Counter Selection Substrate Library Strategy for Developing Specific Protease Substrates and Probes

Article abstract of DOI:10.1016/j.chembiol.2016.05.020

Legumain (AEP) is a lysosomal cysteine protease that was first characterized in leguminous seeds and later discovered in higher eukaryotes. AEP upregulation is linked to a number of diseases including inflammation, arteriosclerosis, and tumorigenesis. Thu

Full text of DOI:10.1016/j.chembiol.2016.05.020

Resource
Counter Selection Substrate Library Strategy for
Developing Specific Protease Substrates and Probes
Graphical Abstract
Authors
Marcin Poreba, Rigmor Solberg,
Wioletta Rut, ..., Boris Turk,
Guy S. Salvesen, Marcin Drag
Correspondence
marcin.poreba@pwr.edu.pl (M.P.),
rigmor.solberg@farmasi.uio.no (R.S.),
marcin.drag@pwr.edu.pl (M.D.)
In Brief
Poreba et al. developed a new chemical
strategy (Counter Selection Substrate
Library, CoSeSuL) to obtain very selective
legumain substrates and activity-based
probe (MP-L01) for the direct inhibition
and visualization of this protease in living
cells. MP-L01 does not cross-react with
caspases.
Highlights
d
d
d
CoSeSuL approach as a new tool for developing legumain-
selective substrates and ABPs
Unnatural amino acids increase selectivity of synthetic
legumain substrates
MP-L01 is a potent and selective cell-permeable legumain
probe
Poreba et al., 2016, Cell Chemical Biology 23, 1–13
August 18, 2016 ª 2016 Elsevier Ltd.
http://dx.doi.org/10.1016/j.chembiol.2016.05.020

Please cite this article in press as: Poreba et al., Counter Selection Substrate Library Strategy for Developing Specific Protease Substrates and Probes,
Cell Chemical Biology (2016), http://dx.doi.org/10.1016/j.chembiol.2016.05.020
Cell Chemical Biology
Resource
Counter Selection Substrate Library
Strategy for Developing Specific
Protease Substrates and Probes
Marcin Poreba,1 * Rigmor Solberg, * Wioletta Rut, Ngoc Nguyen Lunde, Paulina Kasperkiewicz, Scott J. Snipas,
,2,
2,3,
1
3
1
2
4
4
4
2
1,5,
Marko Mihelic, Dusan Turk, Boris Turk, Guy S. Salvesen, and Marcin Drag
*
Department of Bioorganic Chemistry, Faculty of Chemistry, Wroclaw University of Technology, Wyb.Wyspianskiego 27, 50-370 Wroclaw,
1
Poland
2Program in Cell Death and Survival Networks, Sanford Burnham Prebys Medical Discovery Institute, La Jolla, CA 92037, USA
3Department of Pharmaceutical Biosciences, School of Pharmacy, University of Oslo, 0316 Oslo, Norway
4Department of Biochemistry and Molecular and Structural Biology, Jo ꢀz ef Stefan Institute, 1000 Ljubljana, Slovenia
5Lead Contact
*Correspondence: marcin.poreba@pwr.edu.pl (M.P.), rigmor.solberg@farmasi.uio.no (R.S.), marcin.drag@pwr.edu.pl (M.D.)
http://dx.doi.org/10.1016/j.chembiol.2016.05.020
SUMMARY
Legumain (asparaginyl endopeptidase [AEP]; EC 3.4.22.34)
displays high selectivity and cleaves substrates after Asn (P1
Legumain (AEP) is a lysosomal cysteine protease that position) (Chen et al., 1997, 2000; Dall and Brandstetter, 2016).
was first characterized in leguminous seeds and later However, at acid pH this enzyme can also hydrolyze peptide
bonds after Asp, which overlaps with caspases substrate speci-
ficity (Dall and Brandstetter, 2012). Moreover, it was also reported
discovered in higher eukaryotes. AEP upregulation is
linked to a number of diseases including inflamma-
tion, arteriosclerosis, and tumorigenesis. Thus this
protease is an excellent molecular target for the
development of new chemical markers. We deployed
a hybrid combinatorial substrate library (HyCoSuL)
approach to obtain P1-Asp fluorogenic substrates
and biotin-labeled inhibitors that targeted legumain.
thatthisenzymeis inhibited by commonly usedsyntheticcaspase
inhibitors containing Asp residue at P1 (Rozman-Pungercar et al.,
2003). Legumain was first identified and characterized in 1993 as
a cysteine protease in leguminous seed (Kembhavi et al., 1993),
and its paralog was first identified in kidney (Chen et al., 1997).
Today it is known that legumain is expressed in diverse cell types
and is mainly located in late endosomes and lysosomes, since
Since this approach led to probes that were also acidic pH is a prerequisite for prolegumain activation (Dall and
recognized by caspases, we introduced a Counter Brandstetter, 2012; Hashimoto et al., 1999; Lecaille et al., 2004).
Selection Substrate Library (CoSeSuL) approach Unexpectedly, legumain can also be active extracellularly when
stabilized by integrins on the cell surface (Dall and Brandstetter,
that biases the peptidic scaffold against caspases,
2
013). Legumain intra- and extracellular activity is controlled by
endogenous cysteine protease inhibitors called type 2 cystatins
Alvarez-Fernandez et al., 1999; Briggs et al., 2010; Smith et al.,
012). Although displaying a caspase-like fold, legumain is not
inhibited by natural caspase inhibitors (Dall and Brandstetter,
016; Snipas et al., 2001). Legumain is considered to play an
thus delivering highly selective legumain probes.
The selectivity of these tools was validated using
M38L and HEK293 cells. We also propose that the Co-
SeSuL methodology can be considered as a general
principle in the design of selective probes for other
protease families where selectivity is difficult to
achieve by conventional sequence-based profiling.
(
2
2
important role in living cells, since it is the only enzyme known in
humans that is able to recognize Asn in the P1 position of peptide
substrates, indicating regulatory functions (Rawlings et al., 2014).
It has been demonstrated that this enzyme can participate in a
number of biological events including extracellular matrix remod-
eling (Morita et al., 2007), inhibition of osteoclast formation (Choi
INTRODUCTION
Cysteine proteases constitute one of the largest groups of pro- et al., 1999), major histocompatibility complex class II-mediated
teolytic enzymes, consisting of several separate clans (Rawlings antigen presentation (Manoury et al., 2003), conversion of pro-
et al., 2014). One of these is clan CD, which includes legumain, matrix metalloproteinase-2 (proMMP2) into active MMP2 (Chen
clostripains, caspases, paracaspases, metacaspases, gingi- et al., 2001), and processing of lysosomal cysteine cathepsins
pains, and separases/separins (Chen et al., 1998). Clan CD such as cathepsins B, H, and L (Mattock et al., 2010; Shira-
members are distinct from other groups of proteases in their hama-Noda et al., 2003). It is also required for normal kidney
stringent preference for a single amino acid in the P1 position physiology and homeostasis (Miller et al., 2011). Legumain over-
(
McLuskey and Mottram, 2015). Together, the human members expression has been linked with inflammation (Chan et al., 2009),
of clan CD participate in a number of critical cellular pathways atherosclerosis (Clerin et al., 2008), and tumorigenesis. Several
including cell survival and cell death signaling, inflammation, recent studies have reported that overexpression of legumain
antigen presentation, and cell-cycle regulation.
canbeusedasa prognosticfactorinvarioushumancancertypes,
Cell Chemical Biology 23, 1–13, August 18, 2016 ª 2016 Elsevier Ltd.
1

Please cite this article in press as: Poreba et al., Counter Selection Substrate Library Strategy for Developing Specific Protease Substrates and Probes,
Cell Chemical Biology (2016), http://dx.doi.org/10.1016/j.chembiol.2016.05.020
including breast (Gawenda et al., 2007), colorectal (Haugen et al., various unnatural amino acids in each position, making it an
2
013), gastric (Li et al., 2013), glioma (Qiu et al., 2008), ovarian excellent tool for the determination of broad specificity of prote-
Wang et al., 2012), and prostate (Ohno et al., 2013) cancers. ases (Kasperkiewicz et al., 2014; Poreba et al., 2014a). Using this
(
Therefore, legumain is a good candidate for molecular targeting approach we have been able to identify substrates that can
in cancer tumor therapy (Bajjuri et al., 2011; Liu et al., 2003; Luo distinguish individual caspases (Poreba et al., 2014a). This li-
et al., 2006). On the other hand, it has been reported that legumain brary contains Asp in P1, which makes it suitable also for
deficiency is involved in disorders resembling hemophagocytic legumain screening. The challenge thus becomes to develop
syndrome (Chan et al., 2009). Moreover, parasite legumain plays substrates that can distinguish caspases from legumain. To
a significant role in the development of parasitic infections, for this end, we developed a novel counterselection screen whereby
example schistosomatosis, which along with malaria is the most instead of exploiting optimal substrate/enzyme interactions, we
widespread parasitic disease (Gotz et al., 2008; James et al., defined non-productive interactions that would allow for the gen-
2
003; Ovat et al., 2009). eration of highly selective substrates. This Counter Selection
Because it is the activity of a protease that produces a biolog- Substrate Library (CoSeSuL) approach allowed us to identify a
ical event, small-molecule substrates, inhibitors, and activity- peptidic motif that is highly preferred by legumain. The major
based probes (ABPs) that can interrogate this activity provide benefit of this approach is that this peptide ligand can be used
excellent tools for studying protease functions in living organ- to simultaneously obtain legumain-selective substrates, inhibi-
isms (Kasperkiewicz et al., 2014; Powers et al., 2002; Sadaghiani tors, and ABPs. In this example the selectivity is generated by
et al., 2007). The first step in the design of these tools is to probe the P4-P1 peptide, not by the fluorescent tag or warhead, mak-
protease substrate specificity in detail, allowing the generation of ing this approach universal in the context of developing new low
structural motifs that are selective for the protease of interest molecular weight legumain tools. In this study, we describe the
(
Poreba and Drag, 2010). The substrate specificity of legumain step-by-step procedure for development and validation of new
was first characterized by Mathieu et al. (2002), revealing an legumain P1-Asp fluorogenic substrates and ABPs that are
optimal sequence in the P3-P2-P1 positions (Pro-Thr-Asn). able to enter living cells and selectively label active legumain.
Shortly afterward the Kalbacher group, using a simple peptide li-
brary, demonstrated that legumain has a broad substrate spec- RESULTS
0
0
ificity in the P1 position, and that peptides with Pro at P1 were
poorly hydrolyzed (Schwarz et al., 2002). While these results Legumain P1 Specificity
shed some light on legumain specificity, none of the sequences To examine the dual nature of legumain’s Asp and Asn
identified were used in synthesis of legumain-selective sub- specificity, we synthesized two 7-amino-4-carbamoylmethyl-
strates. The first legumain fluorogenic substrate Cbz-Ala-Ala- coumarin (ACC)-labeled substrates, Z-AAN-ACC and Z-AAD-
Asn-AMC was developed by Kembhavi et al. (1993), based on ACC, and measured the kinetics of their hydrolysis by
the P3-P2-P1 sequence Ala-Ala-Asn. The presence of Asn in legumain. Z-AAN-ACC was cleaved ꢀ22 times faster (kcat
/
ꢁ
1
ꢁ1
M ) than its P1-Asp analog (kcat/K
P1 made this substrate specific and very promising in the
K
M
= 36,100 s
ꢁ1
m
=
). Legumain’s ability to recognize Asp in P1 over-
ꢁ1
context of designing selective legumain inhibitors. However, it 1,600 M
s
was later reported that the close proximity of the Asn residue laps with caspases; however, the mechanisms for Asp recogni-
and an electrophile warhead induced intramolecular cyclization, tion are different because these enzymes are active in different
making the inhibitor inactive (Loak et al., 2003). Several types of pH windows. Cytosolic caspases display the highest catalytic
aza-Asn legumain inhibitors have been described to overcome activity at neutral pH range, varying from 6.5 to 8.0 (Asp residue
this problem, including Michael acceptors (Ekici et al., 2004; is deprotonated and occupies the positively charged S1 pocket
Lee and Bogyo, 2012), epoxides (Asgian et al., 2002; Lee and of caspases) (Garcia-Calvo et al., 1999; Stennicke and Salvesen,
Bogyo, 2012), and halomethylketones (Niestroj et al., 2002). All 1997). On the other hand, legumain prefers the acidic environ-
of these compounds inhibited legumain in vitro, but their potency ment of lysosomes and endosomes (below pH 6.0) (Dall and
and selectivity in vivo have never been tested. A new group of le- Brandstetter, 2013). Under these conditions legumain can
gumain inhibitors and ABPs was developed by the Bogyo group cleave substrates not only after Asn but also after Asp, since
(
Edgington et al., 2013; Lee and Bogyo, 2010; Sexton et al., below pH 6.0 the aspartic acid side chain is partially protonated.
007). These compounds contain Asp in the P1 position, which The maximal legumain activity toward a P1-Asp substrate has
2
makes them more stable. One of these probes (LE-28 with the been observed at pH 4.0 (Dall and Brandstetter, 2012), which
P3-P1 sequence Glu-Pro-Asp) is also legumain selective in vivo opened the doors for the development of novel P1-Asp probes
(Edgington et al., 2013). However, this selectivity was not potentially useful for selective legumain detection.
obtained by the Glu-Pro-Asp sequence itself (which is also
recognized by caspases), but by introducing bulky groups The Influence of Peptide Length on Legumain Activity
(
Cy5, a fluorophore and QSY21, a fluorescence quencher), The ultimate goal of this work was to obtain a legumain-selective
directing the probe to the lysosomes through endocytosis/ peptidic motif that could be used as a scaffold for the synthesis
macropinocytosis.
of substrates, inhibitors, and ABPs selectively targeting only this
The design of legumain-selective peptidic probes with Asp in enzyme. Therefore, we first examined the influence of substrate
the P1 position is very challenging due to the fact that most of length for legumain activity. To do this, we synthesized two
these peptides are recognized by caspases. This problem can series of ACC-labeled substrates in which the peptide chains
be overcome by using our previously reported hybrid combina- were systematically elongated (Ac-TN-ACC, Ac-GTN-ACC,
torial substrate library (HyCoSuL) containing more than 100 Ac-DGTN-ACC [series 1] and Ac-SN-ACC, Ac-DSN-ACC,
2
Cell Chemical Biology 23, 1–13, August 18, 2016

Please cite this article in press as: Poreba et al., Counter Selection Substrate Library Strategy for Developing Specific Protease Substrates and Probes,
Cell Chemical Biology (2016), http://dx.doi.org/10.1016/j.chembiol.2016.05.020
Figure 1. P4-P2 Legumain Substrate Specificity Profile Determined with the Use of Two Fluorogenic Substrate Libraries, P1-Asp and P1-Asn
The y axis represents amino acid activity (optimal residue set to 100%) and the x axis presents amino acids.
Ac-YDSN-ACC [series 2]) (Figure S1). These sequences were Asn library was less specific, and 13 out of 20 amino acids were
selected based on the legumain substrate specificity profile ob- recognized very well (>60% compared with the best, Gly). The
M
tained through PS-SCL profiling. Kinetic analysis (kcat/K mea- specificity in the P4 position also differed depending on the li-
surements) revealed that there is no correlation between the brary; however, in both cases this enzyme showed almost no
length of the peptide chain and legumain activity. All substrates specificity, which suggested that legumain does not have a
displayed similar kinetic parameters (Table S1). Taking into ac- well-defined S4 pocket (Figure 1).
count that caspases display almost no activity toward dipeptide
substrates and inhibitors, at first glance it would seem rational to Legumain Substrate Specificity Using the HyCoSuL
use a short scaffold (P2-Asp-warhead/fluorophore) to construct Approach
legumain-selective chemical tools. However, this is not appro- HyCoSuL is a fluorophore-labeled tetrapeptide library that in the
priate when substrates are converted to ABPs, since simple bio- P4-P2 positions contains, in addition to natural amino acids,
tinylated acyloxymethylketone containing only one amino acid more than 100 unnatural amino acids of different structures
(
biotin-linker-Asp-AOMK) are able to bind both legumain and and physicochemical properties, thus providing an excellent
caspases (Kato et al., 2005). We therefore postulated that biotin tool for studying proteolytic enzymes displaying overlapping
and a linker would decrease the ABP potency toward legumain substrate specificity profiles (i.e., legumain and caspases). The
versus caspases. In an attempt to develop novel legumain-se- tetrapeptide architecture of HyCoSuL is based on the assump-
lective ABPs, we decided to dissect legumain’s preferences in tion that a protease would have a well-defined S4 pocket, but
the P4-P2 positions using a traditional PS-SCL approach as this is not the case for legumain (Figure S2) (Dall and Brandstet-
well as the recently developed HyCoSuL strategy to define those ter, 2013). The best amino acid recognized by legumain in P2 po-
amino acids that discriminate legumain from the caspase family. sition was still the natural Thr (Figure 2); however, this enzyme
also tolerated small/medium-sized hydrophobic (Phg, Chg,
Legumain Specificity for Natural Amino Acids
Abu, Ser(Ac)) or neutral (Cit, hCit, Lys(tfa)) amino acids (for unnat-
We employed two tetrapeptide combinatorial fluorogenic sub- ural amino acids abbreviation list, see Data S2). Legumain poorly
strate libraries, Ac-X-X-X-Asp-ACC and Ac-X-X-X-Asn-ACC recognized large hydrophobic amino acids such as Bip, Bpa,
(
Figure S2). Bearing in mind that legumain activity is generated 1-Nal, and substituted Phe, and did not recognize D-amino
by a pH shift, we decided to test legumain specificity at two acids. Legumain specificity was more liberal in the P3 position
different pH values, pH 4.5 (Asp specific) and pH 5.8 (Asn spe- recognizing most amino acids; however, the best were Cys(Bzl),
cific) (Figure S3). We found that legumain cleaved P1-Asn sub- Cys(MeBzl), Cys(Me-O-Bzl), and secondary Tic. The analysis of
strates approximately 20–30 times more efficiently than the the P4 position confirmed the observation from the natural amino
P1-Asp counterparts. Kinetic analysis of both libraries revealed acids library analysis that legumain had no preferences in the S4
that legumain preferred Thr (100%) in the P2 position, although pocket. Indeed, almost all amino acids (even those with D-ste-
other amino acid residues (Ser, Ala, Val, Nle) were also recog- reochemistry) were recognized at the same level. The detailed
nized (30%–70% of Thr). The P3 specificity depended on the li- analysis of legumain preferences and the comparison of its
brary used for screening. In the P1-Asp library only a few amino specificity matrix with that previously obtained with caspases al-
acid residues were recognized at a substantial level: Asp (100%), lowed us to extract some differences and design short peptide
Ala (50%), Gly (46%), and Glu (40%). The P3 position in the P1- sequences recognized only by legumain.
Cell Chemical Biology 23, 1–13, August 18, 2016
3

Please cite this article in press as: Poreba et al., Counter Selection Substrate Library Strategy for Developing Specific Protease Substrates and Probes,
Cell Chemical Biology (2016), http://dx.doi.org/10.1016/j.chembiol.2016.05.020
Figure 2. Substrate Specificity Profile of Legumain Obtained via HyCoSuL
The figure presents legumain preferences in P4, P3, and P2 positions. The y axis represents amino acid activity (optimal residue set to 100%) and the x axis
presents amino acids (natural and unnatural) used for screening.
4
Cell Chemical Biology 23, 1–13, August 18, 2016

Please cite this article in press as: Poreba et al., Counter Selection Substrate Library Strategy for Developing Specific Protease Substrates and Probes,
Cell Chemical Biology (2016), http://dx.doi.org/10.1016/j.chembiol.2016.05.020
Figure 3. Two Novel Legumain-Selective ACC-Labeled Fluorogenic Substrates Containing Asp in P1 Position
The optimal pH for legumain cleavage at Asp was 4.5, and pH for caspases was 7.4. SEs calculated from three measurements were in the range of 5%–15%.
The Influence of pH on Legumain Activity
cellular system where other enzymes (i.e., caspases) are pre-
Legumain activity is pH dependent. To examine this phenome- sent. Through profiling we obtained several P1-Asp legumain-
non, we synthesized substrates with Asp or Asn in the P1 selective substrates, which were in the next step used to design
position (Ac-D-Tyr-Tic-Thr-Asn-ACC and Ac-D-Tyr-Tic-Thr- legumain-selective probes. From six substrate candidates two
Asp-ACC), revealing that the optimal pH for legumain activity champions were selected: Ac-D-Tyr-Tic-Ser-Asp-ACC and Ac-
toward the P1-Asn substrate was 5.8, whereas the pH optimum D-Arg-Tic-Ser-Asp-ACC (Figures 3 and S1; Table S3). These
for the P1-Asp substrate was shifted to 4.5 (Table S2), which is two substrates were highly sensitive to legumain and most selec-
likely due to protonation of the Asp residue at lower pH (Dall tive for this enzyme compared with apoptotic caspases (Figure 3;
and Brandstetter, 2012, 2013, 2016).
Table S3). Based on the aforementioned sequences we
synthesized three biotin-labeled ABPs. The first probe biotin-6-
ahx-D-Tyr-Tic-Ser-Asp-AOMK (MP-L01) had a biotin tag on
Design of Legumain-Selective Substrates: CoSeSuL
To design new legumain-selective P1-Asp fluorogenic sub- the N terminus and acyloxymethylketone (AOMK) as a warhead
strates, we analyzed legumain and caspase specificity profiles (Figure 4). The aminohexanoic linker (6-ahx) was used, as it was
and selected some amino acid candidates for use as substrate shown previously that this spacer increases the potency of
building blocks. The best amino acid for legumain in the P2 po- inhibitors toward endogenous legumain (Sexton et al., 2007).
sition is Thr; however, it is also recognized by caspases so we The kinetic analysis with recombinant enzymes revealed that
selected the less active, but more selective Ser (Poreba et al., this probe was very potent toward legumain (pH 5.8), and
2014a). Another reason for using Ser at P2 instead of Thr (the showed slight off-target activity against caspases when used
best amino acid from library screening) is because inhibitors at pH 7.4 and elevated concentrations (Table 1).
with Thr at P2 have decreased potency toward legumain for un-
known reasons (Sexton et al., 2007).
Two other legumain-selective probes were synthesized with
methylation in the P1 position (biotin-6-ahx-D-Tyr-Tic-Ser-
In the P3 position we selected the preferred Cys(Bzl), Hyp(Bzl), Asp(Me)-AOMK [MP-L02] and biotin-6-ahx-D-Arg-Tic-Ser-
and Tic as the most promising candidates. Finally, we looked Asp(Me)-AOMK [MP-L03]), since methylation of carboxylic acid
toward preferences in the S4 pocket, and exploited the dif- residues is a common strategy to improve cell permeability (Fig-
ferential requirement for occupancy of this pocket between ure 4). Once methylated probes enter the cell, endogenous es-
caspases and legumain. Legumain had no preferences in this terases might remove the methyl group and make the probe
position and also tolerated some D-amino acids (D-Arg, D-Tyr, ready for enzyme targeting. The use of D-Arg in the P4 position
and D-Phg), whereas caspases did not (although caspase 8 of MP-L03 was designed to enhance probe solubility, allowing
displayed a very minimal tolerance for these D-amino acids) it to be used at a higher concentration. Addition of MP-
(
Poreba et al., 2014a). After selection of the amino acid building L01, -L02, and -L03 (0–10 mM) to equal amounts of lysates
blocks, we synthesized several substrates and measured (3.5 mg total protein) from the stable monoclonal legumain-over-
kinetic parameters of their hydrolysis by legumain and caspases expressing M38L cell line (Smith et al., 2012) revealed that all
in their optimal buffer and pH conditions (Figures 3 and S1; three probes bound and inhibited active legumain (intermediate
Table S3).
46-kDa and mature 36-kDa forms) in a concentration-dependent
manner (Figure 5). The potency of MP-L01 was approximately
100-fold higher compared with either methylated probe (MP-
L02 or -L03), which showed similar legumain binding and inhibi-
Design and Binding of Legumain-Selective
Activity-Based Probes
One of the main aims of this work was to find irreversible ABPs tion. These observations were reflected in both legumain activity
that could selectively label the active form of legumain in a measurements (Figure 5A) and visualized by comigration of
Cell Chemical Biology 23, 1–13, August 18, 2016
5

Please cite this article in press as: Poreba et al., Counter Selection Substrate Library Strategy for Developing Specific Protease Substrates and Probes,
Cell Chemical Biology (2016), http://dx.doi.org/10.1016/j.chembiol.2016.05.020
Figure 4. Synthesis and Structures of the Legumain-Selective Activity-Based Probes
(A–D) The synthesis of MP-L01 probe (A–C) is presented in detail, and probes with methylated P1-Asp (MP-L02 and MP-L03; D) were synthesized in a similar
manner using NH -Asp(Me)-AOMK instead of NH -Asp(Bzl)-AOMK. MP-L01 has a free aspartic acid, whereas MP-L02 and MP-L03 have an aspartic acid methyl
2
2
ester at P1. The only difference between the two methylated probes is the P4 position (D-Tyr in MP-L02 and D-Arg in MP-L03).
legumain and each probe on immunoblots (Figure 5B). Interest- thepsins were detected by the probes, suggesting no probe
ingly, although an equal amount of lysate was used in the cross-reactivity with cathepsins (Figure 5B).
incubations, addition of increasing concentrations of probes
The most potent MP-L01 probe (0–100 nM) was incubated
indicated an accumulation of mature legumain (36 kDa) as well with the M38L lysate, and legumain activity in the lysate showed
as the intermediate (46/47 kDa) forms (data not shown). None a dose-dependent decrease with increasing probe concentra-
of the probes bound the proform of legumain in the cell lysate. tion (Figure S4A). Binding of probe to active legumain (36 kDa)
Previously, expression and increased processing of cathepsin was confirmed by immunoblotting (Figure S4B). Secreted prole-
B and L from single-chain to double-chain active forms have gumain (56 kDa) in serum-free conditioned M38L cell medium
been found in the legumain-overexpressing M38L cells (Smith (CM) was allowed to autoactivate at pH 4.0 to the active interme-
et al., 2012). However, no immunobands corresponding to ca- diate legumain form (46 kDa), which was found to bind MP-L01 in
6
Cell Chemical Biology 23, 1–13, August 18, 2016

Please cite this article in press as: Poreba et al., Counter Selection Substrate Library Strategy for Developing Specific Protease Substrates and Probes,
Cell Chemical Biology (2016), http://dx.doi.org/10.1016/j.chembiol.2016.05.020
Table 1. Selectivity of MP-L01 toward Legumain Compared with Caspases
%
Enzyme Inhibition (Enzyme Concentration 10 nM)
MP-L01 (nM)
Legumain
100
100
100
99
Caspase-2
Caspase-3
Caspase-6
Caspase-7
Caspase-8
Caspase-9
Caspase-10
3
1
8
4
2
1
5
20
60
0
–
–
–
–
–
–
–
14
6
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
9
–
–
–
–
–
–
7
–
–
–
–
–
28
21
6
0
–
–
0
94
–
–
0
95
–
–
53
–
–
–
ꢂ
Blank entries (ꢁ) designate no observable inhibition. The enzymes were incubated with MP-L01 probe for 30 min at 37 C. Optimal enzyme buffers
(
pH 5.8 for legumain and pH 7.4 for caspases) were used. The experiment was performed in triplicate and average results are presented (SD < 10%).
a similar manner to mature active legumain (36 kDa) at pH 5.8 zymes (Table 1), none of the probes bound to caspases, indi-
Figure S4C). Autoactivation of recombinant human prolegumain cating legumain selectivity. High molecular weight proteins
(
was also performed at pH 4.0, showing similar MP-L01 binding (>80 kDa) detected by streptavidin were most pronounced in
and inhibition (not shown). Importantly, the probe did not bind the samples harvested in SDS-PAGE sample buffer, since direct
either cellular, secreted (Figures S4B and S4C), or recombinant lysis in this buffer allowed loading of higher total protein concen-
prolegumain (56 kDa). Using probe detection by streptavidin, a trations on the gels. Legumain activity measurements of the
high molecular weight biotinylated protein (approximately samples in legumain lysis buffer reflected the immunoblotting
8
0 kDa) was detected in the cell lysate but not in the conditioned observations, suggesting that the probes entered, bound, and
medium, and no proteins of the size of other active cysteine pro- inhibited intracellular legumain with a similar potency (Figure 6B).
teases such as cathepsins (23–30 kDa) or caspases (18–23 kDa). Whether the methylated probes were demethylated by intracel-
Determination of the inhibition rate constants (kobs/I) of MP- lular esterases before legumain binding is probable, although
L01, -L02, and -L03 was performed with autoactivated recombi- there is no formal proof.
nant human legumain (R&D Systems) using Z-Ala-Ala-Asn-AMC
To investigate the minimum time needed for probe cell
m
as substrate (Table S4). K for the Z-Ala-Ala-Asn-AMC legumain entrance and detection, we conducted a time-course (0–24 hr)
substrate used in these calculations as well as in the cell exper- experiment, which showed maximal detection of the MP probes
iment analyses was determined to be 25.7 mM. Comparison of (MP-L01, -L02, -L03) after 8–24 hr of incubation (Figure S6). The
the kobs/I values (by the method of Salvesen and Nagase, P1 methylated probes seemed to be detectable in the cells
2
001) for recombinant legumain revealed that the MP-L01 probe slightly earlier than the non-methylated one. Overall, the selected
had an inhibition rate constant approximately 100-fold higher 24-hr incubation for the dose-response experiments was
than the two methylated probes (MP-L02 and -L03), which corre- reasonable and practical, although 8–12 hr of incubation seemed
lated with the observed legumain inhibition pattern in M38L cell to be sufficient. In this experiment, the caspase-3/-7 ABP biotin-
lysate, whereby the unmethylated MP-L01 probe inhibited the 6-ahx-DEVD-AOMK (DEVD) was included, and incubation of
enzyme activity more efficiently (Figure 5A).
M38L cells with the DEVD probe and harvesting in SDS sample
buffer showed that this probe did not bind any intracellular pro-
teins in these cells or was not cell permeable. Low permeability
was probably due to three carboxylic groups, which are depro-
Selectivity of Legumain Activity-Based Probes in Living
Cells
Intracellular levels of the three probes were studied by treating tonated and make the probe hydrophilic. Surprisingly, cells incu-
M38L cells for 24 hr with 0–100 mM probe. Parallel wells were bated with the DEVD probe and harvested in legumain lysis
harvested in either legumain lysis buffer (pH 5.8) or SDS-PAGE buffer showed that this probe bound to and inhibited active
sample buffer. SDS-PAGE sample buffer was used to avoid legumain (36 kDa), probably due to probe attachment to the
possible false positives resulting from unspecific binding of well plastic or cell membrane and binding after lysis (not
probes to cell surfaces or well plastic and subsequent binding shown). Thus, the inhibition rate constants (kobs/I) for the DEVD
to legumain after cell lysis. Lysates in each buffer were analyzed probe was measured and calculated (Table S4), showing that
by immunoblotting and showed similar results, confirming cell this probe was inhibiting recombinant legumain approximately
entrance and probe binding to legumain in the cells before lysis. 10-fold better than the methylated legumain probes (MP-L02
All three probes were detected by immunoblotting at probe con- and -L03) but with 20-fold lower efficiency than the best, and
centrations R10 mM (Figures 6A and S5). As with direct addition non-methylated, legumain probe (MP-L01).
of probes to cell lysate (Figure 5), treatment of living cells resulted
in probe binding to and accumulation of mature active legumain The Use of MP-L01 in Apoptosis Models
(
36 kDa). Furthermore, probe binding and inhibition of legumain To investigate in detail the MP-L01 selectivity for legumain
resulted in an accumulation and decreased autoprocessing of versus caspases and compare with the caspase-3 biotin-6-
prolegumain (56 kDa; Figures 6A and S5). As revealed by probe ahx-DEVD-AOMK probe (DEVD), we used two different
incubation directly in cell lysates (Figure 5) or with whole cells apoptosis models. In the first model, M38L cells were
(
Figures 6 and S5), as well as kinetic analyses using pure en- cotransfected with caspase-8 and its physiological substrate
Cell Chemical Biology 23, 1–13, August 18, 2016
7

Please cite this article in press as: Poreba et al., Counter Selection Substrate Library Strategy for Developing Specific Protease Substrates and Probes,
Cell Chemical Biology (2016), http://dx.doi.org/10.1016/j.chembiol.2016.05.020
Figure 5. Direct Probe Binding and Inhibition
A
of Legumain in Cell Lysates
MP-L01, -L02, or -L03 (0–10 mM) was added to
equal volumes of M38L cell lysate (3.5 mg of total
protein) and incubated in legumain assay buffer
ꢂ
(
pH 5.8) (total volume 50 mL) for 30 min at 30 C.
(
A) Residual legumain activity measurements as
percentage of untreated controls ± SD. Note that
the x axis is logarithmic.
(
B) Detection of legumain (red), probes (green), and
overlay (yellow) by immunoblotting (upper panels).
b-Tubulin (50 kDa) was used as loading control
(
lower panels). Lane 1, control; lane 2, 1-nM probe;
lane 3, 10-nM probe; lane 4, 100-nM probe; lane 5,
-mM probe; lane 6, 10-mM probe; -, empty lane; L,
1
Broad-Range SDS-PAGE Standard Ladder (Bio-
Rad). Representative overlay blots are shown.
n = 3–7.
B
MP-L01 at pH 5.8 and not at pH 7.2,
whereas the DEVD probe bound cas-
pase-3 at both pH values (Figure S7).
DISCUSSION
Unnatural (non-proteinogenic) amino acids
offer an excellent platform for the design of
new substrates and ABPs for protease
investigations. The broad structural and
chemical diversity of these amino acids is
especially useful when analyzing proteases
displaying overlapping substrate speci-
ficity. Recently we reported that the
HyCoSuL approach (substrate libraries
employing natural and unnatural amino
caspase-3 to obtain a maximal apoptotic signal. After 24 hr, cells acids) is suitable for development of tetrapeptide substrates dis-
were harvested in legumain lysis buffer (pH 5.8) or caspase lysis tinguishing apoptotic caspases within the group (Poreba et al.,
buffer (pH 7.4), incubating with MP-L01 or DEVD probes and 2014a). In the present study we took advantage of the HyCoSuL
immunoblotting (Figure 7). MP-L01 only bound legumain, while to determine the substrate specificity profile of human legumain.
the DEVD probe labeled both legumain and caspase-3. Legu- We found that the best recognized amino acid in the P2 position is
main binding of MP-L01 was only detected at pH 5.8 and not natural Thr, although other small amino acids were also well toler-
at pH 7.4, which is in agreement with decreased activity of legu- ated. The P3 position was more liberal recognizing most amino
main above pH 6 (Chen et al., 1997). In control cells (non-co- acids; however, the best were cysteine-benzyl derivatives.
transfected) only legumain activity was detected. As a second Importantly, the P4 position displayed no preferences, tolerating
model we employed cytochrome c-mediated programmed almost all amino acids, including those with D-stereochemistry.
in vitro apoptosis (Denault et al., 2007; Poreba et al., 2014a). Based on these results we concluded that the first two positions
Hypotonic extracts from HEK293 FreeStyle cells were activated in substrates (P2 and P1) are crucial for legumain activity (Table
ꢂ
by adding 10 mM cytochrome c and 1 mM dATP at 37 C, driving S1), and the P3 position could be important for legumain selec-
caspase-9, -3, and -7 activation. Subsequently, aliquots of the tivity. Because P4 had a wide tolerance we were able to employ
activated cell extract were directly incubated with MP-L01 or the counterselection strategy (CoSeSuL) to select for probes that
DEVD probe (0–1 mM) in either legumain or caspase activity would not react with caspases and showed exquisite selectivity
ꢂ
buffer (pH 5.8 and pH 7.2, respectively) for 5 or 30 min at 30 C. for legumain. To test this hypothesis, in the next step we used
Immunoblotting showed that procaspase-3 (35 kDa) was acti- the P4 and P3 legumain specificity matrices to obtain the first
vated to the 17- and 12-kDa forms (Figure S7). Whereas the P1-Asp legumain-selective substrates that were not recognized
DEVD probe strongly bound active caspase-3 both at pH 5.8 by apoptotic caspases (Figure 3), and to design a sequence-
and pH 7.2 at all concentrations tested, MP-L01 only weakly basedlegumain-selectivebiotin-labeledABP(MP-L01)(Figure4).
bound to caspase-3 (Figure S7). When incubating the probes This probe was equipped with a biotin tag (N termini) and
(
(
10 nM) with 50 nM recombinant autoactivated legumain electrophilic acyloxymethyl group (AOMK), which inhibits and la-
46 kDa) or active caspase-3 (17 kDa), legumain was bound by bels active legumain by forming the covalent bond with the
8
Cell Chemical Biology 23, 1–13, August 18, 2016

Please cite this article in press as: Poreba et al., Counter Selection Substrate Library Strategy for Developing Specific Protease Substrates and Probes,
Cell Chemical Biology (2016), http://dx.doi.org/10.1016/j.chembiol.2016.05.020
A
B
Figure 6. MP-L01, -L02 and -L03 Bind and Inhibit Legumain in Living
Cells
Figure 7. MP-L01 Does Not Bind Caspases in Caspase-Cotrans-
fected M38L Cells
Legumain-overexpressing M38L cells (25,000 cells/well; 48-well plate) were
incubated with the probes (0–100 mM) for 24 hr before parallel wells were
harvested in either 50 mL of legumain lysis buffer or SDS-PAGE sample buffer
Apoptosis was induced in M38L cells (500,000/six-well) by cotransfection of
caspase-8 and caspase-3 cDNA (C-8/C-3) for 24 hr before harvesting in le-
gumain (pH 5.8) or caspase (pH 7.4) lysis buffer. Aliquots of lysates were
incubated with MP-L01 or biotin-6-ahx-DEVD-AOMK (DEVD) probes (1 mM) at
(
(
Figure S5) and analyzed by immunoblotting.
A) One representative immunoblot of lysates in legumain lysis buffer from cells
incubated with MP-L01 is shown detecting probe binding (left panel, green in
right overlay panel), legumain (middle panel, red in right overlay panel), and
overlay panel (right). Equal amounts of lysates (13.3 mL) were loaded per lane.
For blots for all probes in both buffers, see Figure S5. Lane 1, control; lane 2,
ꢂ
pH 5.8 or pH 7.4 for 30 min at 30 C before addition of SDS-PAGE sample
buffer and immunoblotting. Equal amounts of lysates (20 mL) were loaded per
lane. The immunoblots show legumain (red) and probe (green) overlay (top
panel), as well as individual blots of MP-L01 and DEVD probes (detected by
streptavidin). Full-length (p32) and active caspase-3 (p17) were detected with
caspase-3 antibody. b-Tubulin served as loading control (n = 2). Asterisks
indicate background bands of unknown identity.
1-mM probe; lane 3, 10-mM probe; lane 4, 25-mM probe; lane 5, 50-mM probe;
lane 6, 75-mM probe; lane 7, 100-mM probe; L, Broad-Range SDS-PAGE
Standard Ladder (Bio-Rad).
(B) Legumain activity per microgram of total protein ± SD in all probe-treated
cell lysates. Normalized values ± SEM are presented. n = 3–6.
toward legumain in living cells strongly depends on two factors:
active-site cysteine residue. We also synthesized two other (1) their potency and (2) their cell permeability. Even if P1-Asp-
probes (MP-L02 and MP-L03) with the methylated P1 aspartic methylated derivatives are more cell permeable, once they reach
acid residue to make the compounds more cell permeable. Using the cytosol they must undergo the P1-Asp demethylation carried
the lysates from the stable monoclonal legumain-overexpressing out by esterases to function as legumain inhibitors.
M38L cell line, we demonstrated that all three probes bind effec-
Finally, we demonstrated MP-L01 selectivity for legumain
tively to active legumain and label neither the prolegumain form versus caspases following in vitro and in cellular models of
nor cysteine cathepsins. To further explore the utility of our apoptotic caspase induction (Figures 7 and S7). After cotrans-
probes in biological studies, we then used them to detect the fection of M38L cells with caspase-8 and full-length caspase-3
active legumain in living cells. We demonstrated that all three cDNA, we observed robust activation of caspase-3 using a
probes bind and inhibit legumain in M38L cells (Figure 6); how- DEVD probe; however, MP-L01 could not detect such activity
ever, the most potent was the non-P1-Asp-methylated MP- even after adding 1 mM directly to the lysate. We validated the
L01, which was thought to be less cell permeable than MP-L02 MP-L01 selectivity by employing a well-established cell-free
and MP-L03. It must be stressed here that the probes’ potency model of apoptosis, in which executioner caspases are activated
Cell Chemical Biology 23, 1–13, August 18, 2016
9

Please cite this article in press as: Poreba et al., Counter Selection Substrate Library Strategy for Developing Specific Protease Substrates and Probes,
Cell Chemical Biology (2016), http://dx.doi.org/10.1016/j.chembiol.2016.05.020
through cytochrome c addition. The comparison of MP-L01 and
biotin-6-ahx-DEVD-AOMK probes revealed that MP-L01 dis-
plays high selectivity toward legumain over caspases.
pH 5.8 (optimal pH for legumain activity). The synthesis and structural charac-
teristic of HyCoSuL has been described previously (Poreba et al., 2014a). This
library contains three sublibraries (P4, P3, and P2), and into each sublibrary
1
09 unnatural amino acids were incorporated. Prolegumain was autoactivated
as described in the previous section. The legumain assay buffer was 40 mM
citric acid, 1 mM EDTA, 120 mM Na HPO , and 10 mM DTT (pH 5.8). The re-
SIGNIFICANCE
2
4
action volume was 100 mL, the total final substrate mixture concentration was
1
00 mM, and the legumain concentration in the assay was 350 nM. Kinetic
ꢂ
Using the HyCoSuL approach, we determined the substrate
specificity of human legumain and in parallel used CoSeSuL
to develop new fluorogenic substrates and activity-based
probes that selectively target this protease. The utility of
the new activity-based probes were confirmed by labeling
active legumain in lysates from M38L cells and also in living
cells, providing a clear path toward investigating legumain
activity in cell-based studies. Finally, we showed that our
combined HyCoSuL/CoSeSuL approach led to the genera-
tion of a legumain probe (MP-L01) not binding to caspases,
which is of great significance as these enzymes display
overlapping substrate specificity. The new compounds
developed in this study can be used for tracking active legu-
main in various biological models in which caspases are also
active, and provide a novel paradigm in the development of
highly selective protease probes.
measurements were performed at 37 C and the data analysis was carried
out as described above.
Determination of Kinetic Parameters kcat, K
Legumain-Selective Substrates
M M
, and kcat/K for New
To determine whether the obtained substrates were legumain selective and
not also hydrolyzed by caspases, we determined their kinetic parameters to-
ward legumain and caspases 3, 6, 7, 8, 9, and 10. The detailed protocol for
the determination of these kinetic parameters can be found in Supplemental
Experimental Procedures.
Determination of Direct Legumain Binding and Inhibition by MP-L01,
MP-L02, and MP-L03 in Cell Lysates
The HEK293 monoclonal legumain-overexpressing M38L cell line was used in
this study and is thoroughly characterized elsewhere (Smith et al., 2012). Cells
were cultured in DMEM supplemented with 10% fetal bovine serum, L-gluta-
mine, and antibiotics (100 U/mL penicillin, 100 mg/mL streptomycin, and
8
00 mg/mL G418). To obtain CM for prolegumain autoactivation (see below),
we cultured the cells for 4 days in serum-free DMEM before CM harvesting.
Cells were washed in PBS before total cell lysates were obtained in a lysis
buffer maintaining legumain activity (100 mM sodium citrate, 1 mM diso-
dium-EDTA, 1% n-octyl-b-glucopyranoside [pH 5.8]), and three cycles of
EXPERIMENTAL PROCEDURES
Preparation of Recombinant Caspases and Legumain
The detailed description of expression and purification of caspases can be
found elsewhere (Stennicke and Salvesen, 1999). The detailed protocol for
cloning, expression, and purification of human legumain can be found in Sup-
plemental Experimental Procedures.
ꢂ
ꢂ
freezing (ꢁ70 C) and thawing (30 C) before centrifugation at 10,000 3 g for
min. Total protein concentrations in the lysates were measured by a Pierce
5
BCA Protein Assay Kit and performed according to the manufacturer in a mi-
croplate reader (Spectra Max 190, Molecular Devices) measuring absorbance
at 562 nm. Bovine g-globulin (Pierce) was used to establish a standard curve
for the calculation of the total protein concentrations. First, aliquots of M38L
cell lysate (3.5 mg of total protein) was directly added to various concentrations
Enzymatic Kinetic Studies
These studies were carried out using an fMax spectrofluorimeter (Molecular
Devices) operating in the kinetic mode in 96-well Corning plates (Corning).
The excitation wavelength for ACC-labeled substrates was 355 nm and the
emission wavelength was 460 nm (cutoff at 455 nm). Prolegumain was first
activated in the following buffer: 100 mM sodium acetate, 10 mM DTT,
(
0–10 mM) of the MP-L01, -L02, or -L03 probes and incubated in legumain
assay buffer (39.5 mM citric acid, 121 mM Na
2
4 2
HPO , 1 mM Na EDTA, 0.1%
ꢂ
CHAPS, and 1 mM DTT, total volume 50 mL) at 30 C for 30 min before legumain
activity measurements and immunoblotting. Legumain activity measurements
were performed on cell lysates using the peptide substrate Z-Ala-Ala-Asn-
1
mM EDTA (pH 4.5). The legumain autoactivation was measured using the flu-
orogenic substrate Ac-DGTN-ACC. After 150 min legumain was fully active.
The active legumain concentration was determined by titration using Z-ATN-
AOMK irreversible inhibitor. In the next step active legumain was diluted in
legumain assay buffer containing 40 mM citric acid, 1 mM EDTA, 120 mM
AMC (Bachem) and a kinetic protocol described elsewhere (Johansen et al.,
ꢂ
1
999). Increase in fluorescence over 15 min at 30 C was determined in a
microplate spectrofluorimeter (Spectra Max Gemini EM, Molecular Devices)
using excitation wavelength 355 nm and emission wavelength 460 nm (cutoff
2 4
Na HPO , 10 mM DTT, at pH 4.5 (buffer 1) or pH 5.8 (buffer 2). To determine
4
55 nm). Gel electrophoresis and immunoblotting of cell lysates and CM were
whether legumain substrate specificity depended on the pH, we examined
this enzyme in both buffers. First, legumain substrate specificity was deter-
mined using two tetrapeptide libraries (of only natural amino acids, a traditional
Positional Scanning Substrate Combinatorial Library [PS-SCL] approach) con-
performed using Bolt 4%–12% Bis-Tris Plus Gels (Novex, Life Technologies)
and nitrocellulose membranes (0.2 mM; Bio-Rad). Ponceau staining verified
equal loading. Membranes were blocked with 2% BSA in Tris-buffered saline
with Tween 20 (TBS-T) (v/v). Legumain was detected by subsequent goat
anti-human legumain (1:1,000; R&D Systems) and fluorescent donkey anti-
goat IRDye 680LT (1:10,000; LI-COR Biosciences; red) in TBS-T. Fluorescent
streptavidin IRDye 800CW (1:10,000; LI-COR; green) was used to visualize the
biotinylated probes, and yellow bands indicated legumain and probe comigra-
tion. Fluorescence was scanned at 700 nm (red) and 800 nm (green) in an Od-
yssey fluorescence imaging system (LI-COR). Second, M38L cells were
cultured for 4 days in serum-free medium and the CM containing large
amounts of secreted prolegumain was collected and centrifuged for 5 min at
taining Asp or Asn in P1 position. Kinetic measurements were performed at
ꢂ
37 C. The library concentration in each well was 100 mM, the reaction volume
was 100 mL, and the legumain concentration was 80 nM for the P1-Asn library
and 350 nM for the P1-Asp library. The total assay time was 15–30 min, and the
linear portion of each progress curve (about 10 min) was used to calculate sub-
strate hydrolysis velocity. All experiments were performed at least three times
and the results are presented as mean values. The SE for each sample mea-
surement was less than 10%. Finally, the substrate specificity matrix was
made based on the RFU/s value (relative fluorescence units per second) for
each substrate, setting the highest value from each sublibrary to 100% and ad-
justing the other substrates accordingly. The precise procedure for the deter-
mination of protease substrate specificity profile using the PS-SCL approach
can be found elsewhere (Poreba et al., 2014b).
1
,000 3 g to remove cell debris contaminations. The supernatant was put
on a PD-10 column for sampling in legumain activation buffer (200 mM acetic
acid, 4 mM Na
2
-EDTA [pH 4.0]) as previously described (Smith et al., 2012). Af-
ꢂ
ter incubation and autoactivation at 30 C for 4–6 hr, aliquots was added
various concentrations of MP-L01, -L02, or -L03 (0–1 mM) and incubated for
ꢂ
an additional 30 min at 30 C. The final incubates were analyzed by legumain
Characterization of Legumain Specificity Using HyCoSuL
activity measurements and immunoblotting as above. Third, recombinant hu-
Since legumain displayed the same specificity at pH 4.5 and pH 5.8, its
full substrate specificity using the HyCoSuL approach was determined at
man prolegumain (2.045 mM; R&D Systems) was autoactivated by incubation
ꢂ
in legumain activation buffer (pH 4.0) at 37 C for 2 hr according to the
10 Cell Chemical Biology 23, 1–13, August 18, 2016

Please cite this article in press as: Poreba et al., Counter Selection Substrate Library Strategy for Developing Specific Protease Substrates and Probes,
Cell Chemical Biology (2016), http://dx.doi.org/10.1016/j.chembiol.2016.05.020
manufacturer, and aliquots (200 nM) were further incubated for 20 min with and
without MP-L01 (0–1 mM), before legumain activity measurements and immu-
noblotting were performed as already described.
SUPPLEMENTAL INFORMATION
Supplemental Information includes Supplemental Experimental Procedures,
seven figures, four tables, and two data files can be found with this article
online at http://dx.doi.org/10.1016/j.chembiol.2016.05.020.
Determination of Inhibition Kinetics kobs/I for the Activity-Based
Probes MP-L01, MP-L02, and MP-L03
For the following studies, recombinant human prolegumain (rhLeg) was pur-
chased from R&D Systems and autoactivated in legumain activation buffer
AUTHOR CONTRIBUTIONS
(
pH 4.0) as described above. A constant amount of 2.5 nM enzyme was added
M.P., R.S., W.R., G.S.S., and M.D. designed research; M.P., R.S., W.R.,
N.N.L., P.K., and S.J.S. performed research; M.M., D.T., and B.T. contributed
new reagents/analytical tools; M.P., R.S., W.R., N.N.L., P.K., S.J.S., G.S.S.,
and M.D. analyzed data; and M.P., R.S., G.S.S., and M.D. wrote the paper.
to various concentrations of probes (at least 5-fold excess) and 100 mM Z-Ala-
Ala-Asn-AMC substrate. Probes and substrate were incubated together in le-
ꢂ
gumain assay buffer (pH 5.8) at 30 C for 10 min before addition to the enzyme
and fluorescence measurements every second for 30 min. The reactions were
repeated three times for each probe and kobs/I calculated as previously
ACKNOWLEDGMENTS
M
described (Kasperkiewicz et al., 2014). In addition, K for legumain cleavage
of the legumain substrate Z-Ala-Ala-Asn-AMC was measured using 2.5 nM au-
toactivated rhLeg and various concentrations of substrate (1.565–200 nM) at
the same conditions as above.
This work was supported by the Ministry of Science and Higher Education in
Poland (grant Iuventus Plus IP2012 040172 to M.P., a statutory activity subsidy
for the Faculty of Chemistry at Wroclaw University of Technology to M.P. and
M.D.), University of Oslo, Norway (R.S. and N.N.L.), the grants P1-140 and P1-
0048 from the Research Agency of Slovenia (B.T. and D.T.), and the grants
2011/03/B/ST5/01048 (National Science Center) and FOCUS (Foundation for
Polish Science) to M.D. M.P. and P.K. are beneficiaries of START scholarship
from the Foundation for Polish Science. This project has received funding from
the European Union’s Horizon 2020 research and innovation program under
the Marie Sk1odowska-Curie grant agreement No. 661187.
Entrance of Activity-Based Probes into Living Cells
M38Lcells (100,000 or 25,000) were seeded in 12- or 48-well plates, respec-
tively, and incubated with various concentrations (0–100 mM) of MP-
L01, -L02, or -L03 for 0–24 hr. In addition, the caspase activity-based probe
biotin-6-ahx-DEVD-AOMK was included for comparison. DMSO was added
to control cells. The cells were washed in PBS and parallel treated wells
were lysed in either legumain lysis buffer (pH 5.8) or SDS-PAGE sample buffer.
The cell lysates in lysis buffer were subjected to three circles of freezing-thaw-
ing before centrifugation and legumain activity measurements. Lysates in
SDS-PAGE sample buffer were sonicated before 10 min of boiling to ensure
complete solubilization. All lysate samples were analyzed by immunoblotting
as above.
Received: March 1, 2016
Revised: May 11, 2016
Accepted: May 30, 2016
Published: July 28, 2016
REFERENCES
Caspase-8 and Full-Length Caspase-3 Cotransfection into M38L
Cells
Alvarez-Fernandez, M., Barrett, A.J., Gerhartz, B., Dando, P.M., Ni, J., and
Abrahamson, M. (1999). Inhibition of mammalian legumain by some cystatins
is due to a novel second reactive site. J. Biol. Chem. 274, 19195–19203.
To investigate the MP-L01 selectivity toward legumain over caspase-3
and -8 in a competitive manner, we cotransfected legumain-overexpressing
M38L cells with full-length caspase-8 in pcDNA3 (Boatright et al., 2003) and
Asgian, J.L., James, K.E., Li, Z.Z., Carter, W., Barrett, A.J., Mikolajczyk, J.,
Salvesen, G.S., and Powers, J.C. (2002). Aza-peptide epoxides: a new class
of inhibitors selective for clan CD cysteine proteases. J. Med. Chem. 45,
0
0
caspase-3 cloned into pcDNA3 with 5 EcoRI and 3 XbaI. We seeded M38L
cells (500,000 cells/well, 2 mL) on six-well plates. After 24 hr we mixed 2.5 mg
of caspase-8 plasmid, 2.5 mg of caspase-3 plasmid, and 25 mg of Lipofect-
amine 2000 (Invitrogen) in 300 mL Opti-MEM (OM, Reduced Serum Media,
Thermo Fisher) buffer and incubated the mixture for 30 min at room temper-
ature. After 30 min, we added the mixture dropwise into the cells (300 mL per
well). We also included a negative control (25 mg of Lipofectamine 2000 in
4
958–4960.
Bajjuri, K.M., Liu, Y., Liu, C., and Sinha, S.C. (2011). The legumaisn protease-
activated auristatin prodrugs suppress tumor growth and metastasis without
toxicity. ChemMedChem 6, 54–59.
3
00 mL of Opti-MEM buffer without caspase cDNA plasmids). After 24 hr,
Boatright, K.M., Renatus, M., Scott, F.L., Sperandio, S., Shin, H., Pedersen,
I.M., Ricci, J.E., Edris, W.A., Sutherlin, D.P., Green, D.R., et al. (2003). A unified
model for apical caspase activation. Mol. Cell 11, 529–541.
we harvested cells and labeled the lysates with MP-L01 and DEVD probes
as described above under ‘‘The Use of MP-L01 in Apoptosis Models.’’ Legu-
main lysis/reaction buffer: 40 mM citric acid, 120 mM Na
2 4
HPO , 1 mM EDTA,
Briggs, J.J., Haugen, M.H., Johansen, H.T., Riker, A.I., Abrahamson, M.,
Fodstad, O., Maelandsmo, G.M., and Solberg, R. (2010). Cystatin E/M sup-
presses legumain activity and invasion of human melanoma. BMC Cancer
1% IGEPAL, 10 mM DTT (pH 5.8); caspase lysis/reaction buffer: 10% (w/v)
sucrose, 20 mM PIPES, 10 mM NaCl, 1 mM EDTA, 1% IGEPAL, 10 mM
DTT (pH 7.4).
1
0, 17.
Chan, C.B., Abe, M., Hashimoto, N., Hao, C., Williams, I.R., Liu, X., Nakao, S.,
Yamamoto, A., Zheng, C., Henter, J.I., et al. (2009). Mice lacking asparaginyl
endopeptidase develop disorders resembling hemophagocytic syndrome.
Proc. Natl. Acad. Sci. USA 106, 468–473.
Selectivity of MP-L01 and DEVD Probe for Legumain versus
Caspases, Respectively
To finally demonstrate that MP-L01 was not binding to cellular caspases, we
obtained cytosolic extracts of FreeStyle 293-F cells (Life Technologies) in a hy-
potonic buffer as described previously (Denault et al., 2007). Each 100-mL
portion of FreeStyle293-F cytosolic extract was activated by addition of 1 mL
of dATP (12 mM in water) and 2 mL of horse cytochrome c (1 mM in water). Acti-
vated extract was then incubated with MP-L01 or DEVD probe (0–1 mM) for 5 or
Chen, J.M., Dando, P.M., Rawlings, N.D., Brown, M.A., Young, N.E., Stevens,
R.A., Hewitt, E., Watts, C., and Barrett, A.J. (1997). Cloning, isolation, and
characterization of mammalian legumain, an asparaginyl endopeptidase.
J. Biol. Chem. 272, 8090–8098.
Chen, J.M., Rawlings, N.D., Stevens, R.A., and Barrett, A.J. (1998).
Identification of the active site of legumain links it to caspases, clostripain
and gingipains in a new clan of cysteine endopeptidases. FEBS Lett. 441,
3
0 min in legumain (pH 5.8) or caspase (pH 7.2) activity buffer, respectively,
before the incubates were analyzed by immunoblotting using caspase-3 anti-
body (1:1,000; Cell Signaling #9662) and fluorescent donkey anti-rabbit IRDye
3
61–365.
680LT (1:10,000, LI-COR; red), in addition to streptavidin IRDye 800CW.
Recombinant autoactivated legumain (50 nM) and recombinant human active
caspase-3 (50 nM) were incubated with 10-nM probes and included as
controls.
Chen, J.M., Fortunato, M., and Barrett, A.J. (2000). Activation of human prole-
gumain by cleavage at a C-terminal asparagine residue. Biochem. J. 352,
327–334.
Cell Chemical Biology 23, 1–13, August 18, 2016 11

Please cite this article in press as: Poreba et al., Counter Selection Substrate Library Strategy for Developing Specific Protease Substrates and Probes,
Cell Chemical Biology (2016), http://dx.doi.org/10.1016/j.chembiol.2016.05.020
Chen, J.M., Fortunato, M., Stevens, R.A., and Barrett, A.J. (2001). Activation of
progelatinase A by mammalian legumain, a recently discovered cysteine pro-
teinase. Biol. Chem. 382, 777–783.
Kembhavi, A.A., Buttle, D.J., Knight, C.G., and Barrett, A.J. (1993). The two
cysteine endopeptidases of legume seeds: purification and characterization
by use of specific fluorometric assays. Arch. Biochem. Biophys. 303, 208–213.
Choi, S.J., Reddy, S.V., Devlin, R.D., Menaa, C., Chung, H., Boyce, B.F., and
Roodman, G.D. (1999). Identification of human asparaginyl endopeptidase (le-
gumain) as an inhibitor of osteoclast formation and bone resorption. J. Biol.
Chem. 274, 27747–27753.
Lecaille, F., Muno, D., Kominami, E., and Ishidoh, K. (2004). Proteinases
participating in the processing and activation of prolegumain in primary
cultured rat macrophages. Biol. Chem. 385, 511–516.
Lee, J., and Bogyo, M. (2010). Development of near-infrared fluorophore
(NIRF)-labeled activity-based probes for in vivo imaging of legumain. ACS
Chem. Biol. 5, 233–243.
Clerin, V., Shih, H.H., Deng, N., Hebert, G., Resmini, C., Shields, K.M.,
Feldman, J.L., Winkler, A., Albert, L., Maganti, V., et al. (2008). Expression of
the cysteine protease legumain in vascular lesions and functional implications
in atherogenesis. Atherosclerosis 201, 53–66.
Lee, J., and Bogyo, M. (2012). Synthesis and evaluation of aza-peptidyl inhib-
itors of the lysosomal asparaginyl endopeptidase, legumain. Bioorg. Med.
Chem. Lett. 22, 1340–1343.
Dall, E., and Brandstetter, H. (2012). Activation of legumain involves proteolytic
and conformational events, resulting in a context- and substrate-dependent
activity profile. Acta Crystallogr. Sect. F Struct. Biol. Cryst. Commun. 68,
Li, N., Liu, Q., Su, Q., Wei, C., Lan, B., Wang, J., Bao, G., Yan, F., Yu, Y., Peng,
B., et al. (2013). Effects of legumain as a potential prognostic factor on gastric
cancers. Med. Oncol. 30, 621.
24–31.
Dall, E., and Brandstetter, H. (2013). Mechanistic and structural studies on le-
gumain explain its zymogenicity, distinct activation pathways, and regulation.
Proc. Natl. Acad. Sci. USA 110, 10940–10945.
Liu, C., Sun, C., Huang, H., Janda, K., and Edgington, T. (2003).
Overexpression of legumain in tumors is significant for invasion/metastasis
and a candidate enzymatic target for prodrug therapy. Cancer Res. 63,
2
957–2964.
Loak, K., Li, D.N., Manoury, B., Billson, J., Morton, F., Hewitt, E., and Watts, C.
2003). Novel cell-permeable acyloxymethylketone inhibitors of asparaginyl
Dall, E., and Brandstetter, H. (2016). Structure and function of legumain in
health and disease. Biochimie 122, 126–150.
(
Denault, J.B., Eckelman, B.P., Shin, H., Pop, C., and Salvesen, G.S. (2007).
Caspase 3 attenuates XIAP (X-linked inhibitor of apoptosis protein)-mediated
inhibition of caspase 9. Biochem. J. 405, 11–19.
endopeptidase. Biol. Chem. 384, 1239–1246.
Luo, Y., Zhou, H., Krueger, J., Kaplan, C., Lee, S.H., Dolman, C., Markowitz,
D., Wu, W., Liu, C., Reisfeld, R.A., et al. (2006). Targeting tumor-associated
macrophages as a novel strategy against breast cancer. J. Clin. Invest. 116,
Edgington, L.E., Verdoes, M., Ortega, A., Withana, N.P., Lee, J., Syed, S.,
Bachmann, M.H., Blum, G., and Bogyo, M. (2013). Functional imaging of legu-
main in cancer using a new quenched activity-based probe. J. Am. Chem. Soc.
2
132–2141.
Manoury, B., Mazzeo, D., Li, D.N., Billson, J., Loak, K., Benaroch, P., and
Watts, C. (2003). Asparagine endopeptidase can initiate the removal of the
MHC class II invariant chain chaperone. Immunity 18, 489–498.
1
35, 174–182.
Ekici, O.D., Gotz, M.G., James, K.E., Li, Z.Z., Rukamp, B.J., Asgian, J.L.,
Caffrey, C.R., Hansell, E., Dvorak, J., McKerrow, J.H., et al. (2004). Aza-pep-
tide Michael acceptors: a new class of inhibitors specific for caspases and
other clan CD cysteine proteases. J. Med. Chem. 47, 1889–1892.
Mathieu, M.A., Bogyo, M., Caffrey, C.R., Choe, Y., Lee, J., Chapman, H., Sajid,
M., Craik, C.S., and McKerrow, J.H. (2002). Substrate specificity of schisto-
some versus human legumain determined by P1-P3 peptide libraries. Mol.
Biochem. Parasitol. 121, 99–105.
Garcia-Calvo, M., Peterson, E.P., Rasper, D.M., Vaillancourt, J.P., Zamboni,
R., Nicholson, D.W., and Thornberry, N.A. (1999). Purification and catalytic
properties of human caspase family members. Cell Death Differ. 6, 362–369.
Mattock, K.L., Gough, P.J., Humphries, J., Burnand, K., Patel, L., Suckling, K.E.,
Cuello, F., Watts, C., Gautel, M., Avkiran, M., et al. (2010). Legumain and
cathepsin-L expression in human unstable carotid plaque. Atherosclerosis
Gawenda, J., Traub, F., Luck, H.J., Kreipe, H., and von Wasielewski, R. (2007).
Legumain expression as a prognostic factor in breast cancer patients. Breast
Cancer Res. Treat. 102, 1–6.
2
08, 83–89.
McLuskey, K., and Mottram, J.C. (2015). Comparative structural analysis of
Gotz, M.G., James, K.E., Hansell, E., Dvorak, J., Seshaadri, A., Sojka, D.,
Kopacek, P., McKerrow, J.H., Caffrey, C.R., and Powers, J.C. (2008). Aza-
peptidyl Michael acceptors. A new class of potent and selective inhibitors of
asparaginyl endopeptidases (legumains) from evolutionarily diverse patho-
gens. J. Med. Chem. 51, 2816–2832.
the caspase family with other clan CD cysteine peptidases. Biochem. J. 466,
2
19–232.
Miller, G., Matthews, S.P., Reinheckel, T., Fleming, S., and Watts, C. (2011).
Asparagine endopeptidase is required for normal kidney physiology and ho-
meostasis. FASEB J. 25, 1606–1617.
Hashimoto, S., Suzuki, T., Dong, H.Y., Yamazaki, N., and Matsushima, K.
Morita, Y., Araki, H., Sugimoto, T., Takeuchi, K., Yamane, T., Maeda, T.,
Yamamoto, Y., Nishi, K., Asano, M., Shirahama-Noda, K., et al. (2007).
Legumain/asparaginyl endopeptidase controls extracellular matrix remodeling
through the degradation of fibronectin in mouse renal proximal tubular cells.
FEBS Lett. 581, 1417–1424.
(1999). Serial analysis of gene expression in human monocytes and macro-
phages. Blood 94, 837–844.
Haugen, M.H., Johansen, H.T., Pettersen, S.J., Solberg, R., Brix, K., Flatmark,
K., and Maelandsmo, G.M. (2013). Nuclear legumain activity in colorectal can-
cer. PLoS One 8, e52980.
Niestroj, A.J., Feussner, K., Heiser, U., Dando, P.M., Barrett, A., Gerhartz, B.,
and Demuth, H.U. (2002). Inhibition of mammalian legumain by Michael accep-
tors and AzaAsn-halomethylketones. Biol. Chem. 383, 1205–1214.
James, K.E., Gotz, M.G., Caffrey, C.R., Hansell, E., Carter, W., Barrett, A.J.,
McKerrow, J.H., and Powers, J.C. (2003). Aza-peptide epoxides: potent and
selective inhibitors of Schistosoma mansoni and pig kidney legumains (aspar-
aginyl endopeptidases). Biol. Chem. 384, 1613–1618.
Ohno, Y., Nakashima, J., Izumi, M., Ohori, M., Hashimoto, T., and Tachibana,
M. (2013). Association of legumain expression pattern with prostate cancer
invasiveness and aggressiveness. World J. Urol. 31, 359–364.
Johansen, H.T., Knight, C.G., and Barrett, A.J. (1999). Colorimetric and fluori-
metric microplate assays for legumain and a staining reaction for detection of
the enzyme after electrophoresis. Anal. Biochem. 273, 278–283.
Ovat, A., Muindi, F., Fagan, C., Brouner, M., Hansell, E., Dvorak, J., Sojka, D.,
Kopacek, P., McKerrow, J.H., Caffrey, C.R., et al. (2009). Aza-peptidyl
Michael acceptor and epoxide inhibitors—potent and selective inhibitors of
Schistosoma mansoni and Ixodes ricinus legumains (asparaginyl endopepti-
dases). J. Med. Chem. 52, 7192–7210.
Kasperkiewicz, P., Poreba, M., Snipas, S.J., Parker, H., Winterbourn, C.C.,
Salvesen, G.S., and Drag, M. (2014). Design of ultrasensitive probes for human
neutrophil elastase through hybrid combinatorial substrate library profiling.
Proc. Natl. Acad. Sci. USA 111, 2518–2523.
Poreba, M., and Drag, M. (2010). Current strategies for probing substrate
specificity of proteases. Curr. Med. Chem. 17, 3968–3995.
Kato, D., Boatright, K.M., Berger, A.B., Nazif, T., Blum, G., Ryan, C., Chehade,
K.A., Salvesen, G.S., and Bogyo, M. (2005). Activity-based probes that target
diverse cysteine protease families. Nat. Chem. Biol. 1, 33–38.
Poreba, M., Kasperkiewicz, P., Snipas, S.J., Fasci, D., Salvesen, G.S., and
Drag, M. (2014a). Unnatural amino acids increase sensitivity and provide for
12 Cell Chemical Biology 23, 1–13, August 18, 2016

Please cite this article in press as: Poreba et al., Counter Selection Substrate Library Strategy for Developing Specific Protease Substrates and Probes,
Cell Chemical Biology (2016), http://dx.doi.org/10.1016/j.chembiol.2016.05.020
the design of highly selective caspase substrates. Cell Death Differ. 21, 1482–
492.
Schwarz, G., Brandenburg, J., Reich, M., Burster, T., Driessen, C., and
1
Kalbacher, H. (2002). Characterization of legumain. Biol. Chem. 383, 1813–
1
816.
Poreba, M., Szalek, A., Kasperkiewicz, P., and Drag, M. (2014b). Positional
scanning substrate combinatorial library (PS-SCL) approach to define cas-
pase substrate specificity. Methods Mol. Biol. 1133, 41–59.
Sexton, K.B., Witte, M.D., Blum, G., and Bogyo, M. (2007). Design of cell-
permeable, fluorescent activity-based probes for the lysosomal cysteine pro-
tease asparaginyl endopeptidase (AEP)/legumain. Bioorg. Med. Chem. Lett.
Powers, J.C., Asgian, J.L., Ekici, O.D., and James, K.E. (2002). Irreversible in-
1
7, 649–653.
hibitors of serine, cysteine, and threonine proteases. Chem. Rev. 102, 4639–
4
750.
Shirahama-Noda, K., Yamamoto, A., Sugihara, K., Hashimoto, N., Asano, M.,
Nishimura, M., and Hara-Nishimura, I. (2003). Biosynthetic processing of ca-
thepsins and lysosomal degradation are abolished in asparaginyl endopepti-
dase-deficient mice. J. Biol. Chem. 278, 33194–33199.
Qiu, J., Ai, L., Ramachandran, C., Yao, B., Gopalakrishnan, S., Fields, C.R.,
Delmas, A.L., Dyer, L.M., Melnick, S.J., Yachnis, A.T., et al. (2008). Invasion
suppressor cystatin E/M (CST6): high-level cell type-specific expression in
normal brain and epigenetic silencing in gliomas. Lab. Invest. 88, 910–925.
Smith, R., Johansen, H.T., Nilsen, H., Haugen, M.H., Pettersen, S.J.,
Maelandsmo, G.M., Abrahamson, M., and Solberg, R. (2012). Intra- and extra-
cellular regulation of activity and processing of legumain by cystatin E/M.
Biochimie 94, 2590–2599.
Rawlings, N.D., Barrett, A.J., and Bateman, A. (2014). Using the MEROPS
database for proteolytic enzymes and their inhibitors and substrates. Curr.
Protoc. Bioinformatics 48, 1.25.1–1.25.33.
Snipas, S.J., Stennicke, H.R., Riedl, S., Potempa, J., Travis, J., Barrett, A.J.,
and Salvesen, G.S. (2001). Inhibition of distant caspase homologues by natural
caspase inhibitors. Biochem. J. 357, 575–580.
Rozman-Pungercar, J., Kopitar-Jerala, N., Bogyo, M., Turk, D., Vasiljeva, O.,
Stefe, I., Vandenabeele, P., Bromme, D., Puizdar, V., Fonovic, M., et al.
(2003). Inhibition of papain-like cysteine proteases and legumain by cas-
pase-specific inhibitors: when reaction mechanism is more important than
Stennicke, H.R., and Salvesen, G.S. (1997). Biochemical characteristics of
specificity. Cell Death Differ. 10, 881–888.
caspases-3, -6, -7, and -8. J. Biol. Chem. 272, 25719–25723.
Sadaghiani, A.M., Verhelst, S.H., and Bogyo, M. (2007). Tagging and detection
Stennicke, H.R., and Salvesen, G.S. (1999). Caspases: preparation and char-
strategies for activity-based proteomics. Curr. Opin. Chem. Biol. 11, 20–28.
acterization. Methods 17, 313–319.
Salvesen, G.S., and Nagase, H. (2001). Inhibition of proteolytic enzymes. In
Proteolytic Enzymes: A Practical Approach, Second Edition, J.S. Bonf and
R.J. Beynon, eds. (Oxford: Oxford University Press), pp. 105–130.
Wang, L., Chen, S., Zhang, M., Li, N., Chen, Y., Su, W., Liu, Y., Lu, D., Li, S.,
Yang, Y., et al. (2012). Legumain: a biomarker for diagnosis and prognosis of
human ovarian cancer. J. Cell Biochem. 113, 2679–2686.
Cell Chemical Biology 23, 1–13, August 18, 2016 13