In silico design, synthesis, and assays of specific substrates for proteinase 3: Influence of fluorogenic and charged groups

Article abstract of DOI:10.1021/jm401559r

Neutrophil serine proteases are specific regulators of the immune response, and proteinase 3 is a major target antigen in antineutrophil cytoplasmic antibody-associated vasculitis. FRET peptides containing 2-aminobenzoic acid (Abz) and N-(2,4-dinitrophenyl)ethylenediamine (EDDnp) as fluorophore and quencher groups, respectively, have been widely used to probe proteases specificity. Using in silico design followed by enzymatic assays, we show that Abz and EDDnp significantly contribute to substrate hydrolysis by PR3. We also propose a new substrate specific for PR3.

Full text of DOI:10.1021/jm401559r

Brief Article
pubs.acs.org/jmc
In Silico Design, Synthesis, and Assays of Specific Substrates for
Proteinase 3: Influence of Fluorogenic and Charged Groups
Shailesh Narawane,†,§ Adnan Budnjo,‡,§ Cedric Grauffel,†,⊥ Bengt Erik Haug,*,‡ and Nathalie Reuter*,†,⊥
́
†Department of Molecular Biology, University of Bergen, 5008, Bergen, Norway
‡Department of Chemistry and Centre for Pharmacy, University of Bergen, Alleg
́
aten 41, 5007, Bergen, Norway
⊥Computational Biology Unit, Department of Informatics, University of Bergen, 5008, Bergen, Norway
S
* Supporting Information
ABSTRACT: Neutrophil serine proteases are specific regulators of the
immune response, and proteinase 3 is a major target antigen in
antineutrophil cytoplasmic antibody-associated vasculitis. FRET peptides
containing 2-aminobenzoic acid (Abz) and N-(2,4-dinitrophenyl)ethylene-
diamine (EDDnp) as fluorophore and quencher groups, respectively, have
been widely used to probe proteases specificity. Using in silico design
followed by enzymatic assays, we show that Abz and EDDnp significantly
contribute to substrate hydrolysis by PR3. We also propose a new substrate
specific for PR3.
the proteolytic hydrolysis of substrates with sequences
INTRODUCTION
■
extending on both sides of the cleavable bond. Yet their use
introduces a limitation on the substrate concentration that can
be used for enzymatic assays to avoid inner-filter effect arising
from intermolecular quenching. Moreover, since PR3 and HNE
have extended substrate binding sites, the terminal groups
might bind to the enzyme and possibly influence enzyme
activity. We have used a combination of in silico design, peptide
synthesis, and enzymatic assays to investigate (i) the influence
of the FRET groups Abz/EDDnp on the hydrolysis of
octapeptide substrates by PR3 or HNE and (ii) the effect of
charged amino acids at P2, P2′, and P3′. The use of high
performance liquid chromatography (HPLC), unlike fluores-
cence measurements, allows us to compare the cleavage
efficiency of the enzyme for substrate sequences with and
without the FRET groups and work at higher substrate
concentration than reported in previous studies.
Neutrophil serine proteases (NSPs) proteinase 3 (PR3, EC
3.4.21.76) and human neutrophil elastase (HNE) are
therapeutic targets in a number of chronic inflammatory
diseases.1 In particular PR3 is a major target antigen in
antineutrophil cytoplasmic antibody (ANCA) associated
vasculitis, a life-threatening condition for which molecular
mechanisms remain to be uncovered. Moreover the role of each
of the three NSPs in diseases and inflammation is only
emerging and there is a need for specific substrates that can be
used for in vitro and cellular assays.2 Rational design of specific
peptides is also a natural step toward the design of druggable
low-molecular weight compounds. Achieving specific targeting
of either of these proteases is challenging, as the mature forms
of PR3 and HNE share a high sequence identity (56%) and
structural similarity.3 We and others have shown that
differences between PR3 and HNE in the nature of the S2,
S1′, S2′, and S3′ substrate binding sites (Schechter
convention4) can be exploited to design specific substrates
for PR3.5,6 Because the charged amino acids Lys99 (S2), Asp61
(S1′ and S3′), and Arg143 (S2′) have polar or hydrophobic
counterparts in HNE (Asn, Leu, and Leu, respectively),
peptides with negatively charged residues at P2 and P2′ and
positively charged residues at P1′ and P3′ are highly specific for
PR3.5,6 Most recent investigations of PR3 specificity have been
conducted using fluorescence resonance energy transfer
(FRET) peptide substrates containing N-terminal 2-amino-
benzoic acid (Abz) as a fluorophore and C-terminal N-(2,4-
dinitrophenyl)ethylenediamine (EDDnp) as a quencher.7,8
These groups have been instrumental for the investigation of
RESULTS AND DISCUSSION
■
Substrate Design and Modeling. We have earlier
demonstrated that the number and strength of the enzyme−
substrate interactions for a given peptide in the Michaelis
complexes of PR3 (and HNE) constituted a good predictor of
the ability of the enzyme to efficiently cleave that substrate.6
We here first investigate the interactions of PR3 and HNE with
a peptide, Abz-VADnVADYQ-EDDnp (nV = norvaline),
reported to be specific for PR3.5,8 We used molecular docking
Received: September 18, 2013
Published: January 7, 2014
© 2014 American Chemical Society
1111
dx.doi.org/10.1021/jm401559r | J. Med. Chem. 2014, 57, 1111−1115

Journal of Medicinal Chemistry
Brief Article
and molecular dynamics (MD) simulations to build structural
models of the Michaelis complexes and further analyzed the
MD trajectories by free energy decomposition to quantify the
relative contribution of the different chemical groups to the
overall binding affinity and the energetic costs of replacing key
amino acids. Another advantage of MD simulations over static
structural considerations or rigid molecular docking is that it
takes into account the interdependence between pockets and
their plasticity.6 We then varied the nature of the amino acids at
P2, P2′, and P3′, as well the N- and C-terminal groups
(presence or absence of Abz and EDDnp). Herein, the
substrates will be referred to depending on the nature of P2-
P2′P3′: VADnVADYQ-NH2 as D-DY and Abz-VADnVADYQ-
EDDnp as D-DYFRET, etc. Detailed results of the MM/PBSA
free energy decomposition are given as Supporting Information
(Table S1 and Figures S1−S3).
Figure 1. Structural model of the D-DRFRET/PR3 Michaelis complex
from MD simulations: PR3 in gray, peptide backbone in ochre,
catalytic triad in magenta, charged residues in green or red; pink
spheres, PR3 residues mediating the strongest van der Waals
interactions.
The simulations reproduce the expected structural features
for chymotrypsin-like serine proteases including the antiparallel
β-sheet between backbones of the P1-P3 residues and of amino
acids 214−216. Moreover, as expected, the side chain of P1-
nVal mediates extensive hydrophobic contacts with residues of
the S1 site. Interestingly the simulations of PR3 complexed
with the fluorogenic substrates show that the EDDnp and Abz
groups are in contact with the enzyme. EDDnp essentially
interacts with the loop Met35-Ser39 (side chains of Met35 and
Asn38A, backbones of Pro38B and Gly38C), while Abz
interacts with Trp218 (a glycine in HNE). Although the loop
Met35-Ser39 is shorter in HNE than PR3, most interactions
mediated by EDDnp are conserved between PR3 and HNE.
The free energy decompositions of the five complexes between
PR3 and the FRET substrates show that the two groups are
involved in van der Waals interactions. In particular EDDnp
exhibits binding energies with PR3 (−6.6 to −4.6 kcal/mol)
comparable to that of the P1 side chain (nVal) (from −5.8 to
−5.5 kcal/mol), which is the side chain of the substrates with
the highest contribution. The simulations indicate that EDDnp
and Abz contribute to the stability of the Michaelis complexes,
thus leading to tighter complexes than with the substrates not
containing the fluorogenic groups.
substrates) and experimentally investigate their hydrolysis by
PR3 and HNE.
Synthesis. All substrates listed in Table 1 were then
synthesized using Fmoc solid phase peptide synthesis. EDDnp
Table 1. Percentage of Hydrolysis of Substrates by PR3 and
a
HNE Measured by HPLC
PR3
HNE
peptide name
(P2-P2′P3′)
b
c
b
c
H/NH2
FRET
H/NH2
FRET
D-DY
D-DR
N-ER
N-EY
A-ER
4.8 1.3
99.7 0.1
93.4 2.7
99.2 0.4
94.2 3.7
99.9 0.1
1.6 1.2
52.8 5.7
7.6 2.6
d
d
nh
nh
d
d
nh
nh
10.9 2.4
75.5 10.0
66.1 2.2
0.9 0.9
0.7 1.2
3.4 1.4
d
nh
a
Peptide names reflect the amino acids at P2, P2′, and P3′. The
remaining amino acids in the sequence are identical for all peptides
(Val-Ala-P2-nVal-Ala-P2′-P3′-Gln). Values reported are the mean and
b
standard deviation of three experiments. Results for peptides with a
c
free N-terminal amine and a C-terminal amide. Fluorogenic substrates
d
(Abz-sequence-EDDnp). nh, not hydrolyzed; no product could be
detected.
The MD simulations show a favorable contribution of P3′ to
the interaction between HNE and P3′-Tyr (−3.4 2.7 kcal/
mol) in D-DYFRET, which indicates that D-DYFRET might be
able to significantly bind to HNE. The corresponding P3′
contribution in PR3/D-DYFRET is also favorable (−0.7 1.5
kcal/mol). To improve the specificity of the substrate for PR3,
we replaced the tyrosine by an arginine (P3′-Arg in D-DRFRET),
yielding as expected an unfavorable P3′-S3′ interaction with
HNE (4.0 3.8 kcal/mol) but a favorable one with PR3 (−1.8
4.4 kcal/mol). In D-DRFRET the three sites (S2-S2′S3′) are
charged amino acids and engage in electrostatic interactions
with PR3 and hydrogen bonds with K99, R143, and D61,
respectively (Figure 1). We then exchanged the charged P2-Asp
for a neutral but polar amino acid: asparagine (P2-Asn in N-
ERFRET). The only significant change in energy upon this
modification is a moderate favorable increase of the van der
Waals contribution of P2-Asn in PR3 and HNE, indicating that
the N-ER sequence could bind preferentially to PR3 compared
to HNE. Unfortunately a small hydrophobic amino acid in P2
(P2-Ala in A-ERFRET) does not contribute any unfavorable
contribution in HNE and thus might increase the overall
binding affinity for the latter. We next proceeded to synthetize
these five sequences (with and without EDDnp, resulting in 10
was incorporated into the substrates by coupling of Fmoc-Glu-
EDDnp to a Rink amide resin through the side chain carboxylic
acid, which ensured formation of the desired glutamine side
chain upon cleavage of the peptides from the solid support. The
required Fmoc-Glu-EDDnp was prepared (Scheme 1) by
Scheme 1. Synthesis of Fmoc-Glu-EDDnp
HCTU-mediated coupling of Fmoc-Glu(OtBu)-OH with N-
(2,4-dinitrophenyl)ethylenediamine, which in turn was pre-
pared by reacting ethylenediamine dihydrochloride with 2,4-
dinitrochlorobenzene in refluxing ethanol. Subsequent depro-
tection of the tert-butyl ester using TFA gave the desired
material in 93% yield over the two steps. In addition to the
substrates listed in Table 1, we also synthetized substrates
containing two variants of the VADnVADYQ-NH2 sequence,
1112
dx.doi.org/10.1021/jm401559r | J. Med. Chem. 2014, 57, 1111−1115

Journal of Medicinal Chemistry
Brief Article
either with only Abz (Abz-VADnVADYQ-NH2, D-DYAbz) or
only EDDnp (VADnVADYQ-EDDnp, D-DYEDDnp).
Table 2. Hydrolysis of D-DY, D-DR, and N-ER by PR3:
Michaelis (Km) and Specificity (kcat/Km) Constants Obtained
Using a Michaelis−Menten or a Substrate Inhibition
Kinetics Model
Influence of the FRET Groups on Substrate Hydrol-
ysis. The percentage of hydrolysis of each peptide by PR3 and
by HNE in vitro was measured using HPLC analysis of the
reaction products (Table 1). We also analyzed the products
resulting from reactions of each fluorogenic substrate with PR3
and HNE (when relevant) using liquid chromatography−mass
spectrometry (LCMS) and found that the expected products
were formed upon hydrolysis of FRET substrates by both
enzymes (cleavage between P1-nVal and P1′-Ala). However,
HNE yielded additional hydrolysis products (cleavage between
P1′-Ala and P2′, detailed in Supporting Information).
name
Km (μM)
kcat/Km (M−1·s−1
)
b
c
c
c
b
c
D-DYFRET
D-DRFRET
35.1 /27.4
615921 /651275
a
b
b
c
33.7 /21.4
548753 /649976
a
b
b
c
N-ERFRET
16.5 /11.5
99848 /120047
a
No hydrolysis by HNE under the same experimental conditions.
Substrate inhibition model. Michaelis−Menten kinetics model.
b
c
nonpolar alanine (cf. A-ERFRET in Table 1) also resulted in an
increased cleavage by HNE.
The fluorogenic substrates were all more extensively
hydrolyzed by PR3 and HNE than the nonfluorogenic
substrates. None of the five substrates without Abz/EDDnp
showed a significant degree of proteolysis by PR3 or HNE
under the experimental conditions. This difference is in
agreement with the predictions from MD simulations and
free energy decompositions. Interestingly adding EDDnp only
to the D-DY sequence (D-DYEDDnp) increased the percentage
of hydrolysis by PR3 (79.5 9.7%) while 99.8 0.2% of D-
DYAbz is cleaved. This indicates an equal contribution of the
fluorophore and the quencher to the increased hydrolysis,
although the simulations predicted a stronger effect of EDDnp
compared to Abz (cf. Table S1 in Supporting Information).
This discrepancy could be a consequence of the fact that the
simulations, by considering only the interactions in the
Michaelis complexes, miss the importance of interactions in
the reaction intermediates. Moreover it is important to consider
that strong interactions might stabilize the Michaelis complex
or reaction intermediates and thereby influence the activation
energy of the corresponding reaction step.
Hydrolysis of D-DYFRET (10 μM) by PR3 (5 nM) was
slowed by a factor of 2 in the presence of an excess of D-DY
(200 μM) compared to the same experiment with only D-
DYFRET (data not shown), indicating that the unmodified
peptide does bind to PR3 and to the same binding sites as its
FRET counterpart, albeit with a lower affinity. This result
further confirms that the Abz and EDDnp groups contribute to
the hydrolysis rate of the substrates by PR3 and HNE.
Influence of P2, P2′, and P3′ on Substrate Hydrolysis.
In what follows, we focus on the fluorogenic peptides. The
percentages of hydrolysis reported in Table 1 show that two of
the five FRET substrates are preferentially cleaved by PR3. D-
DRFRET is cleaved by PR3 (93.4 2.7% hydrolysis) but hardly
by HNE (7.6 2.6%). N-ERFRET, the result of our attempts to
reduce the number of charged amino acids in the substrates, is
also preferentially cleaved by PR3 with a measured percentage
of hydrolysis of 99.2 0.4% and only 10.9 2.4% hydrolysis
by HNE. Although these differences indicate only a moderate
selectivity, it is important to keep in mind that these results are
obtained at high enzyme concentrations in order to detect even
low level of hydrolysis of the non-FRET substrates.
Determination of Km and Specificity Constants for D-
DYFRET, D-DRFRET, and N-ERFRET. The specificity (kcat/Km) and
Michaelis (Km) constants were determined for D-DYFRET and
the two substrates identified as specific for PR3 from the
percentages of cleavage analyzed by HPLC. We used two
kinetic models, either substrate inhibition or Michaelis Menten
(cf. Supporting Information Figure S4). The results are
reported in Table 2. The specificity constants of these two
FRET substrates are comparable, while N-ERFRET yields the
lowest Km and specificity constant of the three substrates. The
low kcat/Km measured for this substrate indicates a kcat
significantly lower than for the other two substrates. Among
the three substrates N-ERFRET is the one with the highest
affinity for the enzyme. Working at lower substrate
concentrations than for the determination of the percentages
of hydrolysis, we do not observe any hydrolysis of D-DRFRET
and N-ERFRET by HNE. We found kcat/Km values that are about
1 order of magnitude lower for D-DYFRET and D-DRFRET than
previously reported values (7 000 000 and 3 400 000 M−1·s−1,
respectively), using lower substrate concentrations (<1 μM)
than in the present study (2−256 μM).5 This discrepancy is
possibly due to the differences in the enzyme concentrations in
the earlier experimental setup and the assumptions used for
kcat/Km determination.
CONCLUSION
■
Molecular dynamics simulations of the substrate−enzyme
complexes and subsequent free energy decompositions led us
to suspect that EDDnp and Abz favorably interact, via van der
Waals interactions, with PR3 and HNE. Using enzymatic
assays, we show that N- and C-terminal FRET groups
significantly contribute to enzymatic hydrolysis. We would
therefore recommend a cautious approach while using Abz/
EDDnp substrates in fluorescence assays and especially so for
the determination of kinetic constants, as these might be biased
by the low substrate concentrations used to avoid inner-filter
effect. On the other hand and because of their improved
interaction with PR3, the introduction of Abz and EDDnp
groups in inhibitors of PR3 for in vitro diagnostic tools is an
advantage over simpler substrates.2 Other FRET groups
containing aromatic rings will have the same ability to mediate
van der Waals interactions and should yield the same effect.
This study also reports a FRET substrate (N-ERFRET) with high
affinity for PR3, not cleaved by HNE, and that can form the
basis for the development of peptidomimetic inhibitors.
Moreover, the results of this study are relevant for the
development of novel FRET substrates for the detection of
NSPs in diagnoses of chronic inflammatory diseases.2,9 Finally
this study illustrates how MD simulations of enzyme/substrates
Interestingly the substrate D-DYFRET has been earlier
reported to be specific for PR3,5 in a study where rather low
substrate concentrations (below 1 μM) were used for the
fluorescence activity assays. In the present work, using higher
substrate concentrations (90 μM, near saturation; see Table 2),
we observe a significant cleavage of D-DYFRET by HNE (52.8
5.7%), in agreement with MD simulations. This discrepancy
could be due to the difference in the experimental conditions
between the two studies. Attempts to replace P2-Asp by a small
1113
dx.doi.org/10.1021/jm401559r | J. Med. Chem. 2014, 57, 1111−1115

Journal of Medicinal Chemistry
Brief Article
Km and kcat/Km Determination. An experimental setup similar to
the one described above was used to determine Michaelis constants
(Km) for the fluorogenic substrates. We used eight different substrate
concentrations ranging from 2 to 256 μM, while the final PR3
concentration was 0.4 nM. Aliquots were taken for analysis
immediately after addition of the enzyme (t = 0) and every 6 min
until t = 42 min. The product formation was determined by HPLC for
each concentration and time point. This was then used to determine
the reaction velocity and further Km values using GraphPad Prism 6
(GraphPad Software, Inc., CA, USA). The reduction in reaction
velocity at high substrate concentrations led us to evaluate the
Michaelis constants by fitting to a model of substrate inhibition.
Molecular Dynamics Simulations. We built the enzyme−
peptide complexes using X-ray structures of PR3 and HNE.6 The
fluorogenic groups were added using the program Autodock 4.2.12
Enzymes/substrate complexes were submitted to energy minimization
followed by MD simulations with NAMD2.913 with the Charmm27
force field.14 Sets of Charmm force field parameters for Abz and
EDDnp were generated and validated as described in the Supporting
Information. The simulations were performed in the NPT ensemble at
a temperature of 300 K, with an integration time step of 1 fs. These
consisted of four successive heating phases (10, 100, 200, and 300 K),
a 150 ps equilibration phase, and a production phase of 2.5 ns. The last
2 ns of five production runs (five replicas with different initial
velocities) are used for analysis. The protocol used to obtain the
energetic contribution of all amino acids to the formation of the
complexes is based on the MM/PBSA approach15 (described in
Supporting Information).
Michaelis complexes followed by MM/PBSA free energy
decompositions can qualitatively guide substrate design and
evaluation of the energetic cost of replacing key charged amino
acids by noncharged groups.
EXPERIMENTAL SECTION
■
Synthesis of Substrates. Rink amide MBHA (100−200 mesh,
typically 0.400 g, 0.590 mmol/g loading) was swelled in a peptide
synthesis vessel using 3 times bed volume of DMF and gentle agitation
for ∼30 min. The solvent was removed by applying vacuum suction,
and a solution of 20% piperidine in DMF was added covering the
resin, followed by agitation for 5 min and draining by vacuum suction.
This procedure was repeated twice more, and the deprotected Rink
amide was washed by adding 3 times bed volume of DMF and
agitating for 1 min and draining of the solvent. The washing procedure
was repeated five times. Simultaneous with the washing, a solution of
Fmoc-Gln-OH or Fmoc-Glu-EDDnp (4 equiv) and HBTU (3.9
equiv) in a minimum amount of DMF was prepared (Fmoc-Glu-
EDDnp was prepared by modification of literature procedures,10 and
full experimental details are in the Supporting Information). To this
solution was added DIPEA (8 equiv) followed by throughout mixing
and immediate addition to the resin. The reaction mixture was agitated
for 1 h after which the solvent was drained and the resin was washed
five times with DMF and treated with 20% piperidine in DMF as
described above. Subsequent coupling of the remaining amino acids
using 4 equiv of the Fmoc-protected amino acid, 3.9 equiv of HBTU,
and 8 equiv of DIPEA was performed following the same procedure.
The deprotection of Fmoc-Asp(OtBu)-OH and Fmoc-Asn(Trt)-OH
and all subsequent Fmoc deprotections were carried out using a 20%
piperidine in DMF solution containing 0.1 M HOBt in order to
suppress aspartimide formation.11 After the coupling and deprotection
of the last amino acid/fluorophore the resin was dried-down by
washing five times with DMF, CH2Cl2, MeOH, and hexane,
respectively before it was dried by vacuum suction for ∼30 min.
The dry resin was treated in the peptide synthesis vessel with 2 times
bed volume of a mixture of TFA, TIS, and water (95:2.5:2.5, v/v)
under gentle agitation for 3 h. The TFA mixture was drained off, and
the resin was washed three times with fresh portions of TFA. The
combined TFA solution was concentrated by rotatory evaporation,
and the residue was precipitated by the addition of cold diethyl ether.
The ether was removed and the residue triturated twice more with
cold diethyl ether. The crude peptides were purified by reverse phase
HPLC, and the combined fractions were lyophilized to give the pure
peptides as fluffy white materials. Peptides containing the Abz and
EDDnp groups were isolated as fluffy yellow materials. All substrates
were found to be of >95% purity (HPLC 220 nm).
ASSOCIATED CONTENT
* Supporting Information
■
S
MM/PBSA decompositions, enzyme kinetics, LC−MS results,
and synthesis procedures. This material is available free of
charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Authors
*B.E.H.: phone, (+47) 55583468; e-mail, Bengt-Erik.Haug@
farm.uib.no.
■
*N.R.: phone, (+47) 55584040; e-mail, Nathalie.Reuter@mbi.
uib.no.
Author Contributions
§S.N. and A.B. contributed equally.
Notes
Proteolytic Analysis Using HPLC. The enzymes PR3 (EC
3.4.21.76) and HNE were purchased from Athens Research &
Technology, Inc.. Purity was checked using SDS−PAGE gel, and PR3
was titrated with α1-proteinase inhibitor. Their activity was verified by
analyzing the hydrolysis of Boc-Ala-Pro-nVal p-chlorothiobenzyl ester,
in which cleavage was detected using 5,5-dithiobis(2-nitrobenzoic
acid) and measuring absorbance at 412 nm. Lyophilized substrates
were dissolved in 30% v/v DMF. The stock solution was further
diluted with 50 mM Hepes, pH 7.4, 750 mM NaCl, supplemented
with 0.05% (v/v) Igepal CA-630. For the proteolytic analysis, 90 μM
substrate was incubated with 100 nM PR3 or HNE in 60 μL final
reaction volume. Briefly, after a 30 min incubation at 37 °C the
reaction was stopped by adding 5 μL of 10% trifluoroacetic acid (TFA)
and incubating on ice for 10 min. This was followed by centrifugation
for 10 min at 13000g at 4 °C. The supernatant was then transferred to
HPLC vials. Hydrolysis products were separated and analyzed by
reverse-phase HPLC using Shimadzu Prominence module HPLC
instrument fitted with a Machery-Nagel C18 HD column. Samples
were eluted using different gradients of water/acetonitrile/TFA
(0.01% TFA) mobile phase for 55 min. Percentages of hydrolysis
were calculated from relative areas under the curve of the (uncleaved)
substrate peak. Each experiment was conducted three times with three
different enzyme and substrate preparations.
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
Parallab and NOTUR are thankfully acknowledged for
provision of CPU time. The authors thank Prof. M. Ziegler
for insightful discussions and Dr. B. Holmelid and J. Johansen
for performing MS analyses. This work was supported by grants
from the Bergen Research Foundation and the Meltzer
Foundation to N.R., C.G., and S.N. and by a grant from
BTO-Visjon vest to N.R. and B.E.H.
ABBREVIATIONS USED
■
Abz, 2-aminobenzoic acid; EDDnp, N-(2,4-dinitrophenyl)-
ethylenediamine; PR3, proteinase 3; HNE, human neutrophil
elastase; MD, molecular dynamics
REFERENCES
■
(1) Korkmaz, B.; Horwitz, M. S.; Jenne, D. E.; Gauthier, F.
Neutrophil elastase, proteinase 3, and cathepsin G as therapeutic
targets in human diseases. Pharmacol. Rev. 2010, 62, 726−759.
1114
dx.doi.org/10.1021/jm401559r | J. Med. Chem. 2014, 57, 1111−1115

Journal of Medicinal Chemistry
Brief Article
(2) Wysocka, M.; Lesner, A.; Gruba, N.; Korkmaz, B.; Gauthier, F.;
Kitamatsu, M.; Legowska, A.; Rolka, K. Three wavelength substrate
system of neutrophil serine proteinases. Anal. Chem. 2012, 84, 7241−
7248.
(3) Hajjar, E.; Broemstrup, T.; Kantari, C.; Witko-Sarsat, V.; Reuter,
N. Structures of human proteinase 3 and neutrophil elastaseso
similar yet so different. FEBS J. 2010, 277, 2238−2254.
(4) Schechter, I.; Berger, A. On the size of the active site in proteases.
I. Papain. Biochem. Biophys. Res. Commun. 1967, 27, 157−162.
(5) Korkmaz, B.; Hajjar, E.; Kalupov, T.; Reuter, N.; Brillard-Bourdet,
M.; Moreau, T.; Juliano, L.; Gauthier, F. Influence of charge
distribution at the active site surface on the substrate specificity of
human neutrophil protease 3 and elastase. A kinetic and molecular
modeling analysis. J. Biol. Chem. 2007, 282, 1989−1997.
(6) Hajjar, E.; Korkmaz, B.; Gauthier, F.; Brandsdal, B. O.; Witko-
Sarsat, V.; Reuter, N. Inspection of the binding sites of proteinase 3 for
the design of a highly specific substrate. J. Med. Chem. 2006, 49, 1248−
1260.
(7) Juliano, M. A.; Filira, F.; Gobbo, M.; Rocchi, R.; Del, N. E.;
Juliano, L. Chromogenic and fluorogenic glycosylated and acetylgly-
cosylated peptides as substrates for serine, thiol and aspartyl proteases.
J. Pept. Res. 1999, 53, 109−119.
(8) Korkmaz, B.; Attucci, S.; Moreau, T.; Godat, E.; Juliano, L.;
Gauthier, F. Design and use of highly specific substrates of neutrophil
elastase and proteinase 3. Am. J. Respir. Cell Mol. Biol. 2004, 30, 801−
807.
(9) Popow-Stellmaszyk, J.; Wysocka, M.; Lesner, A.; Korkmaz, B.;
Rolka, K. A new proteinase 3 substrate with improved selectivity over
human neutrophil elastase. Anal. Biochem. 2013, 442, 75−82.
(10) Hirata, I. Y.; Cezari, M. H. S.; Nakaie, C. R.; Boschcov, P.; Ito,
A. S.; Juliano, M. A.; Juliano, L. Internally quenched fluorogenic
protease substrates: solid-phase synthesis and fluorescence spectros-
copy of peptides contain ortho-aminobenzoyl/dinitrophenyl groups as
donor−acceptor pairs. Lett. Pept. Sci. 1995, 1, 299−308.
(11) Lauer, J. L.; Fields, C. G.; Fields, G. B. Sequence dependence of
aspartimide formation during 9-fluorenylmethoxycarbonyl solid-phase
peptide synthesis. Lett. Pept. Sci. 1995, 1, 197−205.
(12) Goodsell, D. S.; Olson, A. J. Automated docking of substrates to
proteins by simulated annealing. Proteins 1990, 8, 195−202.
(13) Phillips, J. C.; Braun, R.; Wang, W.; Gumbart, J.; Tajkhorshid,
E.; Villa, E.; Chipot, C.; Skeel, R. D.; Kale, L.; Schulten, K. Scalable
molecular dynamics with NAMD. J. Comput. Chem. 2005, 26, 1781−
1802.
(14) Mackerell, A. D.; Feig, M.; Brooks, C. L. Extending the
treatment of backbone energetics in protein force fields: limitations of
gas-phase quantum mechanics in reproducing protein conformational
distributions in molecular dynamics simulations. J. Comput. Chem.
2004, 25, 1400−1415.
(15) Lafont, V.; Schaefer, M.; Stote, R.; Altschuh, D.; Dejaegere, A.
Protein−protein recognition and interaction hot spots in an antigen−
antibody complex: free energy decomposition identifies efficient
amino acids. Proteins 2007, 67, 418−434.
1115
dx.doi.org/10.1021/jm401559r | J. Med. Chem. 2014, 57, 1111−1115