Research Overview

Computational Protein Design / Protein-Protein Interactions / Protein Switches / Structural Biology
What is protein design?

Most ambitiously it is the creation of novel proteins to perform useful tasks. At a more modest level it might be identifying amino acid mutations that enhance protein stability, alter binding specificity, or increase solubility.

How do we design proteins?

We have developed a computer program that identifies low energy sequences for a target structure or interface. In essence, it is like solving a jigsaw puzzle.  The pieces, in this case amino acids, must fit together so that there are no overlaps and little empty space.  In addition, specific interactions such as hydrogen bonds must be satisfied.

What have we designed in the past?

In the past we have used our program to enhance protein stability, design a protein with a topology that has not been observed in nature, enhance protein-protein binding affinities, and design a protein conformational switch..

What would we like to design in the future?

Currently we are focusing on a variety of design goals including the creation of protein conformational switches, the rewiring of protein signal transduction pathways, and the design of protein biosensors for live cell imaging.

Current Projects

De novo design of protein-protein interfaces. Protein-protein interactions are essential to life. The ability to rationally design proteins that bind to target proteins would allow for the creation of new therapeutics, biosensors and tools for regulating cell biology. The de novo design of interfaces is an extremely challenging goal because it requires creating proteins that interact favorably with the polar amino acids typically found on protein surfaces. To meet this goal we are developing protocols that iterate between sequence design, backbone refinement and rigid body docking. We are currently investigating three methods of attack on this problem.  In one method, we are trying to seed a de novo interface by grafting residues from a binding partner of a target onto a new scaffold.  In another method, we are integrating metal ligand sides (for example tetrahedral Zn sites) whose sidechains are split between halves of the interface.  In a third method, we are attempting to use exposed surface beta-strands as a handhold, and bind a new protein to the target via starting with the exposed strand's hydrogen bonds.  

    1) Anchor-based interface design.

We propose a new method for designing novel protein reagents that combines advantages of redesign and de novo methods and allows for extensive backbone motion. This method requires a bound structure of a target and one of its natural binding partners. A key interaction in this interface, the anchor, is computationally grafted out of the partner and into a surface loop on the design scaffold. The design scaffold's surface is then redesigned with backbone flexibility to create a new binding partner for the target.

    2) Metal-mediated protein interface design.  
A) In designing new protein:protein interactions, it is a major challenge to achieve tight binding between the wild-type target and surface-designed scaffold.  To promote tight binding, a metal coordination site will be incorporated at the interface, as formation of these coordinating bonds is very energetically favorable.  In particular, zinc binding sites commonly feature four His/Cys residues arranged with tetrahedral geometry.  In what we call the 3-by-1 approach, three of the coordinating bonds are designed at the scaffold surface (green), and the target provides the fourth bond (blue) to complete the metal site.  Scaffold interface residues (gray sticks) are then designed to make favorable steric and electrostatic contacts with the target surface.  Heterodimeric interactions are tested using fluorescence polarization, isothermal titration calorimetry, NMR, and crystallography.

B) As another approach to metal-mediated interface design, we are designing a monomeric protein to form a symmetric zinc-mediated homodimer.  One of these designs forms a high-affinity homodimer (dissociation constant <30 nM), and crystallography was used to confirm that the binding orientation matches the computational model.  The model features two interface zinc sites coordinated by four histidines (left).  The crystal structure (right, cyan) shows that the binding orientation matches the computational model.


-strand mediated interface design.

Main chain hydrogen bonds formed between solvent exposed 

-strands form the basis of many protein-protein interactions and help drive complex formation. 

-strands are naturally prone to assemble and could help form the basis of novel protein-protein interactions.

This figure shows the conceptual interaction between exposed 

-strands of a 

protein of interest (target protein) and another protein that has been designed to bind to it (scaffold protein).

 This figure shows the result of a β
-strand mediated protein interface design. The computational model is shown in purple and green. The crystal structure of the designed interaction is shown in cyan. The interacting 
-strands between the two monomers help drive formation of a symmetric homodimer.

Protein design methods

De novo protein design, the design of proteins from scratch, is a rigorous test of our understanding of protein folding and protein stability. Conceptually, de novo protein design has three steps: creation of starting structures, amino acid sequence design, and experimental testing of de novo designed proteins. Helix bundle proteins are an interesting model system because of their abundance in nature and the amazing diversity of helix bundle folds.

Explicit consideration of hydrophobic patch size in RosettaDesign.
One strategy for promoting solubility is to disallow hydrophobic residues on the protein surface during design. However, naturally occurring proteins often have hydrophobic amino acids on their surface that contribute to protein stability via the partial burial of hydrophobic surface area or play a key role in the formation of protein-protein interactions. A less restrictive approach for surface design that is used by the modeling program Rosetta is to parameterize the energy function so that the number of hydrophobic amino acids designed on the protein surface is similar to what is observed in naturally occurring monomeric proteins.designed proteins.

Design of protein conformational switches. 

One amazing property of proteins is that they are often able to adopt multiple specific structures, each with its own functional importance. The relative stability of these states is typically regulated by ligand binding or post-translational modification. To design new proteins that switch conformation we have modified Rosetta so that it can search for a sequence that is simultaneously good for multiple target structures. We have used this algorithm do design a protein that can switch between a coiled-coil and a zinc finger. To add functionality to this switch we are now adding unique DNA binding properties to each of the two structures.

Redesign and de novo design of β -sheet proteins.

Despite the large number of all-β proteins in nature, the de novo design of a β -sheet protein has alluded protein designers. To redesign and de novo design a β -protein we are using a protocol that iterates between structure refinement and sequence optimization. The challenge with β -sheet design is creating proteins that do not aggregate. We are currently focused on adding negative design elements to our sequences that prevent aggregation.

Supercharging protein surfaces for reversibility of folding.

Supercharging has been shown to prevent aggregation of the unfolded state, making the original protein more robust. A Rosetta-based approach explicity considers protein stability and surface interactions when choosing mutations to increase net charge.  See our web-based server:

Peptide-protein interface design with non-natural amino acids. Peptide-protein interactions are of great therapeutic interest, because high-affinity peptides can modulate bioactivity when targeted toward binding sites normally occupied by other proteins or peptides. For many peptide-protein interfaces evolution has probably selected for sequences that are already close to optimal. To enhance peptide binding, therefore, it may be advantageous to consider amino acids that nature did not have at its disposal. Hydrophobic amino acids with novel geometries should help create new packing configurations while the inclusion of non-natural polar amino acids may help create more ideal hydrogen bonds. An additional feature of some non-natural amino acids is that they are more resistant to proteases, and peptides made from them may have better bioavailability.

Probing ubiquitination specificity with redesigned ubiquitin conjugation enzymes and ligases.
Tagging specific proteins with ubiquitin is one of the primary methods used by cells to target particular proteins for degradation, and therefore this modification is essential for many cellular processes including cell cycle control and cellular stress response. Ubiquitin is attached to proteins by a cascade of enzymatic reactions involving the E1 ubiquitin-activating enzyme, the E2 ubiquitin-conjugating enzymes, and the E3 ubiquitin ligases. The substrate specificity of the pathway is conveyed by the E3s of which there are many in the human genome. A key question in the field is which E3s target which proteins for ubiquitination. The goal of this project is to develop a new method for probing the specificity of E3s. The interface between the E3, E6AP, and its E2 partner, UbcH7, will be redesigned so that the new variants bind each other, but no longer have appreciable affinity for the wild type proteins. The substrate specificity of E6AP will then be probed by loading the redesigned E2 with labeled ubiquitin in vitro and then adding the loaded E2 to cellular extract containing the redesigned E6AP. Only proteins that are substrates for E6AP should be modified with the labeled ubiquitin.