Thioamide-based fluorescent protease sensors

Article abstract of DOI:10.1021/ja412297x

Thioamide quenchers can be paired with compact fluorophores to design "turn-on" fluorescent protease substrates. We have used this method to study a variety of serine-, cysteine-, carboxyl-, and metallo-proteases, including trypsin, chymotrypsin, pepsin,

Full text of DOI:10.1021/ja412297x

Article
pubs.acs.org/JACS
Thioamide-Based Fluorescent Protease Sensors
Jacob M. Goldberg,†,§ Xing Chen,† Nataline Meinhardt,‡ Doron C. Greenbaum,‡
and E. James Petersson*,†
†Department of Chemistry, University of Pennsylvania, 231 South 34th Street, Philadelphia, Pennsylvania 19104-6323, United States
‡Department of Pharmacology, Perelman School of Medicine, University of Pennsylvania, 433 South University Avenue, Philadelphia,
Pennsylvania 19104-6019, United States
S
* Supporting Information
ABSTRACT: Thioamide quenchers can be paired with compact
fluorophores to design “turn-on” fluorescent protease substrates. We
have used this method to study a variety of serine-, cysteine-,
carboxyl-, and metallo-proteases, including trypsin, chymotrypsin,
pepsin, thermolysin, papain, and calpain. Since thioamides quench
some fluorophores red-shifted from those naturally occurring in
proteins, this technique can be used for real time monitoring of
protease activity in crude preparations of virtually any protease. We
demonstrate the value of this method in three model applications:
(1) characterization of papain enzyme kinetics using rapid-mixing
experiments, (2) selective monitoring of cleavage at a single site in a peptide with multiple proteolytic sites, and (3) analysis of
the specificity of an inhibitor of calpain in cell lysates.
derivatives bearing free amines, and enzymatic hydrolysis of the
amide bond can result in a substantial gain in fluorescence.10−13
Unfortunately, this type of probe must be installed directly at
the site of proteolysis and requires the protease to tolerate a
large aromatic dye in its active site. Furthermore, since these
probes must be installed at the P1′ site, the specificity of the S1′
pocket cannot be easily exploredspecificity can only be
conferred from the S1, S2, S3, and other pockets. Although
these constraints can be relaxed to some extent by the use of 6-
amino-1-naphthalenesulfonamides, which can allow for the
incorporation of alkyl chains as short spacers, these methods
are limited in scope by the fact that the P1′ site must be a
fluorogenic probe.14,15
INTRODUCTION
■
Proteases are a diverse class of enzymes that catalyze the
hydrolysis of peptide bonds with varying degrees of specificity.
They serve important physiological roles, especially in
metabolic pathways, signaling cascades, and regulatory
processes, though they have also been implicated in a variety
of disease pathologies, including cancer, hypertension, viral
infections (e.g., hepatitis C, HIV, and malaria), and neuro-
degeneration.1−4 As such, proteases have received a consid-
erable amount of clinical, commercial, and academic attention,
and numerous techniques have been developed to monitor
their activity or probe their specificity.5,6 Chief among these
techniques is fluorescence spectroscopy, which can provide
results in real time and is amenable to high throughput
methodology, although radioactive, chromatographic, and
colorimetric assays are also common.7
Fluorescence experiments often require that a peptide or
protein substrate be labeled with at least one fluorophore that is
selectively excitable in the presence of Trp and Tyr, since these
residues are common in proteolytic enzymes. The proteolysis
of substrates bearing a single fluorescent label can be monitored
with fluorescence anisotropy measurements if the signal of the
intact substrate is sufficiently different from that of the cleavage
product. For this method to be useful, the absolute difference in
size between the substrate and cleavage products needs to be
rather substantial; short peptides are often too small to provide
accurate results.8 An alternative strategy is based on fluorogenic
molecules that can be covalently installed at the P1′ site of a
protease substrate.9 Some amidated fluorophores, such as 2-
napthylamides, 4-methyl-7-coumarylamides, rhodamine 110,
and various anilides are quenched relative to the corresponding
Protease substrates labeled with two chromophores can be
used to overcome this restriction. Changes in distance-
dependent energy transfer between the labelseither through
Forster resonance energy transfer (FRET) or photoinduced
̈
electron transfer (PET) mechanismscan be used to monitor
protease activity.9,16 Typically, these probes are installed on
opposite ends of a short peptide sequence and interact through
FRET or PET. Upon cleavage of the intervening sequence, the
fluorophore interaction is lost as the fragments diffuse through
solution. FRET based sensors can sometimes be used as
ratiometric probes if both chromophores are fluorescent; PET-
quenched substrates typically give a fluorogenic response. A
major limitation of these methods, however, is the requirement
that the substrate be labeled with two probes, which are often
Received: December 3, 2013
Published: January 28, 2014
© 2014 American Chemical Society
2086
dx.doi.org/10.1021/ja412297x | J. Am. Chem. Soc. 2014, 136, 2086−2093

Journal of the American Chemical Society
Article
bulky.17−19 The large size of the requisite probes may influence
or interfere with the kinetics of proteolysis.
To eliminate the problems associated with bulkier probes,
this sort of profluorescent reporter design can be adapted for
use with small thioamide quenchers (Figure 1). Previously, we
substrates to study a variety of serine-, cysteine-, carboxyl-, and
metallo-proteases. Furthermore, we demonstrate the value of
this method by showing that it can be used in rapid-mixing
experiments, monitoring of cleavage at selective sites, and
tracking specific activity in cell lysates.
RESULTS AND DISCUSSION
■
We began our investigation by preparing short target peptides
labeled with N-terminal thioamides and C-terminal Mcm.
Intervening amino acid sequences were chosen such that the
peptides would be recognized and cleaved by a variety of
proteases, including chymotrypsin, papain, pepsin, thermolysin,
and trypsin (Figure 2 and Figure S1, Supporting Information).
Figure 1. Profluorescent thiopeptides for monitoring protease activity.
Thioamide (denoted by the one or three letter code of the
corresponding natural amino acid with a prime symbol, e.g., L′)
substrates can be prepared on solid phase from benzotriazole
precursors and fluorescent amino acids such as 7-methoxycoumarin-
4-ylalanine (μ). Incubation of a coumarin/thioamide labeled peptide
with a protease results in cleavage and a concomitant gain of
fluorescence.
have shown that thioamides quench a variety of fluorophores,
including 7-methoxycoumarin and fluorescein, through a PET
mechanism.20−23 As probes, thioamides can be installed in
peptide backbones as single-atom substitutions in amide bonds,
and they are much smaller than almost any other conventional
quencher or fluorophore used in this type of experiment. In this
way, thioamides can be incorporated in positions where larger
probes would not be well tolerated by a protease. In principle,
thioamides could be scanned though an entire candidate
substrate sequence, thus providing more thorough, and perhaps
more accurate, information about proteolysis than other
methods would allow.
It is important to note that thioamide replacement of the
scissile bond in the substrate might affect proteolysis, so care
must be taken when examining this position.24−34 This was
seen in attempts to use thioamides as protease inhibitors and in
the two previous cases in which thioamides have been used as
probes of proteolysis. Bond et al. inserted a thioamide at the
scissile bond and monitored carboxypeptidase A activity by
disappearance of the thiocarbonyl absorbance at 270 nm.27 Cho
et al. tracked papain activity by the increase in fluorescence
from cleavage of a thioacyl anilide.29 In both cases, placing the
thioamide at the site of cleavage disrupted protease activity.
Our PET-based design allows one to place the fluorophore and
thioamide probes at sites that flank the scissile bond, and is thus
less perturbing and very general.
Figure 2. Protease substrate trials. Representative results for
chymotrypsin, pepsin, papain, and thermolysin experiments. The
fluorescence of the thiopeptide in the presence (red trace) and absence
(orange trace) of protease is shown with the corresponding
oxopeptide also in the presence (green trace) and absence (blue
trace) of protease. Chymotrypsin: fluorescence of 8.3 μM AAFAμ or
A′AFAμ in 100 mM Tris-HCl, pH 7.8, at 30 °C in the presence and
absence of 0.2 mg/mL chymotrypsin. Pepsin: fluorescence of 0.5 μM
AAFAμ or A′AFAμ in 10 mM HCl at 37 °C in the presence and
absence of 1.5 mg/mL pepsin. Papain: fluorescence of 8.3 μM
LLKAAμ or L′LKAAμ in 2.0 mM EDTA, 5.00 mM L-cysteine, 300
mM NaCl, pH 6.2, at 25 °C in the presence and absence of 2.5 μg/mL
papain. Thermolysin: fluorescence of 8.3 μM AAFAμ or A′AFAμ in 2
mM calcium acetate 10 mM sodium acetate, pH 7.5 at 25 °C in the
presence and absence of 6 μg/mL thermolysin.
In a typical experiment, we measured the fluorescence of the
peptide in the presence and absence of a protease as a function
of time. As a control, we also synthesized the corresponding
oxoamide version of each peptide and measured the
fluorescence of these samples as well. The concentrations of
peptide and protease were adjusted so that proteolysis reached
completion within 1 to 2 h. In the absence of any proteases, the
thioamide peptide A′AFAμ was quenched ∼65% relative to the
all oxoamide control peptide (AAFAμ) and the thioamide
peptide L′LKAAμ was quenched ∼40% relative to the control
peptide, LLKAAμ. Thioamide amino acid analogs are denoted
by the one or three letter code of the corresponding natural
amino acid with a prime (′) symbol. The fluorescence of both
We have recently reported an initial trial of this design using
fluorescein as a fluorescent donor.20 Here, we show that this
general strategy can be used to track the activity of a variety of
proteases using 7-methoxycoumarin-4-yl-alanine (Mcm; μ) as a
donor fluorophore. Mcm is only slightly larger than Trp, and
can be selectively excited at 325 nm in the presence of
endogenous Trp and Tyr residues in the analyte. Mcm is
superior to fluorescein as a donor because it is smaller, can be
easily incorporated during solid phase synthesis, and is more
quenched by thioamides.22 We have used Mcm/thioamide
2087
dx.doi.org/10.1021/ja412297x | J. Am. Chem. Soc. 2014, 136, 2086−2093

Journal of the American Chemical Society
Article
peptides remained constant over the timecourse of experiments
in the absence of protease, with slight decreases attributed to
photobleaching. We observed an immediate gain in fluo-
rescence upon addition of an appropriate protease to the
thiopeptides, but almost no change in fluorescence in the
oxoamide control experiments.
To ensure that the increases in fluorescence were the result
of protease activity, we repeated these experiments with heat-
deactivated proteases. The fluorescence of the thiopeptide in
the presence of a heat-deactivated protease was identical to that
of the peptide in pure buffer (Figure S3, Supporting
Information). To further confirm that the observed changes
were the result of proteolysis and not some other coincidental
process, and to demonstrate that the oxoamide control peptides
underwent the same proteolytic degradation, we analyzed the
papain reactions by high-performance liquid chromatography
(HPLC). Here, we sampled each reaction at various time
points. After precipitating the protein, we analyzed the
supernatant by HPLC (Figure S2, Supporting Information).
We saw essentially no change in the chromatograms of the
peptides that were not treated with papain. For the peptides
that were treated with papain, we saw a decrease in starting
material signal and an increase in signal for the expected
cleavage products, which were identified by matrix-assisted laser
desorption ionization mass spectrometry (MALDI MS).
Cleavage occurred at approximately the same rate in the
thiopeptide and in the oxoamide control peptide. Additional
experiments with labeled peptide substrates that were not
expected to be cleaved by papain provided further evidence that
the presence of a thioamide probe does not stimulate
proteolysis (Figure S4, Supporting Information).
Figure 3. Papain kinetics as determined from stopped-flow
fluorescence measurements. Top: kinetic scheme for papain catalysis
in terms of enzyme (E), substrate (S), enzyme−substrate complexes
(ES and ES′), and products (P and Q) with rate constants as defined.
Left: primary stopped-flow data for the proteolysis of L′LKAAμ by
papain (2.5 μg/mL) in 2.0 mM EDTA, 5.00 mM L-cysteine, 300 mM
NaCl, pH 6.2, at 25 °C. Only a subset of data for selected
concentrations are shown for clarity. The inset shows a representative
fit to the burst equation for the 60 μM data set. Fits for each data set
are shown in the Supporting Information. Right: Michaelis−Menten
plot of initial rates determined from primary data. Data are colored
according to initial substrate concentration: pink 1.0 μM; magenta 2.5
μM; purple 5.0 μM; azure 8.0 μM; royal blue 10 μM; green 20 μM;
yellow 30 μM; orange 40 μM; red 60 μM.
after the Phe residue should lead to a turn on of fluorescence,
but trypsin cleavage after the Lys residue should not, since
intramolecular quenching would be maintained in the resulting
GL′AAFAμ fragment. This is indeed what was observed, as
incubation with chymotrypsin gave a robust and rapid increase
in fluorescence intensity until it reached the levels observed in
the oxopeptide, AKGLAAFAμ (Figure 4, left). As before, this
occurred only for the thiopeptide, and only in the presence of
the protease. Incubation in the presence of sequencing grade
trypsin produced no change in fluorescence over 2 h (Figure 4,
right). However, HPLC and MALDI MS analysis of the
To demonstrate the suitability of this method for studying
enzyme kinetics, we conducted stopped-flow experiments with
papain and variable initial concentrations of L′LKAAμ. Papain
catalysis can be described by a three step reaction scheme
(Figure 3, top). An expression for the rate of proteolysis can be
developed from this scheme and globally fit to the series of
stopped flow experiments (Figure S6, Supporting Information).
We obtained the following kinetic parameters: KS = k1/k−1
=
921 μM, k2 = 0.92 s−1, k3 = 0.057 s−1. These can be used to
determine kcat = 0.054 s−1 and KM,app = 53.9 μM. The
concentration dependence of papain activity can be seen in
plots of the stopped-flow primary data and the initial rate
(Figure 3, bottom). Nearly identical values for kcat and KM of
0.057 0.001 s−1 and 50.2 1.1 μM, respectively, can also be
obtained by fitting this initial rate data to the standard
Michaelis−Menten equation (Supporting Information). The
enzyme efficiency, kcat/KM = 1.00 × 103 M−1·s−1, is similar to
that previously reported for papain for the substrate
lysylnitroanilide (2.57 × 103 M−1·s−1).35 Wide variations in
rates are often observed for fluorogenic substrates used in
studying the kinetics of the same enzyme.36−38 The small size
of the thioamide allows us to place it at a variety of locations
while keeping our reporter peptides nearly identical to the
actual substrate, which should ensure that the rates obtained in
rapid mixing experiments are relevant to the actual protein
substrates.
Figure 4. Specific monitoring of chymotrypsin activity in a dual
substrate peptide. Incubation of a thioamide/7-methoxycoumarin
labeled peptide with both chymotrypsin and trypsin cleavage sites
results in a gain of fluorescence only in the presence of chymotrypsin.
Representative results for chymotrypsin and trypsin experiments. The
fluorescence of the thiopeptide in the presence (red trace) and absence
(orange trace) of protease is shown with the corresponding
oxopeptide also in the presence (green trace) and absence (blue
trace) of protease. Chymotrypsin: fluorescence of 8.3 μM
AKGLAAFAμ or AKGL′AAFAμ in 100 mM Tris-HCl, pH 7.8, at
25 °C in the presence and absence of 5 μg/mL chymotrypsin. Trypsin:
fluorescence of 8.3 μM AKGLAAFAμ or AKGL′AAFAμ in 67 mM
sodium phosphate, pH 7.6, at 25 °C in the presence and absence of 10
μg/mL trypsin.
The ability to place the thioamide at different locations
without disrupting the native peptide fold can be particularly
valuable where one would like to distinguish cleavage at one of
two sites, both of which may be bound near the protease active
site. To test our ability to carry out such a study, the peptide
AKGL′AAFAμ was labeled such that chymotrypsin cleavage
2088
dx.doi.org/10.1021/ja412297x | J. Am. Chem. Soc. 2014, 136, 2086−2093

Journal of the American Chemical Society
Article
reaction mixtures indicated that cleavage had indeed taken
place at the expected location (Figure S8, Supporting
Information). HPLC analysis of both chymotrypsin and trypsin
cleavage showed that the thiopeptide and oxopeptide were
cleaved at roughly the same rates, but that only the intended
chymotrypsin activity was detectable by a fluorescence increase.
It is interesting to note that when lower grade trypsin was used,
trace chymotrypsin activity reported by the manufacturer could
be observed as a turn on of fluorescence (Figure S7, Supporting
Information). The results observed with AKGL′AAFAμ should
be general, in that thiopeptides could be labeled so that a
specific cleavage event can be monitored in real time in an
otherwise native sequence.
As a final demonstration of the use of thioamides as protease
probes, we evaluated a fluorogenic substrate for monitoring
calpain activity in cell lysates. Cell lysates contain a host of
different proteases, challenging the specificity of our probes.
Insoluble or fluorescent material can contribute to background
signal, which demands robustness from the assay. Thus, we
viewed this as a stringent test of the utility of the fluorophore/
thioamide strategy. We designed a thiopeptide, L′PLFAERμ, to
serve as a reporter of calpain activity and to be used in
screening calpain inhibitor efficacy in crude cell lysates.
Calpains 1−15 are Ca2+-activated cysteine proteases that have
been implicated in several neuronal processes, insulin secretion,
cell survival, and the regulation of blood vessels.39−42 It has
recently been shown that malarial parasites highjack calpain
activity in order to digest heme proteins to serve as metabolites
during their exit from host red blood cells.43 Calpain activity
has also been tied to Alzheimer’s disease via the regulation of
APP proteolysis and to breast cancer.44,45 As a consequence,
there is substantial current interest in discovering reporters and
modulators of calpain activity.46−48 We felt that calpains would
be a valuable test of our technology as the substrate peptide
adopts an unusual helical conformation in the active site,
different than the extended structures typical of many
proteases. We therefore believed that taking advantage of the
natural substrate sequences to confer specificity for calpains on
our probes would show their value.
Prior to any usage of thiopeptide protease sensors in cell
lysates, we performed important control experiments that
evaluated the stability of thioamides toward nonproteolytic
degradation. Although backbone thioamides are found in
several peptide natural products and have in fact shown some
resistance to metabolism, we were able to find only one
previous assessment of their stability that we considered
applicable to our studies.49−51 Here, we refer to only acyclic
thioamides of the type incorporated in our peptides, not the
well-studied thiazolene rings in molecules such as thiostrepton,
which are also commonly referred to as “thiopeptides.”52 To
evaluate thioamide stability in cell lysates, we synthesized a
biotin-labeled thiopeptide composed of D-amino acids,
Biotin∼F′aa, which should not be recognized by proteases,
but should report on any nonstereo-specific metabolic
degradation of the thioamide bond. When this probe was
incubated with PBS at 25 °C, recovered using neutravidin
beads, and analyzed by HPLC, it was found to be completely
completed in 0.5−2 h. While we continue to pursue
experiments to understand the nature of this 15−20%
degradation, we take confidence from these experiments that
thiopeptide applications in cell lysates are viable.
During initial proteolytic testing using purified calpain 1 in
buffer, we only observed an increase in fluorescence when the
L′PLFAERμ thiopeptide was incubated with calpain that had
been activated by the addition of Ca2+. No increase was
observed in the oxopeptide (LPLFAERμ), or in the thiopeptide
when calpain activity was inhibited by the inclusion of the Ca2+
chelator EDTA in the buffer (Figure S11, Supporting
Information). We then tested the probes using mouse
embryonic fibroblast (MEF) cells, which endogenously express
calpains 1 and 2. When the cell lysates were incubated with the
thiopeptide L′PLFAERμ in the presence of added Ca2+, we
indeed observed an increase in fluorescence (Figure 5, right).
Figure 5. Inhibition of calpain by calpastatin peptide in cell lysate.
Top: scheme for experimental setup wherein MEF cell lysate is
incubated with a reporter peptide (either A′AFAμ or L′PLFAERμ) in
the presence or absence (EDTA) of Ca2+, used to activate calpain
proteases. Ca2+-activated samples were incubated in the presence or
absence of a peptide inhibitor derived from calpastatin. Left: relative
fluorescence of A′AFAμ in MEF cell lysate in the presence of Ca2+ and
inhibitor peptide (A, yellow), in the presence of Ca2+ (B, red), or in
the presence of EDTA (C, green). The increase of fluorescence of 7-
methoxycoumarin in Ca2+-loaded lysate is also shown (μ, black).
Right: relative fluorescence of L′PLFAERμ in MEF cell lysate in the
presence of Ca2+ and inhibitor peptide (D, purple), in the presence of
Ca2+ (E, pink), or in the presence of EDTA (F, blue). All traces are
normalized to the fluorescence of the peptide at time 0 min.
When the lysates were incubated with the thiopeptide in a
buffer containing EDTA, no increase in fluorescence was
observed (Figure 5, right). However, the interpretation of these
results is not quite as simple as it might seem; a small increase
in fluorescence was also observed for the 7-methoxycoumarin
fluorophore itself, induced by nonspecific aggregation in the
Ca2+-loaded lysates (Figure 5, left and right). When a saturating
concentration of a high affinity calpain-specific peptide inhibitor
(a 27mer peptide derived from the endogenous inhibitor
calpastatin) was added to the lysates,53 the timecourse of the
fluorescence increase was nearly identical to the nonspecific
increase observed for 7-methoxycoumarin alone, showing that
all calpain activity had been inhibited. (Figure 5, Right)
intact after 1 h and 94
2% recoverable after 24 h (as
compared to a biotinylated coumarin control dipeptide)
(Figure S10, Supporting Information). Moreover, when the
same experiments were performed with mouse serum, the
peptide was recovered in 85 2% yield after 1 h and 80 2%
yield after 4 h. We note that most of our protease assays are
2089
dx.doi.org/10.1021/ja412297x | J. Am. Chem. Soc. 2014, 136, 2086−2093

Journal of the American Chemical Society
Article
In contrast, when the nonspecific thiopeptide reporter
A′AFAμ was used, an increase in fluorescence was again
observed upon addition to lysates, but only a small portion of
this increase could be inhibited by the calpastatin peptide
(Figure 5, left). Moreover, a substantial increase in fluorescence
was observed in the lysates even in the presence of EDTA,
indicating clearly that non-calpain proteases were also acting on
the A′AFAμ thiopeptide (Figure 5, left). Taken together, these
data show that thiopeptides can be used as fluorescence
reporters even in cell lysates and that through judicious
sequence design, they can be made highly specific for use in
monitoring the activity of only a select protease toward an
essentially native sequence.
thioamides is their small size, which we believe is a significant
improvement over current methods. Although there are
certainly situations in which the placement of the fluorophores
on the substrate does not appear to interfere with enzyme
kinetics, there are also many examples, such as with papain, in
which fluorophore placement disturbs the observed kinetics,
sometimes by as much as 50-fold.55−57 Placing the fluorophores
far away from the scissile bond can obviate such problems, but
results in a loss of information about which bond is being
hydrolyzed. Here, we show that thioamides can be incorporated
at positions close to the site of cleavage to retain some of this
information that would otherwise be lost. The sensitivity of
fluorescent protease sensing is not changed by our innovation,
but its resolving power is increased substantially. Ultimately,
thioamide probes will not be appropriate for every experiment,
but should find use in a number of applications in which the use
of larger probes might interfere with enzyme activity.
Finally, while short reporter peptides of the type used here
can be valuable for simple assays, we also envision future
applications of this technology to analyzing proteolysis in full-
sized proteins. We have shown that native chemical ligation
(NCL) reactions can be applied to synthesize full-sized proteins
containing thioamides, taking advantage of ligations to
expressed proteins to minimize unnecessary protein synthesis.58
Furthermore, we have recently shown that NCL can be coupled
to unnatural amino acid (Uaa) mutagenesis to generate double
labeled proteins wherein the fluorophore is incorporated during
ribosomal protein synthesis.59 We have applied a combination
of NCL and Uaa methods to the synthesis of labeled versions of
α-synuclein (αS), a neuronal protein that forms fibrillar
aggregates that are believed to contribute to Parkinson’s
disease pathology. Several proteases have been reported to act
on αS, and cleavage of the C-terminal tail of αS has been shown
to accelerate fibrillization.60−63 Moreover, there has been a
report of autoproteolysis of αS.64 It is difficult to imagine
studying the mechanism of such a reaction without access to
labeled versions of the full-length protein. Clearly short
peptides would not be appropriate model systems. Synthesis
of appropriately labeled versions of αS using NCL and Uaa
techniques will allow us to investigate such questions using
thioamide-quenching methods. Many other proteins undergo
autoproteolysis, or are only cleaved in multiprotein complexes.
These proteins also lend themselves to investigation with
fluorophore/thioamide pairs where probes could be placed at
appropriate locations to monitor specific proteolytic events for
which maintaining the native conformation of the protein is
important. We are pursuing such applications while we further
study fundamental aspects of fluorophore/thioamide placement
to design optimal probes for certain types of cleavage events.
CONCLUSIONS
■
In summary, we have shown that thioamides can be used to
prepare quenched substrates that become fluorescent upon
cleavage by a protease. These probes can be used to monitor
kinetics in real time and should be compatible with most
proteolytic enzymes. Since thioamides are small, and since they
can conceivably be scanned through a protein backbone with
minimal perturbation to native structure, thioamide-based
probes should allow investigators to study protease activity
with more detail than current methods allow. However, the
design of such substrates requires two important consider-
ations. First, that quenching by PET is distance dependent so
that the fluorophore and thioamide must be placed relatively
close to each other in three-dimensional space. We have shown
that they can be placed at least six amino acids apart, and we
will continue to explore the trade-off between fluorophore/
thioamide spacing and the increase in fluorescence upon
cleavage. The second consideration is one of the nonperturbing
nature of the thioamide substitution. Here, we have shown that
in the case of the papain substrate, placement of the thioamide
two residues away from the scissile bond does not impact the
cleavage rate relative to the corresponding oxoamide peptide.
Further study of thioamide placement with papain and other
proteases will allow us to provide more general guidance on
how closely one can place the thioamide to the cleavable bond.
Given the success of our cell lysate experiments, we expect to
be able to extend this method to applications in cell culture and
we are currently exploring strategies for delivering appropriately
labeled peptides into cells.
Since there are many existing strategies for monitoring
proteolysis through fluorescence, a comparison of these
approaches to our thioamide method is valuable. A primary
concern in any fluorescent assay is sensitivity, especially with
respect to signal/noise, signal/background, and the limit of
detection. For our steady state (i.e., not rapid mixing)
measurements, we found signal/noise ratios that were typically
>100 and signal/background ratios that were typically >150
(See Supporting Information). These values are comparable to
most other fluorophore reporting systems. Turn-on values for
typical FRET or PET probes routinely vary by substrate and
range from ∼2 to >190-fold.9 Most of the thioamide examples
presented here have a ∼2-fold fluorescence increase, which is at
the low end of the observed range, but more than sufficient to
extract reliable kinetic data given the low background and high
signal/noise of our probes. In principle, thioamide quenchers
should have limits of detection similar to other fluorescent
probes, given that they can be paired with very bright
fluorophores such as Alexa Fluor 488 that can even be used
in single molecule studies.20,54 Of course, the chief advantage of
EXPERIMENTAL PROCEDURES
■
Peptide Synthesis. Peptides were synthesized on solid phase
using standard Fmoc chemistry and purified to homogeneity by
reverse-phase high performance liquid chromatography (HPLC).
Thioamide benzotriazole precursors were either commercially available
or synthesized according to literature precedent.21 Explicit protocols
are provided in the Supporting Information.
Fluorescence Spectroscopy. Steady-state fluorescence measure-
ments were collected with a Tecan M1000 plate reader, the kinetics
module of a Varian Cary Eclipse fluorometer, or a Photon
Technologies International QuantaMaster fluorometer outfitted with
multicell Peltier sample holders. For the fluorometers, the excitation
wavelength was 325 nm and the excitation slit width was 5 nm. The
emission wavelength was 391 nm and the slit width was 5 nm. The
2090
dx.doi.org/10.1021/ja412297x | J. Am. Chem. Soc. 2014, 136, 2086−2093

Journal of the American Chemical Society
Article
averaging time was 0.100 s. Samples were stirred in 1.00 cm quartz
cuvettes. Peptide concentrations were determined by absorbance at
325 nm (ε325 = 12 000 M−1·cm−1).65 Plate reader measurements were
recorded with black plates with optically transparent bottoms (Greiner
Bio-One No. 675096; Frickenhausen, Germany) as the average of nine
20 μs reads per well with an excitation wavelength of 330 nm and
emission wavelength of 385 nm with 5 nm slit widths and a 400 Hz
flash frequency with 30 s shaking before each kinetics run unless
otherwise indicated.
and the peptide concentration was 8.3 μM. Fluorescence was
monitored as described above at 25 °C.
The reactions (120 μL total volume) were quenched at 1 h
(chymotrypsin) or 2 h (trypsin) by addition of 1080 μL of cold
ethanol to precipitate the protease. The samples were then cooled at
−20 °C for 1 h and then centrifuged at 13 200 rpm at 4 °C for 20 min.
The supernatant was transferred to a clean microcentrifuge tube and
dried in a vacuum centrifuge. Each sample was brought up in 800 μL
H2O and analyzed by HPLC (Figure S8, Supporting Information).
Cell Culture. Mouse embryonic fibroblast (MEF) cells were grown
in Dulbecco’s modified eagle medium (DMEM) (Gibco, Life
Technologies, Grand Island, NY, USA) supplemented with 10% fetal
bovine serum (FBS) (Gibco) and 1:100 penicillin and streptomycin
(Gibco) at 37 °C. Cells were grown to 90% confluency then
trypsinized, pelleted, and frozen.
Calpain Cell Lysate Assay. The MEF cells were both hypotoni-
cally lysed and lysed via freeze/thaw (3×) in liquid nitrogen in
hypotonic lysing buffer (10 mM dithiothreitol (DTT), 5 mM
KH2PO4, and 6 mM EGTA, pH 7.5).66 Total protein concentration
was ∼4 mg/mL as determined by bicinchoninic acid assays (Thermo
Scientific/Pierce; Rockford, IL, USA).67 The LPLFAERμ and
L′PLFAERμ substrate concentrations were 8 μM. The inhibitor,
calpastatin peptide (sequence: Biotin-Asp-Pro-Met-Ser-Thr-Tyr-Ile-
Glu-Glu-Leu-Gly-Lys-Arg-Glu-Val-Thr-Ile-Pro-Pro-Lys-Tyr-Arg-Glu-
Leu-Leu-Ala-NH2), was used at 0 and 25 μM concentrations. Calpains
were activated by the addition of CaCl2 via a multichannel pipet to a
final concentration of 10 mM. In negative control experiments, buffer
without calcium was added instead. All assays were done at a total well
volume of 100 μL in 96-well black flat bottom plates (Greiner Bio-One
No. 655209). Fluorescence was read in a Tecan M1000 plate reader
over 30 min. The excitation wavelength was 330 nm and the emission
wavelength was 390 nm. All other settings were as described above.
Steady-State Protease Assays. At least three independent trials
of each assay were conducted to ensure reproducibility (Figure S1,
Supporting Information). Concentrated stock solutions of each
peptide and protease in the appropriate buffer were used to prepare
fresh samples immediately prior to each experiment. Chymotrypsin
assay (30 °C): Chymotrypsin (type II from bovine pancreas; ≥40
units/mg protein) was dissolved in 1 mM HCl, 10 mM CaCl2 and
diluted to a working concentration of 0.2 mg/mL immediately prior to
use such that each sample contained 8.3 μM AAFAμ or A′AFAμ in
100 mM Tris-HCl, pH 7.8. Pepsin assay (37 °C): Pepsin (from
porcine gastric mucosa; 3200−4500 units/mg protein) was diluted to
a working concentration of 1.5 mg/mL in samples that were 0.5 μM
AAFAμ or A′AFAμ in 10 mM HCl. Papain assay (25 °C): Papain
(crude, from papaya latex; 1.5−10 units/mg solid) was diluted to a
working concentration of 2.5 μg/mL such that samples also contained
8.3 μM LLKAAμ or L′LKAAμ in 2.0 mM EDTA, 5.00 mM L-cysteine,
300 mM NaCl, pH 6.2. Thermolysin assay (25 °C): Thermolysin
(from Bacillus thermoproteolyticus rokko; 50−100 units/mg protein)
was diluted to a working concentration of 6 μg/mL such that the
samples contained 8.3 μM AAFAμ or A′AFAμ in 2 mM calcium
acetate, 10 mM sodium acetate, pH 7.5. Trypsin assay (25 °C): Stock
solutions of trypsin (type II from porcine pancreas; 1000−2000 units/
mg dry solid) in cold 1 mM HCl were diluted with 67 mM sodium
phosphate, pH 7.6, to a working concentration of 25 μg/mL such that
the samples also contained 8.3 μM LLKAAμ or L′LKAAμ.
ASSOCIATED CONTENT
■
Stopped-Flow Kinetics. Stopped-flow kinetics measurements
were obtained with a KinTek SF-120 instrument (KinTek
Corporation; Austin, TX, USA). One syringe was filled with 2 μg/
mL papain in 4.0 mM EDTA, 10 mM L-cysteine, 600 mM NaCl, pH
6.2. The second syringe was filled with LLKAAμ or L′LKAAμ in water
at concentrations ranging from 2 to 200 μM, as determined by
absorbance at 325 nm. After mixing, data were collected for 2 min at a
rate of 5,000 points/min. The excitation wavelength was 310 nm and
the emission was recorded as the integrated signal for wavelengths
>320 nm. Data sets for each concentration were collected as the
average of at least three trials. The determination of the concentrations
of the cleaved thiopeptides as a function of time and the kinetic
analysis of the data are presented in Supporting Information. Primary
stopped flow data are shown with the results of global fitting (Figure
S6, Supporting Information).
Multisite Substrate Assay. Multisite substrate assays were
conducted in a fashion similar to that described for other steady-
state measurements in the presence and absence of enzyme.
Concentrated stocks of the peptides AKGLAAFAμ and AKGL′AAFAμ
were prepared in 67 mM sodium phosphate buffer, pH 7.6 for
experiments with trypsin. Sequencing grade modified trypsin
(Promega; Madison, WI, USA) was dissolved in 50 mM acetic acid
and stored at −80 °C. For the trypsin assay, samples were prepared
such that the enzyme concentration was 1 μg/mL and the peptide
concentration was 8.3 μM. Fluorescence was monitored as described
above at 30 °C. Preliminary experiments with technical grade trypsin
(25 μg/mL) resulted in a slight turn on of fluorescence for the
AKGL′AAFAμ sample in the presence of enzyme (Figure S7,
Supporting Information). We attribute this observation to chymo-
trypsin-like activity of the lower grade enzyme as indicated by the
manufacturer; this activity was not observed in with sequencing-grade
enzyme. For chymotrypsin experiments, concentrated stocks of
AKGLAAFAμ and AKGL′AAFAμ were prepared in 100 mM Tris
buffer, pH 7.8, and lyophilized chymotrypsin was dissolved in a
solution of 1 mM HCl and 10 mM CaCl2 immediately prior to use.
For the chymotrypsin assay, the enzyme concentration was 5 μg/mL
S
* Supporting Information
Additional data for protease screening, detailed peptide
synthesis information, HPLC analysis of protease reactions
and thiopeptide stability, details of papain kinetic modeling.
This material is available free of charge via the Internet at
http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
ejpetersson@sas.upenn.edu
Present Address
§J.M.G.: Department of Chemistry, Massachusetts Institute of
Technology, Cambridge, MA 02139 USA.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by funding from the Searle Scholars
Program (10-SSP-214 to E.J.P.), the Alfred P. Sloan
Foundation (BR2012-085 to E.J.P.), the National Institutes of
Health (NS081033 to E.J.P. and AI09727301 to D.C.G.), and
the Penn Genome Frontiers Institute (to D.C.G.). MALDI MS
instrumentation was supported by NSF MRI-0820996. We
thank Jeff Saven for use of the fluorometer. We also thank John
Farr and the KinTek Corporation for use of the stopped-flow
fluorometer. J.M.G. thanks the University of Pennsylvania for a
dissertation completion fellowship. N.M. thanks the NIH for
funding through the Chemistry-Biology Interface Training
Program (T32 GM07133). MEF cells were generously
provided by Dr. Peter Greer (Queen’s University).
2091
dx.doi.org/10.1021/ja412297x | J. Am. Chem. Soc. 2014, 136, 2086−2093

Journal of the American Chemical Society
Article
(38) Szabelski, M.; Rogiewicz, M.; Wiczk, W. Anal. Biochem. 2005,
342, 20−27.
REFERENCES
■
(1) Cory, S.; Adams, J. M. Nat. Rev. Cancer 2002, 2, 647−656.
(2) Nicholson, D. W.; Thornberry, N. A. Trends Biochem. Sci. 1997,
22, 299−306.
(3) Selkoe, D. J. Physiol. Rev. 2001, 81, 741−766.
(4) Visse, R.; Nagase, H. Circ. Res. 2003, 92, 827−839.
(5) Edgington, L. E.; Verdoes, M.; Bogyo, M. Curr. Opin. Chem. Biol.
2011, 15, 798−805.
(6) Serim, S.; Haedke, U.; Verhelst, S. H. L. ChemMedChem 2012, 7,
1146−1159.
(7) Barrett, A. J. Handbook of Proteolytic Enzymes; Academic Press:
San Diego, CA, 2002.
(8) Maeda, H. Anal. Biochem. 1979, 92, 222−227.
(9) Knight, C. G. Methods Enzymol. 1995, 248, 18−34.
(10) Barrett, A. J.; Kirschke, H. Methods Enzymol. 1981, 80, 535−
561.
(39) Goll, D. E.; Thompson, V. F.; Li, H. Q.; Wei, W.; Cong, J. Y.
Physiol. Rev. 2003, 83, 731−801.
(40) Sreenan, S. K.; Zhou, Y. P.; Otani, K.; Hansen, P. A.; Currie, K.
P. M.; Pan, C. Y.; Lee, J. P.; Ostrega, D. M.; Pugh, W.; Horikawa, Y.;
Cox, N. J.; Hanis, C. L.; Burant, C. F.; Fox, A. P.; Bell, G. I.; Polonsky,
K. S. Diabetes 2001, 50, 2013−2020.
(41) Tan, Y.; Wu, C.; De Veyra, T.; Greer, P. A. J. Biol. Chem. 2006,
281, 17689−17698.
(42) Zatz, M.; Starling, A. New Engl. J. Med. 2005, 352, 2413−2423.
(43) Chandramohanadas, R.; Davis, P. H.; Beiting, D. P.; Harbut, M.
B.; Darling, C.; Velmourougane, G.; Lee, M. Y.; Greer, P. A.; Roos, D.
S.; Greenbaum, D. C. Science 2009, 324, 794−797.
(44) Kulkarni, S.; Reddy, K. B.; Esteva, F. J.; Moore, H. C. F.; Budd,
G. T.; Tubbs, R. R. Oncogene 2010, 29, 1339−1350.
(45) Liang, B.; Duan, B. Y.; Zhou, X. P.; Gong, J. X.; Luo, Z. G. J.
Biol. Chem. 2010, 285, 27737−27744.
(11) Leytus, S. P.; Melhado, L. L.; Mangel, W. F. Biochem. J. 1983,
209, 299−307.
(46) Kelly, J. C.; Cuerrier, D.; Graham, L. A.; Campbell, R. L.;
Davies, P. L. Biochim. Biophys. Acta, Proteins Proteomics 2009, 1794,
1505−1509.
(12) Leytus, S. P.; Patterson, W. L.; Mangel, W. F. Biochem. J. 1983,
215, 253−260.
(13) Zimmerman, M.; Yurewicz, E.; Patel, G. Anal. Biochem. 1976,
70, 258−262.
(47) Jo, H.; Meinhardt, N.; Wu, Y. B.; Kulkarni, S.; Hu, X. Z.; Low,
K. E.; Davies, P. L.; DeGrado, W. F.; Greenbaum, D. C. J. Am. Chem.
Soc. 2012, 134, 17704−17713.
(14) Butenas, S.; Orfeo, T.; Lawson, J. H.; Mann, K. G. Biochemistry
1992, 31, 5399−5411.
(48) Pietsch, M.; Chua, K. C. H.; Abell, A. D. Curr. Top. Med. Chem.
2010, 10, 270−293.
(15) Butenas, S.; Ribarik, N.; Mann, K. G. Biochemistry 1993, 32,
6531−6538.
(16) Jones, L. J.; Upson, R. H.; Haugland, R. P.; Panchuk-Voloshina,
N.; Zhou, M.; Haugland, R. P. Anal. Biochem. 1997, 251, 144−152.
(17) Chevalier, A.; Massif, C.; Renard, P. Y.; Romieu, A. Chem.Eur.
J. 2013, 19, 1686−1699.
(49) Banala, S.; Sussmuth, R. D. ChemBioChem 2010, 11, 1335−
1337.
(50) Zacharie, B.; Lagraoui, M.; Dimarco, M.; Penney, C. L.; Gagnon,
L. J. Med. Chem. 1999, 42, 2046−2052.
(18) Neefjes, J.; Dantuma, N. P. Nat. Rev. Drug Discovery 2004, 3,
58−69.
(51) Zhang, W.; Li, J.; Liu, L. W.; Wang, K. R.; Song, J. J.; Yan, J. X.;
Li, Z. Y.; Zhang, B. Z.; Wang, R. Peptides 2010, 31, 1832−1838.
(52) Bagley, M. C.; Dale, J. W.; Merritt, E. A.; Xiong, A. Chem. Rev.
2005, 105, 685−714.
(19) Shi, L. F.; De Paoli, V.; Rosenzweig, N.; Rosenzweig, Z. J. Am.
Chem. Soc. 2006, 128, 10378−10379.
(20) Goldberg, J. M.; Batjargal, S.; Chen, B. S.; Petersson, E. J. J. Am.
Chem. Soc. 2013, 135, 18651−18658.
(53) Hanna, R. A.; Campbell, R. L.; Davies, P. L. Nature 2008, 456,
409−U478.
(21) Goldberg, J. M.; Batjargal, S.; Petersson, E. J. J. Am. Chem. Soc.
2010, 132, 14718−14720.
(54) Torbeev, V. Y.; Myong, S. A.; Ha, T.; Kent, S. B. H. Isr. J. Chem.
2011, 51, 960−967.
(22) Goldberg, J. M.; Speight, L. C.; Fegley, M. W.; Petersson, E. J. J.
Am. Chem. Soc. 2012, 134, 6088−6091.
(55) Gauthier, F.; Moreau, T.; Lalmanach, G.; Brillardbourdet, M.;
Ferrerdimartino, M.; Juliano, L. Arch. Biochem. Biophys. 1993, 306,
304−308.
(23) Goldberg, J. M.; Wissner, R. F.; Klein, A. M.; Petersson, E. J.
Chem. Commun. 2012, 48, 1550−1552.
(56) Pennington, M. W.; Thornberry, N. A. Peptide Res. 1994, 7, 72−
76.
(57) Serveau, C.; Juliano, L.; Bernard, P.; Moreau, T.; Mayer, R.;
(24) Yao, S.; Zutshi, R.; Chmielewski, J. Bioorg. Med. Chem. Lett.
1998, 8, 699−704.
(25) Bartlett, P. A.; Spear, K. L.; Jacobsen, N. E. Biochemistry 1982,
21, 1608−1611.
Gauthier, F. Biochimie 1994, 76, 153−158.
(58) Batjargal, S.; Wang, Y. J.; Goldberg, J. M.; Wissner, R. F.;
Petersson, E. J. J. Am. Chem. Soc. 2012, 134, 9172−9182.
(59) Wissner, R. F.; Batjargal, S.; Fadzen, C. M.; Petersson, E. J. J.
Am. Chem. Soc. 2013, 135, 6529−6540.
(26) Beattie, R. E.; Elmore, D. T.; Williams, C. H.; Guthrie, D. J. S.
Biochem. J. 1987, 245, 285−288.
(27) Bond, M. D.; Holmquist, B.; Vallee, B. L. J. Inorg. Biochem. 1986,
28, 97−105.
(60) Li, W. X.; West, N.; Colla, E.; Pletnikova, O.; Troncoso, J. C.;
Marsh, L.; Dawson, T. M.; Jakala, P.; Hartmann, T.; Price, D. L.; Lee,
M. K. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 2162−2167.
(61) Sevlever, D.; Jiang, P. Z.; Yen, S. H. C. Biochemistry 2008, 47,
9678−9687.
(62) Mishizen-Eberz, A. J.; Guttmann, R. P.; Giasson, B. I.; Day, G.
A.; Hodara, R.; Ischiropoulos, H.; Lee, V. M. Y.; Trojanowski, J. Q.;
Lynch, D. R. J. Neurochem. 2003, 86, 836−847.
(63) Choi, D. H.; Kim, Y. J.; Kim, Y. G.; Joh, T. H.; Beal, M. F.; Kim,
Y. S. J. Biol. Chem. 2011, 286, 14168−14177.
(64) Vlad, C.; Lindner, K.; Karreman, C.; Schildknecht, S.; Leist, M.;
Tomczyk, N.; Rontree, J.; Langridge, J.; Danzer, K.; Ciossek, T.; Petre,
A.; Gross, M. L.; Hengerer, B.; Przybylski, M. ChemBioChem 2011, 12,
2740−2744.
(65) Brun, M. P.; Bischoff, L.; Garbay, C. Angew. Chem., Int. Ed.
2004, 43, 3432−3436.
(66) Millholland, M. G.; Chandramohanadas, R.; Pizzarro, A.; Wehr,
A.; Shi, H.; Darling, C.; Lim, C. T.; Greenbaum, D. C. Mol. Cell.
Proteomics 2011, 10, No. M111.010678.
(28) Campbell, P.; Nashed, N. T. J. Am. Chem. Soc. 1982, 104, 5221−
5226.
(29) Cho, K. Anal. Biochem. 1987, 164, 248−253.
(30) Foje, K. L.; Hanzlik, R. P. Biochim. Biophys. Acta, Gen. Subj.
1994, 1201, 447−453.
(31) Maziak, L.; Lajoie, G.; Belleau, B. J. Am. Chem. Soc. 1986, 108,
182−183.
(32) Mock, W. L.; Chen, J. T.; Tsang, J. W. Biochem. Biophys. Res.
Commun. 1981, 102, 389−396.
(33) Schutkowski, M.; Jakob, M.; Landgraf, G.; Born, I.; Neubert, K.;
Fischer, G. Eur. J. Biochem. 1997, 245, 381−385.
(34) Stockel-Maschek, A.; Mrestani-Klaus, C.; Stiebitz, B.; Demuth,
H. U.; Neubert, K. Biochim. Biophys. Acta, Protein Struct. Mol. Enzymol.
2000, 1479, 15−31.
(35) Angelides, K. J.; Fink, A. L. Biochemistry 1979, 18, 2363−2369.
(36) Harris, J. L.; Backes, B. J.; Leonetti, F.; Mahrus, S.; Ellman, J. A.;
Craik, C. S. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 7754−7759.
(37) Melo, R. L.; Alves, L. C.; Del Nery, E.; Juliano, L.; Juliano, M. A.
Anal. Biochem. 2001, 293, 71−77.
2092
dx.doi.org/10.1021/ja412297x | J. Am. Chem. Soc. 2014, 136, 2086−2093

Journal of the American Chemical Society
Article
(67) Gu, C.; Shannon, D. A.; Colby, T.; Wang, Z. M.; Shabab, M.;
Kumari, S.; Villamor, J. G.; McLaughlin, C. J.; Weerapana, E.; Kaiser,
M.; Cravatt, B. F.; van der Hoorn, R. A. L. Chem. Biol. 2013, 20, 541−
548.
2093
dx.doi.org/10.1021/ja412297x | J. Am. Chem. Soc. 2014, 136, 2086−2093