Substituent effects on the turn-on kinetics of rhodamine-based fluorescent pH probes

Article abstract of DOI:10.1039/c3ob42089b

Fluorescent turn-on probes based on a rhodamine spirolactam (RSL) structure have recently become a popular means of detecting pH, metal ions, and other analytes of interest. RSLs are colorless and non-fluorescent until the target analyte induces opening of the spirocyclic ring system, revealing the fully conjugated and highly fluorescent rhodamine dye. Among RSLs opened by acid, we have observed wide variation in the kinetics of the fluorescence turn-on process such that some probes would not be usable in situations where a rapid reading is desired or the pH fluctuates temporally. Herein we present a systematic investigation of the fluorescence turn-on kinetics of RSLs to probe the hypothesis that the reaction rates are influenced by the electronic properties of the spirolactam ring system. A series of 8 aniline-derived RSLs with para substituents ranging from electron-donating to electron-withdrawing was prepared from rhodamine B. The fluorescence turn-on rates are observed to increase by a factor of four as the substituent is tuned from methoxy to nitro. This effect is explained in terms of the destabilization of the reaction intermediate by the substituent. As the reaction rates increase across the series, a concomitant increase in fluorescence intensity is also observed. This result is attributed to an increase in the concentration of the fluorescent form of the dye and is consistent with the expected equilibrium properties of this system. These findings are applied to the design of a faster-reacting and more intensely fluorescent RSL pH probe.

Full text of DOI:10.1039/c3ob42089b

Organic &
Biomolecular Chemistry
View Article Online
View Journal
PAPER
Substituent effects on the turn-on kinetics of
rhodamine-based fluorescent pH probes†
Cite this: DOI: 10.1039/c3ob42089b
William L. Czaplyski, Grace E. Purnell, Courtney A. Roberts, Rebecca M. Allred and
Elizabeth J. Harbron*
Fluorescent turn-on probes based on a rhodamine spirolactam (RSL) structure have recently become a
popular means of detecting pH, metal ions, and other analytes of interest. RSLs are colorless and non-
fluorescent until the target analyte induces opening of the spirocyclic ring system, revealing the fully con-
jugated and highly fluorescent rhodamine dye. Among RSLs opened by acid, we have observed wide vari-
ation in the kinetics of the fluorescence turn-on process such that some probes would not be usable in
situations where a rapid reading is desired or the pH fluctuates temporally. Herein we present a systematic
investigation of the fluorescence turn-on kinetics of RSLs to probe the hypothesis that the reaction rates
are influenced by the electronic properties of the spirolactam ring system. A series of 8 aniline-derived
RSLs with para substituents ranging from electron-donating to electron-withdrawing was prepared from
rhodamine B. The fluorescence turn-on rates are observed to increase by a factor of four as the substitu-
ent is tuned from methoxy to nitro. This effect is explained in terms of the destabilization of the reaction
intermediate by the substituent. As the reaction rates increase across the series, a concomitant increase in
fluorescence intensity is also observed. This result is attributed to an increase in the concentration of the
fluorescent form of the dye and is consistent with the expected equilibrium properties of this system.
These findings are applied to the design of a faster-reacting and more intensely fluorescent RSL pH
probe.
Received 20th October 2013,
Accepted 21st November 2013
DOI: 10.1039/c3ob42089b
www.rsc.org/obc
Introduction
The use of fluorescent probes to measure pH is particularly
suited to samples in which the pH may be spatially or tem-
porally heterogeneous, such as cells. Fluorescent probes for
intracellular environments have been designed to respond to
small differences in pH levels within two distinct regions,
neutral (ca. pH 7) and acidic (pH 4.5–6). While a wide variety
of fluorescent probes for the neutral region exist, many fewer
have been developed for the acidic region.1 Organelles includ-
ing endosomes and lysosomes function within this acidic pH
Scheme 1
range, and fluorescent probes specifically designed for acidic such as rhodamine 6G and rhodamine B. As prepared, RSLs
pH levels are required to track processes within these are colorless and non-fluorescent due to the separation of the
organelles.2
lower xanthene core from the upper ring system by a tetra-
We are interested in developing fluorescent pH probes for hedral spiro carbon. Acid induces a ring-opening reaction that
the acidic region based on a rhodamine spirolactam (RSL) yields a completely conjugated rhodamine dye structure. The
structure. RSLs such as 1c–8c (Scheme 1, where c denotes ring-opened form is intensely colored and possesses the favor-
closed) are easily prepared from commercially available dyes able fluorescence characteristics of a rhodamine dye3 includ-
ing high fluorescence quantum yields and good photostability.
Thus, RSLs function as fluorescent turn-on probes that switch
Department of Chemistry, The College of William and Mary, Williamsburg, Virginia
from non-fluorescent to fluorescent in response to acid
23187-8795, USA. E-mail: ejharb@wm.edu
†Electronic supplementary information (ESI) available: Derivation of eqn (2) and
stimulus.1
In recent years RSL-based pH probes have begun to appear
in the literature and have thus far been applied to live cell
(4), absorption and fluorescence spectra and titration curves for probes 1–9,
1H- and 13C NMR spectra of probes 1c–9c. See DOI: 10.1039/c3ob42089b
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
imaging,4–12 detection of acid vapors,13 detection of aluminum
corrosion,14 and solution-based pH measurements.15–19
Several studies have shown that RSLs become fluorescent in
the acidic environment within lysosomes and can be used for
lysosomal imaging in live cells.4,5,8,11,12 For example, Han and
coworkers demonstrated that an RSL derivative selectively
stains lysosomes and has good enough photostability and
retention times in cells that it can be used to monitor changes
in lysosome morphology in apoptotic cells.8 Peng and co-
workers also used an RSL derivative to detect pH changes in
lysosomes during apoptosis and found the performance of the
RSL dye to be superior to commercially available probes.11
These studies demonstrate the utility of RSL-based pH probes.
In our preliminary investigations of the fluorescence turn-
on behavior of RSLs, we observed that some derivatives
respond instantaneously to acid while others take seconds,
minutes, or even longer to reach maximum fluorescence inten-
sity at a given pH. The turn-on kinetics are critical to probe
design as a slow reaction with acid renders an otherwise useful
pH probe essentially unusable for many practical applications.
Slow turn-on kinetics have been discussed for some RSL-based
metal ion sensors,20 but the subject has received little atten-
tion in the emerging literature of RSL pH probes. Bossi and co-
workers recently explored the kinetics of ring-opening for
photochromic RSLs21 that become fluorescent in the presence
of both acid and ultraviolet light.22 They measured different
thermal ring-opening rates in the presence of acid for structu-
rally different RSLs in agreement with our own observations.
We hypothesize that the rate of acid-induced ring opening of
RSLs is influenced by electronic factors. To probe the role of
substituent electronics in turn-on kinetics, we designed a
series of aniline-derived RSLs with substituent properties
ranging from electron-donating to electron-withdrawing.
Herein we present the results of kinetic studies of this series
as well as application of our findings to the design of a much
faster-reacting probe.
Fig. 1 Absorption (A) and fluorescence (B) of compound 8 in 1 : 1 (v/v)
ethanol–water as pH is decreased from 5.96 to 3.26.
column chromatography.25 The spiro carbon’s unique reson-
ance in the 13C NMR spectrum, observed between 67.2 and
67.6 ppm for 1c–8c, provided verification of the spirocyclic
structure.
In the as-prepared spirocyclic form, 1c–8c absorb ultraviolet
light, with well-defined peaks at 315 nm and 275 nm in a 1 : 1
(v/v) solution of ethanol and water. As expected based on their
non-conjugated structures, they do not absorb or fluoresce in
the visible region of the spectrum. Upon introduction of HCl,
a new absorbance band with a maximum around 562 nm
appears, and excitation into this band yields a fluorescence
spectrum with a maximum intensity around 583 nm. Fig. 1
illustrates the increase in visible absorbance and fluorescence
for 8 as increasing amounts of acid are added and 8c is con-
verted to 8o, where the o denotes the open form of the dye.
Derivatives 1o–8o have essentially identical absorption and
fluorescence spectra, with λmax,abs and λmax,fl values each
varying less than 2 nm across the series with no discernible
trend with regard to substituent electronics (spectra are shown
in the ESI†). The absorption and fluorescence spectra of 1o–8o
are slightly red-shifted from the parent rhodamine B dye,
which has a λmax,abs of 552 nm and a λmax,fl of 580 nm in the
same solvent system.
Results and discussion
In addition to having nearly indistinguishable absorption
and fluorescence spectra, 1–8 also give virtually the same fluo-
rescence titration curves when titrated with HCl. From these
titration curves, the pKa of each derivative can be calculated by
determining the pH at which probe fluorescence is half of its
maximum value.6 Fig. 2 shows the normalized titration curves
for the electronic extremes of the series, 1 (OMe) and 8 (NO2),
which have pKa values of 4.16 and 4.11, respectively. The
average pKa for the series was 4.14 0.04, and the narrow dis-
tribution of values demonstrates that there are no electronic
substituent effects on the pKa. This result was somewhat sur-
prising as we originally expected that both the pKa and the
kinetics of ring opening would be influenced by the electronic
properties of the spirolactam. The uniformity in pKa values
may be a fortuitous consequence of the relationship between
fluorophore concentration and rate constant, which is explored
further below.
RSL derivatives 1c–8c (Scheme 1) are designed to probe the
influence of electronic factors on the acid-induced spirocyclic
ring-opening reaction. The electronics of the spirolactam ring
system are manipulated via para substitution of an aniline-
derived phenyl ring that is attached to the spirolactam nitro-
gen. The selected aniline substituents span a range of elec-
tronic properties and possess Hammett σp constants between
−0.27 (OMe) and 0.78 (NO2).23 Related aniline-derived rhod-
amine deoxy-lactams were reported by Scott and coworkers as
the present manuscript was in final preparation.24 As shown in
Scheme 1, 1c–8c were prepared from commercially available
rhodamine B and the indicated anilines. The reaction pro-
ceeded via the rhodamine B acid chloride, which was gene-
rated in situ by addition of phosphorus oxychloride to a stirred
solution of rhodamine B and the desired aniline. The products
were obtained in 40–92% yield following purification by flash
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
Fig. 2 Normalized fluorescence intensity at 585 nm as a function of pH
for compounds 1 (circles) and 8 (squares) in 1 : 1 v/v ethanol–water.
With the similarities across the series established, we
measured the kinetics of the spirocyclic ring-opening reaction
by monitoring the fluorescence intensity at 585 nm as HCl was
introduced to a stirring solution of each derivative. All deriva-
tives exhibited an increase in fluorescence intensity that was
consistent with first order kinetics. Fig. 3A shows kinetic
traces for 1–8, plotted in the form ln(I∞ − It) versus time, where
I∞ and I0 are the fluorescence intensities at times infinity and
t, respectively. When plotted in this manner, first order kinetic
data are linear with slope = −k.26 Fig. 3A illustrates both the
first order nature of the data as well as the clear variation in
rate across the series. Rate constants for 1–8 are given in
Table 1. The slowest kinetics are observed for 1 (OMe), and the
rate of reaction increases across the series as the electron-with-
drawing nature of the substituent increases. The rate constant
of the fastest derivative, 8 (NO2), is more than four times
greater than that of 1. The substituent electronics clearly
impact the rate of the spirolactam ring-opening reaction.
The observed reaction kinetics can be explained by the
mechanism shown in Scheme 2, which is comprised of separ-
ate protonation and ring-opening steps. Protonation of the
spirolactam carbonyl yields a resonance-stabilized intermedi-
ate in which the positive charge is delocalized over oxygen,
carbon, and nitrogen. Electronic rearrangement breaks down
this intermediate, opening the spirolactam and producing a
fully conjugated rhodamine dye. The rate constants involved in
the two steps are defined in eqn (1), where CF and OF denote
the closed and open forms of the dye, respectively. With 3 dye
species and 4 rate constants involved, the reaction kinetics
cannot be described by a simple first order expression.
Fig. 3 (A) Kinetic data for compounds 1–8 in 1 : 1 (v/v) ethanol–water.
Acid was added to the solutions at time = 0 s. (B) Hammett plot of the
log of the relative rate constant versus σ−.
Table 1 Rate constants and relative fluorescence intensity and
absorbance
Derivative
σ−
kR (s−1
)
IR/IH
AR/AH
1 –OCH3
2 –t-C4H9
3 –CH3
4 –H
5 –Cl
6 –CF3
7 –CN
−0.27
−0.20
−0.17
0
0.25
0.65
1
0.0240 0.0009
0.0250 0.0016
0.0264 0.0008
0.0323 0.0007
0.0386 0.0012
0.0518 0.0017
0.0745 0.0017
0.1041 0.0021
0.23
0.61
0.72
1.00
1.34
4.96
5.27
3.89
0.53
0.53
0.70
1.00
1.39
4.84
5.06
5.60
8 –NO2
1.27
k1
k2
þ
þ
CF þ Hþ )ꢀ CFH )ꢀ OFH
ð1Þ
ꢀ*
ꢀ*
k
ꢀ2
k
ꢀ1
Eqn (2) demonstrates that the observed first order rate con-
stant, k, incorporates all 4 rate constants from eqn (1) as well
as the concentration of acid (i.e., k = Q + k−2). Given that k1 and
k−1 are expected to be rapid22 and essentially the same for all
derivatives and [H+] is identical in all kinetic experiments, the
observed differences in k arise from structure-dependent
differences in k2 and k−2. Bossi and coworkers previously esti-
mated k−2 for a different series of RSLs and found it to be
Rather, the rate expression for the production of the open,
fluorescent form of the dye is given by eqn (2), which is
derived in the ESI.†
Q½Dꢁ
ꢀ2Þt
½OFHþꢁ ¼
ð1 ꢀ eꢀðQ þ k
Þ
ð2Þ
Q þ k
ꢀ2
k2k1½Hþꢁ
where Q ¼
and [D] = [CF] + [CFH+] + [OFH+].
k
ꢀ1 þ k1½Hþꢁ
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
Scheme 2
approximately equal for their compounds.22 If the same is true
here, then differences in the observed k would be dominated
by k2.
A Hammett analysis further demonstrates the relationship
between substituent electronics and ring-opening reaction
Fig. 4 Chemical shift of the carbonyl carbons of 1c–8c in CDCl3
versus σ−.
kinetics in this system. We initially plotted the log of the rela- σ− < 0.4. The chemical environment thus appears to be
tive rate constant versus σp and observed a linear correlation of affected by steric interactions in addition to the expected influ-
rate constants with electron-donating and weakly electron- ence of substituent electronics; these complex influences on
withdrawing substituents (not shown). Deviations from linear- the chemical environment may explain the scatter observed in
ity were observed for the strongly electron-withdrawing substi- Fig. 4. When 2c is excluded from the analysis, a linear fit of
tuents, which exhibited rate constants that were larger than chemical shift versus σ− gives r = 0.959. Overall, the general
predicted. Fig. 3B shows the revised Hammett plot with the correlation of carbonyl chemical shift with σ− demonstrates
log of the relative rate constant versus σ−, the Hammett substi- that spirolactam electronics can be manipulated by the substi-
tuent constant used when resonance between the reaction site tuents in a predictable way.
and electron-withdrawing substituents is important.23,27 This
The consequences of the tunable ring-opening kinetics
plot is linear across all substituents (r = 0.997) and gives reac- observed here extend beyond just reaction rates. The ideal fluo-
tion constant ρ = 0.40 0.01. A positive reaction constant for rescent pH probe would not only have a fast response to pH
this system indicates that positive charge is diminished during change but also be bright and, therefore, easy to detect.
the course of the reaction. This information is consistent with Table 1 presents the peak fluorescence intensities at pH 3 of
Scheme 2, where the positively-charged intermediate is trans- 1o–8o relative to unsubstituted 4o, and the data demonstrate
formed to the rhodamine form of the dye during the rate- surprising variation across the series. The most intense deriva-
determining reaction step. Electron-withdrawing substituents tive, 7o (CN), is nearly 23 times more intense than the
destabilize the positively-charged intermediate both induc- dimmest, 1o (OMe). With the exception of 8o (NO2), the fluo-
tively and via resonance, where possible, while electron-donat- rescence intensities increase as the electron-withdrawing
ing substituents have the opposite effect. Electronic nature of the substituent increases. Thus, the fastest-reacting
destabilization of the intermediate is expected to accelerate k2.
probes are also the brightest. The reason for this dramatic vari-
The Hammett analysis implies that the RSL substituents ation in fluorescence intensity becomes apparent upon exam-
affect the rate of ring-opening by altering the electronics of the ination of the relative peak absorbance values for 1o–8o, also
spirolactam ring. Additional support for this idea comes from measured at pH 3 and shown in Table 1. The fluorescence
NMR evidence that the electronic environment of the spirolac- intensities of derivatives 2o–7o are highly correlated with their
tam ring does indeed change as a function of the substituent. absorbances. Derivatives 1o (OMe) and 8o (NO2) both have
Fig. 4 shows the 13C NMR chemical shifts of the carbonyl lower fluorescence intensities than predicted by their absor-
carbon for each derivative in its closed form (1c–8c) versus σ−. bances, an effect that may be due to substituent-specific
The chemical shift values are clearly correlated with the substi- effects on the fluorescence quantum yield. Absorbance varies
tuent constants: the carbonyl carbon becomes deshielded as by a factor of 10 across the series, and, according to Beer’s
the electron-withdrawing power of the substituent increases. Law, this trend can only be due to a substituent-dependent
The trend is not as linear as in Fig. 3B, and substituents tert- variation in extinction coefficient or concentration. Given the
butyl (2) and methyl (3) were included in the series so that the improbability that substitution of the same chromophore
potential role of steric interactions in the scatter could be could cause such a large difference in extinction coefficients,
explored. In terms of the ring-opening reaction, steric effects concentration differences emerge as the most likely expla-
derived from the para substituents do not appear to play a role nation for this trend.
as 2 and 3 both yielded nearly identical rate constants consist-
The apparent increase in probe concentration with increas-
ent with their extremely similar σ− values (Fig. 3B, Table 1). In ing electron-withdrawing nature of the substituent is un-
contrast, the carbonyl chemical shift of 2c deviates noticeably expected as 1–8 all have essentially the same pKa. According to
from 3c, which is more in line with the other substituents with the standard Henderson-Hasselbalch equation (eqn (3)),
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
concentrations of the two forms of an acid-base indicator, In
and HIn+, are equal when pH = pKa. Our observed results
clearly fail to match this expectation as the variation in absor-
bance is the same whether measured at pH = pKa or other pH
values.
ꢀ
ꢁ
½lnꢁ
pH ¼ pKa þ log
ð3Þ
½H lnþꢁ
Eqn (3) applies to a typical pH indicator that has only 2
forms, one colorless and one colored. As shown in Scheme 2
and eqn (1), RSL-based indicators are more complex, with two
colorless species (CF and CFH+) and one colored species
(OFH+). Given the Scheme 2 mechanism and standard equili-
brium assumptions, a new expression relating pH, pKa, and
probe concentrations can be derived for this 3-species scenario
(derivation shown in ESI†). The result is given as eqn (4),
where the terms were defined in eqn (1).
Fig. 5 Kinetic data for compounds 8 (squares) and 9 (circles) in 1 : 1
(v/v) ethanol–water. Acid was added to the solutions at time = 0 s. Inset:
normalized peak fluorescence intensity at 585 nm as a function of pH
for compounds 8 (squares) and 9 (circles).
ꢀ
ꢁ
k2½CFꢁ
series. At early times following the addition of acid, the kine-
tics of 9 are well-described by first order kinetics, as can be
seen by the linear nature of the trace in Fig. 5. The kinetics at
later times deviate from first order, an effect that could be due
to the additional steric constraints of 9 as compared to 1–8. A
first order fit of the early-time data for 9 yields a rate constant
that is more than 16 times faster than 8. In practical terms,
the observed kinetics represent the difference between an
acceptably fast ring-opening for 8 with half of the fluorescence
appearing in just over 6 s and an almost instantaneous reac-
tion for 9 with half of the fluorescence appearing in under
1 s. In addition to rapid ring-opening kinetics, 9o also possesses
a fluorescence intensity exceeding that of 1o–8o. Relative to 4o,
its fluorescence intensity is 6.67, which makes it the brightest
of the RSLs studied here. The concentration of 9o as expressed
by its absorbance is even higher at 10.43 relative to 4o. It
appears at the highest concentration and greatest fluorescence
intensity of all of the derivatives though, like 8o (NO2), its fluo-
rescence is less intense than predicted by absorbance.
Our finding that electron withdrawing substituents acceler-
ate the ring-opening kinetics of 9 indicates that the powerful
electron-withdrawing effect of the one nitro and two chloro
substituents is at least partially responsible for its fast reac-
tion. However, unfavorable steric interactions due to the ortho
substituents in 9 most likely also play a role in the rapid kine-
tics. Titration of 9 reveals that it has a pKa of 5.43, which is dis-
tinctly higher than the average pKa of 4.14 observed for 1–8.
The Fig. 5 inset demonstrates the significant difference in
titration response between 9 and 8, again selected as a refer-
ence. Lin and coworkers previously demonstrated that RSL pKa
values can be increased by functionalizing the spirolactam
nitrogen with substituents that are extremely sterically
demanding, such as adamantyl.7 Although pKa and kinetics
are separate issues, it seems likely that unfavorable steric inter-
actions that are significant enough to shift the pKa of 9 also
play a role in its rapid ring-opening kinetics. We can estimate
the extent to which 9’s fast response is due to electronic versus
steric factors by evaluating 9 in terms of the Hammett plot for
pH ¼ pKa þ log
ð4Þ
k
½OFHþꢁ
ꢀ2
Not only species concentrations but also rate constants
appear in this expression. As discussed previously, our
measured rate constants represent the sum of k−2 and an
expression involving k2, which complicates comparison of
experimental data to eqn (4), in which k2 and k−2 appear as a
simple ratio. The observed rate constant increases by a factor
of 4 as the concentration of OFH+, represented by absorbance,
increases by a factor of 10 across the series 1–8. This result is
in agreement with the correlated increase of k2 and concen-
tration of open form predicted by eqn (4). The variation in
magnitude of increase between rate constant and absorbance
is most likely due to the complexity of what the measured k
represents and/or small variations in extinction coefficients
among the derivatives. Eqn (4) supports our conclusion that
the fastest-reacting derivatives are also the brightest. Corre-
lation of rate constant with equilibrium constant for another
series of RSL derivatives has also been observed by Bossi.22
Having established that the fastest-reacting and brightest
probes are those with electron-withdrawing substituents on
the spirolactam ring, we next applied this finding to the
design of an RSL pH probe with superior properties. Probe 9,
shown in its open form below, is derived from 2,6-dichloro-4-
nitroaniline, which is expected to exert a much stronger elec-
tron-withdrawing substituent effect on the spirolactam moiety
than 8, which bears only the nitro substituent.
Fig. 5 shows the kinetic trace for 9 along with 8, the fastest-
reacting of the original series, for reference. The kinetics of 9
are significantly faster than those observed for the original
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
1–8. An estimated Hammett constant for the 2,4-dichloro-6- containing a stirring solution of 1c–9c while the fluorescence
nitro moiety was calculated by adding together Traynham’s σo intensity at 585 nm was monitored upon 535 nm excitation.
value of 0.50 for each of the o-chloro substituents and the σ− The acid concentration in the cuvette was 6.6 mM, and the pH
value for the p-nitro substituent to yield 2.27.28 We used this was 2.7. Reported rate constants are the average of a minimum
Hammett constant value and the ρ calculated for Fig. 3B to of 9 total kinetic runs from at least 3 different samples.
determine an expected rate constant for 9 based solely upon Absorption and fluorescence titrations were conducted by
the electronics of its substituents. This analysis yielded a value adding small aliquots of 2 M HCl to a stirring stock solution
of 0.25 s−1 for the rate constant, while the actual value of 1.74 (100 mL) of 1c–9c. Aliquots were removed for spectroscopic
s−1 is nearly 7 times greater in magnitude. Thus, the majority measurement (fluorescence λexc = 535 nm) after the pH stabil-
of 9’s fast response is likely due to unfavorable steric factors ized and were then returned to the stock solution.
while a smaller proportion is attributable to the strong elec-
Materials and methods
tron-withdrawing effect of the substituents. Nevertheless, our
findings regarding the kinetics of spirolactam ring opening All reactions were carried out with flame-dried glassware
enabled us to design the fastest-reacting and most intensely under an argon atmosphere. Reagents were purchased from
1
fluorescent probe of this series.
Acros or Sigma-Aldrich and used as received. H and proton-
decoupled 13C NMR spectra were recorded on a Varian
Mercury 400 (1H 400 MHz, 13C 100 MHz) and referenced to
TMS (1H) or CDCl3 at 77.0 ppm (13C). Chemical shifts are
reported in ppm and coupling constants in Hz. Mass spectra
Conclusions
Rate of response to the desired analyte is a key figure of merit were measured through positive electrospray ionization
for any fluorescent turn-on probe. We have demonstrated that (w/NaCl) on a Bruker 12 Tesla APEX-Qe FTICR-MS with and
pH probes based on a rhodamine spirolactam structure Apollo II ion source.
convert from non-fluorescent to fluorescent at a rate that
depends on the electronics of the spirolactam ring, with the
General procedure for the synthesis of compounds 1c–9c
fastest fluorescence turn-on behavior observed for dyes with To a stirred solution of rhodamine B (0.21 mmol) and the
electron-withdrawing substituents. In addition to reaction rate, desired aniline (0.63 mmol) in 1,2-dichloroethane (5 mL) at
fluorescence intensity was also observed to be correlated with 0 °C, phosphorus oxychloride (0.25 mmol) was added drop-
substituent electronics. The fastest-reacting derivatives were wise under argon atmosphere. The mixture was stirred at 0 °C
also the most intensely fluorescent, and this effect was due to for 15 min, heated to 85 °C for 5 h, and then cooled to room
an increase in fluorescent dye concentration. This variation in temperature. The solution was diluted with chloroform
concentration was shown to originate in the 3-species equili- (20 mL) and acidified with 2 M HCl (30 mL). The organic layer
brium for the RSL pH probes, which creates a scenario in was washed with additional 2 M HCl (2 × 30 mL), 2 M NaOH
which larger rate constants yield larger concentrations of the (3 × 30 mL), and brine (30 mL). The organic layer was dried
fluorescent form of the dye. These findings were used to with Na2SO4 and concentrated in vacuo.
design a faster-reacting, more intensely fluorescent pH probe
1c: Compound 1c was reported previously.24 The brown oil
derived from 2,4-dichloro-6-nitroaniline. This probe also had a was purified by flash column chromatography on silica gel
much higher pKa than those with only para substituents on (elution: 40% to 60% EtOAc in hexanes) to afford 1 (72.8 mg,
the aniline (5.43 vs. 4.14) in agreement with the findings of 64% yield) as a pale tan solid: TLC (40% EtOAc in hexanes)
Lin that sterically demanding substituents raise the pKa of Rf: 0.45 (UV); 1H NMR (400 MHz, CDCl3) 8.01 (m, 1H),
RSL-based pH probes.7 Collectively, these results demonstrate 7.50–7.48 (t, J = 3.5 Hz, 2H), 7.18–7.16 (m, 1H), 6.64–6.60 (m,
that RSLs can be easily functionalized to furnish rapid-reacting 6H), 6.33–6.30 (dd, J = 8.8, 2.5 Hz, 2H), 6.25–6.24 (d, J = 2.3 Hz,
pH probes that are responsive within the so-called acidic 2H), 3.67 (s, 3H), 3.35–3.29 (qd, J = 7.0, 2.7 Hz, 8H), 1.16–1.13
window.
(t, J = 7.0 Hz, 12H); 13C NMR (100 MHz, CDCl3) 167.5, 158.1,
153.1, 152.9, 148.6, 132.6, 131.3, 128.95, 128.85, 128.0, 124.0,
123.2, 113.8, 107.9, 106.2, 97.6, 67.3, 55.1, 44.2, 12.5; HRMS
(ES+): Exact mass calcd for C35H37N3O3 [M + Na]+, 570.2727.
Found 570.2729.
Experimental section
Absorbance and fluorescence studies
2c: The brown solid was purified by flash column chromato-
Absorbance and fluorescence measurements were made on a graphy on silica gel (elution: 5% to 40% EtOAc in hexanes)
Varian Cary50 and Varian Eclipse, respectively. Stock solutions and trituration from hexanes to afford 2 (111.4 mg, 94% yield)
of 1c–9c were prepared in 1 : 1 (v/v) ethanol–water at a concen- as a white powder: TLC (20% EtOAc in hexanes) Rf: 0.35 (UV);
tration of 7.7 μM (1–5) or 0.077 μM (6–9) and were stirred at 1H NMR (400 MHz, CDCl3) 7.99 (m, 1H), 7.47–7.45 (t, J =
room temperature for 1 h prior to absorbance and fluo- 3.5 Hz, 2H), 7.13–7.11 (m, 3H), 6.78–6.76 (d, J = 8.2 Hz, 2H),
rescence studies.16 Data for compounds 6–9 were corrected for 6.66–6.64 (d, J = 8.6 Hz, 2H), 6.32–6.29 (d, J = 9.0 Hz, 2H), 6.28
the additional dilution to facilitate comparison with 1–5. For (s, 2H), 3.32–3.31 (m, 8H), 1.20 (s, 9H), 1.16–1.13 (t, J = 7.0 Hz,
kinetic measurements, 2 M HCl was added to a cuvette 12H); 13C NMR (100 MHz, CDCl3) 167.8, 153.5, 152.9, 148.9,
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
148.6, 133.8, 132.7, 130.7, 128.8, 127.9, 126.3, 125.4, 123.8, 97.7, 67.3, 44.2, 12.4; HRMS (ES+): Exact mass calcd for
123.2, 108.0, 105.5, 97.7, 67.2, 44.2, 34.3, 31.2, 12.5, 1.0; HRMS C35H34F3N3O2 [M + Na]+, 608.2495. Found 608.2497.
(ES+): Exact mass calcd for C38H43N3O2 [M + Na]+, 596.3247.
Found 596.3250.
7c: The magenta oil was purified by flash column chrom-
atography on silica gel (elution: 0% to 5% MeOH in CHCl3) to
3c: Compound 3c was reported previously.24 The red oil was afford 7 (208.3 mg, 92% yield) as a tan solid: TLC (2% MeOH
1
purified by flash column chromatography on silica gel in CHCl3) Rf: 0.30 (UV); H NMR (400 MHz, CDCl3) 7.98 (d, J =
(elution: 0% to 2% MeOH in CHCl3) to afford a brown solid 6.3 Hz, 1H), 7.45–7.43 (m, 2H), 7.37–7.36 (d, J = 7.0 Hz, 2H),
residue. Trituration with Et2O yielded 3 (89.1 mg, 40% yield) 7.28–7.26 (d, J = 8.6 Hz, 2H), 7.08 (d, J = 6.3 Hz, 1H), 6.58–6.56
as a pale tan powder: TLC (2% MeOH in CHCl3) Rf: 0.25 (UV); (d, J = 9.0 Hz, 2H), 6.34–6.33 (d, J = 2.0 Hz, 2H), 6.27–6.24 (d,
1H NMR (400 MHz, CDCl3) 8.02–7.99 (m, 1H), 7.50–7.46 (m, J = 9.0 Hz, 2H), 3.32–3.27 (q, J = 6.6 Hz, 8H), 1.15–1.11 (t, J =
2H), 7.16–7.15 (m, 1H), 6.93–6.91 (d, J = 7.0 Hz, 2H), 6.66–6.64 6.6 Hz, 12H); 13C NMR (100 MHz, CDCl3) 168.0, 153.5, 152.5,
(d, J = 8.2 Hz, 4H), 6.33–6.30 (d, J = 8.6 Hz, 2H), 6.26 (s, 2H), 148.7, 141.5, 133.4, 132.3, 129.1, 128.1, 128.0, 124.9, 123.7,
3.34–3.29 (q, J = 7.0 Hz, 8H), 2.21 (d, J = 1.2 Hz, 3H), 1.17–1.13 123.3, 118.7, 108.4, 108.1, 105.6, 97.6, 67.3, 44.1, 12.4; HRMS
(t, J = 7.0 Hz, 12H); 13C NMR (100 MHz, CDCl3) 167.6, 153.2, (ES+): Exact mass calcd for C35H34N4O2 [M + Na]+, 565.2574.
153.1, 148.6, 136.3, 133.7, 132.6, 131.1, 129.2, 128.9, 128.0, Found 565.2576.
127.3, 123.9, 123.3, 108.0, 106.4, 97.7, 67.2, 44.3, 21.1, 12.5;
HRMS (ES+): Exact mass calcd for C35H37N3O2 [M + Na]+, graphy on silica gel (elution: 40% to 50% EtOAc in hexanes) to
554.2778. Found 554.2779. afford 8 (97.2 mg, 83% yield) as a yellow solid: TLC (40%
8c: The red oil was purified by flash column chromato-
4c: Compound 4c was reported previously.22,24 The brown EtOAc in hexanes) Rf: 0.70 (UV); 1H NMR (400 MHz, CDCl3)
solid was purified by flash column chromatography on silica 8.01–7.99 (m, 3H), 7.53–7.46 (m, 2H), 7.40–7.37 (dd, J = 7.0,
gel (elution: 40% to 50% EtOAc in hexanes) to afford 4 2.0 Hz, 2H), 7.11 (d, J = 6.6 Hz, 1H), 6.58–6.56 (d, J = 8.6 Hz,
(58.4 mg, 54% yield) as a white powder: TLC (40% EtOAc in 2H), 6.34–6.33 (d, J = 2.7 Hz, 2H), 6.27–6.24 (dd, J = 8.8, 2.5 Hz,
hexanes) Rf: 0.60 (UV); 1H NMR (400 MHz, CDCl3) 8.00 (m, 2H), 3.34–3.29 (q, J = 7.0 Hz, 8H), 1.17–1.13 (t, J = 7.0 Hz,
1H), 7.50 (m, 2H), 7.17–7.16 (m, 2H), 7.12 (m, 3H), 6.81–6.78 12H); 13C NMR (100 MHz, CDCl3) 168.4, 153.8, 152.6, 148.9,
(d, J = 6.6 Hz, 2H), 6.65–6.63 (dd, J = 8.6/1.6 Hz, 2H), 6.31–6.29 144.4, 143.6, 133.7, 129.0, 128.3, 128.1, 124.4, 124.0, 123.8,
(d, J = 9.0 Hz, 2H), 6.25 (s, 2H), 3.34–3.29 (q, J = 7.2 Hz, 8H), 123.5, 108.2, 105.7, 97.7, 67.6, 44.3, 12.5; HRMS (ES+):
1.16–1.3 (t, J = 7.0 Hz, 12H); 13C NMR (100 MHz, CDCl3) 167.6, Exact mass calcd for C34H34N4O4Na [M + Na]+, 585.2472.
153.2, 153.1, 148.7, 136.6, 132.8, 131.0, 128.8, 128.5, 128.1, Found 585.2474.
127.3, 126.6, 124.0, 123.3, 108.0, 106.4, 97.7, 67.4, 44.3, 12.5;
HRMS (ES+): Exact mass calcd for C34H35N3O2 [M + Na]+, atography on silica gel (elution: 0% to 3% MeOH in CHCl3) to
540.2621. Found 540.2623. afford 9 (161.3 mg, 61% yield) as an orange solid: TLC (2%
9c: The purple solid was purified by flash column chrom-
5c: Compound 5c was reported previously.24 The brown oil MeOH in CHCl3) Rf: 0.30 (UV); 1H NMR (400 MHz, CDCl3) 8.07
was purified by flash column chromatography on silica gel (d, J = 7.8 Hz, 1H), 8.0 (d, J = 1.2 Hz, 1H), 7.71 (t, J = 7.4 Hz,
(elution: 0% to 4% MeOH in CHCl3) to afford 5 (114.6 mg, 1H), 7.65 (t, J = 7.4 Hz, 1H), 7.41 (d, J = 7.4 Hz, 1H), 7.26 (d, J =
50% yield) as a tan solid: TLC (2% MeOH in CHCl3) Rf: 0.25 6.6 Hz, 1H), 6.61–6.58 (d, J = 9.0 Hz, 2H), 6.31–6.28 (dd, J = 9.0,
(UV); 1H NMR (400 MHz, CDCl3) 8.00 (m, 1H), 7.50–7.48 (m, 2.3 Hz), 6.23 (d, J = 2.0 Hz, 2H), 3.36–3.30 (q, J = 7.0 Hz, 8H),
2H), 7.15 (m, 1H), 7.11–7.08 (dd, J = 8.8, 2.5 Hz, 2H), 6.80–6.77 1.16–1.12 (t, J = 6.8 Hz, 12H); 13C NMR (100 MHz, CDCl3)
(dd, J = 8.8, 2.1 Hz, 2H), 6.63–6.60 (dd, J = 8.8, 2.1 Hz, 2H), 196.0, 155.4, 149.5, 149.3, 146.9, 138.7, 138.4, 133.0, 132.0,
6.32–6.28 (m, 4H), 3.35–3.30 (q, J = 6.9 Hz, 8H), 1.17–1.13 (t, 130.3, 128.9, 124.9, 124.0, 123.1, 107.6, 107.4, 97.9, 71.2, 44.4,
J = 7.0 Hz, 12H); 13C NMR (100 MHz, CDCl3) 167.7, 153.2, 12.4; HRMS (ES+): Exact mass calcd for C34H32Cl2N4O4
153.0, 148.7, 135.3, 133.0, 132.1, 130.6, 128.73, 128.67, 128.3, [M + Na]+, 653.1693. Found 653.1698.
128.2, 124.0, 123.3, 108.1, 106.0, 97.7, 67.4, 44.3, 12.5; HRMS
(ES+): Exact mass calcd for C34H34ClN3O2 [M + Na]+, 574.2232.
Found 574.2234.
Acknowledgements
6c: Compound 6c was reported previously.24 The red oil was
purified by flash column chromatography on silica gel We gratefully acknowledge support of this work by the
(elution: 0% to 2% MeOH in CHCl3) to afford 6 (209 mg, 85% Camille and Henry Dreyfus Foundation through
a
yield) as a pale tan solid: TLC (2% MeOH in CHCl3) Rf: 0.35 Henry Dreyfus Teacher-Scholar Award and by the William and
(UV); 1H NMR (400 MHz, CDCl3) 8.01 (m, 1H), 7.47 (t, J = Mary Women in Scientific Education (WISE) Initiative
3.3 Hz, 2H), 7.39–7.37 (d, J = 8.6 Hz, 2H), 7.18–7.16 (d, J = supported by the National Science Foundation. R.M.A. thanks
8.2 Hz, 2H), 7.12 (m, 1H), 6.64–6.62 (d, J = 9.0 Hz, 2H), the Charles Center of the College of William and Mary for
6.33–6.32 (d, J = 2.3 Hz 2H), 6.31–6.28 (dd, J = 9.0/2.3 Hz, 2H), support via an Honors Fellowship. Partial support for
3.32–3.29 (q, J = 6.8 Hz, 8H), 1.16–1.13 (t, J = 7.0 Hz, 12H); this work was provided by student research fellowships to
13C NMR (100 MHz, CDCl3) 168.0, 153.5, 152.7, 148.7, 140.3, W.L.C. and C.A.R. from the Howard Hughes Medical Institute
133.2, 129.7, 128.4, 128.2, 127.5 (2J = 32.4 Hz), 125.7, 125.6 Undergraduate Science Education grant to the College of
(3J = 3.7 Hz), 124.0 (1J = 272 Hz), 123.8, 123.3, 108.1, 105.8, William and Mary.
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
14 A. Augustyniak and W. Ming, Prog. Org. Coat., 2011, 71,
References
406.
1 X. Chen, T. Pradhan, F. Wang, J. S. Kim and J. Yoon, Chem. 15 M. Tian, X. Peng, J. Fan, J. Wang and S. Sun, Dyes Pigm.,
Rev., 2012, 112, 1910.
2012, 95, 112.
2 J. Han and K. Burgess, Chem. Rev., 2009, 110, 2709.
3 M. Beija, C. A. M. Afonso and J. M. G. Martinho, Chem. Soc.
Rev., 2009, 38, 2410.
4 H. Li, H. Guan, X. Duan, J. Hu, G. Wang and Q. Wang, Org.
Biomol. Chem., 2013, 11, 1805.
16 Q. A. Best, R. Xu, M. E. McCarroll, L. Wang and D. J. Dyer,
Org. Lett., 2010, 12, 3219.
17 V. B. Bojinov, A. I. Venkova and N. I. Georgiev, Sens. Actua-
tors, B, 2009, 143, 42.
18 J. Hu, L. Dai and S. Liu, Macromolecules, 2011, 44, 4699.
5 T. Hasegawa, Y. Kondo, Y. Koizumi, T. Sugiyama, 19 M. Adamczyk and J. Grote, Bioorg. Med. Chem. Lett., 2003,
A. Takeda, S. Ito and F. Hamada, Biorg. Med. Chem., 2009,
17, 6015.
6 W. Zhang, B. Tang, X. Liu, Y. Liu, K. Xu, J. Ma, L. Tong and
G. Yang, Analyst, 2009, 134, 367.
13, 2327.
20 K. Sreenath, R. J. Clark and L. Zhu, J. Org. Chem., 2012, 77,
8268.
21 V. N. Belov and M. L. Bossi, Isr. J. Chem., 2013, 53, 267.
7 L. Yuan, W. Lin and Y. Feng, Org. Biomol. Chem., 2011, 9, 22 H. Montenegro, M. Di Paolo, D. Capdevila, P. F. Aramendia
1723.
and M. L. Bossi, Photochem. Photobiol. Sci., 2012, 11, 1081.
23 C. Hansch, A. Leo and R. W. Taft, Chem. Rev., 1991, 91,
165.
24 Q. A. Best, C. Liu, P. D. van Hoveln, M. E. McCarroll and
C. N. Scott, J. Org. Chem., 2013, 78, 10134.
25 W. C. Still, M. Kahn and A. Mitra, J. Org. Chem., 1978, 43,
2923.
26 K. A. Connors, Chemical Kinetics: The Study of Reaction
Rates in Solution, VCH, New York, 1990.
27 H. H. Jaffé, Chem. Rev., 1953, 53, 191.
8 Z. Li, S. Wu, J. Han and S. Han, Analyst, 2011, 136, 3698.
9 Z.-Q. Hu, M. Li, M.-D. Liu, W.-M. Zhuang and G.-K. Li, Dyes
Pigm., 2013, 96, 71.
10 N. B. Yapici, S. R. Mandalapu, T.-L. Chew, S. Khuon and
L. Bi, Bioorg. Med. Chem. Lett., 2012, 22, 2440.
11 H. Zhu, J. Fan, Q. Xu, H. Li, J. Wang, P. Gao and X. Peng,
Chem. Commun., 2012, 48, 11766.
12 S. Wu, Z. Li, J. Han and S. Han, Chem. Commun., 2011, 47,
11276.
13 S. Kang, S. Kim, Y.-K. Yang, S. Bae and J. Tae, Tetrahedron 28 M. T. Tribble and J. G. Traynham, J. Am. Chem. Soc., 1969,
Lett., 2009, 50, 2010.
91, 379.
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2014