The Folding Pathway of Spectrin R17 from Experiment and Simulation: Using Experimentally Validated MD Simulations to Characterize States Hinted at by Experiment Kathryn A. Scott 1,2 , Lucy G. Randles 1 , Stephen J. Moran 1 Valerie Daggett 2 * and Jane Clarke 1 * 1 MRC Centre for Protein Engineering, University of Cambridge Chemical Laboratory, Lensfield Road Cambridge CB2 1EW, UK 2 Department of Medicinal Chemistry, University of Washington, Seattle, WA 98195-7610, USA We present an experimental and computational analysis of the folding pathway of the 17th domain of chicken brain a-spectrin, R17. Wild-type R17 folds in a two-state manner and the chevron plot (plot of the logarithm of the observed rate constant against concentration of urea) shows essentially linear folding and unfolding arms. A number of mutant proteins, however, show a change in slope of the unfolding arm at high concentration of denaturant, hinting at complexity in the folding landscape. Through a combination of mutational studies and high temperature molecular dynamics simulations we show that the folding of R17 can be described by a model with two sequential transition states separated by an intermediate species. The rate limiting transition state for folding in water has been characterized both through experimental F-value analysis and by simulation. In contrast, a detailed analysis of the transition state predicted to dominate under highly denaturing conditions is only possible by simulation. q 2006 Elsevier Ltd. All rights reserved. Keywords: protein folding; chevron; F-value; spectrin; molecular dynamics *Corresponding authors Introduction The combination of F-value analysis and mol- ecular dynamics (MD) simulation in understanding protein folding pathways has proved to be enormously successful. Properly validated simu- lations can both reveal intermediate species that are not observable in experiment and be used to aid the understanding of experimental results. Two such examples are described here; the forced unfolding of the beta-sandwich fibronectin type III (fnIII) domain and the thermal/denaturant unfolding of the three-helix bundle, cellular myleoblastosis protein c-Myb. FnIII domains occur in a number of proteins whose function involves resisting or responding to force, and the response of these domains to force has been studied by several different groups. 1–10 Both biased and constant force simulations of the forced unfolding of several fnIII domains have shown there to be intermediates on the forced unfolding pathway. 4,6,9,10 However, it remained unclear which transitions were dominant under experimental conditions. Using a combination of protein engineering and simulation Clarke and co-workers addressed this question for the third fnIII domain of tenascin. 1 Deletion mutations were made in each strand of the b-sheet to assess the importance of each region of the protein in resisting force. Two mutations were made in strand A, one, at position 2, towards the N terminus the other, at position 8, near the middle of the strand. The mutation at position 2 had no effect on the force required to unfold the protein, whilst the other mutation significantly lowered the unfolding force. From these results, it would usually be inferred that the mutation at position 2 was in a region that was fully structured in the transition state, whilst that at position 8 was in a region that was much less structured in the transition state. Given the structure of the protein this was not, however, a reasonable inference. The simplest explanation for the apparent discrepancy 0022-2836/$ - see front matter q 2006 Elsevier Ltd. All rights reserved. Present address: S. J. Moran, Department of Biochemistry, University of Oxford, South Parks Road, Oxford OX1 3QU, UK. Abbreviations used: MD, molecular dynamics; TS, transition state; H, helix. E-mail addresses of the corresponding authors: daggett@u.washington.edu; jc162@cam.ac.uk doi:10.1016/j.jmb.2006.03.011 J. Mol. Biol. (2006) 359, 159–173