Primary Photophysical Processes in Chromoproteins - Fluorescent Proteins: From Understanding to Design - page 49

quantum yield. When the HBDI absorption spectra are measured in a range of
solvents, no single solvent function (e.g., various polarity functions, , E T , etc. [ 30 ])

adequately fits all the data. However, in a detailed study, Dong et al. were able to
obtain a good fit to the spectral shift data in a range of solvents using a combination
of solvent acidity, basicity and polarity functions [ 57 ].
Recently, an attempt to reproduce the range of electronic transition energies
found in CPs by tuning the electronic structure of HBDI with different substituents
was reported [ 58 ]. Even very weak electron-donating substituents such as the
methyl group were found to cause a dramatic shift in the spectrum of the HBDI
anion when placed at the 3,5 positions on the phenyl ring (i.e., ortho to the OH
group). The shifts for the neutral form are much smaller. The red-shift of the anion
spectrum for the dimethyl derivative matches the protein shift between HBDI and
avGFP even in the moderately polar solvent ethanol. Replacement of methyl with
the somewhat more strongly electron-donating t -butyl substituent shifts the anion
transition even further to the red, matching the transition energy seen in YFPs.
These results suggest that even quite modest perturbations of the electronic struc-
ture of the chromophore are sufficient to cause large spectral shifts. The origin of
this effect seems likely to lie in the charge transfer character of the transition, but
confirmation will require quantum chemical calculations on these derivatives.

p

3.2 HBDI Photodynamics

The transient behaviour of HBDI following electronic excitation has been investi-
gated in detail [ 59 - 65 ]. Single colour pump-probe polarisation spectroscopy and
broadband transient absorption experiments showed that ground state repopulation
occurs on an ultrafast timescale. This is consistent with the very weak fluorescence
and suggests that the radiationless process is IC. The excited state fluorescence
decay time has been measured in a range of solvents at room temperature. The
lifetime of the neutral form is sub-picosecond and is slightly lengthened in non-
polar solvents. The anionic form of HBDI has a slightly but consistently longer
decay time than the neutral. In aqueous solution, both neutral and anionic forms
have faster decay times than in other polar solvent, suggesting an enhanced
quenching [ 62 ]. However, the solvent and charge effects are slight; the HBDI
decay time is never longer than a few picoseconds, and it can be concluded that
in all fluid solvents IC dominates the excited state decay. The observed excited state
decay times are on the same order or slightly faster than the ground state recovery
times, suggesting that a short-lived dark intermediate (perhaps simply a vibration-
ally hot ground state) may be involved in the relaxation pathway.
To provide information on the coordinate promoting IC, the effect of solvent
viscosity on the decay rate has been investigated [ 61 , 63 , 65 ]. One plausible
mechanism for radiationless decay in HBDI is excited state isomerisation. For a
number of related molecules in solution, such behaviour is well characterised, for
example stilbenes, azobenzenes and cyanine dyes

(which are structurally