Conformationally locked chromophores as models of excited-state proton transfer in fluorescent proteins

Article abstract of DOI:10.1021/ja3010144

Members of the green fluorescent protein (GFP) family form chromophores by modifications of three internal amino acid residues. Previously, many key characteristics of chromophores were studied using model compounds. However, no studies of intermolecular

Full text of DOI:10.1021/ja3010144

Article
pubs.acs.org/JACS
Conformationally Locked Chromophores as Models of Excited-State
Proton Transfer in Fluorescent Proteins
Mikhail S. Baranov,† Konstantin A. Lukyanov,† Alexandra O. Borissova,‡ Jordan Shamir,§,⊥
Dmytro Kosenkov,∥ Lyudmila V. Slipchenko,∥ Laren M. Tolbert,§ Ilia V. Yampolsky,*,†
and Kyril M. Solntsev*,§
†Institute of Bioorganic Chemistry, Russian Academy of Sciences, Miklukho-Maklaya 16/10, 117997 Moscow, Russia
‡Nesmeyanov Institute of Organoelement Compounds, Russian Academy of Sciences, Vavilova 28, 119991 Moscow, Russia
§School of Chemistry and Biochemistry, Georgia Institute of Technology, 901 Atlantic Drive, Atlanta, Georgia 30332-0400, United
States
∥Department of Chemistry, Purdue University, 560 Oval Drive, West Lafayette, Indiana 47907-2084, United States
S
* Supporting Information
ABSTRACT: Members of the green fluorescent protein
(GFP) family form chromophores by modifications of three
internal amino acid residues. Previously, many key character-
istics of chromophores were studied using model compounds.
However, no studies of intermolecular excited-state proton
transfer (ESPT) with GFP-like synthetic chromophores have
been performed because they either are nonfluorescent or lack
an ionizable OH group. In this paper we report the synthesis
and photochemical study of two highly fluorescent GFP
chromophore analogues: p-HOBDI-BF2 and p-HOPyDI:Zn. Among known fluorescent compounds, p-HOBDI-BF2 is the
closest analogue of the native GFP chromophore. These irrreversibly (p-HOBDI-BF2) and reversibly (p-HOPyDI:Zn) locked
compounds are the first examples of fully planar GFP chromophores, in which photoisomerization-induced deactivation is
suppressed and protolytic photodissociation is observed. The photophysical behavior of p-HOBDI-BF2 and p-HOPyDI:Zn
(excited state pKa’s, solvatochromism, kinetics, and thermodynamics of proton transfer) reveals their high photoacidity, which
makes them good models of intermolecular ESPT in fluorescent proteins. Moreover, p-HOPyDI:Zn is a first example of “super”
photoacidity in metal−organic complexes.
INTRODUCTION1
Members of the green fluorescent protein (GFP) family are
The chromophore in GFP is a 5-(4-hydroxybenzylidene)-3,5-
■
dihydro-4H-imidazol-4-one.3 This bicyclic structure originates
from the six-membered aromatic ring of the Tyr66 and a five-
now extensively used in experimental biology as genetically
encoded fluorescent markers.2 In contrast to other natural
membered heterocycle formed by condensation of the carbonyl
carbon of a residue at position 65 with the nitrogen of Gly67. It
pigments whose biosyntheses involve multiple enzymes and
was found that Tyr66 can be mutated to aromatic residues with
cofactors, GFP-like proteins form chromophore groups by
formation of blue-shifted chromophores.4 In particular, cyan
modifications of three internal amino acid residues. These
modifications are fully catalyzed by the fluorescent protein
(FP) itself and require no external enzymatic activities and
and blue mutants of A. victoria GFP carry Trp66 and His66,
respectively. The most blue-shifted spectra were achieved in the
Phe66-containing protein Sirius with excitation at 355 and
cofactors except for molecular oxygen. The unique capability of
emission at 424 nm.5 Further chemical modifications of residue
unassisted chromophore formation is based on the spatial
65 of the core green chromophore can occur in natural GFP-
structure of fluorescent proteins. All GFP-like proteins share
like proteins.2 These modifications extend a system of
the same fold, which includes an 11-sheet β-barrel with a
distorted α-helix inside the barrel. A chromophore is formed on
conjugated double bonds and result in strongly red-shifted
spectra in yellow, orange, and red FPs and purple-blue
this helix by cyclization and oxidation of the protein backbone
chromoproteins.
at positions 65−67 (e.g., Ser65-Tyr66-Gly67 in GFP from
jellyfish Aequorea victoria; here and below the numbering is in
accordance with A. victoria GFP) and is located in the
geometrical center of the β-barrel. Buried residues surrounding
Importantly, ionization of the phenolic hydroxyl of
chromophore Tyr66 changes the spectra dramatically. FPs
with a protonated (neutral) green chromophore possess an
the chromophore (mainly those located in the middle of the β-
strands) catalyze chromophore formation and participate in
fine-tuning of its spectral properties.
Received: January 31, 2012
Published: March 8, 2012
© 2012 American Chemical Society
6025
dx.doi.org/10.1021/ja3010144 | J. Am. Chem. Soc. 2012, 134, 6025−6032

Journal of the American Chemical Society
Article
Chart 1. Locked (in the Box) and Unlocked GFP Chromophores Discussed in the Text
absorption peak at about 400 nm, while chromophore
deprotonation leads to an 80−90 nm absorption red shift. A
similar dependence is observed for red FPs: they absorb at
about 450 and 550−600 nm with protonated and deprotonated
chromophores, respectively. Protonated chromophores can
produce a corresponding short-wavelength emission (blue for
green FPs or green for red FPs), but more typically they
undergo ultrafast excited-state proton transfer (ESPT) and emit
at longer wavelengths, similar to emission of the corresponding
anionic chromophore.6 Perhaps the most studied example of
ESPT in GFP-like proteins is the A. victoria wild-type (wt)
GFP. The absorption spectrum of this protein possesses a
major peak at 398 nm (protonated chromophore) and a minor
peak at 478 nm (deprotonated chromophore).7 Notably, due to
efficient ESPT, excitation of the wtGFP at either peak produces
green fluorescence with only slightly different maxima (508 and
503 nm for excitation at 398 and 482 nm, respectively). Time-
resolved spectroscopy demonstrates that excitation at 398 nm
gives off a short-lived blue emission at about 460 nm which
has ever been observed in p-HOBDI in all solvents studied.
Other model chromophores carry various substituents on this
core or represent a modified artificial core similar, but not
identical, to the chromophores in native fluorescent proteins.
Using model compounds, many key characteristics of
chromophores have been studied, including spectral properties
of charged and neutral (protonated/deprotonated) states of
chromophores, the influence of solvents and various sub-
stituents on the chromophore spectra,13 the dependence of
FQY on steric hindrance,13b,14 and the formation of mature
chromophores from biomimetic precursors.15 At the same time,
no studies of ESPT with GFP-like synthetic chromophores
have been performed to date, because they either are
nonfluorescent13b or lack an ionizable (−OH) group.14a
The Georgia Tech team has studied intermolecular ESPT in
various protonation states of m-HOBDI.16 Depending on pH of
the solution, photoinduced deprotonation of the hydroxyl
group and protonation of the imidazolone were observed. The
characteristic times of this process were in the range of several
picoseconds. However, most of the ESPT steps were diabatic
and did not lead to fluorescent products.
converts into green emission on a picosecond time scale.6a
A
pathway of proton migration within the GFP β-barrel which
includes a water molecule, Ser205, and Glu222 has been
proposed.8 According to an alternative model, a proton from
the excited chromophore can exit to the protein surface via
Thr203, while reprotonation of the chromophore can occur
due to a long proton entry pathway starting from Glu5 at the
protein surface and continuing with several inner water
molecules and residues including Glu222 and Ser205.9
A study of intramolecular proton transfer in o-HOBDI has
been published recently.17 However, such intramolecular ESPT
results in strongly different spectral characteristics compared to
those of p-HOBDI and wtGFP.
Here we report the synthesis and examine the photophysical
behavior (including intermolecular ESPT) of two highly
fluorescent GFP chromophore analogues. The first compound
is “irreversibly locked” p-HOBDI analogue, locked by a BF2
group ((5Z)-5-[(2-difluoroboryl-4-hydroxyphenyl)-
methylidene]-2,3-dimethyl-3,5-dihydro-4H-imidazol-4-one, p-
HOBDI-BF2 , Chart 1), similar to the strategy of Burgess.14a
Among known fluorescent compounds, p-HOBDI-BF2 is
closest to the native GFP chromophore. The second is a
hydroxy derivative (p-HOPyDI) of the fluorophore PyDI
([(Z)-1,2-dimethyl-4-(pyridin-2-ylmethylene)]-1H-imidazol-
5(4H)-one, Chart 1) studied recently by the Georgia Tech
group.14c PyDI becomes fluorescent upon complexation with
Zn2+ or Cd2+, therefore introducing a “reversible locking”
strategy. Upon metal binding, no new emission bands (MLCT,
etc.) were observed, confirming the absence of photoinduced
electron transfer between the d10 Zn2+ ion and the PyDI ligand.
Here we demonstrate the enhanced photoacidity of the p-
HOPyDI:Zn metal complex.
The chemical and physical principles of chromophore
formation and functioning in GFP-like proteins have generated
keen interest. A number of different approaches, including
crystallography, biochemical studies of native and hydrolyzed
fluorescent proteins, directed and random mutagenesis, and
steady-state and time-resolved spectroscopy, have been applied
to study these issues.2−9 Among others, chemical synthesis of
model chromophores was found to be a useful approach to
confirm or refute structures suggested from structural studies,
as well as to clarify details of chemical and spectral behavior of
fluorescent proteins’ chromophores.10 Our groups have
developed and studied extended libraries of the GFP and
RFP synthetic chromophores.11 The simplest compound
identical to the native GFP core is 4-(4-hydroxybenzylidene)-
1,2-dimethyl-1H-imidazol-5(4H)-one (p-HOBDI, Chart 1).
The fluorescence properties of this model chromophore differ
dramatically from those of the wtGFP. Due to the very efficient
photoisomerization-induced deactivation, its fluorescence
quantum yield (FQY) is <10−4, and the fluorescence lifetime
in most solvents is <1 ps.12 As a result, no intermolecular ESPT
RESULTS
■
Synthesis and Characterization of GFP Synthetic
Chromophores. An elegant borylation of aromatic com-
6026
dx.doi.org/10.1021/ja3010144 | J. Am. Chem. Soc. 2012, 134, 6025−6032

Journal of the American Chemical Society
Article
pounds with boron tribromide, directed by a neighboring
nitrogen atom, was recently described.18 We have modified this
method by using a “nonbasic base”molecular sievesinstead
of N,N-diisopropylethylamine (DIPEA) as HBr scavenger to
introduce a bridging BF2 group into the bicyclic structure of p-
HOBDI. p-HOBDI was silyl-protected at phenol and then
reacted with BBr3 and finally worked up with tetrabutylammo-
nium fluoride (TBAF), killing two birds with one stone:
removing the protecting group and converting the dibromo-
boryl fragment into the more stable difluoroboryl analogue
(Figure 1). Full synthetic procedures and characterization of p-
HOBDI-BF2 are presented in the Supporting Information.
planar: all the atoms except fluorine and the hydrogen atoms of
the methyl groups occupy special positions on the mirror plane.
A certain degree of delocalization manifests in C−C bond
lengths; e.g., C(4)−C(5) and C(3)−C(4) bond lengths
(1.439(3) and 1.355(3) Å, respectively) deviate from both
classical single bond (1.48 Å) and double bond (1.34 Å)
distances. p-HOBDI-BF2 molecules in the crystal are linked in
chains by relatively strong O(16)−H(16)···O(15) hydrogen
bonds (O···O distance is 2.693(3) Å, O−H···O angle is 175°).
In turn, these chains are organized in stacks by strong π-
stacking interactions with π···π distance ca. 3.3 Å. The parallel
orientation of planar moieties is extremely favorable for strong
stacking interactions. Stacks are linked by C−H···F interactions
to form a 3D network.
p-HOPyDI was recently synthesized and characterized by
Baldridge et al. as described in ref 19.
Steady-State and Time-Resolved Prototropic Behav-
ior of p-HOBDI-BF2. The absorption and emission spectra of
p-HOBDI-BF2 in various solvents resembled those of p-
HOBDI but exhibited a 30−40 nm bathochromic shift (see
Table S1). Indeed, only the neutral and anionic forms of the
chromophores can be compared since the cationic form of p-
HOBDI-BF2 cannot be produced. However, the most
important difference was the dramatic fluorescence turn-on of
the locked p-HOBDI-BF2. Its FQY in acetonitrile was 0.73, and
the fluorescence lifetime in the absence of ESPT (τ0) was 3.2
ns! This amazingly high value of FQY is close to those of
wtGFP (0.79) and another boron-locked compound reported
by Wu and Burgess (0.81).14a
We20 and others21 have successfully used the Kamlet−Taft
multivariant approach22 for various hydroxyaromatic com-
pounds, including p-HOBDI.23 This approach correlates the
spectral shift ν of the solute with the solvent parameters that
are responsible for its acidic (α), basic (β), and polar solvating
(π*) properties.
v(1/cm) = v + pπ* + aα + bβ
(1)
0
It allows a straightforward separation of selective (H-
bonding) and nonselective (dipole−dipole interaction) sol-
vation. We reported that the solvatochromic behavior of p-
HOBDI and its derivatives upon absorption is governed by both
polar and acid/base properties of the solvents. The magnitude
of the solute parameters p can be related to the relative dipole
moments of the molecules, while proton susceptibility
parameters a and b reflect relative proton basicity and acidity
of the chromophore. The magnitudes and directions of the
solvatochromic shifts strongly depended on the protonation
state of the solute. Our analysis clearly demonstrated the
increase of the dipole moment of p-HOBDI upon ionization
and showed its amphoteric behavior. In this work, an
exceptionally strong fluorescence of p-HOBDI-BF2 allowed
solvatochromic analysis of both absorption and emission
spectra. As in p-HOBDI, the absorption of the neutral form
of p-HOBDI-BF2 has a weak solvent dependence, but the
Kamlet−Taft linear regression analysis was unsatisfactory, with
the correlation coefficient <0.5. In contrast, such analysis
resulted in good to very good fits for the anion absorption and
all emission data (Table 1)
Figure 1. (Top) Synthesis of p-HOBDI-BF2: (i) TBDPSCl, imidazole,
DIPEA, THF; (ii) BBr3, MS 4 Å, DCM; (iii) TBAF·EtOAc. (Bottom)
General view of p-HOBDI-BF2 with atoms represented as thermal
ellipsoids at 50% probability level.
Crystals of p-HOBDI-BF2 (C12H10BF2N2O2, M = 263.03)
grown from MeCN were light yellow, orthorhombic, space
group Pnma. X-ray diffraction data were collected using a
Bruker SMART APEX2 CCD diffractometer (λ(Mo Kα) =
0.71073 Å, graphite monochromator) at 100(2) K: a =
17.743(5), b = 6.6328(17), and c = 9.856(3) Å. Intensities of
11 350 reflections were measured, and 1378 independent
reflections (Rint = 0.0757) were used in further refinement.
Initially spherical atom refinements were undertaken with
SHELXTL PLUS 5.0,10 using the full-matrix least-squares
method. All non-hydrogen atoms were allowed to have an
anisotropic thermal motion. The refinement converged to wR2
= 0.0977 and GOF = 1.000 for all independent reflections (R1
= 0.0397 was calculated against F for 928 observed reflections
with I > 2σ(I)). Atomic coordinates, bond lengths, bond angles,
and thermal parameters have been deposited at the Cambridge
Crystallographic Data Center (CCDC) with number 829645.
The crystallographic data confirm the desired structure of the
locked chromophore (Figure 1). The molecule is strictly
The spectral shift of the anionic absorption was governed
solely by the acidity of the solvent. It is interesting that the
magnitude of this interaction decreased 2-fold in the excited
state, showing the weakening of the anion basicity (or the
6027
dx.doi.org/10.1021/ja3010144 | J. Am. Chem. Soc. 2012, 134, 6025−6032

Journal of the American Chemical Society
Article
Table 1. Solvatochromic Parameters (in 103/cm) Used in
Multivariable Regression Fits of Absorption and Emission
Data for Neutral and Anionic Forms of p-HOBDI-BF2 and p-
HOBDI (Anion abs Only) According to Eq 1
a
band
ν0
p
a
b
R
anion abs
18.7
22.9
22.1
18.4
−0.1
−1.4
−1.1
−0.2
1.5
1.7
0
0.95
0.94
0.83
0.86
b
p-HOBDI
0.53
−0.5
0
neutral fluor
anion fluor
−0.2
0.7
a
b
Correlation coefficient. Measured in MeOH/H2O 1/1 v/v. From ref
23.
increase of the neutral acidity) upon excitation. The same effect
was observed by us for a number of photoacids.20
Figure 3. Fluorescence and absorption pH titration curves of p-
HOBDI-BF2 in water. φ/φo (A/Ao) and φ′/φ′ox (A′/A′o) are the
normalized fluorescence (absorption) intensities of R*OH (ROH)
and R*O− (RO−). Note that the intensity of R*O− (φ′ox) measured at
the end of the titration curves is not the “real” one, since hydrolysis of
p-HOBDI-BF2 is observed. See Figure S3.
In contrast to p-HOBDI, p-HOBDI-BF2 was readily soluble
in water. Absorption maxima of p-HOBDI-BF2 in neutral and
basic water exhibited bathochromic shifts as compared to those
of p-HOBDI and were much closer to the absorption maxima
of p-HOBDIMe+ at the same conditions. As expected, upon an
increase in pH the peak at 400 nm decreased, while the new
peak at 485 nm increased (Figure 2a). These bands
interconverted with an isosbestic point at 425 nm and a pKa
of 6.4.
Utilization of the modified Forster cycle equation25 resulted in
̈
pKa* = 0.6. Such moderate discrepancy between these methods
is a common situation in the photoacid prototropic analysis and
is usually associated with the errors in the spectral 0−0 energy
transitions, as well as the presence of diabatic processes.
Therefore, for the first time, we experimentally determined an
excited-state pKa of a close analogue of GFP chromophore.
Analysis of the fluorescence decay curves of R*OH and
R*O− in water resulted in an apparent dissociation rate
constant kESPT = 0.45 1/ns.25 In the past, one of us26 has
analyzed the relationship between the kinetics and the
thermodynamics of proton transfer from various hydroxyar-
omatic compounds in both the ground and excited states. All
data were fitted by the unified Brønsted-type equation (Figure
4). In the current work we did not analyze this dependence but
Figure 2. Absorption and emission of p-HOBDI-BF2 and p-HOPyDI
in various solvents. Absorbance spectra of p-HOBDI-BF2 in water at
pH 1 and 6 are identical.
Figure 4. Relationship between the protolytic dissociation rate
constants of various hydroxyaromatic compounds and their pKa in
water (Brønsted plot). Both ground- and excited-state proton transfer
data are presented. Solid line is a fitting curve described in ref 26.
In both protonated (R*OH) and deprotonated (R*O−)
states, p-HOBDI-BF2 fluoresced brightly. Excitation of the
neutral absorption peak at 400 nm resulted in dual emission at
485 and 527 nm (Figure 2b). This phenomenon is a well-
known intermolecular ESPT to water.24 Our data are the first
observation of the pronounced adiabatic ESPT in the GFP
chromophore analogue, so we studied this reaction in detail.
We found that interconversion of the blue and green emission
peaks occurred at low pH with pKa* = 2.125 (Figure 3).
used it as a guideline. The ESPT data for p-HOBDI-BF2 fit this
plot perfectly, and this locked GFP chromophore falls into the
range of moderate strength photoacids.
The excited-state behavior of p-HOBDI-BF2 as a well-
behaved moderate photoacid was further investigated by ESPT
studies in methanol/water mixtures and by reaction with an
6028
dx.doi.org/10.1021/ja3010144 | J. Am. Chem. Soc. 2012, 134, 6025−6032

Journal of the American Chemical Society
Article
nitrate, we observed a 73 nm bathochromic shift in the
absorption spectrum (Figure 2c).25 We attribute this band to
the O-deprotonated state of the complex. At the same time, a
moderate 10-fold fluorescence increase with a 62 nm
bathochromic shift was observed in the emission spectrum
(Figure 2d). Such modest fluorescence turn-on was much
smaller than the 150-fold fluorescence increase in PyDI:Zn. We
hypothesize that the moderate fluorescence turn-on in p-
HOPyDI:Zn in aqueous solutions can be associated with the
quenching by water. The latter may involve a non-adiabatic
ESPT to water, as proposed by Manca31 for p-HOBDI.
To exclude the possibility of quenching by water, and to
check the occurrence of “super” photoacidity27 in p-
HOPyDI:Zn, we studied the formation and properties of this
complex in nonaqueous media. The presence of 30 mM zinc
triflate (ZnTf2) in acetonitrile shifted the absorption band of
the chromophore from 362 nm to the red-shifted band with
vibronic maxima at 372 and 393 nm (Figure S4). The intensity
of the single-band fluorescence of the p-HOPyDI:Zn complex
in acetonitrile (λmax = 426 nm, φ = 0.029, τo = 145 and 640 ps,
amplitude ratio 7.8) was 120 times greater than that of a free
ligand. We have not determined the structure of the complex,
but by analogy with structurally similar PyDI:Zn14c we assume
that the stoichiometry of the p-HOPyDI:Zn complex is 1:1.
Upon gradual addition of water, the vibronic structure in the
absorption spectra became less pronounced, the emission of the
426 nm peak decreased, and the very weak emission band at
510 nm appeared due to the ground-state deprotonation and
(to a minor extent) to ESPT to water (Figure S3). While the
emission at 426 nm was quenched dramatically by water, the
concomitant increase of the phenolate emission was negligible.
This confirmed our hypothesis of the predominantly diabatic
ESPT character in water.
Efficient formation of the highly fluorescent p-HOPyDI:Zn
complex was also observed in dry methanol. Its absorption
spectrum was similar to that of acetonitrile solution (Figures 2e
and S4). However, in contrast to the latter, a pronounced two-
band emission with maxima at 431 and 507 nm was observed in
methanol (Figure 2f). Excitation spectra monitored at these
bands were almost identical and perfectly matched the
absorption spectrum (Figure 2e). This is evidence of efficient
ESPT. To our knowledge, this is the first example of “super”
photoacidity (i.e., ESPT to nonaqueous solvents)27 of inorganic
complexes. Similarly to acetonitrile/water mixtures, upon
gradual addition of water to the methanol solution, the
emission intensity of the R*OH form decreased, while the
emission of the anion practically did not change.
Figure 5. Effect of proton acceptor on the ESPT kinetics of p-HOBDI-
BF2. Fluorescence spectra of p-HOBDI-BF2 in methanol/water
mixtures (a) and in water at various concentrations of sodium acetate
(c). Quadratic dependence of the ESPT rate constant MeOH/H2O on
[H2O] (b). Stern−Volmer dependence (d).
external base. There is no ESPT from p-HOBDI-BF2 in pure
methanol, as predicted for the photoacids with pKa* > 0.27
Upon addition of water, the ESPT product band appeared in
the emission spectra (Figure 5a). The ESPT rate constant had a
quadratic dependence on water concentration (Figure 5b), as
observed earlier for the photoacids with pKa* > 0.28 Addition of
an external base (acetate ion) to the aqueous solution of p-
HOBDI-BF2 accelerated the observed ESPT (Figure 5c).24
Analysis of the emission data using the modified Stern−Volmer
equation (Figure 5d) resulted in the apparent bimolecular rate
3.8 M−1 ns−1. This diffusion-limited exergonic proton transfer
from the excited GFP chromophore to the carboxylate does
proceed on the ultrafast time scale in proteins, where the
diffusion of the reagents is absent and the short proton wire
supports a concerted proton transfer. It is interesting to draw
parallels between the fluorescence spectrum of p-HOBDI-BF2
in 3 M sodium acetate solution and that of the H148D/S65T
mutant29 of the wtGFP. In both cases only the R*O− emission
is observed. Therefore, a 3 M concentration of proton acceptor
in solution may mimic the chromophore−Asp interaction in
H148D/S65T derived from the X-ray analysis.
Keeping in mind that the complex may be unstable in
aqueous solution, we diluted the methanol solution of p-
HOPyDI:Zn with water up to 1/1 v/v and performed pH
titration (Figure S5). The complex was stable only in a quite
narrow pH range: 3.0 < pH < 6.2. Nevertheless, the pKa of the
phenolic moiety can be estimated as 5.3 from both emission
Steady-State and Time-Resolved Prototropic Behav-
ior of p-HOPyDI:Zn. Addition of the para-hydroxy group to
the PyDI molecule resulted in the rich ground- and excited-
state properties of p-HOPyDI. The latter molecule has five
acid−base groups, but, to our surprise, the pH dependence of
the absorption spectra resembled that of 3-hydroxypyridine.30
This bifunctional compound exhibits a complicated pH
dependence showing the equilibrium between the neutral,
tautomeric (or zwitterionic), anionic, and cationic species.
Similar prototropic behavior of p-HOBDI indeed influences its
ability to bind zinc ions and moderates the ground- and excited-
state properties of the resulting complex.
and absorption data. Using the Forster cycle equation (see
̈
above) and the emission/absorption maxima for the neutral
(385/433 nm) and the anion (432/506 nm) of p-HOPyDI:Zn
in methanol/water, we estimated the pKa* as −1.3, confirming
our classification of p-HOPyDI:Zn as a “super” photoacid.
The fluorescence lifetime of the p-HOPyDI:Zn neutral form
in the presence of ESPT (τ) in methanol is 56 ps; its
conjugated base has a rise time of 53 ps and decays with 110 ps
lifetime. It is important to realize that the decay of the neutral
form reflects the sum of adiabatic and diabatic ESPT.32 Because
First, we produced the Zn complex of p-HOPyDI using the
method successfully utilized for PyDI:Zn synthesis.14c Upon
mixing of a very weakly fluorescent acetonitrile (or methanol)
solution of p-HOPyDI with a 60 mM aqueous solution of zinc
6029
dx.doi.org/10.1021/ja3010144 | J. Am. Chem. Soc. 2012, 134, 6025−6032

Journal of the American Chemical Society
Article
of p-HOPyDI:Zn instability at the high or lower pH when the
acid−base equilibrium is shifted to pure neutral or anionic, it is
impossible to separate the rates of adiabatic and diabatic ESPT,
so only the upper limit of the adiabatic kESPT can be
determined.
Therefore, we have demonstrated the first example of the
enormous photoacidity in an inorganic complex. This area is
relatively undeveloped,33 and usually the ΔpKa between the
ground and the excited states does not exceed 5 pKa units.34 It
would be intriguing to synthesize complexes of p-HOPyDI with
heavy metals such as Pt that would promote efficient
intersystem crossing. The resulting long-lived triplet state
may exhibit a dramatic photoacidity as well.24 Similarly, we can
propose that the metal-dependent “turn-on” of photoacidity
may have mechanistic and diagnostic consequences.
Table 2. Comparison of p-HOBDI-BF2 and p-HOPyDI:Zn
Kinetic and Thermodynamics Data: Fluorescence Quantum
Yields, ESPT Rate Constants, pKa, and pKa*
p-HOBDI-BF2
p-HOPyDI:Zn
Experimental Data
a
FQY
0.73
3.2
0.029
a
τo, ns
τ, ns
<0.21>
b
c
1.45
0.056
d
b
c
kESPT, 1/ns
pKa
0.45
<13
b
e
6.4
5.3
bf
,
eg
,
pKa*
2.1
−1.3
Computational Results
h
qo, au
−0.63
−8.07
5.62
−0.57
−8.32
2.59
ΔEex,solv, kcal/mol
pKa
bi
,
Theoretical Computations of p-HOBDI-BF2 and p-
HOPyDI:Zn Ground- and Excited-State Acidity. A deeper
understanding of the ground- and excited- state acidity of novel
compounds could be achieved using high-level calculations.
Theoretical computations of pKa* are generally very daunting.
A simple scheme has been employed in this work, with a goal to
provide qualitative insight into the origins of differences in the
pKa* between p-HOBDI-BF2 and p-HOPyDI:Zn. The pKa of
the ground state of an acid (AH) can be calculated using the
Gibbs free energy of a dissociation reaction AH → A− + H+ in a
solvent: pKa = ΔGaq/ln(10)RT. The Gibbs free energy ΔGaq of
the reaction can be evaluated using the Born−Haber cycle35 as
bj
,
pKa*
−0.30
−3.52
a
b
c
d
In ACN. In water. In MeOH. Determined as kESPT = 1/τ − 1/τo.
e
f
g
In MeOH/H2O 2/1 v/v. From fluorescence pH titration. From
h
Forster equation. The Mulliken charges on hydroxyl oxygen in
̈
i
j
protonated species. From eq 2. From eq 3.
in p-HOPyDI:Zn versus 3737 cm−1 in p-HOBDI-BF2). This
suggests that the positively charged Zn substitution weakens
the OH bond strength and promotes deprotonation. The
calculated ground-state pKa values are in agreement with this
qualitative analysis. The acidity of both compounds decreases
by 5.9−6.1 units upon excitation, placing them as “super”
photoacids. However, taking into account that the computed
ground-state pKa values are probably underestimated, there is a
nice agreement between theory and experiment in both the
relative acidity of p-HOPyDI:Zn and p-HOBDI-BF2 and the
increase of the acidity upon excitation.
−
AH→A
a sum of the gas-phase electronic energy ΔE
, zero-
g
−
AH→A
point vibrational energy ΔZPE
, Gibbs free energy
g
AH→A−
g,0→298K
AH→A−
ΔG
, and solvation energy ΔG
differences between
solv
protonated (AH) and deprotonated (A−) species:
−
−
−
AH→A
AH→A
AH→A
ΔG = ΔE
aq
+ ΔZPE
+ ΔG
CONCLUSIONS
g
g
g,0→298K
■
−
+
+
We have demonstrated the first known examples of fully planar
synthetic GFP chromophores, in which photoisomerization-
induced deactivation is suppressed and efficient intermolecular
ESPT is observed. A pronounced difference in the photoacidity
of p-HOBDI-BF2 and p-HOPyDI:Zn was established. Of
course, we note that these compounds carry oppositely charged
substituents in the position meta to the reactive hydroxyl group.
Such meta-effects are known to influence the excited-state
acidity dramatically.38 We propose that the pKa* values of these
compounds may serve as the lower and upper estimates for the
pKa* of unsubstituted p-HOBDI. The theoretical pKa* of p-
HOBDI (0.1 in water)39 calculated by Scharnagl and Raupp-
Kossmann fits nicely within these boundaries.
AH→A
H
H
+ ΔE
+ ΔG
+ ΔG
solv
g
solv
(2)
+
H
where ΔG
= −6.28 kcal/mol is the standard Gibbs free
g
H+
solv
energy of the proton in the gas phase and ΔG
is the
solvation free energy of the proton, estimated theoretically as
−262.23 kcal/mol.36 Gas-phase geometry optimizations, ZPE,
and Gibbs free energies were computed at the B3LYP/6-
31G(*) level of theory. Gibbs free energies in an aqueous
solvent were estimated using PBE0/6-31+G(*) and a polar-
izable continuum model (PCM).37 The excited-state pKa*
values were obtained using the Forster cycle equation as a sum
of the ground-state pKa and the excitation energy difference
̈
AH→A−
ASSOCIATED CONTENT
* Supporting Information
Synthetic procedures, additional spectral data, and pH titration
curves of p-HOBDI-BF2. This material is available free of
charge via the Internet at http://pubs.acs.org.
ΔE
between protonated and deprotonated species in a
ex,solv
■
solvent:
S
−
AH→A
pK * = pK + ΔE
/ln(10)RT
a
a
ex,solv
(3)
The excitation energies in a solvent were computed using
time-dependent density functional theory (TD-DFT) at the
PBE0/6-31+G(*) level of theory and PCM. Results of these
calculations are summarized in Table 2.
In agreement with experimental findings, the optimized
structures of p-HOBDI-BF2 and p-HOPyDI:Zn are planar both
in the gas phase and in aqueous solvent. Substituents in the
meta-position with respect to the hydroxyl group influence a
charge distribution on hydroxyl oxygen, as well as the
vibrational frequencies of the OH stretch mode (3713 cm−1
AUTHOR INFORMATION
Corresponding Author
ivyamp@gmail.com; solntsev@gatech.edu; http://ww2.
chemistry.gatech.edu/solntsev
■
Notes
The authors declare no competing financial interest.
⊥Summer program participant (high school student) from
Lawrenceville School, 2500 Main St., Lawrenceville, NJ 08648
6030
dx.doi.org/10.1021/ja3010144 | J. Am. Chem. Soc. 2012, 134, 6025−6032

Journal of the American Chemical Society
Article
(15) (a) Kojima, S.; Hirano, T.; Niwa, H.; Ohashi, M.; Inouye, S.;
Tsuji, F. I. Tetrahedron Lett. 1997, 38, 2875. (b) Ivashkin, P. E.;
Lukyanov, K. A.; Lukyanov, S.; Yampolsky, I. V. J. Org. Chem. 2011,
76, 2782.
(16) (a) Dong, J.; Solntsev, K. M.; Poizat, O.; Tolbert, L. M. J. Am.
Chem. Soc. 2007, 129, 10084. (b) Solntsev, K. M.; Poizat, O.; Dong, J.;
Rehault, J.; Lou, Y.; Burda, C.; Tolbert, L. M. J. Phys. Chem. B 2008,
112, 2700.
(17) Hsieh, C.-C.; Chou, P.-T.; Shih, C.-W.; Chuang, W.-T.; Chung,
M.-W.; Lee, J.; Joo, T. J. Am. Chem. Soc. 2011, 133, 2932.
(18) Ishida, N.; Moriya, T.; Goya, T.; Murakami, M. J. Org. Chem.
2010, 75, 8709.
ACKNOWLEDGMENTS
■
Financial support from the National Science Foundation
(CHE-0809179 for K.M.S, and Career CHE-0955419 for
L.V.S.) is gratefully acknowledged. We thank the Department
of Energy, Office of Basic Science (DE-FG02-04ER46141), for
generous financial support to L.M.T. L.V.S. acknowledges
support from Purdue University. The Moscow group acknowl-
edges grants by the Molecular and Cell Biology Program of the
Russian Academy of Sciences, and Ministry of Education and
Science of the Russian Federation (16.740.11.0367 and
16.512.11.2139). We thank Anthony Baldridge for providing
p-HOPyDI, and Julia Weinstein for stimulating discussions.
(19) Lee, J.-S.; Baldridge, A.; Feng, S.; SiQiang, Y.; Kim, Y. K.;
Tolbert, L. M.; Chang, Y.-T. ACS Comb. Sci. 2011, 13, 32.
(20) (a) Solntsev, K. M.; Huppert, D.; Tolbert, L. M.; Agmon, N. J.
Am. Chem. Soc. 1998, 120, 7981. (b) Solntsev, K. M.; Huppert, D.;
Agmon, N. J. Phys. Chem. A 1998, 102, 9599. (c) McGrier, P. L.;
Solntsev, K. M.; Miao, S.; Tolbert, L. M.; Miranda, O. R.; Rotello, V.
M.; Bunz, U. H. F. Chem.Eur. J. 2008, 14, 4503.
REFERENCES
■
(1) This is part of our series “Photochemistry of “super” photoacids”.
For the previous paper, see: Gould, E.-A.; Popov, A. V.; Tolbert, L. M.;
Presiado, I.; Erez, Y.; Huppert, D.; Solntsev, K. M. Phys. Chem. Chem.
Phys. 2012, DOI: 10.1039/c2cp23891h.
(2) Chudakov, D. M.; Matz, M. V.; Lukyanov, S.; Lukyanov, K. A.
Physiol. Rev. 2010, 1103.
(3) Cody, C. W.; Prasher, D. C.; Westler, W. M.; Prendergast, F. G.;
Ward, W. W. Biochemistry 1993, 32, 1212.
(4) Heim, R.; Prasher, D. C.; Tsien, R. Y. Proc. Natl. Acad. Sci. U.S.A.
1994, 91, 12501.
(5) Tomosugi, W.; Matsuda, T.; Tani, T.; Nemoto, T.; Kotera, I.;
Saito, K.; Horikawa, K.; Nagai, T. Nat. Methods 2009, 6, 351.
(6) (a) Chattoraj, M.; King, B. A.; Bublitz, G. U.; Boxer, S. G. Proc.
Natl. Acad. Sci. U.S.A. 1996, 93, 8362. (b) Kogure, T.; Karasawa, S.;
Araki, T.; Saito, K.; Kinjo, M.; Miyawaki, A. Nat. Biotechnol. 2006, 24,
577. (c) Henderson, J. N.; Osborn, M. F.; Koon, N.; Gepshtein, R.;
Huppert, D.; Remington, S. J. J. Am. Chem. Soc. 2009, 131, 13212.
(d) Piatkevich, K. D.; Malashkevich, V. N.; Almo, S. C.; Verkhusha, V.
V. J. Am. Chem. Soc. 2010, 132, 10762.
̌
̌ ́ ́ ̌ ́ ́ ́
(21) (a) Mijin, D. Z.; Uscumlic, G. S.; Perisic-Janjic, N. U.; Valentic,
N. V. Chem. Phys. Lett. 2006, 418, 223. (b) Reta, M.; Cattana, R.;
Silber, J. J. J. Solution Chem. 2001, 30, 237. (c) Magnes, B.-Z.; Pines,
D.; Strashnikova, N.; Pines, E. Solid State Ionics 2004, 168, 225.
(d) Vokin, A. I; Sherstyannikova, L. V.; Kanitskaya, L. V.; Abzaeva, K.
A.; Lopyrev, V. A.; Turchaninov, V. K. Russ. J. Gen. Chem. 2001, 71,
1708.
(22) Kamlet, M. J.; Abboud, J. L. M.; Abraham, M. H.; Taft, R. W. J.
Org. Chem. 1983, 48, 2877.
(23) Dong, J.; Solntsev, K. M.; Tolbert, L. M. J. Am. Chem. Soc. 2006,
128, 12038.
(24) (a) Yu, I.; Martynov, A. B.; Demyashkevich, B. M. U.; M. G., K.
Russ. Chem. Rev. 1977, 46, 1. (b) Weller, A. Prog. React. Kinet. 1961, 1,
189. (c) Arnaut, L.; Formosinho, S. J. J. Photochem. Photobiol. A 1993,
75, 1.
(25) See Supporting Information for more details.
(26) Solntsev, K. M.; Al-Ainain, S. A.; Il’ichev, Y. V.; Kuzmin, M. G. J.
Phys. Chem. A 2004, 108, 8212.
(7) Ward, W. W.; Bokman, S. H. Biochemistry 1982, 21, 4535.
(8) (a) Palm, G. J.; Zdanov, A.; Gaitanaris, G. A.; Stauber, R.;
Pavlakis, G. N.; Wlodawer, A. Nat. Struct. Biol. 1997, 4, 361. (b) Brejc,
K.; Sixma, T. K.; Kitts, P. A.; Kain, S. R.; Tsien, R. Y.; Ormo, M.;
Remington, S. J. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 2306.
(27) Tolbert, L. M.; Solntsev, K. M. Acc. Chem. Res. 2002, 35, 19.
(28) Solntsev, K. M.; Huppert, D.; Agmon, N.; Tolbert, L. M. J. Phys.
Chem. A 2000, 104, 4658.
̈
(29) (a) Leiderman, P.; Genosar, L.; Huppert, D.; Shu, X.;
Remington, S. J.; Solntsev, K. M.; Tolbert, L. M. Biochemistry 2007,
46, 12026. (b) Stoner-Ma, D.; Jaye, A. A.; Ronayne, K. L.; Nappa, J.;
Meech, S. R.; Tonge, P. J. J. Am. Chem. Soc. 2008, 130, 1227.
(30) (a) Metzler, D. E.; Harris, C. M.; Johnson, R. J.; Siano, D. B.;
Thomson, J. A. Biochemistry 1973, 12, 5377. (b) Reiter, J.; Beyer, A.;
Potschka, M.; Schuster, P.; Winkler, H.; Ebeling, H.; Franck, E. U. J.
Phys. Chem. 1989, 93, 442. (c) D’Angel, J. C.; Collette, T. W. Anal.
Chem. 1997, 69, 1642.
(31) Manca, C. Chem. Phys. Lett. 2007, 443, 173.
(32) Solntsev, K. M.; Al-Ainain, S. A.; Il’ichev, Y. V.; Kuzmin, M. G. J.
Photochem. Photobiol. A 2005, 175, 178.
(33) (a) Vos, J. G. Polyhedron 1992, 11, 2285. (b) Liu, W.; Thorp, H.
H. Adv. Transition Met. Coord. Chem. 1996, 1, 187.
(9) Agmon, N. Biophys. J. 2005, 88, 2452.
(10) (a) Niwa, H.; Inouye, S.; Hirano, T.; Matsuno, T.; Kojima, S.;
Kubota, M.; Ohashi, M.; Tsuji, F. I. Proc. Natl. Acad. Sci. U.S.A. 1996,
93, 13617. (b) Bell, A. F.; He, X.; Wachter, R. M.; Tonge, P. T.
Biochemistry 2000, 39, 4423. (c) Voityuk, A. A.; Kummer, A. D.;
Michel-Beyerle, M.-E.; Rosch, N. Chem. Phys. 2001, 269, 83.
̈
(d) Yampolsky, I. V.; Remington, S. J.; Martynov, V. I.; Potapov, V.
K.; Lukyanov, S.; Lukyanov, K. A. Biochemistry 2005, 44, 5788.
(11) (a) Ivashkin, P. E.; Yampolsky, I. V.; Lukyanov, K. A. Russ. J.
Bioorg. Chem. 2009, 35, 652. (b) Tolbert, L. M.; Baldridge, A.;
Kowalik, J.; Solntsev, K. M. Acc. Chem. Res. 2012, 45, 171.
(12) Mandal, D.; Tahara, T.; Meech, S. R. J. Phys. Chem. B 2004, 108,
1102.
(13) (a) Kojima, S. Tetrahedron Lett. 1998, 39, 5239. (b) Niwa, H.;
Inouye, S.; Hirano, T.; Matsuno, T.; Kojima, S.; Kubota, M.; Ohashi,
M.; Tsuji, F. I. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 13617. (c) He, X.;
Bell, A. F.; Tonge, P. J. Org. Lett. 2002, 4, 1523. (d) Yampolsky, I. V.;
Remington, S. J.; Martynov, V. I.; Potapov, V. K.; Lukyanov, S.;
Lukyanov, K. A. Biochemistry 2005, 44, 5788. (e) Yampolsky, I. V.;
Kislukhin, A. A.; Amatov, T. T.; Shcherbo, D.; Potapov, V. K.;
Lukyanov, S.; Lukyanov, K. A. Bioorg. Chem. 2008, 36, 96.
(f) Yampolsky, I. V.; Balashova, T. A.; Lukyanov, K. A. Biochemistry
2009, 48, 8077.
(34) (a) Giordano, P. J.; Bock, C. R.; Wrighton, M. S. J. Am. Chem.
Soc. 1978, 100, 6960. (b) Cattaneo, M.; Fagalde, F.; Katz, N. E. Inorg.
Chem. 2006, 45, 6884. (c) Su, C.-H.; Chen, H.-Y.; Tsai, K. Y.-D.;
Chang, I.-J. J. Phys. Chem. B 2007, 111, 6857. (d) Han, M.-J.; Chen, Y.-
M.; Wang, K.-Z. New J. Chem. 2008, 32, 970.
̀
(35) (a) Jacquemin, D.; Perpete, E. A.; Ciofini, I.; Adamo, C. J. Phys.
Chem. A 2008, 112, 794. (b) Jang, Y. H.; Goddard, W. A. III; Noyes,
K. T.; Sowers, L. C.; Hwang, S.; Chung, D. S. J. Phys. Chem. B 2003,
107, 344.
(36) Tawa, G. J.; Topol, I. A.; Burt, S. K.; Caldwell, R. A.; Rashin, A.
A. J. Chem. Phys. 1998, 109, 4852.
(37) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. Rev. 2005, 105,
(14) (a) Wu, L.; Burgess, K. J. Am. Chem. Soc. 2008, 130, 4089.
(b) Conyard, J.; Kondo, M.; Heisler, I. A.; Jones, G.; Baldridge, A.;
Tolbert, L. M.; Solntsev, K. M.; Meech, S. R. J. Phys. Chem. B 2011,
115, 1571. (c) Baldridge, A.; Solntsev, K. M.; Song, C.; Tanioka, T.;
Kowalik, J.; Hardcastle, K.; Tolbert, L. M. Chem. Commun. 2010, 46,
5686.
2999.
(38) (a) Zimmerman, H. E. J. Am. Chem. Soc. 1995, 117, 8988.
(b) Lewis, F. D.; Yang, J.-S. J. Am. Chem. Soc. 1997, 119, 3834.
6031
dx.doi.org/10.1021/ja3010144 | J. Am. Chem. Soc. 2012, 134, 6025−6032

Journal of the American Chemical Society
Article
(39) Scharnagl, C.; Raupp-Kossmann, R. A. J. Phys. Chem. B 2004,
108, 477.
6032
dx.doi.org/10.1021/ja3010144 | J. Am. Chem. Soc. 2012, 134, 6025−6032