Printer-friendly versionPrinter-friendly version

Stochasticity Open Seminar June 27, 2017, 12:00

Open Seminars of the Research Group on


Stochasticity and Control in the Dynamics and Diversity of Immune Repertories: an Example of Multi-Cellular Co-Operation


Tuesday, June 27, 2017, 12:00

at the Israel Institute for Advanced Studies, Room 128




Learning from antibodies how to design new protein functions


Sarel Fleishman (Weizmann)


Computational protein design holds the promise of programming a new generation of enzymes and therapeutics with desired qualities such as affinity, specificity, and stability. Protein design studies have so far mostly focused on the fundamental question of how to design new protein folds using atomistic design simulations . Success in designing new molecular functions, however, has been much more limited. To address this gap, we have looked to immune-system antibodies, which are the most versatile class of binding molecules in nature, and inferred general principles for designing stable and specific binders. Our design strategy is unique in combining evolutionary principles with atomistic modeling; it uses commonly observed backbone conformations and sequence patterns in order to design new antibodies . The resulting antibodies, though very different from any mammalian germline, show the same desirable features, such as high stability, affinity, and specificity. Furthermore, experimental structures show that the design models are atomically accurate, suggesting that future antibodies may be designed completely by computer. We moreover show that the design principles we inferred are general, and can be applied to substantially improve the stability and activity of challenging eukaryotic proteins, including the leading malaria vaccine candidate, the Plasmodium falciparum RH5 protein . In unpublished results, we have also designed new enzymes and novel high-dimensionality specificity networks. We therefore conclude that the synergy between evolutionary principles and atomistic design calculations can resolve some of the most recalcitrant problems in protein engineering and design.



Related Research Questions


  1. What is the molecular architecture of a protein active site (in enzymes or binders), and what differentiates it from other parts of the protein?
  2. Why is it so much more difficult to design active sites than other parts of the protein?
  3. Why are natural proteins in mesophiles often only marginally stable, whereas thermophiles exhibit homologous and functionally equivalent proteins that are hyperstable?


Suggested Reading


Role of the biomolecular energy gap in protein design, structure, and evolution

Fleishman, S.J. & Baker, D. (2012).

Cell 149, 262–273


Why reinvent the wheel? Building new proteins based on ready-made parts

Khersonsky, O. & Fleishman, S.J. (2016).

Protein Sci. 25, 1179–1187


Automated Structure-and Sequence-Based Design of Proteins for High Bacterial Expression and Stability

Goldenzweig, A., Goldsmith, M., Hill, S.E., Gertman, O., Laurino, P., Ashani, Y., Dym, O., Unger, T., Albeck, S., Prilusky, J., Lieberman, R.L., Aharoni, A., Silman, I., Sussman, J.L., Tawfik, D.S. & Fleishman, S.J. (2016).

Mol. Cell 63​, 337–346 



> back to seminars