Photoconversion of the Green Fluorescent Protein and Related Proteins - Fluorescent Proteins: From Understanding to Design - page 182

et al. [ 24 ]. These proposals were based on the analysis of the X-ray structure and the
details of the H-bonding interactions in the chromophore region. In 1997, Brejc
et al. proposed that the fluorescence photocycle of the protonated chromophore
involves the ultrafast proton transfer from the phenolic oxygen to the carboxylate of
Glutamic acid 222 [ 5 ].
Evidence from ultrafast infrared spectroscopy later supported this mechanism
in detail. Femtosecond optical excitation of the GFP A state leads to absorption
changes of an isolated band, in H 2 O at 1,725 cm 1 [ 25 - 28 ] which is a specific
frequency for protonated carboxylic acid. Based on the single wavelength
measurements at a lower frequency of 1,706 cm 1 , the assignment of this
transient to protonated Glutamate 222 was proposed [ 29 ]onthebasisofthe
earlier proposal from Brejc et al. [ 5 ]. The assignment was made on the basis of
the following four observations. (1) The isotope substitution effect of the mode
observed at 1,712 cm 1 in D 2 O and at 1,725 cm 1 in H 2 O supports assignment
to protonated glutamic acid [ 28 ]. (2) The corresponding H/D KIE of the
observed ESPT equals 5 [ 28 ], which fits the KIE determined by fluorescence
upconversion experiments for the ESPT reaction [ 16 ]. (3) Spectral and kinetic
changes occur in a mutant Glu222Asp, in which the proton acceptor is replaced
with a functional Aspartic acid 222 group [ 27 , 28 ]. In this mutant, ESPT was
delayed relative to the wild type with 13 and 125 ps time constants in H 2 Oand
D 2 O, respectively, and the COOH/D stretching mode of the carboxylate accep-
tor showed a reduced amplitude and downshiftedfrequencycomparedtothe
wild type [ 27 , 28 ]. (4) The infrared anisotropy for the COOD band [ 27 , 29 ],
which was corrected for the finite bleach [ 27 ], agreed with the coordinates from
the X-ray frame.
The results mentioned above are taken as strong evidence for the model put
forward by Brejc et al. already in 1997. This is testimony for the important insight
that was obtained from the functional analysis of X-ray structural information of
GFP. It is interesting to note that an alternative proposal that is not supported by
recent evidence invoking His 148 as the transient ESPT acceptor [ 6 , 30 ] was also
based on X-ray analysis, but made on the basis of different data.
In addition to predicting the identity of the excited state transient proton accep-
tor, Brejc et al. additionally proposed that subsequent syn-anti isomerisation of
transiently protonated Glu222 would stabilise the anionic species and explain the
photochromic behaviour [ 5 ] (Fig. 4 ; red). However, this was not supported by FTIR
measurements which established the absence of protonation changes of COOH/D
groups in GFP upon phototransformation [ 20 ]. It was later shown that electron
transfer is at the molecular basis of photoconversion rather than structural rearran-
gements in the chromophore-binding site [ 19 ] (see Sect. 2). Instead, syn-anti
isomerisation of Glu222 has been invoked in the fluorescence photocycle of GFP
[ 21 - 23 ].
More recently, ESPT has also been reported in the red fluorescent protein mutant
mKeima from the stony coral Montipora sp. [ 31 ] and in the LSSmKate1 and
LSSmKate2 red fluorescent protein mutants from the sea anemone Entacmaea
quadricolor [ 32 ].