Primary Photophysical Processes in Chromoproteins - Fluorescent Proteins: From Understanding to Design - page 51

about three possible coordinates for three different charge states [ 78 ]. They found
that only twisting about the imidazolinone double bond was barrierless for both
neutral and ionic states (consistent with experiment), but it did not lead to a crossing
of ground and excited states for the anion. They also considered the so-called hula
twist mechanism. This is a volume conserving route to isomerisation, and thus
consistent with experiment. However, it was found that this coordinate only leads to
an S 0 /S 1 crossing in the HBDI anion via a significant energy barrier.
More recent high-level quantum chemical calculations are consistent with a major
contribution to the coordinate promoting IC arising from zero or low barrier single
bond rotation, possibly coupled with a degree of pyramidalisation at the bridging
carbon atom [ 79 - 81 ]. Pathways involving the volume conserving 'hula twist' motion
are calculated to lead to fast IC but again only via an energy barrier [ 78 , 80 ]. There are
conflicting conclusions as to which single bond rotation is dominant in the excited
state isomerisation. Olivucci and co-workers using solvent-free conditions found that
two coordinates are important in achieving close approach of S 0 and S 1 in anionic
HBDI - a fast stretching coordinate, corresponding to reduced bond order, and a
slower rotation about the phenolic single bond [ 80 ]. This interpretation is consistent
with the model proposed by Huppert and co-workers for the temperature-dependent
HBDI fluorescence [ 59 ]. Gas phase calculations have recently been extended to
neutral and cationic states, and in those case significant barriers are found to isomer-
isation [ 82 ]. This is inconsistent with experiment, where neutral HBDI has a shorter
lifetime than the anion. This result suggests that details of the reactive potential
surfaces may be modified in the condensed phase.
Altoe et al. included a polarisable continuum model of the solvent in their
calculations for the HBDI anion and found that rotation about the imidazolinone
double bond was the most significant in promoting IC, in contrast to the gas phase
calculation [ 79 ]. Martinez and co-workers also identified an important role for the
medium in their study of a molecule similar to HBDI [ 81 , 83 ]. In vacuum, the
excited state dynamics primarily involved twisting about the bridging double bond
again accompanied by a large excursion in the phenyl torsion. These calculations,
however, predicted a significant lifetime on the excited state surface, which is not
consistent with recent gas phase measurements [ 55 ]. However, in the calculation for
a water solvated chromophore, fast barrierless rotation about the double bond
was predicted, leading to an S 1 /S 0 conical intersection, and a sub-picosecond
excited state lifetime, as found experimentally. These results therefore point to an
important role for the medium in determining the coordinate leading to IC. This is in
itself an important conclusion, as it suggests the means by which the protein can
modulate the chromophore's photophysics. They also suggest that rotation
about either or both bridging bonds may be important, with the mechanism being
dependent on the environment. Significantly, Olsen [ 84 , 85 ] has shown in cal-
culations comparing HBDI and a model of the red-emitting CP chromophore
(i.e., HBDI with an N -acylimine substitution) that a switch in the nature of the
radiationless decay (and presumably isomerisation) coordinate from double bond
to phenoxy bond rotation occurs on substitution. Recently, it was observed experi-
mentally that the excited state decay of HBDI is somewhat dependent on the nature