De novo design of a homo-trimeric amantadine-binding protein

Jooyoung Park, Brinda Selvaraj, Andrew C. McShan, Scott E. Boyken, Kathy Y. Wei, Gustav Oberdorfer, William F. DeGrado, Nikolaos G. Sgourakis, M.j. Cuneo, Dean A. A. Myles, David BakereLife 2019 · DOI

TL;DR

Researchers have successfully designed a novel homo-trimeric protein (ABP) from scratch that specifically binds the small molecule amantadine. This work represents a significant step forward in designing protein assemblies that interact with non-metal small molecules, paving the way for new applications in areas like cellular engineering.

What Problem Does This Paper Address?

Designing proteins that can bind to specific small molecules is a crucial area of research with applications ranging from drug development to biosensing. However, a major challenge has been the de novo design of protein assemblies (oligomers) that bind to symmetry-matched small molecules. Specifically, the authors aimed to create a protein trimer where each of the three identical protein subunits makes the exact same interactions with a C3-symmetric small molecule. This approach offers a more efficient design strategy by reusing the same binding interface multiple times. Furthermore, achieving such ligand-induced trimerization is a stepping stone towards creating controllable molecular systems for cellular engineering, which are currently lacking for trimeric assemblies.

Methods

The core of this work relies on computational protein design using the Rosetta software suite. The researchers started with a pre-existing, parametrically generated C3-symmetric helical bundle scaffold. They then employed RosettaDesign to optimize the amino acid sequences and conformations within a 12.5 Å radius of the amantadine molecule, aiming for high-affinity binding. Crucially, the design process enforced C3 symmetry, ensuring that each monomer in the trimer would be identical and interact with the amantadine in the same way.

To validate their computational design, the researchers utilized a multi-pronged experimental approach:

  • SEC-MALS: To confirm the oligomeric state of the expressed protein.
  • Thermofluor assay: To assess protein stability and the effect of amantadine binding on its thermal stability.
  • Circular Dichroism (CD) spectroscopy: To probe the secondary structure and its stability.
  • Solution NMR (Nuclear Magnetic Resonance): Specifically, 2D ¹H-¹⁵N SOFAST-HMQC and ¹H-¹³C SOFAST-HMQC experiments were used to study the protein's structure and dynamics in solution, and to characterize the binding interaction with amantadine through titration experiments. ¹³C-methyl labeling was employed to specifically observe key residues.
  • X-ray crystallography: To obtain a high-resolution static structure of the ABP-amantadine complex.
  • Neutron diffraction: This technique was used to directly visualize the hydrogen-bond networks, providing experimental validation of the designed interactions, particularly those involving water molecules.

Datasets

The paper does not describe the use of large datasets for training or evaluation in the traditional machine learning sense. Instead, the primary "dataset" is the computational model generated by RosettaDesign. Experimental data were collected from purified ABP protein and amantadine. The structural data (X-ray and neutron diffraction) were deposited in the RCSB Protein Data Bank under accession codes 6N9H and 6NAF.

Key Insights and Ideas

  • Symmetry-guided design is powerful: The successful design of a symmetric protein that binds a symmetric small molecule demonstrates the efficacy of enforcing symmetry constraints in computational protein design.
  • Water molecules are critical for binding: While the initial design did not explicitly include water molecules, the crystal structure revealed their crucial role in mediating hydrogen bonds between the protein and amantadine. This highlights the need to incorporate solvent effects more explicitly in future design efforts.
  • Neutron diffraction validates hydrogen bonds: The use of neutron diffraction provided direct experimental evidence for the designed hydrogen-bond networks, a powerful validation for computational models.
  • Designing inducible systems is challenging: The authors attempted to create amantadine-inducible trimerization by destabilizing the monomer, but this proved difficult. The inherent stability of the trimeric interface and the small size of amantadine were identified as key limitations.

Key Results

  • A homo-trimeric protein (ABP) was successfully designed de novo using RosettaDesign.
  • ABP exhibits regular three-fold symmetry in solution, as confirmed by NMR.
  • ABP binds amantadine with low micromolar affinity (estimated KD < 55 µM).
  • The X-ray crystal structure of ABP in complex with amantadine closely matches the computational design, with an RMSD of 0.63 Å.
  • The neutron structure confirmed the presence of designed hydrogen-bond networks and revealed the role of ordered water molecules in the binding site.
  • Attempts to create amantadine-inducible trimerization were unsuccessful, with designed variants either being poorly expressed, losing amantadine binding, or remaining constitutive trimers.

Limitations and Future Work

The authors acknowledge several limitations and suggest future directions:

  • Inducible Trimerization: The primary limitation is the difficulty in designing a system where amantadine binding induces trimerization. This is attributed to:
  • Amantadine's small size not providing a strong enough driving force for assembly.
  • The trimeric interface having significant buried non-polar surface area, making monomers poorly behaved and unstable in the absence of the ligand.
  • Water Molecules in Design: The design did not explicitly account for water molecules, which were found to be critical for binding. Future designs should incorporate solvent effects more rigorously.
  • Future Directions: To overcome these limitations, the authors suggest:
  • Designing systems with smaller subunit interfaces and higher affinity binding sites.
  • Potentially using larger C3 symmetric molecules as ligands to provide a stronger driving force for assembly.
  • Further exploration of computational strategies to design inducible protein assembly systems.

Summary updated

Abstract

The computational design of a symmetric protein homo-oligomer that binds a symmetry-matched small molecule larger than a metal ion has not yet been achieved. We used de novo protein design to create a homo-trimeric protein that binds the C3 symmetric small molecule drug amantadine with each protein monomer making identical interactions with each face of the small molecule. Solution NMR data show that the protein has regular three-fold symmetry and undergoes localized structural changes upon ligand binding. A high-resolution X-ray structure reveals a close overall match to the design model with the exception of water molecules in the amantadine binding site not included in the Rosetta design calculations, and a neutron structure provides experimental validation of the computationally designed hydrogen-bond networks. Exploration of approaches to generate a small molecule inducible homo-trimerization system based on the design highlight challenges that must be overcome to computationally design such systems.
BibTeX
@inproceedings{park2019de,
  author = {Jooyoung Park and Brinda Selvaraj and Andrew C. McShan and Scott E. Boyken and Kathy Y. Wei and Gustav Oberdorfer and William F. DeGrado and Nikolaos G. Sgourakis and M.J. Cuneo and Dean A. A. Myles and David Baker},
  title = {{De novo design of a homo-trimeric amantadine-binding protein}},
  booktitle = {eLife},
  year = {2019},
  doi = {10.7554/elife.47839},
}
Figure