(D-157) Computational Design of Specific Protein Binders with Varying Affinities for Interleukin-4

Introduction:: Immunomodulatory cytokines are small secreted proteins that influence the body’s responses to infection, physical trauma, and cancer. Dysregulation of cytokines can result in sepsis, pain, incomplete healing of injuries, or cancer¹. Interleukin 4 (IL-4) is an important cytokine that is responsible for multiple aspects of the immune response, including polarization of pro-inflammatory macrophages (M1) to anti-inflammatory macrophages (M2)². Sustained delivery of IL-4 can mediate dysregulated inflammatory responses in injury sites to alleviate symptoms and improve healing outcomes. To address this challenge, our lab is developing an affinity-based hydrogel delivery system, a hydrophilic polymer-based biomaterial functionalized with affinity ligands, to provide sustained release of IL-4. Our lab uses directed evolution to identify affibodies, which are protein binders ranging from 50-65 amino acids with a three-helical structure, as affinity ligands for our target proteins. We have identified two affibodies that bind to IL-4 via directed evolution: one that is high affinity and specific for IL-4 and one that is low affinity but lacks specificity. Directed evolution is a time-consuming and resource intensive process that is not always successful; however, computational protein design can generate a new protein sequence, structure, or model within a few minutes, target specific regions of a protein for binder design, and only requires computational resources³. Therefore, the goal of this work was to use computational protein design to obtain specific protein binders with varying affinity strengths for IL-4.

Materials and Methods:: Starting with one of our existing affibody sequences that binds to IL-4 with low affinity (dissociation constant, KD = 92 μM), I designed new affibody sequences using AlphaFold, ZDock, and Rosetta Commons that are expected to have stronger binding affinities to IL-4 than the original sequence. First, I used AlphaFold to predict the affibody structure. Next, I used ZDock to sample docking conformations of the affibody with an IL-4 X-ray diffraction structure from the Protein Data Bank (PDB: 1IAR Chain A). Using Rosetta score functions, a collection of energy terms for calculating atomic interactions in proteins such as physical forces and statistical terms, and visual inspection, I identified the most stable and probable (i.e., lowest scoring) affibody-IL-4 docking complex. Then, I used Rosetta FastDesign to generate new sequences by mutating amino acids at the affibody-IL-4 binding interface to create more stable binding interactions. Six mutated affibodies out of 500 displayed notably better binding conformations (i.e., lower energies) than the rest of the mutant affibodies. I used Rosetta Dock to score these affibody-IL-4 docking complexes and verify binding conformations. I selected the two lowest scoring mutant affibodies to express and purify. I designed double-stranded DNA fragments (gBlocks) for the two mutant affibodies modified with a C-terminal hexahistidine tag and terminal cysteine for purification and conjugation to the hydrogel. Then, I cloned the dsDNA fragments into a pET-28 b (+) expression plasmid, transformed the plasmid into BL21 E. coli, and sequenced the plasmid to verify sequence was successfully cloned and transformed.

Results, Conclusions, and Discussions:: In silico modeling of IL-4 with IL-4 receptor alpha indicated that binding is likely primarily driven by electrostatic interactions (i.e., hydrogen bonding and polar contacts). Compared to the existing low affinity affibody, I was able to increase the number of electrostatic interactions in the computationally designed affibodies. I have successfully expressed one of the two mutant affibodies in E. coli, and I am currently in the process of purifying this affibody using metal affinity chromatography. My next steps will include characterizing the dissociation constant (i.e., binding affinity strength) of the affibody-IL-4 interaction using biolayer interferometry (BLI) and secondary structure of the affibody using circular dichroism.

Recent massive developments in computational protein design software and protocols have allowed for more robust and efficient design of protein binders. In the future, I will implement these new improved software and protocols into my own work to design better protein binders for IL-4 and other target proteins.

Acknowledgements (Optional): :

References (Optional): :

1. Deckers, J., Anbergen, T., Hokke, A. M., de Dreu, A., Schrijver, D. P., de Bruin, K., ... & Mulder, W. J. (2023). Engineering cytokine therapeutics. Nature Reviews Bioengineering, 1-18.


2. Lin T, Kohno Y, Huang JF, Romero-Lopez M, Maruyama M, Ueno M, Pajarinen J, Nathan K, Yao Z, Yang F, Wu JY. Preconditioned or IL4-secreting mesenchymal stem cells enhanced osteogenesis at different stages. Tissue Engineering Part A. 2019 Aug 1;25(15-16):1096-103.


3. Kuhlman, B., & Bradley, P. (2019). Advances in protein structure prediction and design. Nature Reviews Molecular Cell Biology, 20(11), 681-697.