1. Introduction 2

2. The macroscopic and microscopic views of protein folding 2

2.1 The macroscopic view: the experimental folding free-energy profile 2

2.2 The microscopic view: an underlying energy landscape 3

3. The micro to macro projection: from an energy landscape to a free-energy profile 6

4. Global features of the protein folding transition-state ensemble 12

4.1 Overall transition state location β[Dagger]: a measure of compactness 12

4.2 What makes folding so robust ? 13

5. Structural characterization of the transition-state ensemble 16

5.1 Insights from ϕ-value analysis 16

6. Deviations from ideality 20

6.1 β[Dagger] shifts along seemingly robust trajectories 21

6.2 Anomalous ϕ values, frustration and inhomogeneities 25

7. Intermediates 28

8. Detours, traps and frustration 29

8.1 Premature collapse and non-native trapping 29

9. Diffusion on the energy landscape and the elementary events of protein folding 30

10. Malleability of folding routes: changes of the dominant collective coordinates for folding 33

11. The evolution of the shape of the energy landscape 35

11.1 Negative design: the hidden dimension of the folding code 35

12. Mechanistic multiplicity and evolutionary choice 36

13. Acknowledgements 37

14. References 38

We review what has been learned about the protein-folding problem from experimental kinetic studies. These studies reveal patterns of both great richness and surprising simplicity. The patterns can be interpreted in terms of proteins possessing an energy landscape which is largely, but not completely, funnel-like. Issues such as speed limitations of folding, the robustness of folding, the origin of barriers and cooperativity and the ensemble nature of transition states, intermediate and traps are assessed using the results from several experimental groups highlighting energy-landscape ideas as an interpretive framework.