12a Sunday, February 18, 2018 Symposium: Biophysical Mechanisms of 68-Symp An Alternative Strategy to Generate Binding Proteins Molecular Evolution Andreas Plueckthun. Biochemistry, University of Zurich, Zurich, Switzerland. 65-Symp The most frequently used binding proteins in research are monoclonal anti- Structural and Functional Constraints on Protein Evolution bodies, made by the >40-year-old hybridoma technology, some with question- Claus O. Wilke. able performance [1]. More recently, recombinant antibodies and non-antibody Department of Integrative Biology, The University of Texas at Austin, scaffolds, selected from synthetic libraries, have started to provide access to Austin, TX, USA. molecularly defined molecules [2]. Nonetheless, all of these approaches require Proteins evolve under constraints determined by their structural and functional one to treat every target as a completely new project. This is unavoidable for properties. These constraints are visible in particular at buried sites in proteins, folded proteins. However, we hypothesized that for unfolded proteins or sites in or near active sites in enzymes, and sites involved in protein-protein in- unfolded stretches (tags, posttranslationally modified tails, denatured proteins terfaces. All these sites experience increased purifying selection and evolve on western blots), termed ‘‘peptides’’ for simplicity, the regularity of the pep- more slowly than do other sites. Here, I will discuss the relative strengths of tide main chain can be exploited. If true, a modular detection system can be these effects, and I will show that even though most active sites of enzymes devised, which would ultimately allow one to generate a sequence-specific are located in the protein core, we can disentangle conservation due to solvent binding protein without experimentation. accessibility and conservation due to enzymatic activity. We also find that cat- The basis of our approach are Armadillo Repeat Proteins [3-9], which bind pep- alytic residues in enzymes exert a long-range effect, causing increased conser- tides in a completely extended way, providing a pocket for each side chain, and vation of residues throughout 80% of a typical enzyme structure. Finally, I will thus access to a modular approach. Combining evolutionary engineering, show that protein-protein interfaces show surprisingly little sequence conserva- NMR, X-ray crystallography and structure-based computation, we have now tion, and that interfaces can diverge substantially yet retain the ability to bind to achieved well crystallizing ArmRPs with bound peptides, picomolar affinities, ancestral binding partners. and a well functioning selection and evolution technology, as well as many biochemical and biophysical analysis technologies for the engineered ArmRPs. 66-Symp Progress in the various aspects will be summarized. Molecular Ensembles Shape Evolutionary Trajectories 1. Bradbury and Pluckthun€ (2015) and 110 co-signatories, Nature 518, 27. Michael J. Harms, Zachary R. Sailer, Lucas C. Wheeler. 2. Pluckthun€ (2015). Annu. Rev. Pharmacol. Toxicol. 55, 489. Institute of Molecular Biology, Department of Chemistry and Biochemistry, 3. Reichen et al., (2016). J. Mol. Biol. 428, 4467 University of Oregon, Eugene, OR, USA. 4. Hansen et al. (2016). J. Am. Chem. Soc. 138, 352 Evolutionary prediction is of deep practical and philosophical importance. One 5. Reichen et al., (2016). Acta Crystallogr. D72, 168. avenue for such a prediction would be to measure the fitness effects of all mu- 6. Reichen et al., (2014). Protein Science 23, 1572. tations to a protein, and then use this information to predict evolution. We at- 7. Reichen et al., (2014). J. Struct. Biol. 185, 147. tempted such a prediction for the simple case of evolving increased 8. Alfarano et al., (2012). Protein Science 21, 1298. thermodynamic stability in a simple protein lattice model. Surprisingly, we 9. Madhurantakam et al., (2012). Protein Science 21, 1015. were unable to predict evolutionary trajectories, even given complete knowl- edge of the effects of all mutations to the ancestral protein. This is a direct consequence of the ensemble of similar structures populated by proteins. The Symposium: DNA Supercoiling effect of a mutation depends on the relative probabilities of conformations in the ensemble, which in turn depend on the exact amino acid sequence of the 69-Symp protein. Accumulating substitutions alter the relative probabilities of conforma- Seeing Supercoiled DNA with Atomistic Simulation: A New Twist on a tions, thereby changing the effects of future mutations. This manifests itself as Familiar Structure subtle, but pervasive, multi-way interactions between mutations (high-order Sarah A. Harris1, Agnes Noy2, Thana Sutthibutpong3. epistasis). As mutations accumulate, uncertainty in the predicted effect of 1School of Physics and Astronomy, University of Leeds, Leeds, United each mutation accumulates and undermines prediction. We then developed Kingdom, 2School of Physics, University of York, York, United Kingdom, computational and statistical tools to look for evidence of this high-order epis- 3Theoretical and Computational Science Center, King Mongkut’s University tasis in experimentally determined genotype-fitness maps. We find that high- of Technology Thonburi, Bangkok, Thailand. order epistasis is ubiquitous. This work reveals that evolutionary unpredict- The discovery of the structure of duplex DNA revealed how cells store genetic ability arises—even for fixed environments and under strong natural selec- information. However, we are far from understanding the more complex biolog- tion—as a direct consequence of the thermodynamic ensemble populated by ical question of how this information is regulated and processed by the cell. DNA proteins. supercoiling is generated whenever a gene is transcribed, and complex cellular machinery, such as toposisomerases, are required to modulate the effect of this 67-Symp induced torsional stress. Supercoiling has been implicated in the packaging and Cellular Consequences of Systematic Perturbations of a Highly Conserved 3D arrangement of both prokaryotic and eukaryotic DNA, which in turn has Biological Switch fundamental consequences for transcription regulation and genome stability. Tanja Kortemme. In spite of the ubiquity of supercoiled DNA in cells, no experimental tool has University of California, San Francisco, San Francisco, CA, USA. been able to capture atomically detailed structural information. Small DNA cir- Cellular protein-protein interactions can be highly interconnected. Because of cles containing between 100 and 400 base pairs, however, offer a controllable this complexity, it is often difficult to extract quantitative information on how model system for the systematic exploration of the dependence of DNA structure each interaction contributes to distinct or overlapping cellular functions, and, on supercoiling through cryo-electron microscopy, atomic force microscopy, moreover, how changes to individual interactions result in altered function or and computer modelling. We use atomistic molecular dynamics simulations disease. We are developing an experimental platform for studying perturba- to explore the supercoiling-dependent conformation of small DNA circles. tions to multi-functional network ‘‘hub’’ proteins by combining high- We show that kinks and denaturation bubbles are generated in the DNA by throughput in vivo genetic interaction screening technology (Epistatic MiniAr- high torsional stress, that the compaction of the DNA is highly dependent on ray Profile (E-MAP)) with mass-spectrometry and biophysical assays. Our case salt, and that the DNA adopts writhed structures that are highly dynamic and study protein is the highly conserved multi-functional Gsp1/Ran GTPase which offer additional opportunities for DNA/protein interactions in 3D space. switch that controls key eukaryotic processes. The approach first engineers We then offer an atomistic interpretation for the growing experimental data that defined perturbations to Gsp1/Ran protein-protein interactions by amino acid shows the regulatory role proposed for supercoiling in the genome. point mutations (‘‘edge perturbations’’). The second step determines the func- tional effects of these perturbations at the cellular and organism level in the 70-Symp model S. cerevisiae. We find that E-MAPs have a resolution that enables us Protein-Mediated Loops in Supercoiled DNA Generate Large, Dynamic to identify quantitative functional differences in vivo between individual point Topological Domains mutations, even those between different amino acid substitutions of the same Laura Finzi. residue. Our analysis reveals several classes of observed phenotypes that could Physics, Emory University, Atlanta, GA, USA. be explained by the underlying biophysical perturbations of the on/off balance Proteins that act at a distance along DNA by binding at one site and contacting of the fundamental GTPase switch and considerable allosteric effects in the another create loops that form topological domains and influence regulation. system. Such loops are affected by the torsional state of DNA, which dramatically BPJ 8532_8543 Sunday, February 18, 2018 13a modulates topology, driving the DNA from extended and accessible, to more localization microscopy as a function of the measured photophysical parame- compact and genetically secured forms. Furthermore, the response of the ters of the probe such as photobleaching quantum yield, count rate per mole- DNA filament to supercoiling