Tuning activity-based probe selectivity for serine proteases by on-resin 'click' construction of peptide diphenyl phosphonates

Article abstract of DOI:10.1039/c3ob40907d

Activity-based probes (ABPs) are powerful tools for functional proteomics studies. Their selectivity can be influenced by modification of a recognition element that interacts with pockets near the active site. For serine proteases there are a limited numb

Full text of DOI:10.1039/c3ob40907d

Organic &
Biomolecular Chemistry
View Article Online
View Journal | View Issue
PAPER
Tuning activity-based probe selectivity for serine
proteases by on-resin ‘click’ construction of peptide
diphenyl phosphonates†
Cite this: Org. Biomol. Chem., 2013, 11,
5714
Sevnur Serim, Susanne V. Mayer and Steven H. L. Verhelst*
Activity-based probes (ABPs) are powerful tools for functional proteomics studies. Their selectivity can be
influenced by modification of a recognition element that interacts with pockets near the active site. For
serine proteases there are a limited number of simple and efficient synthetic procedures for the develop-
ment of selective probes. Here we describe a new synthetic route combining solid and solution phase
chemistries to generate a small library of diphenyl phosphonate probes. Building blocks carrying a P1
recognition element and an electrophilic phosphonate warhead were prepared in solution and ‘clicked’
on-resin onto a tripeptide. We show the ability to modulate the activity and selectivity of diphenyl phos-
phonate ABPs and demonstrate activity-dependent labeling of endogenous proteases within a tissue
proteome. The herein described synthetic approach therefore serves as a valuable method for rapid diver-
sification of serine protease ABPs.
Received 1st May 2013,
Accepted 9th July 2013
DOI: 10.1039/c3ob40907d
www.rsc.org/obc
proteases, general5 and selective ABPs6,7 have been designed
based on the epoxysuccinate and vinyl sulfone electrophile.
Introduction
Activity-based protein profiling (ABPP) has played an increas- Both general8 and selective9 acyloxymethyl ketone-based
ing role in the functional analysis of enzymes within complex probes have been reported for caspases. The aforementioned
biological systems.1,2 With the ever-growing applicability of probes can all be constructed on solid support, which accele-
this technique there is constant interest in designing activity- rates the synthesis and optimization of the selectivity. This
based probes (ABPs), which are the workhorses of ABPP. The convenient strategy has also been employed for enzymes other
ability to modulate probe selectivity is crucial for their applica- than cysteine proteases.10
bility: while general probes are sufficient for some analysis
For serine proteases, general fluorophosphonate11,12 and
methods, other techniques require probes with high selecti- sulfonyl fluoride probes13 and more selective isocoumar-
vity.3 Most ABPs derive their selectivity from two probe fea- ins14,15 and peptidyl diphenyl phosphonates16–18 (DPPs) are
tures. (1) An electrophilic warhead, which covalently binds to a available (Fig. 1). However, these have been made using solu-
nucleophile in the active site. Depending on the nature of the tion phase chemistries.
electrophile, this warhead can react with a variety of nucleo-
We implemented a solid phase strategy for the synthesis
philic amino acid residues or can be specific for a certain type and selectivity modulation of serine protease ABPs. We chose
of side chain, such as the catalytic serine or cysteine in serine diphenyl esters of α-aminophosphonates because they are low
and cysteine proteases, respectively. (2) A recognition element molecular weight, irreversible inhibitors exhibiting explicit
that modulates the affinity of the probe and directs the selectivity19 for serine proteases and therefore represent useful
warhead to a subset of enzymes within the targeted class.
warheads for ABPs. DPPs have been validated as covalent irre-
ABPP has been especially important for the study of pro- versible inhibitors of serine proteases and they can be fine-
teases4 since their activities are tightly regulated by several tuned both via modification of the group in the P1 position,
post-translational mechanisms. For cysteine cathepsin which interacts with the enzyme S1 pocket and via introduc-
tion of additional recognition sites that interact with more
distal non-primed site pockets. The herein reported approach
Lehrstuhl für Chemie der Biopolymere, Technische Universität München,
allows rapid diversification of the recognition moiety and con-
Weihenstephaner Berg 3, 85354 Freising, Germany. E-mail: verhelst@wzw.tum.de
venient synthesis of DPP ABPs. In addition, their applicability
†Electronic supplementary information (ESI) available: Graphical representation
for selective labeling of endogenous proteases within complex
proteomes is demonstrated.
of probe 10 docked into bovine beta-trypsin, and copies of spectra and chroma-
tograms. See DOI: 10.1039/c3ob40907d
5714 | Org. Biomol. Chem., 2013, 11, 5714–5721
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
Scheme 1 (A) Synthesis of building blocks with hydrophobic side chains. (a)
CbzNH2, P(OPh)3, AcOH, 80 °C. (b) 33% HBr–AcOH. (c) Propiolic acid, DIC (pre-
activation in THF, 0 °C), TEA, DMF. (B) Synthesis of building blocks with basic side
chains. (d) Fe powder, AcOH, 70 °C. (e) 1,3-Di-Boc-2-methylisothiourea, HgCl2,
TEA, DCM.
Fig. 1 Influencing the selectivity of phosphonate ABPs. The warhead (triangle)
and spacer or recognition element (black line) determine the selectivity of the
probe. Fluorophosphonates (A) are general probes that label most serine hydro-
lases. Peptide diphenyl phosphonates (B and C) label serine proteases. The
selectivity is steered by the side chain in the P1 position and optional additional
elements at distal sites.
to 86%. The introduction of an alkyne handle to form the
building blocks 3a–e was achieved by coupling of propiolic
acid after Cbz deprotection of the phosphonates (Scheme 1A).
As a basic P1 residue we chose a diphenyl phosphonate with a
p-guanidinophenyl side chain, which was made by using two
additional reaction steps.28 The p-nitrophenyl moiety of 2e was
reduced to an aniline (2f) using iron powder in acetic acid in
quantitative yield. The aniline was then transformed into a
Boc-protected guanidine in a HgCl2-catalyzed reaction with
1,3-di-Boc-2-methylisothiourea giving compound 2g in 57%
yield (Scheme 1B). The introduction of the propiolic acid to
both 2f and 2g was then performed as for the other building
blocks.
The extended recognition elements of the probes were gene-
rated by solid phase peptide synthesis on a Rink amide resin
(Scheme 2). All probes contain an L-propargylglycine to enable
visualization of target proteases after labeling using click
chemistry. This gives the flexibility to choose between different
reporter tags such as biotin or fluorophores. Further elonga-
tion took place by coupling of two amino acids which form the
P4 and P3 positions. We chose either two alanine residues or
two residues according to the substrate specificities of
different serine proteases.29,30 The N-terminal amine was then
transformed into an azide (4) via on resin diazo transfer. Sub-
sequently, the alkyne DPP building blocks were reacted with
the azide via on resin Cu(I)-catalyzed 1,3-dipolar cycloaddition.
In principle, cross-coupling of an azide on one peptide chain
and an alkyne on a neighboring peptide chain is possible.
However, after cleavage from the resin, we did not observe any
intramolecular cyclo-addition products, which may be due to
the excess of the alkyne-containing phosphonate building
Results and discussion
Synthetic approach
Diphenyl esters of phosphonates are labile under basic con-
ditions, which limits the reaction conditions that can be used
during the synthesis. Our synthetic strategy towards phospho-
nate ABPs with extended peptide recognition elements there-
fore comprises the connection of a DPP warhead to a solid
support during the last step using mild and chemoselective
click chemistry. In this route, the peptidic portions of the
probes are generated by solid phase peptide synthesis and the
amino group of the last amino acid is converted to an azide by
on-resin diazo transfer.20 The azide allows for the introduction
of the DPP building blocks via on-resin click reaction21,22
forming a 1,4-substituted 1,2,3-triazole. Triazoles have been
used before in selective probe design for cathepsin S.23 More-
over, they are proteolytically and metabolically stable and
provide good analogues of peptide bonds due to their strong
dipole and H-bond accepting and donating properties.24,25 We
anticipated that the interactions at the P3 and P4 positions
would provide the desired modulation of selectivity.
We synthesized alkynylated versions of diphenyl α-amino-
alkylphosphonates in three easy steps. Cbz-protected deriva-
tives with hydrophobic side chains or a hydrogen in the P1
position, such as glycine (2a), valine (2b), leucine (2c), phenyl-
alanine (2d) and p-nitrophenylglycine (2e), were prepared
from commercially available aldehydes (1) as described by
Oleksyszyn et al.26,27 The condensation products were obtained
as pure solids after crystallization with yields ranging from 26
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem., 2013, 11, 5714–5721 | 5715

View Article Online
Paper
Organic & Biomolecular Chemistry
Scheme 2 Solid phase peptide synthesis. (a) Piperidine–DMF (1/4, v/v); then, Fmoc-aa-OH, HOBt, DIC, DMF. (b) Piperidine–DMF (1/4, v/v); then, TfN3, CuSO4,
DCM–MeOH (9/1, v/v). (c) Building blocks 3, CuSO4, sodium ascorbate, TBTA, DMF–H2O (10/1, v/v). (d) 95% TFA, 2.5% TIS, 2.5% H2O. The abbreviations indicated
behind each compound number correspond to the residues in the P4, P3 and P1 positions.
block in solution or to unfavorable distances between the between the probes and proteases occur in an activity-depen-
resin-bound reaction partners. Final cleavage from the resin dent manner, since pretreatment with an active site-directed
resulted in eight different DPP ABPs (5–11) with yields varying inhibitor competes away the labeling. The influence of the
from 3 to 40% after purification (Scheme 2).
extended peptide recognition element on the probes’ affinity
towards their target proteases can be clearly observed from the
overall increase of the labeling intensities in comparison to
the P1 probes.
Labeling experiments
With both the P1 probes and the extended probes in hand we
set out to test their labeling capability using commercially
available, purified serine proteases of different specificities:
bovine chymotrypsin, known for its P1 specificity for large
hydrophobic residues; human cathepsin G with tryptic–chymo-
tryptic “dual specificity”;31 human neutrophil elastase, known
for its P1 specificity for small hydrophobic residues; bovine
trypsin and human urokinase-type plasminogen activator
(uPA) known for their basic P1 specificity. For comparison we
included a nonselective fluorophosphonate ABP, FP–rhoda-
mine32 (FP–R). Proteases tagged by DPP ABPs 3a–g and 5–12
were visualized by clicking a tetramethylrhodamine (TAMRA)
derivative carrying an azide function, using conditions similar
to those that have been optimized by Speers and Cravatt for
efficient labeling in tandem ABPP.33
The general serine hydrolase ABP FP–R labeled all proteases
regardless of their P1 preferences while a change in the
extended recognition elements resulted in different selecti-
vities of the different probes. Probe 7, with a Phe in the P1
position, mainly labels chymotrypsin. Unexpectedly, a change
in the P3 and P4 position in probe 8, led to labeling of trypsin
and uPA as well. Another example of the modulation of the
selectivity was observed for probes with a basic P1 side chain.
Trypsin, uPA and human cathepsin G were all equally labeled
with the Gua P1 probe (3g) and the extended probe nT-Gua
(11). This result is in line with previously reported phospho-
nate inhibitors of these enzymes.34,35 In contrast, AS-Gua (10)
displayed a high intensity band for trypsin and only a weaker
band for uPA.
To gain insight into the binding mode of the extended
probes that carry a triazole in their backbone, we performed
Both the P1 probes 3a–g and the extended probes 5–12
resulted in labeling of proteases (Fig. 2). The reactions
Fig. 2 Labeling of purified proteases with FP–rhodamine, P1 and extended DPP probes. Probe concentrations: 5 µM for DPPs, 1 µM for FP–R.
5716 | Org. Biomol. Chem., 2013, 11, 5714–5721
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
Fig. 4 In-gel fluorescence (left) of proteases labeled in the context of a pro-
teome (a cell lysate of the human colon adenocarcinoma cell line HT-29). The
Coomassie stain (right) shows protein loading. The apparent lower amount of
protein in the uninhibited lanes for chymotrypsin and trypsin are due to the
high digestive properties of these proteases. Note that for trypsin, there is still
some digestion observed in the inhibited sample compared with the no trypsin
control probably due to incomplete inhibition of the trypsin. Chy.: chymotrypsin;
Try.: trypsin; Cat. G: cathepsin G; u: urokinase-type plasminogen activator.
Fig. 3 Docking of probe 10 bound to Ser195 in bovine beta-trypsin. The
enzyme surface is represented in white, except for the S195 residue, which is
shown as a stick model as are the small molecules. Left: probe 10 (colored by
element: red = oxygen, blue = nitrogen, cyan = carbon, brown = phosphorus)
docks with its para-guanidino-phenyl group into the S1 pocket. Right: probe 10
is colored in magenta and overlaid with a non-covalent inhibitor.34 The triazole
of probe 10 takes a similar position as the proline in the inhibitor, while the
other residues interact with various sites on the non-primed site.
confirmed by pretreatment with chymotryptic inhibitor
DAP22c and tryptic inhibitor TLCK) with probe concentrations
down to 0.1 and 1 µM, respectively (Fig. 5). No labeling was
observed in the case of unactivated or preinhibited lysates, as
well as buffer containing only enterokinase.
molecular docking of a covalently bound probe 10 inside a
bovine trypsin crystal structure. As expected, the p-guanidino-
phenyl ring in the P1 position of the probe occupies the S1
pocket (Fig. 3), where it forms a salt bridge with Asp189
(Fig. S1†). The triazole occupies the S2 site and has a similar
position as a proline in the P2 position of a non-covalent
trypsin inhibitor36 (Fig. 3). The alanine and propargylglycine
interact with the more distal S3/S4 sites. Overall, this structure
suggests that the probes – despite the non-natural triazole and
the reversed polarity of the P3–P5 backbone – interact with the
non-primed site to give increased potency compared to the P1
only probes.
We next investigated whether the extended probes can be
applied within the context of a complex proteome. To demon-
strate that the presence of other proteins does not interfere
with the labeling of the target proteases, we spiked selected
proteases into a cell lysate and incubated with probes that dis-
played the most selective activity-based labeling of the purified
proteases. In-gel fluorescence showed intense activity-depen-
dent bands for each enzyme without much cross-reactivity
with other proteins (Fig. 4). Only nT-Gua (11) when reacted
with cathepsin G displayed some background bands which
were not competed away by pretreatment with PMSF.
Finally, to illustrate the applicability of DPP ABPs in the
labeling of endogenous serine proteases within a complex pro-
teome, we used a rat pancreas lysate that was activated by treat-
ment with enterokinase. Enterokinase activates trypsinogen
into trypsin, the common activator of all pancreatic zymo-
gens.37 Both activated and unactivated proteomes were used in
labeling experiments with decreasing concentrations of probes
that displayed strong activity towards chymotrypsin and
trypsin in the previous experiments. A selective labeling of
chymotryptic and tryptic enzymes was detectable (as
Conclusion
Diphenyl esters of α-aminophosphonates are phosphonic ana-
logues of naturally occurring amino acids. They selectively, co-
valently and irreversibly bind to the active site serine residue
of serine proteases. These features qualify them as attractive
warheads for ABPs. The selectivity of DPPs can be modulated
via the adjustment of the group at the P1 position and other
non-primed site residues. While synthetic strategies to gene-
rate DPPs in solution have been reported, an easy and rapid
way to synthesize DPP ABPs remains a challenge. Here, we pre-
sented a new synthetic route that enables simple modification
of the extended peptide sequence and an overall convenient
construction of DPP ABPs. Key features of the synthesis
include solid phase generation of extended recognition
elements and on-resin click chemistry mediated introduction
of the P1 phosphonate building blocks to form the final
probes. Using this approach we synthesized eight different
DPP ABPs, and labeled serine proteases of different specifici-
ties. We demonstrated that it is possible to tune the activity
and selectivity of the DPP ABPs by varying the extended recog-
nition elements. Molecular modeling suggests that the probes
interact with the non-primed sites, which may explain their
higher potency and capacity to influence the selectivity
towards the protease targets. Furthermore, the probes display
activity-dependent labeling of proteases within complex
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem., 2013, 11, 5714–5721 | 5717

View Article Online
Paper
Organic & Biomolecular Chemistry
Fig. 5 Fluorescent labeling of endogenous proteases in enterokinase-activated rat pancreas lysate by DPP ABPs with extended recognition elements and FP–R. The
relatively unstable inhibitor DFP did not completely inhibit all labeling by FP–R. The higher intensity of the two bands at approximately 42 and 50 kDa in the inhibi-
ted sample may be due to the higher availability of FR–R compared to the enterokinase treated sample, in which FP–R reacts with the activated proteases. Entero-
kinase is not labeled (rightmost lanes for probes 7 and 10, respectively), probably due to the hydrophobic P1 of probe 7 and the preference of enterokinase for
negatively charged residues in the P2 and P3 position, which are absent from the probes used.
proteomes. We envision that this strategy will allow for future General methods for the synthesis of propynoylated diphenyl
probe optimization when selective serine protease ABPs are α-aminoalkylphosphonate building blocks
required.
Birum–Oleksyszyn reaction. Benzyl carbamate (1 eq.) was
combined with triphenylphosphite (1 eq.), the corresponding
aldehyde (1.5 eq.) and acetic acid (0.1–0.2 mL per mmol alde-
hyde) and heated to 82 °C for ca. 2 h. The reaction was moni-
tored by TLC until benzyl carbamate was consumed. Volatile
components were removed and methanol was added to allow
Experimental
General
crystallization at −20 °C overnight. Crystals were collected by
filtration, washed with cold methanol and dried. The resulting
solids were sufficiently pure for the following reactions.
Cbz deprotection. Cbz-protected DPPs were treated with
33% HBr–AcOH solution for 1–2 h at room temperature. The
solvent was removed and the oily residue was dissolved in a
minimal amount of methanol; excess diethylether was added
and overnight storage at −20 °C led to crystallization. The crys-
tals were filtered, washed with cold diethylether and dried.
The purity of the compounds was sufficient for further
reactions.
Coupling of propiolic acid. Propiolic acid (2.6 eq.) was pre-
activated with DIC (1.3 eq.) in THF (0.3 mL per mmol propiolic
acid) for 1 h at 0 °C. The DPP hydrobromide (1 eq.) was dis-
solved in DMF (0.75 mL mmol−1), treated with DIEA (2 eq.)
and then added to the preactivated propiolic acid solution and
stirred overnight at room temperature. Base was neutralized
with AcOH and the volatile components were removed. The
residue was dissolved in ethylacetate and washed 2× with 1 M
HCl, 2× with H2O, 2× with a saturated NaHCO3 solution, and
brine. The organic phase was dried over MgSO4 and concen-
trated at reduced pressure. If necessary, compounds were
further purified via silica column chromatography.
All starting materials, enzymes and inhibitors were obtained
from Sigma-Aldrich with the exception of uPA (VWR) and FP–R
(Thermo Scientific) and used without further purification.
Thin layer chromatography (TLC) was performed on
ALUGRAM sil G/UV254 TLC plates (Carl Roth). Solvents used
for SPPS and purification via column chromatography were
purchased from Applichem. Compounds were separated over
silica gel with a grain distribution of 0.04–0.063 mm and a
pore size of 60 Å (Carl Roth). Creosalus was the supplier for
Fmoc-protected L-amino acids, Rink resin and coupling
reagents for SPPS. LC-MS spectra were recorded by an Agilent
1100 Series LC system coupled to an Agilent 6210 ESI TOF
mass spectrometer, with elution by solvents A (5% ACN–H2O +
0.1% formic acid) and B (95% ACN–H2O + 0.1% formic acid).
A Zorbax SB C18 5 µm (0.5 × 150 mm) column was used to
separate samples at r.t. with a flow rate of 20 µl min−1 and
with a gradient of 2.57% ACN per min starting from A for
35 min. HPLC purifications were carried out on Waters
Xbridge C18 5 µm (4.6 × 150 mm) and a Waters Xbridge
BEH130 Prep C18 5 µm (19 × 150 mm) columns. 1H and
13C NMR spectra were measured on a Bruker 400 MHz DRX
(400; 100 MHz). Chemical shifts (δ) are given in parts per million
(ppm) relative to tetramethylsilane as an internal standard.
5718 | Org. Biomol. Chem., 2013, 11, 5714–5721
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
Transformation of the NO2 group to a Boc-protected guani- (m, 6 H), 6.66 (s, 1 H), 4.11 (dd, J = 12.0 Hz, J = 6.0 Hz, 2 H),
dine. Acetic acid (4 mL mmol−1) was added to a mixture of the 2.88 (s, 1 H).
nitro compound (1 eq.) and Fe powder (9 eq.). The reaction
Diphenyl α-N-(propiolamido)-2-methylpropylphosphonate
mixture was heated to 70 °C and stirred under a nitrogen (3b). The title compound was isolated as yellowish oil in 45%
atmosphere for 2 h. Afterwards, the acid was removed and the yield. ESI-MS: [M + H]+ m/z 358.1237 (found), 358.1202 (calcu-
crude residue was dissolved in EtOAc. Fe2O3 was centrifuged lated). 1H-NMR (400 MHz, CDCl3): δ = 7.40–7.28 (m, 4 H),
down and the supernatant was concentrated to form the crude 7.26–7.09 (m, 6 H), 6.46 (d, J = 10.4 Hz, 1 H), 4.83 (ddd, J =
product. The aniline (1 eq.) was mixed with 1,3-bis-(tert-butoxy- 19.2 Hz, J = 10.4 Hz, J = 4.3 Hz, 1 H), 2.90 (s, 1 H), 2.57–2.37
carbonyl)-2-methyl-2-thiopseudourea (1.1 eq.) and mercury(II) (m, 2 H), 1.19–1.08 (m, 6 H).
chloride (1.2 eq.). Subsequently, DCM and Et3N (3 eq.) were
Diphenyl
α-N-(propiolamido)-3-methylbutylphosphonate
added resulting in a yellowish suspension, which was stirred (3c). The title compound was isolated as yellowish oil in 29%
overnight. DCM was evaporated and the residue was redis- yield. ESI-MS: [M + H]+ m/z 372.1350 (found), 372.1358 (calcu-
solved in EtOAc. The remaining solid was centrifuged down; lated). 1H-NMR (400 MHz, CDCl3): δ = 7.39–7.29 (m, 4 H),
the supernatant was decanted and washed with 1 M KHSO4, 7.26–7.08 (m, 6 H), 6.64 (d, J = 10.2 Hz, 1 H), 5.06–4.87 (m,
saturated NaHCO3 solution and brine. The organic phase was 1 H), 2.84 (s, 1 H), 1.91–1.71 (m, 3 H), 1.06–0.90 (m, 6 H).
dried over MgSO4 and the solvent was removed. Further purifi-
Diphenyl
α-N-(propiolamido)-2-phenylethylphosphonate
cation was carried out via silica column chromatography.
Diphenyl
(3d). The title compound was isolated as a white solid in 52%
α-N-(benzyloxycarbonyl)amino-methylphospho- yield. ESI-MS: [M + H]+ m/z 406.1255 (found), 406.1202 (calcu-
nate (2a). A mixture of benzyl carbamate (1 eq.), acetic anhy- lated). 1H-NMR (400 MHz, CDCl3): δ = 7.49–6.94 (m, 15 H),
dride (1.25 eq.), and paraformaldehyde (1 eq.) in acetic acid 5.38–5.01 (m, 1 H), 3.59–3.27 (m, 1 H), 3.15–2.97 (m, 1 H), 2.74
(100 µL mmol−1) was heated at 65 °C for 3 h. The resulting (s, 1 H).
solution was treated with triphenyl phosphite (1 eq.) and
Diphenyl α-N-(propiolamido)-(4-nitro-phenyl)methanephos-
heated at 115 °C for 2 h. The mixture was concentrated under phonate (3e). The title compound was isolated as an orange
high vacuum, and a small volume of diethyl ether was added solid in 48% yield. ESI-MS: [M + H]+ m/z 437.10 (found),
1
to allow crystallization at −20 °C overnight. The precipitate was 437.08 (calculated). H-NMR (400 MHz, CDCl3): δ = 8.19 (d, J =
then collected by filtration. The resulting solid was sufficiently 8.0 Hz, 2 H), 7.71 (dd, J = 8.0 Hz, J = 2.0 Hz, 2 H), 7.41–7.33 (m,
pure for the following reactions. The title compound was iso- 2 H), 7.28–7.21 (m, 3 H), 7.20–7.12 (m, 3 H), 6.94–6.89 (m,
lated as a white solid in 26% yield. ESI-MS: [M + H]+ m/z 2 H), 6.03 (dd, J = 24.0 Hz, J = 8.0 Hz, 1 H), 2.83 (s, 1 H).
398.1194 (found), 398.1151 (calculated).
Diphenyl α-N-(propiolamido)-(4-amino-phenyl)methanephos-
Diphenyl α-N-(benzyloxycarbonyl)amino-2-methylpropylphos- phonate (3f). The title compound was isolated in 75% yield.
phonate (2b).26 The title compound was isolated as white crys- ESI-MS: [M + H]+ m/z 407.13 (found), 407.11 (calculated).
tals in 52% yield. ESI-MS: [M + H]+ m/z 440.1577 (found), 1H-NMR (400 MHz, DMSO-d6): δ = 10.14 (d, J = 9.7 Hz, 1 H),
440.1621 (calculated).
7.41–7.30 (m, 6 H), 7.24–7.15 (m, 2 H), 7.07 (d, J = 7.5 Hz, 2 H),
Diphenyl α-N-(benzyloxycarbonyl)amino-3-methylbutylphos- 6.94 (d, J = 7.5 Hz, 2 H), 6.74 (d, J = 7.5 Hz, 2 H), 5.77 (dd, J =
phonate (2c).26 The title compound was isolated as white crys- 21.3 Hz, J = 9.7 Hz, 1 H), 4.34 (s, 1 H).
tals in 69% yield. ESI-MS: [M + H]+ m/z 454.1737 (found),
454.1777 (calculated).
Diphenyl α-N-(benzyloxycarbonyl)amino-2-phenylethylphos- (found), 449.1372 (calculated). H-NMR (400 MHz, DMSO-d6):
phonate (2d).26 The title compound was isolated as a white δ = 10.35 (d, J = 10.0 Hz, 1 H), 9.71 (s, 1 H), 7.71 (dd, J = 8.4 Hz,
solid in 40% yield. ESI-MS: [M + H]+ m/z 488.1566 (found), 2.0 Hz, 2 H), 7.46 (s, 3 H), 7.42–7.32 (m, 4 H), 7.29 (d, J =
Diphenyl
α-N-(propiolamido)-(4-guanidinium-phenyl)-
methanephosphonate (3g). ESI-MS: [M + H]+ m/z 449.1413
1
488.1621 (calculated).
Diphenyl
8.4 Hz, 2 H), 7.26–7.19 (m, 2 H), 7.09 (d, J = 8.4 Hz, 2 H), 7.02
α-N-(benzyloxycarbonyl)amino-(4-nitro-phenyl)- (d, J = 8.4 Hz, 2 H), 6.00 (dd, J = 22.6 Hz, J = 10.0 Hz, 1 H), 4.41
methanephosphonate (2e).27 The title compound was isolated (s, 1 H).
as white crystals in 86% yield. ESI-MS: [M + H]+ m/z 519.1321
General methods for solid phase peptide synthesis of peptide
diphenyl α-aminoalkylphosphonates
(found), 519.1315 (calculated).
Diphenyl α-N-(benzyloxycarbonyl)amino-(4-amino-phenyl)-
methanephosphonate (2f).27 The title compound was isolated
Amino acid coupling and Fmoc deprotection. Fmoc groups
as an orange solid in 90% yield. ESI-MS: [M + H]+ m/z 489.1622 were deprotected by 20% piperidine–DMF (v/v) for 20 min. For
(found), 489.1573 (calculated).
Diphenyl
amino acid couplings, Fmoc–aa–OH (3 eq., 0.25 M in DMF),
α-N-(benzyloxycarbonyl)amino-(4-diBocguanidi- DIC (3 eq.) and HOBt (3 eq.) were added to the resin and
nium-phenyl)methanephosphonate (2g).27 The title com- shaken at room temperature for 1–3 h. Completion of the reac-
pound was isolated as a white solid in 57% yield. ESI-MS: tions was checked by the Kaiser test.
[M + H]+ m/z 731.2890 (found), 731.2840 (calculated).
Diazotransfer. Triflyl azide was freshly prepared from
Diphenyl
α-N-(propiolamido)-methylphosphonate
(3a). sodium azide (5 eq.) and triflic anhydride (1 eq.). NaN3 was
ESI-MS: [M + H]+ m/z 316.0735 (found), 316.0732 (calculated). dissolved in H2O (0.16 mL mmol−1). At 0 °C, DCM (0.27 mL
1H-NMR (400 MHz, CDCl3): δ = 7.40–7.32 (m, 4 H), 7.26–7.16 mmol−1) was added to the clear solution under heavy stirring.
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem., 2013, 11, 5714–5721 | 5719

View Article Online
Paper
Organic & Biomolecular Chemistry
Tf2O was added dropwise and the reaction mixture was stirred chain of the bovine trypsin S195 (PDB coordinates: 1MAX,
at 0 °C for 2 h. The water phase was extracted twice with DCM. from which the original covalent phosphonate inhibitor struc-
The combined organic phases were washed once with a satu- ture was deleted). Docking of the inhibitor as a flexible side
rated NaHCO3 solution. The maximal volume of the organic chain of S195 was performed with AutoDock Vina.38 A non-
phase was 2.70 mL mmol−1 TfN3. For the diazotransfer, the covalent inhibitor bound to bovine trypsin was taken as a com-
Fmoc deprotected tripeptide on the resin (1 eq.) was treated parison (PDB coordinates: 2ZFT). Pictures were generated
with TfN3 (0.37 M, 16.7 eq.) in DCM and CuSO4 (12.6 mM, 0.1 using VMD 1.9.39
eq.) in MeOH for 24 h while shaking. Before a Kaiser test was
performed, the resin was washed with NMP (3 × 2 min), 0.5%
Labeling experiments
DIEA–NMP (3 × 2 min), 0.05 M sodium diethyldithiocarbamate
Labeling of purified enzymes. Each enzyme in PBS was
in NMP (3 × 10 min), NMP (5 × 5 min) and DCM (3 × 3 min)
treated for 30 min at room temperature with either a suitable
and Et2O.
inhibitor (0.1 mM DAP22c for chymotrypsin, 1 mM PMSF for
On-resin click reaction. The resin-tripeptide azide conjugate
cathepsin G, 1 mM DFP for uPA and neutrophil elastase,
(1 eq.) was resuspended in DMF and the building block (3 eq.)
1 mM TLCK for trypsin) or DMSO in a final volume of 50 µL.
was added as a solution in DMF (0.45 M), followed by the click
Pretreatment was followed by labeling with ABPs for 30 min at
reagents TBTA (0.2 eq.), CuSO4 (0.1 eq.) and sodium ascorbate
r.t. In the case of DPPs, TAMRA azide (25 µM) was sub-
(3 eq.) in DMF–H2O (10/1, v/v). The reaction was shaken at
sequently clicked onto the alkynylated ABPs in the presence of
room temperature for about 24 h, followed by three washing
CuSO4 (1 mM), TBTA (50 µM) and sodium ascorbate (0.5 mM)
steps each with DCM, DMF, 0.02 M sodium diethyldithiocarba-
for 30 min at room temperature in the dark. The reactions
mate in DMF, DMF, MeOH, DMF, DCM and Et2O.
were stopped with 4× sample buffer. The samples were boiled
Resin cleavage. The resin was treated with 95% TFA, 2.5%
at 95 °C for 3 min. Protein samples were resolved on a 12% AA
H2O and 2.5% TIS for 1 h. Supernatant was collected and com-
SDS gel and gels were visualized by a Typhoon Trio+ scanner
pounds were precipitated by the addition of excess Et2O. The
precipitates were dried under a nitrogen stream and purified
(GE Healthcare). Protein/lane: 100 ng.
Labeling of purified proteases in a proteome background.
by HPLC.
Each protease was diluted in a lysate of an HT-29 cell line
PraAlaGlu(triazole)ValP(OPh)2 (5). The title compound was
(1 mg mL−1) as 1% of total protein. Afterwards the protocol for
isolated as a white solid in 3% yield after HPLC purification.
the labeling of purified enzymes was followed. Protein loading:
ESI-MS: [M + H]+ m/z 696.2496 (found), 696.2541 (calculated).
10 µg per lane.
PraAlaAla(triazole)LeuP(OPh)2 (6). The title compound was
isolated as a white solid in 40% yield after HPLC purification.
Labeling in enterokinase-activated rat pancreas lysate. Rat
pancreas lysate in PBS (concentration of total protein: 3.5 mg
ESI-MS: [M + H]+ m/z 652.2603 (found), 652.2642 (calculated).
mL−1) was activated with enterokinase (1 U mg−1 protein) for
PraAlaAla(triazole)PheP(OPh)2 (7). The title compound was
2 h on ice. Labeling with ABPs was carried out at a protein con-
isolated as a white solid in 7% yield after HPLC purification.
centration of 3 mg ml−1 following the protocol for labeling of
ESI-MS: [M + H]+ m/z 686.2456 (found), 686.2486 (calculated).
purified enzymes. Protein samples were resolved on a 15% AA
PraMetPhe(triazole)PheP(OPh)2 (8). The title compound was
SDS gel. Protein loading: 15 µg per lane.
isolated as a white solid in 4% yield after HPLC purification.
ESI-MS: [M + H]+ m/z 822.2804 (found), 822.2833 (calculated).
PraAlaAla(triazole)GuaP(OPh)2 (9). The title compound was
isolated as a white solid in 8% yield after HPLC purification.
Acknowledgements
ESI-MS: [M + H]+ m/z 729.2622 (found), 729.2656 (calculated).
PraAlaSer(triazole)GuaP(OPh)2 (10). The title compound was
We acknowledge financial support from the DFG (Emmy
Noether program), the Fonds der Chemischen Industrie and
the Center for Integrated Protein Science Munich. We thank
Dr O. Frank for measuring NMR, Eliane Wolf for a critical
reading of the manuscript and Prof. Dr D. Langosch for
general support.
isolated as a white solid in 9% yield after HPLC purification.
ESI-MS: [M + H]+ m/z 745.2548 (found), 745.2605 (calculated).
PraNleThr(triazole)GuaP(OPh)2 (11). The title compound
was isolated as a white solid in 8% yield after HPLC purifi-
cation. ESI-MS: [M + H]+ m/z 801.3216 (found), 801.3231
(calculated).
PraLeuPhe(triazole)GuaP(OPh)2 (12). The title compound
was isolated as a white solid in 7% yield after HPLC purifi-
Notes and references
cation. ESI-MS: [M + H]+ m/z 847.3404 (found), 847.3439
(calculated).
1 M. J. Evans and B. F. Cravatt, Chem. Rev., 2006, 106, 3279–
3301.
2 A. M. Sadaghiani, S. H. L. Verhelst and M. Bogyo, Curr.
Opin. Chem. Biol., 2007, 11, 20–28.
3 U. Haedke, E. V. Kuttler, O. Vosyka, Y. Yang and
S. H. L. Verhelst, Curr. Opin. Chem. Biol., 2013, 17, 102–109.
Docking experiments
The phosphonate 10 (with the same chirality at the P1 position
as a natural substrate) was geometry optimized with an
MMFF94 force field and defined as an extension of the side
5720 | Org. Biomol. Chem., 2013, 11, 5714–5721
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
4 S. Serim, U. Haedke and S. H. L. Verhelst, ChemMedChem, 22 V. Aucagne and D. A. Leigh, Org. Lett., 2006, 8, 4505–
2012, 7, 1146–1159. 4507.
5 D. Greenbaum, K. F. Medzihradszky, A. Burlingame and 23 M. Verdoes, L. E. Edgington, F. A. Scheeren, M. Leyva,
M. Bogyo, Chem. Biol., 2000, 7, 569–581.
G. Blum, K. Weiskopf, M. H. Bachmann, J. A. Ellman and
M. Bogyo, Chem. Biol., 2012, 19, 619–628.
24 Y. L. Angell and K. Burgess, Chem. Soc. Rev., 2007, 36,
1674–1689.
25 D. S. Pedersen and A. Abell, Eur. J. Org. Chem., 2011, 2399–
2411.
6 F. Yuan, S. H. L. Verhelst, G. Blum, L. M. Coussens and
M. Bogyo, J. Am. Chem. Soc., 2006, 128, 5616–5617.
7 S. H. L. Verhelst and M. Bogyo, ChemBioChem, 2005, 6,
824–827.
8 D. Kato, K. M. Boatright, A. B. Berger, T. Nazif, G. Blum,
C. Ryan, K. A. H. Chehade, G. S. Salvesen and M. Bogyo, 26 J. Oleksyszyn, L. Subotkowska and P. Mastalerz, Synthesis,
Nat. Chem. Biol., 2005, 1, 33–38. 1979, 985–986.
9 A. B. Berger, M. D. Witte, J. B. Denault, A. M. Sadaghiani, 27 J. Oleksyszyn and J. C. Powers, Methods Enzymol., 1994,
K. M. Sexton, G. S. Salvesen and M. Bogyo, Mol. Cell, 2006,
23, 509–521.
10 K. A. Kalesh, L. P. Tan, K. Lu, L. Gao, J. Wang and
S. Q. Yao, Chem. Commun., 2010, 46, 589–591.
11 D. Kidd, Y. S. Liu and B. F. Cravatt, Biochemistry, 2001, 40,
4005–4015.
244, 423–441.
28 M. Sienczyk and J. Oleksyszyn, Tetrahedron Lett., 2004, 45,
7251–7254.
29 J. L. Harris, B. J. Backes, F. Leonetti, S. Mahrus,
J. A. Ellman and C. S. Craik, Proc. Natl. Acad. Sci. U. S. A.,
2000, 97, 7754–7759.
12 Y. S. Liu, M. P. Patricelli and B. F. Cravatt, Proc. Natl. Acad. 30 D. N. Gosalia, C. M. Salisbury, J. A. Ellman and
Sci. U. S. A., 1999, 96, 14694–14699. S. L. Diamond, Mol. Cell. Proteomics, 2005, 4, 626–636.
13 D. A. Shannon, C. Gu, C. J. McLaughlin, M. Kaiser, 31 W. W. Raymond, N. N. Trivedi, A. Makarova, M. Ray,
R. A. van der Hoorn and E. Weerapana, ChemBioChem,
2012, 13, 2327–2330.
C. S. Craik and G. H. Caughey, J. Immunol., 2010, 185,
5360–5368.
14 U. Haedke, M. Gotz, P. Baer and S. H. L. Verhelst, Bioorg. 32 M. P. Patricelli, D. K. Giang, L. M. Stamp and
Med. Chem., 2012, 20, 633–640. J. J. Burbaum, Proteomics, 2001, 1, 1067–1071.
15 J. C. Powers, C. M. Kam, L. Narasimhan, J. Oleksyszyn, 33 A. E. Speers and B. F. Cravatt, Chem. Biol., 2004, 11, 535–
M. A. Hernandez and T. Ueda, J. Cell. Biochem., 1989, 39,
33–46.
16 B. F. Gilmore, L. Carson, L. L. McShane, D. Quinn,
W. A. Coulter and B. Walker, Biochem. Biophys. Res.
Commun., 2006, 347, 373–379.
546.
34 M. Sienczyk, A. Lesner, M. Wysocka, A. Legowska,
E. Pietrusewicz, K. Rolka and J. Oleksyszyn, Bioorg. Med.
Chem., 2008, 16, 8863–8867.
35 M. Sienczyk and J. Oleksyszyn, Bioorg. Med. Chem. Lett.,
2006, 16, 2886–2890.
17 S. Mahrus and C. S. Craik, Chem. Biol., 2005, 12, 567–577.
18 Z. Pan, D. A. Jeffery, K. Chehade, J. Beltman, J. M. Clark, 36 T. Brandt, N. Holzmann, L. Muley, M. Khayat,
P. Grothaus, M. Bogyo and A. Baruch, Bioorg. Med. Chem.
Lett., 2006, 16, 2882–2885.
C. Wegscheid-Gerlach, B. Baum, A. Heine, D. Hangauer
and G. Klebe, J. Mol. Biol., 2011, 405, 1170–1187.
19 M. Sienczyk and J. Oleksyszyn, Curr. Med. Chem., 2009, 16, 37 X. L. Zheng, Y. Kitamoto and J. E. Sadler, Front. Biosci.
1673–1687. (Elite Ed), 2009, 1, 242–249.
20 D. T. S. Rijkers, H. H. R. van Vugt, H. J. F. Jacobs and 38 O. Trott and A. J. Olson, J. Comput. Chem., 2010, 31, 455–
R. M. J. Liskamp, Tetrahedron Lett., 2002, 43, 3657–3660. 461.
21 N. G. Angelo and P. S. Arora, J. Org. Chem., 2007, 72, 7963– 39 W. Humphrey, A. Dalke and K. Schulten, J. Mol. Graphics,
7967.
1996, 14, 33–38, 27–38.
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem., 2013, 11, 5714–5721 | 5721