Organic Letters
Letter
that the fluorescence quantum yield decreases from 0.37 for
furane-based dye 5 to 0.20 for thiophene-based dye 6 to 0.06
for selenophene-based dye 7, which is consistent with heavy-
atom effects promoting triplet formation relative to fluo-
rescence.23 Another interesting feature that will be subjected to
additional investigations concerns the indole-based dye 8,
which was found to be almost nonfluorescent (ΦF < 0.01) in
PB, probably as a consequence of photoinduced electron
transfer (PeT) process.24 Conversely, preliminary fluorescence
fixed-cell imaging experiments have shown the emission ability
of this quinoxalin-2(1H)-one in a real biological context and its
superior capability for cellular staining compared with the
phenyl-based dye 2 otherwise characterized by a dramatically
higher ΦF (0.6 in PB) (Figures S68 and S69).
Scheme 1. Synthesis of PGA-Responsive “Covalent-
Assembly” Fluorogenic Probes 12−18
a
Fluorogenic PGA assays and blank experiments were
achieved through time-course measurements.11f,14 In all
cases, only the addition of an enzyme caused a rapid increase
in the blue-cyan fluorescence emission, which reached a
plateau in <10 min (Figure 3 and Figures S143−S156), except
a
O/N = overnight, FC (SiO2) = flash-column chromatography over
silica gel.
the 1,2-phenylenediamine scaffold through a self-immolative
linker (i.e., para-hydroxybenzyl) to lower the steric hindrance
and hence increase the accessibility of the probe to the
enzyme’s active site.22 Indeed, our preliminary investigations
have shown that the direct N-acylation of 9 with phenylacetyl
chloride led to hindered “covalent-assembly” fluorogenic
probes, not recognized by PGA (data not shown). Molecular
diversity was readily achieved by selecting seven different α-
ketoacids including pyruvic acid and 2-oxo-2-(hetero)arylacetic
acids. (See the Supporting Information for their synthesis and
conversion into acyl chlorides.) Practical implementation of
the two distinct N-acylation reactions was performed as
follows: 1,2-Phenylenediamine 9 (2.75 equiv) was reacted with
activated carbonate 10 (1.0 equiv) in dry DMF in the presence
of 1-hydroxybenzotriazole (HOBt, 1.0 equiv) and TEA (2.75
equiv). Thereafter, amidation of the remaining primary aniline
with α-ketoacid chloride (1.2 equiv) was conducted under
nearly identical conditions. Purification by flash-column
chromatography over silica gel provided the PGA-responsive
fluorogenic probes in moderate to good yields (23−69%). All
spectroscopic data (see the Supporting Information), especially
multinuclear nuclear magnetic resonance (NMR) mass
spectrometry data, were in agreement with the structures
assigned. Photophysical measurements also confirmed the lack
of fluorescence emission (vide infra).
We have also prepared the quinoxalin-2(1H)-one fluoro-
phores 1−7 assumed to be formed upon the enzymatic
activation of probes 12−18. Indeed, their use as references in
fluorescence assays and HPLC analyses is essential to clearly
demonstrate the reaction-based sensing mechanism. (See the
Supporting Information for their synthesis and character-
ization.) The synthesis of indole-based derivative 8 was also
achieved, but we failed to obtain the corresponding PGA-
responsive probe owing to its poor stability. To have an
indication of the magnitude of the fluorogenic “OFF−ON”
response arising from the PGA activation of probes 12−18,
photophysical properties of fluorophores were also determined,
especially in phosphate buffer (PB, 100 mM, pH 7.6). (See
Tables S1 and S2 and Figures S120−S135.) Compounds 1−3
and 5 have been identified as the most brilliant dyes within the
blue-cyan spectral range typical for quinoxalin-2(1H)-ones (B
(ε × ΦF) values: 5000−9100 M−1 cm−1). It is worth noting
Figure 3. Fluorescence emission time course (see the Supporting
Information for the detection parameters used) of fluorogenic probes
12−18 (concentration: 1.0 μM) in the presence of PGA (1 U) in PB
at 37 °C. Note that the curves for probes 16 and 18 are
superimposed.
for probe 15. For this compound, the electron-withdrawing
effect of −CF3 may negatively impact the dehydration step,
leading to quinoxalin-2(1H)-one aromatization (Figure 2); this
rate-determining step should be favored by the combined
electron push effects of the N atom and the neighboring alkyl/
aryl substituent toward the carbon atom center undergoing the
loss of a hydroxide group.
Outstanding fluorescence “OFF−ON” responses were
obtained (Figures S144, S146, S148, S150, S152, S154, and
S156). This highlights the main valuable feature of the
“covalent-assembly” approach in a striking manner. Further-
more, the enzymatic activation and fluorogenic response
arising from this were not impacted by the presence of
biological interferents such as glutathione (Figures S157−
S163). To confirm that the intense blue-cyan fluorescence
signal detected was due to the in situ formation of the
quinoxalin-2(1H)-one N-heterocycle, each enzymatic reaction
mixture was subjected to RP-HPLC−fluorescence analyses
(Figure 4A,B for phenyl derivative 13 and Figures S164−
S212). For each sample, a single peak was detected and
unambiguously assigned to the expected fluorophore (co-
injection with authentic samples of quinoxalin-2(1H)-ones 1−
7). To prove the validity of our assumption that the PGA-
initiated domino reaction yielded these benzo-fused N-
heterocycles, the same enzymatic mixtures were next analyzed
by RP-HPLC−MS (Figure 4C,D for phenyl derivative 13 and
C
Org. Lett. XXXX, XXX, XXX−XXX