A high-resolution protein structure solved by crystallography or nuclear magnetic resonance (NMR) spectroscopy can yield a wealth of information about that protein's biological function. These structures in most cases are static pictures of the native, or ground, state—the most stable state of the protein. However, proteins can also exist in higher-energy states, referred to as excited states, which can be crucial for carrying out biological functions.

Despite their importance, excited-state proteins pose a major experimental challenge. These states can be fleeting, lasting for only milliseconds. Only a very small percentage of proteins in a population exists in an excited state at any given time. Because many excited-state proteins cannot be isolated, solving their structures has been impossible using traditional methods. Lewis Kay of the University of Toronto sums up the conundrum: “How do you go about studying things that are essentially invisible to any of the experimental approaches that you have but that you think are going to be very important biologically?”

By putting more than a decade of methodological advances in NMR spectroscopy together, Kay's group recently solved the structure of an excited state of a small protein called the FF domain, corresponding to a folding intermediate (Korzhnev et al., 2010).

Though transient excited-state structures cannot be directly observed, they do leave their mark on an NMR spectrum. They cause chemical shifts in NMR spectra—which provide information about the local environment of an NMR-active nucleus—to broaden. This is usually frustrating for researchers wanting to obtain a native state structure of a protein. But for this project, broadened chemical shifts were exactly what Kay wanted to see because it meant that the protein population was undergoing exchange between the native and excited state. “We like to get good spectra—but not too good, so that we can actually exploit the broadening,” he explains.

To measure chemical shifts for the excited state of the FF domain, Kay's group applied NMR experiments known as relaxation dispersion. These experiments also yielded information about the lifetime and population of the folding intermediate, as well as another parameter called residual dipolar couplings, which provided information about the orientation of the backbone amide bond vectors. The researchers then fed these excited-state data into a powerful tool called CS-Rosetta, a program designed to solve small protein structures using only chemical shift data (Shen et al., 2008).

FF domain structures. The secondary and tertiary structures of the native state (N) differ from the structure of the folding intermediate (I), which is also more disordered. Reprinted with permission from the American Association for the Advancement of Science.

With this strategy, Kay's group solved the structure of the folding intermediate of the FF domain, which has a lifetime of one millisecond and represents 3% of the protein population. To validate the structure, they made a 'trapped' version of the excited-state protein by truncating a portion of the C terminus to destabilize the protein. The chemical shifts of the trapped excited-state version agreed with those from the 3% excited population, providing strong evidence that the intermediate state structure was correct.

Importantly, the excited-state FF domain structure provides experimental proof that metastable folding intermediates do exist, an issue that has been controversial because of the lack of experimental methods to simply detect intermediates, let alone solve their structures. Kay notes that the method could also be used to gain insights into many other interesting questions in biology, ranging from understanding protein misfolding, to molecular recognition, to enzyme catalysis and to how macromolecular machines function.

The future applications of this approach are no doubt 'exciting', but Kay stresses that these are still early days. “There will have to be development of NMR methodologies to keep pace with the biochemical questions that we want to answer,” he notes.