FRET

Following the excitation of a fluorophore, the transition dipole of the excited electron can exert a force on a nearby dipole as the molecule returns to its ground electronic state. Where two fluorophores are very close together, the resulting interaction (coupling) of the two dipoles (one in each fluorophore) can cause the concerted motion (resonance) of the dipoles resulting in the non-radiative transfer of energy from one fluorophore (the donor) to the other (the acceptor). This is known as Förster Resonance Energy Transfer (FRET).

The relative rates of FRET and normal emission from a fluorophore determine what is known as the FRET efficiency, EFRET, the probability of FRET occurring after a photon has been absorbed by the donor fluorophore. The rate of FRET depends on the relative orientation of the

fluorescent centres, the distance between them and the extent to which the emission spectrum of the donor and the absorption spectrum of the acceptor overlap91. FRET is extremely sensitive to small increases in separation distance because the rate of this energy transfer is inversely proportional to the sixth power of the distance between the donor and acceptor. This occurs because the dipole-dipole interaction energy is inversely proportional to the distance cubed and the probability of energy transfer is proportional to the square of this interaction energy92.

46 𝐸_{𝐹𝑅𝐸𝑇} = 1

1 + (_𝑅𝑟

0)

6 _1.6

Where r is the distance between the donor and acceptor and R0 is the distance of 50% FRET efficiency, the Förster distance. In the first part of the FRET work in this project, enhanced green fluorescent protein (eGFP) was used as the donor and enhanced yellow fluorescent protein (eYFP) was used as the acceptor. The Förster distance, the distance at which there is 50% transfer efficiency, is approximately 5.6 nm for this pair93.

𝑅0= √ 9(ln 10)𝜅2_𝑄 𝐷𝐽 128𝑁_𝐴𝜋5_𝑛4 6 1.7

Where QD is the quantum yield of the donor (the proportion of absorbed light reemitted), NA is Avogadro’s number (6.022  1023 mol-1) and κ is the dipole orientation factor, a variable describing the relative orientation of the donor and acceptor that ranges from 0 (dipoles

antiparallel) to 2 (dipoles parallel)94(p13). For a donor and acceptor tumbling rapidly with respect to the time period between absorption and reemission by the donor, the time-averaged

orientation factor is two thirds. J is the spectral overlap.

𝐽 = ∫ 𝜆4_𝐹

𝐷(𝜆)𝜀𝐴(𝜆)d𝜆 ∞

0

1.8 Where FD is the normalised fluorescence of the donor at the wavelength λ. The overlap is scaled by the wavelength to the power 4 because the efficiency of transfer is proportional to λ4. εA is the extinction coefficient of the fluorophore solution. An extinction coefficient describes the rate at which the intensity of light entering a medium is exponentially diminished.

− log₁₀ 𝐼

𝐼0= 𝐴 = 𝜀𝑐𝑙 1.9

Where 𝐼 is the intensity of light a distance 𝑙 into a solution of concentration c, extinction coefficient ε and therefore absorbance A if it enters at an intensity of 𝐼₀.

Unlike the other methods outlined in section 1.3.2.1 above, FRET can be used to carry out in vivo assays since if both the binding components are proteins with a fluorescent protein label, they can be synthesised in a cell. FRET is also very amenable to high-throughput screening. Different compounds can be tested in the wells of microwell plates, which can in theory be pipetted out and tested robotically95. Also, it is possible to perform competition assays in which a ligand competes with a labelled species (p53 for example), to bind to the second labelled molecule (Mdm2 for example). These have the advantage that the ligands do not have to be labelled, a process which might affect the binding capacity of each compound differently. In this project, the hybrid proteins p53-YFP and p53-Cherry were used to investigate the binding of p53 to GFP-Mdm2. Figure 1.8 shows diagrammatically how fluorescent labelling of

47 Mdm2 and p53 can result in FRET. On binding, the YFP labelled C-terminal end of the p53 is

brought into close proximity with the GFP, bringing about FRET.

A B

Figure 1.8: Diagrams showing how fluorescent labelling of Mdm2 and p53 can result in FRET. A) A computer generated structure of Mdm2 produced using PDB structure 1T4F showing the C- terminal end of p53 (dark grey) in red and the N (blue) and C (red) termini of Mdm2. The N-terminal end of Mdm2 was fluorescently labelled because it is closer to p53 than the C-terminal end. B) A diagram showing the principle of the FRET experiment. Binding of p53 and Mdm2 brings the GFP and YFP labels together causing FRET to occur.

Many assumptions are made when carrying out FRET measurements.

In this project, it was assumed that the YFP, GFP and Cherry labels had no effect on the binding of the peptide; however, GFP, YFP and Cherry are all considerably larger than the peptide and could get in the way of binding, reducing the binding affinity. For more details regarding these fluorescent labels see p49.

In a titration monitored using FRET, it is usually presumed that there are two FRET states, one in which the fluorophores are close and there is considerable FRET, and another in which the fluorophores are separated, resulting in little or no FRET. As the titration proceeds, the emission spectrum is assumed to transform from that one of these states to that of the other and between these points, the amount of FRET is assumed to be directly proportional to the number of donor- acceptor complexes remaining. In reality, there could be intermediate states with their own emission spectra.

A further assumption is that the FRET donor and acceptor come together as predicted, which might not be the case. For example, rather than binding with specificity, molecules might become denatured, exposing hydrophobic surfaces, leading to non-specific aggregation.

48 In addition, the conformational changes occurring during FRET experiments may differ from

those expected. For example, differences in FRET could reflect a variation in the relative orientation of the FRET donor and acceptor rather than a change in the distance between the fluorophores. Kon et al.96 describe their use of FRET to monitor conformational change within cytoplasmic dynein, specifically the distance between the tail (labelled with GFP) and the AAA domains of the head (labelled with BFP). They discuss the possibility that some of the FRET change observed might be due to conformational changes within the tail and not a change in the distance between the fluorophores. Furthermore, it is unclear which AAA domain the GFP fluorophore is closest to when the fluorophores are brought together.

In a titration, an assumption is made that the intensity of the fluorescence being measured is unaffected by factors such as the absorbance or direct fluorescence of the titrant, the

concentration of which may be very high by the end of the experiment. Adding large quantities of a ligand and its solvent, for example DMSO in the case of a low solubility compound, can change the properties of the solution being observed. The resulting changes in pH or

hydrophobicity could affect the fluorescence. For example, the presence of DMSO could cause partial unfolding of a protein, increasing the solvent exposure of an attached fluorophore. The volume of solution excited by the incident light and from which fluorescence is detected also has a direct effect on the observed fluorescence. In a spectrofluorimeter cuvette, this volume is unlikely to vary. However, in a plate reader well, the volume excited by the incident light depends on the shape of the meniscus, which is affected by factors such as DMSO concentration.

Kon et al.96 used FRET to investigate the effect of excess ATP on the distance between domains of dynein. To check that ATP did not alter the fluorescence of either fluorophore in isolation, they carried out controls in which ATP was added to each polypeptide-fluorophore complex in the absence of the other.