Restricted state selection in fluorescent protein Förster resonance energy transfer.

The measurement of donor lifetime modification by Förster resonance energy transfer (FRET) is a widely used tool for detecting protein-protein interactions and protein conformation change. Such measurements can be compromised by the presence of a significant noninteracting fraction of molecules. Combining time-resolved intensity and anisotropy measurements gives access to both molecular distance and orientation. Fluorescent proteins frequently used to detect energy transfer in biological systems often exhibit decay characteristics indicative of more than one excited state. However, little attention has thus far been given to the specific modes of energy transfer, in particular, which states are predominantly coupled. Here, we use a previously characterized dimerization system to study energy transfer between EGFP and mCherry. Optically excited EGFP and mCherry both exhibit biexponential decays, and FRET should therefore involve dipole-dipole transfer between these four states. Analysis of the sensitized fluorescence anisotropy and intensity decays indicates that FRET transfer is predominantly from the shorter lived EGFP emitting state (2.43 ns) to the longer lived (ca. 2.77 ns) minority component (ca. 16%) of the optically excited mCherry emission. This high degree of state selection between these two widely used FRET pairs highlights the fundamental differences that can arise between direct optical excitation of an isotropic molecular population and dipole-dipole coupling in a far from isotropic interaction geometry and has consequences regarding the accurate interpretation of fluorescent protein FRET data.

Regulation of 3-phosphoinositide-dependent protein kinase 1 activity by homodimerization in live cells.

3-Phosphoinositide-dependent kinase 1 (PDK1) plays a central role in regulating the activity of protein kinases that are essential for signaling; however, how PDK1 itself is regulated is largely unknown. We found that homodimerization of PDK1 is a spatially and temporally regulated mechanism for controlling PDK1 activity. We used Förster resonance energy transfer monitored by fluorescence lifetime imaging microscopy to observe PDK1 homodimerization in live cells. A pleckstrin homology (PH) domain-dependent, basal dimeric association of PDK1 was increased upon cell stimulation with growth factors; this association was prevented by a phosphatidylinositol 3-kinase inhibitor and by a mutation in, or a complete deletion of, the PH domain of PDK1. The distinct spatial distribution of PDK1 homodimers relative to that of heterodimers of PDK1 and protein kinase B (PKB), and the ability of monomeric mutants of PDK1 to phosphorylate PKB, suggested that the monomer was the active conformation. Mutation of the autophosphorylation residue threonine-513 to glutamate, which was predicted to destabilize the homodimer interface, enhanced the interaction between PDK1 and PKB and the activity of PKB. Through in vitro, time-resolved fluorescence intensity and anisotropy measurements, combined with existing crystal structures and computational molecular modeling, we determined the geometrical arrangement of the PDK1 homodimer. With this approach, we calculated the size of the population of PDK1 dimers in cells. This description of a previously uncharacterized regulatory mechanism for the activation of PDK1 offers possibilities for controlling PDK1 activity therapeutically.

Multiple conformations of full-length p53 detected with single-molecule fluorescence resonance energy transfer.

The tumor suppressor p53 is a member of the emerging class of proteins that have both folded and intrinsically disordered domains, which are a challenge to structural biology. Its N-terminal domain (NTD) is linked to a folded core domain, which has a disordered link to the folded tetramerization domain, which is followed by a disordered C-terminal domain. The quaternary structure of human p53 has been solved by a combination of NMR spectroscopy, electron microscopy, and small-angle X-ray scattering (SAXS), and the NTD ensemble structure has been solved by NMR and SAXS. The murine p53 is reported to have a different quaternary structure, with the N and C termini interacting. Here, we used single-molecule FRET (SM-FRET) and ensemble FRET to investigate the conformational dynamics of the NTD of p53 in isolation and in the context of tetrameric full-length p53 (flp53). Our results showed that the isolated NTD was extended in solution with a strong preference for residues 66-86 forming a polyproline II conformation. The NTD associated weakly with the DNA binding domain of p53, but not the C termini. We detected multiple conformations in flp53 that were likely to result from the interactions of NTD with the DNA binding domain of each monomeric p53. Overall, the SM-FRET results, in addition to corroborating the previous ensemble findings, enabled the identification of the existence of multiple conformations of p53, which are often averaged and neglected in conventional ensemble techniques. Our study exemplifies the usefulness of SM-FRET in exploring the dynamic landscape of multimeric proteins that contain regions of unstructured domains.

Probing nanosecond motions of plasminogen activator inhibitor-1 by time-resolved fluorescence anisotropy.

Plasminogen activator inhibitor-1 (PAI-1) is a member of the serpin (serine protease inhibitor) superfamily. Like most serpins, the inhibitory function of PAI-1 relies on a flexible reactive centre loop (RCL) undertaking a striking conformational transition. We have investigated the conformational dynamics of the RCL of PAI-1 by time-resolved fluorescence anisotropy. A heterogeneous population model with three rotational correlation times has been employed to account for the "dip and rise" observed in fluorescence anisotropy decay curves. The RCL becomes almost fully solvent exposed and exhibits faster rotation when PAI-1 interacts with a RCL-mimicking octapeptide which blocks the loop insertion pathway, indicating that the RCL is well displaced from the protein surface; while the binding of Somatomedin B (SMB) domain of vitronectin, only induces small changes in the RCL. Comparison of the fluorescence lifetime and anisotropy decay of the wild-type PAI-1 with that of the stabilised mutant suggests that there would be no major structural differences between them. Our results indicate that in a native serpin, the P14 residue of the hinge region can flip in and out of the central beta-sheet A more readily than previously thought, which is likely an inherent property for serpins' protease inhibitory function.