Engineering of Reversible Rapamycin-Binding to FKBP and FRB
The FKBP-rapamycin-FRB protein complex is an important tool used to study spatiotemporal protein dynamics in a cell. This tertiary complex, composed of FKBP and the FRB domain of mTOR, forms only in the presence of the immunosuppressant drug rapamycin. FRB and FKBP have been used to create a rapid inducible protein dimerizing system, whereby proteins of interest can be brought together upon the addition of rapamycin. Because FRB and FKBP bind to the same ligand, it is possible to fuse FRB and FKBP to proteins of interest and induce the association of these two proteins of interest by the addition of rapamycin. Figure 1 below, adapted from Dong, et al. 2010, shows how using this protein dimerizer, it is possible to induce translocation of a protein of interest to a subcellular compartment.
One caveat of this system is that it is not reversible. Because FRB and FKBP associate with a high affinity, even if washed with rapamycin-free media, the complex does not fall apart on the time scale of live-cell imaging experiments. We took a computational approach to predict potential mutations that could improve the reversibility of binding by weakening the binding energy of the complex.
We assumed that the most important residues affecting binding affinity would be the 11 residues on FRB located within 4Å of rapamycin, which we identified using PDB entry 1NSG. We then simulated mutating these residues to all other possible amino acids, allowing repacking of the side chains of the other ten candidate residues.
To accomplish this task, we created a PyRosetta script, generate_all_resfiles.py, to generate resfiles for all 20 amino acid mutations at all 11 residues. We then created another PyRosetta script, calculate_dGmut_ddGbind.py, that uses PackRotamersTask to simulate these mutations, repack residues at the binding interface, and minimize energy through side-chain moves while keeping the backbones fixed. After each iteration, calculate_dGmut_ddGbind.py outputs the ΔΔG of binding and the ΔG of mutation for both bound and unbound forms. ΔΔG of binding gives an estimate of change in binding energy between the mutant complex and the wild-type complex, and ΔG of mutation tell whether the change in binding energy is due to a stabilization of the monomers or a destabilization of the complex. We then plotted these values and determined the most relevant mutants which had values of ΔΔG of binding between 1.4 and 4.1 kcal/mol (corresponding to 10-1000 fold decreases in binding affinity) and had modest values of ΔG of mutation.
We were able to identify a panel of mutations that showed a clear decrease in binding affinity without drastically affecting the protein stability shown below in Figure 2.
Mutations in the boxed region are those that fell within the aforementioned selection criteria. Starting with 209 different sequences, we were able to select 23 mutants that are strong candidates for our re-engineered system. In future iterations we hope to improve the predictive power of our algorithm by mutating other residues and examining double mutants. The next step would be to create these 23 mutants through site-directed mutagenesis and to test which ones show weakened binding affinity.
All of the files needed to run the simulations described in this blog post downloaded from: http://www.rosettadesigngroup.com/RDG_Corey_Award/2012_RDG_Corey_Winning_Entry_Materials
About the RDG Corey Award Winners
Brian Ross is a first year doctoral student at Johns Hopkins University studying biomedical engineering. He did his undergraduate in biological engineering at MIT, where he worked with Dr. K. Dane Wittrup in antibody-based cancer therapies. He is currently doing rotations at Johns Hopkins University, and has worked with Dr. Andre Levchenko studying the dynamics of calcium signaling and with Dr. Takanari Inoue studying protein-lipid interactions using the rapamycin-based inducible protein dimerization system. He is currently doing a lab rotation with Dr. Jin Zhang doing work in biosensor design to be able to detect the spatiotemporal dynamics of kinase activities within a cell.
Nick Trenton is a rising senior majoring in chemical/biomolecular engineering, and works in Dr. Denis Wirtz’s lab at Johns Hopkins University on cancer motility-related work. His projects focus on knocking down key motility proteins in different cancerous cell lines and using time-lapse imaging to analyze movement. He will pursue a doctorate after graduating, in bioengineering or biomedical engineering, and is broadly interested in cancer dynamics, computational biology, and molecular evolution.
The RDG Corey Award recognizes exceptional undergraduate and beginning graduate students that demonstrate creativity and boldness in the field of macromolecular structure prediction and design. The award is named in honor of Robert Corey, who, together with Linus Pauling at the California Institute of Technology, predicted the secondary structures of proteins and hydrogen bonding patterns of DNA base pairs. An achievement, which at the time, was comparable to the complex macromolecular structure prediction and design problems faced today. For Trivial Pursuit aficionados, CPK is short for Corey, Pauling, Koltun. The award has a prize of $600 USD, and the recipients get to showcase their work on the Macromolecular Modeling Blog™.
The 2012 RDG Corey Award Competition was held at Johns Hopkins University in conjunction with Dr. Jeffrey Gray’s “ChemBE 416/616: Current Topics in Protein Structure Prediction” course . A special thanks to Dr. Gray and all those who participated!
If you would like to add your course to the competition, have general comments or comments on the current award, or suggestions for future competitions, please contact us at email@example.com
References and further reading
- Characterization of the FKBP.rapamycin.FRB ternary complex. Banaszynski LA, Liu CW, Wandless TJ. J Am Chem Soc. 2005 Apr 6;127(13):4715-21. Erratum in: J Am Chem Soc. 2006 Dec 13;128(49):15928.
- Structure of the FKBP12-rapamycin complex interacting with the binding domain of human FRAP. Choi J, Chen J, Schreiber SL, Clardy J. Science. 1996 Jul 12;273(5272):239-42.
- PI(3,5)P(2) controls membrane trafficking by direct activation of mucolipin Ca(2+) release channels in the endolysosome. Dong XP, Shen D, Wang X, Dawson T, Li X, Zhang Q, Cheng X, Zhang Y, Weisman LS, Delling M, Xu H. Nat Commun. 2010 Jul 13;1:38. doi: 10.1038/ncomms1037.
- Organelle-specific, rapid induction of molecular activities and membrane tethering. Komatsu T, Kukelyansky I, McCaffery JM, Ueno T, Varela LC, Inoue T. Nat Methods. 2010 Mar;7(3):206-8. Epub 2010 Feb 14.
- Controlled dimerization of ErbB receptors provides evidence for differential signaling by homo- and heterodimers. Muthuswamy SK, Gilman M, Brugge JS. Mol Cell Biol. 1999 Oct;19(10):6845-57.
- A ligand-reversible dimerization system for controlling protein-protein interactions. Rollins CT, Rivera VM, Woolfson DN, Keenan T, Hatada M, Adams SE, Andrade LJ, Yaeger D, van Schravendijk MR, Holt DA, Gilman M, Clackson T. Proc Natl Acad Sci U S A. 2000 Jun 20;97(13):7096-101.
Enjoyed this Post ?