Photoconversion of the Green Fluorescent Protein and Related Proteins - Fluorescent Proteins: From Understanding to Design - page 186

400 nm would indicate the presence of a long-lived intermediate, possibly a triplet
state, which has so far not been reported.
One study carefully characterised the photoconversion of a T203V mutant of
GFP that lacks the GFP B anionic species, with femtosecond excitation at 400 and
800 nm (via two photon absorption) [ 39 ]. In the range of 0-5

10 9 Wcm 2 and

10 11 Wcm 2 for 400 and 800 nm 80 fs pulses, the power dependency for
photoconversion was found to have reaction orders with exponents of 2.06 and
2.76, respectively. This quantitative information obtained in the high-power regime
indicated that also with blue excitation a two photon process is more efficient than a
one photon process. Given the large difference in quantum yield for continuous UV
and blue illumination, this result at 400 nm appears to be in agreement with these
observations. An alternate suggestion that for continuous illumination, in the low
power regime, the wavelength dependence of photoconversion reflects two photon
excitation of the chromophore [ 37 ] appears to be inconsistent with the linear power
dependence at nanosecond pulsed excitation at 390 nm [ 19 ] and the observation of
photoconversion with CW illumination at 413 nm [ 35 ]. A higher lying excited state,
probably the S2 state, is considerably more efficient in triggering photoconversion
than the S1 state [ 19 ]. The steady-state fluorescence anisotropy measurements
identified the higher lying optical transition of the chromophore within the excita-
tion envelope in the UV region and explain the high efficiency of photoconversion
with 254 nm excitation via direct excitation of the chromophore [ 19 ]. The reaction
order of 2.7 observed with 800 nm femtosecond pulses [ 39 ] could represent the
effects of three photon excitation, at 267 nm, being more efficient than two photon
excitation, at 400 nm [ 39 ]. In agreement with this proposal, it was observed that one
400 nm pulse in addition to one 800 nm pulse triggered photoconversion efficiently.
Moreover, the photoconversion efficiency of the second pulse tracks the lifetime of
the GFP A * (S1) state, with scanning of the delay between the 400 and 800 nm
pulses between 0 and 50 ps [ 39 ]. These results are fully consistent with the reported
wavelength dependency with CW illumination [ 19 ], but do not explain the apparent
lack of photoconversion reported with low power at 400 nm [ 37 ].
However, a similar behaviour was reported for the photoconversion of DsRed
[ 40 ]. Although thought to involve both chromophore photoisomerisation and oxi-
dative decarboxylation of Glu215 (see Sect. 2.3), evidence for a two photon process
was provided from the power density analysis for pulsed as well as for CW
irradiation. With regard to the latter observation, the involvement of an intermedi-
ate with microsecond life time would be needed to explain such behaviour, sug-
gested to be either a triplet state, an isomerised form or a different ionic state [ 40 ].

0-6

10 3 Fig. 6 (continued) spot size, the observed rate constants were between 2 and

10 2 s 1 , showing a linear dependence of the rate constant with power density. The
calculated photoconversion quantum yield at 390 nm was 1.6

1.6

10 3 [ 36 ]. Note that no isobestic
point is present, indicating the contribution of photobleached product state in the mixture using
pulsed blue excitation. (b) Using a 254 nm UV lamp, photoconversion proceeds with significantly
higher quantum yield, estimated at 0.03 [ 19 , 36 ]. Isobestic points at 325 and 426 nm are observed
with UV illumination