns simulations, but the trajectories did not provide clear indications of a specific mechanism. Later microsecond-scale simulations showed that H3S10phos PTM impedes S10 binding to Glu62 and further weakens electrostatic interactions between R8 and CD by forming an intramolecular R8-S10phos salt bridge [36]. In addition, a study of effects of HP1 N-terminal tail phosphorylation on CD-H3K9 binding via 50 ns replica-exchange enhanced-sampling MD simulations in conjunction with NMR and SAXS have shown that the phosphorylation extends the tail and enhances its interactions with H3K9 peptide [25]. A recent study of full-length HP1 binding to H3K9 methylated dinucleosome using coarse-grained MD simulations [37] suggests that HP1 preference for dinucleosomes over mononucleosomes is due to interaction between the hinge region and DNA joining the nucleosomes, and that the γ isoform binds to dinucleosomes less efficiently than the α isoform because its hinge region is shorter and less basic.

In the study presented here we applied MD techniques to study HP1γ CD's binding to its four known targets. Molecular modelling allowed us to explore the behavior of flexible tails, which were omitted in experimental structures. Focusing on the role of the ARKme3S/T motif and its sequence context in binding, the simulations revealed that cationic residues as far as seven amino acids from Kme3 affect the binding and can compensate for the lack of clamp charge neutralization by the −3 residue in H3 targets. CD dynamics near the clamp also depend on the −4 residue of the bound target, since glutamine stabilizes CD-target interaction. We modelled the CD-H3K23 complex and the results suggest that the H3K23 peptide's interactions are analogous to those of the other targets but involve different residues. Results also neatly explain why H3K27 does not efficiently bind CD, in contrast to H3K23. We also studied effects of two experimentally detected HP1(CD) PTMs on the structure and dynamics of free CD and discuss here their potential effects on peptide binding. In addition, we critically evaluate the performance of AMBER ff19SB/OPC and ff14SB/TIP3P force fields for modelling these quite dynamic protein-protein complexes. Both provide reasonable agreement with available experimental data, but ff19SB provides somewhat better results for the focal systems.

Section snippets

Residue nomenclature and numbering

Here, we use three- and one-letter codes for amino acids of HP1 and its binding targets, respectively. Positions of the target peptide residues are specified relative to Kme3 by a subscript, e.g. R−1. HP1 residue numbering is based on human HP1γ. Kme3 numbering in Fig. 1C is based on human H1.4, H3.1, and EHMT2.1 sequences. The Kme3 number is also included in names of H3 targets to distinguish H3K9, H3K23 and H3K27 (in these abbreviations the sign for K methylation is omitted). If a mutation

Results

We aimed to elucidate effects of the targets' sequence variability on their recognition by the CD by modelling the HP1γ CD domain's interactions with several targets. Effects of threonine PTM modifications in the CD domain and the performance of protein simulation force fields are also addressed. Unless otherwise specified, ff19SB/OPC results are described.

Discussion

We have used MD simulations to explore structural dynamics of complexes formed by the HP1γ CD domain and its diverse IDR targets, which cannot be directly inferred from available ensemble-averaged experimental data. The simulations confirm that the bound state is associated with substantial local dynamics, which allow different residues along the target sequences to form analogous interactions that enhance the sequence flexibility of the CD targets. Besides investigating the role of structural

Conclusions

HP1 is a chromatin hub protein with a CD domain that binds specific Kme3 marks found mainly on histone tails (Fig. 1C). We report here microsecond-scale MD simulations intended to complement available experimental data on CD-target complexes through exploration of their dynamic behavior. In addition to simulations based on experimental structures, we have modelled the CD-H3K23 complex, for which there was no structural information. The simulations show that the bound state is associated with

CRediT authorship contribution statement

Pavlína Pokorná: Conceptualization, Methodology, Investigation, Writing - original draft, Writing - review & editing. Miroslav Krepl: Methodology, Validation. Jiří Šponer: Conceptualization, Resources, Funding acquisition, Writing - review & editing.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgement

We thank Klaudia Mráziková for help with simulation analyses. This work was supported by the Czech Science Foundation (grant number: 18-07384S) and Brno PhD Talent Financial Aid to P.P.

References (95)

  • L. Kaustov et al.

    Recognition and specificity determinants of the human cbx chromodomains

    J. Biol. Chem.

    (2011)
  • J.R. Allison

    Using simulation to interpret experimental data in terms of protein conformational ensembles

    Curr. Opin. Struc. Biol.

    (2017)
  • M. Bonomi et al.

    Principles of protein structural ensemble determination

    Curr. Opin. Struc. Biol.

    (2017)
  • G.V. Papamokos et al.

    Structural ro of RKS Motifs in cromatin iteractions: a molecular dynamics study of HP1 bound to a variably modified histone tail

    Biophys. J.

    (2012)
  • S. Watanabe et al.

    Interactions of HP1 bound to H3K9me3 dinucleosome by molecular simulations and biochemical assays

    Biophys. J.

    (2018)
  • S. Le Grand et al.

    SPFP: speed without compromise - a mixed precision model for GPU accelerated molecular dynamics simulations

    Comput. Phys. Commun.

    (2013)
  • J. Aqvist et al.

    Molecular dynamics simulations of water and biomolecules with a Monte Carlo constant pressure algorithm

    Chem. Phys. Lett.

    (2004)
  • J.P. Ryckaert et al.

    Numerical-integration of cartesian equations of motion of a system with constraints - molecular-dynamics of N-alkanes

    J. Comput. Phys.

    (1977)
  • W. Humphrey et al.

    VMD: visual molecular dynamics

    J. Mol. Graphics Modell.

    (1996)
  • D. Pantoja-Uceda et al.

    The isolated C-terminal nuclear localization sequence of the breast cancer metastasis suppressor 1 is disordered

    Arch. Biochem. Biophys.

    (2019)
  • R. Lu et al.

    Tudor: a versatile family of histone methylation 'readers'

    Trends Biochem. Sci.

    (2013)
  • R. Sanchez et al.

    The PHD finger: a versatile epigenome reader

    Trends Biochem. Sci.

    (2011)
  • F. He et al.

    Structural insight into the zinc finger CW domain as a histone modification reader

    Structure

    (2010)
  • M. Soniat et al.

    Karyopherin-beta2 recognition of a PY-NLS variant that lacks the proline-tyrosine motif

    Structure

    (2016)
  • P.S. Nerenberg et al.

    New developments in force fields for biomolecular simulations

    Curr. Opin. Struc. Biol.

    (2018)
  • R.B. Best

    Computational and theoretical advances in studies of intrinsically disordered proteins

    Curr. Opin. Struct. Biol.

    (2017)
  • R.B. Best et al.

    Residue-specific alpha-helix propensities from molecular simulation

    Biophys. J.

    (2012)
  • D. Matthes et al.

    Secondary structure propensities in peptide folding simulations: a systematic comparison of molecular mechanics interaction schemes

    Biophys. J.

    (2009)
  • A. Kumar et al.

    Heterochromatin protein 1 (HP1): interactions with itself and chromatin components

    Biophys. Rev.

    (2020)
  • Y.H. Li et al.

    Does heterochromatin protein 1 always follow code?

    Proc. Natl. Acad. Sci. U. S. A.

    (2002)
  • K.Y. Huang et al.

    10-year anniversary of a resource for post-translational modification of proteins

    Nucleic Acids Res.

    (2016)
  • E. Minc et al.

    Localization and phosphorylation of HP1 proteins during the cell cycle in mammalian cells

    Chromosoma

    (1999)
  • A.J. Bannister et al.

    Selective recognition of methylated lysine 9 on histone H3 by the HP1 chromo domain

    Nature

    (2001)
  • M. Lachner et al.

    Methylation of histone H3 lysine 9 creates a binding site for HP1 proteins

    Nature

    (2001)
  • H.G. Chin et al.

    Automethylation of G9a and its implication in wider substrate specificity and HP1 binding

    Nucleic Acids Res.

    (2007)
  • E.K. Shanle et al.

    Histone peptide microarray screen of chromo and Tudor domains defines new histone lysine methylation interactions

    Epigenet. Chromatin.

    (2017)
  • D.L. Mendez et al.

    The HP1a disordered C terminus and chromo shadow domain cooperate to select target peptide partners

    Chembiochem

    (2011)
  • A.H. Horakova et al.

    SUV39h-independent association of HP1 beta with fibrillarin-positive nucleolar regions

    Chromosoma

    (2010)
  • C. Maison et al.

    HP1 and the dynamics of heterochromatin maintenance

    Nat. Rev. Mol. Cell Biol.

    (2004)
  • V. Saint-Andre et al.

    Histone H3 lysine 9 trimethylation and HP1 gamma favor inclusion of alternative exons

    Nat. Struct. Mol. Biol.

    (2011)
  • R.R. Meehan et al.

    HP1 binding to native chromatin in vitro is determined by the hinge region and not by the chromodomain

    EMBO J.

    (2003)
  • C. Muchardt et al.

    Coordinated methyl and RNA binding is required for heterochromatin localization of mammalian HP1 alpha

    EMBO Rep.

    (2002)
  • H. Shimojo et al.

    Extended string-like binding of the phosphorylated HP1 alpha N-terminal tail to the lysine 9-methylated histone H3 tail

    Sci. Rep.

    (2016)
  • T. Cheutin et al.

    In vivo dynamics of Swi6 in yeast: evidence for a stochastic model of heterochromatin

    Mol. Cell. Biol.

    (2004)
  • T. Cheutin et al.

    Maintenance of stable heterochromatin domains by dynamic HP1 binding

    Science

    (2003)
  • J. Ruan et al.

    Structural basis of the chromodomain of Cbx3 bound to methylated peptides from histone h1 and G9a

    PLoS One

    (2012)
  • J. Sponer et al.

    RNA structural dynamics as captured by molecular simulations: a comprehensive overview

    Chem. Rev.

    (2018)
  • View full text