Toward Selective Histone Deacetylase Inhibitor Design: Homology Modeling, Docking Studies, and Molecular Dynamics Simulations of Human Class I Histone Deacetylases

Difei Wang, Paul Helquist, Norbert L. Wiech, Olaf WiestJournal of Medicinal Chemistry 2005 · DOI

TL;DR

This paper tackles the challenge of designing selective inhibitors for Class I histone deacetylases (HDACs), crucial targets in cancer therapy. By building homology models of HDAC1, HDAC2, HDAC3, and HDAC8, and then using molecular docking and dynamics simulations, the researchers identified key structural differences that could be exploited for isoform-specific drug design.

What Problem Does This Paper Address?

Histone deacetylases (HDACs) are enzymes that play a critical role in gene regulation. Inhibiting HDACs has shown promise in cancer treatment by inducing cell differentiation and suppressing tumor growth. However, current HDAC inhibitors often lack selectivity, affecting multiple HDAC isoforms and potentially leading to side effects. The Class I HDACs (HDAC1, HDAC2, HDAC3, and HDAC8) are particularly important targets, but their high sequence and structural similarity within the active site makes it difficult to design drugs that specifically target one isoform over others. This paper aims to bridge this gap by providing a detailed structural understanding of these Class I HDACs and their interactions with inhibitors, paving the way for more targeted therapies.

Methods

The researchers employed a multi-pronged computational approach:

  • Homology Modeling: They built three-dimensional structural models for HDAC1, HDAC2, HDAC3, and HDAC8 using the known structure of a related protein, HDLP (histone deacetylase-like protein), as a template. This is a standard technique when experimental structures are unavailable.
  • Model Validation: The homology models were rigorously validated. Crucially, the model for HDAC8 was compared to a recently published X-ray crystal structure of HDAC8, showing excellent agreement. This provided strong confidence in the accuracy of their models.
  • Molecular Docking: They used AutoDock 3.0 to computationally "dock" several known HDAC inhibitors (TSA, SK-683, CG-1521, SAHA, MS-275, and NVP-LAQ824) into the active sites of their HDAC models. This process predicts how the inhibitors might bind and estimates their binding energies.
  • Molecular Dynamics (MD) Simulations: To account for protein flexibility, which is often neglected in standard docking, they performed 1 ns MD simulations of the HDAC-inhibitor complexes. This allowed them to observe how the protein and inhibitor interact over time and assess the stability of the binding. They specifically used a nonbonded model for the zinc ion in the active site to allow for dynamic coordination changes.

Datasets

The study utilized publicly available protein sequences from the NCBI protein sequence database for HDAC1, HDAC2, HDAC3, HDAC8, and the template protein HDLP. They also used existing X-ray crystal structures of HDLP and HDAC8 for model building and validation. The inhibitors used (TSA, SK-683, CG-1521, SAHA, MS-275, and NVP-LAQ824) are well-established compounds from the literature.

Key Insights and Ideas

  • Active Site Similarity is a Major Hurdle: The active sites and the primary channel leading to them are highly conserved across Class I HDACs. This makes it very difficult to achieve isoform selectivity by targeting these regions alone.
  • The "Internal Cavity" as a Potential Target: A conserved 14 Å internal cavity adjacent to the active site, whose function was previously unclear, is highlighted. This cavity's flexibility and exposure to the exterior in some HDAC structures (like HDAC8) suggest it could serve as a secondary binding site, offering a potential avenue for isoform-specific interactions.
  • Protein Flexibility Matters: Standard docking protocols often treat proteins as rigid. The MD simulations revealed that active site residues and surrounding loops exhibit flexibility, which can influence inhibitor binding and is crucial for accurate modeling.
  • Surface Electrostatics Offer Clues: Differences in the electrostatic potential on the protein surface, particularly around the active site entrance, were observed between isoforms. These differences might play a role in recruiting other proteins to form complexes, influencing HDAC function and potentially offering another angle for selective targeting.
  • Computational Approach Validated: The successful reproduction of experimental observations (e.g., relative binding strengths of inhibitors) and the excellent agreement of the HDAC8 homology model with its X-ray structure validate the computational methodology for future drug design efforts.

Key Results

  • Homology Models: Reliable homology models for HDAC1, HDAC2, HDAC3, and HDAC8 were generated and validated.
  • Binding Energy Predictions: Docking studies provided estimated binding free energies for several HDAC inhibitors across the different isoforms. These calculations generally agreed with experimental data, showing TSA and SK-683 binding strongly to HDAC1.
  • Isoform-Specific Binding Modes: While TSA, SK-683, and CG-1521 showed similar binding patterns across HDAC1, HDAC2, and HDAC3, the study identified subtle differences in how NVP-LAQ824 interacts with HDAC1 and HDAC3, suggesting a potential for isoform selectivity. Specifically, a change in a surface residue (Glu98 to Asp92) was implicated in the differential binding of NVP-LAQ824.
  • MS-275 and the Secondary Binding Site: The docking of MS-275 to HDAC8 revealed that it binds to the internal cavity adjacent to the active site, rather than directly to the active site itself. This explains its lack of activity against HDAC8 and highlights the importance of the secondary binding site.
  • MD Simulation Stability: The MD simulations confirmed the stability of the HDAC-inhibitor complexes over 1 ns, with zinc coordination numbers and geometries remaining consistent with experimental observations.

Limitations and Future Work

  • Limitations of Scoring Functions: The authors acknowledge that scoring functions used in docking have inherent errors, and differences in binding energies within this error range are difficult to interpret definitively.
  • Limited Experimental Data for Some Isoforms: While comparisons were made to experimental data where available, the lack of comprehensive experimental data for all isoforms and inhibitors limits direct validation of all findings.
  • Future Work: The paper strongly suggests that the identified differences in the internal cavity and surface electrostatics, along with protein flexibility, are key areas for future exploration in designing isoform-selective HDAC inhibitors. They propose that optimizing interactions in these regions, rather than solely focusing on the metal-binding site, could lead to more targeted drugs. The authors are actively pursuing the design, synthesis, and evaluation of novel HDAC inhibitors based on these principles. The computational approach is also proposed as a valuable tool for screening large virtual libraries to identify new selective inhibitors.

Summary updated

Abstract

Histone deacetylases (HDACs) play an important role in gene transcription. Inhibitors of HDACs induce cell differentiation and suppress cell proliferation in tumor cells. Although many HDAC inhibitors have been designed and synthesized, selective inhibition for class I HDAC isoforms is a goal that has yet to be achieved. To understand the difference between class I HDAC isoforms that could be exploited for the design of isoform-specific HDAC inhibitors, we have built three-dimensional models of four class I histone deacetylases, HDAC1, HDAC2, HDAC3, and HDAC8. Comparison of the homology model of HDAC8 with the recently published X-ray structure shows excellent agreement and validates the approach. A series of HDAC inhibitors were docked to the homology models to understand the similarities and differences between the binding modes. Molecular dynamic simulations of these HDAC-inhibitor complexes indicate that the interaction between the protein surface and inhibitor is playing an important role; also some active site residues show some flexibility, which is usually not included in routine docking protocols. The implications of these results for the design of isoform-selective HDAC inhibitors are discussed.
BibTeX
@article{wang2005toward,
  author = {Difei Wang and Paul Helquist and Norbert L. Wiech and Olaf Wiest},
  title = {{Toward Selective Histone Deacetylase Inhibitor Design: Homology Modeling, Docking Studies, and Molecular Dynamics Simulations of Human Class I Histone Deacetylases}},
  journal = {Journal of Medicinal Chemistry},
  year = {2005},
  doi = {10.1021/jm0505011},
}
Figure