Synthesis, biological evaluation, and docking studies of PAR2-AP-derived pseudopeptides as inhibitors of kallikrein 5 and 6

Article abstract of DOI:10.1515/hsz-2014-0190

A series of protease activated receptor 2 activating peptide (PAR2-AP) derivatives (1-15) were designed and synthesized. The obtained compounds were tested on a panel of human kallikreins (hKLK1, hKLK2, hKLK5, hKLK6, and hKLK7) and were found completely i

Full text of DOI:10.1515/hsz-2014-0190

Biol. Chem. 2015; 396(1): 45–52
Short communication
Beatrice Severino*, Ferdinando Fiorino, Angela Corvino, Giuseppe Caliendo, Vincenzo
Santagada, Diego Magno Assis, Juliana R. Oliveira, Luiz Juliano, Serena Manganelli, Emilio
Benfenati, Francesco Frecentese, Elisa Perissutti and Maria Aparecida Juliano
Synthesis, biological evaluation, and docking
studies of PAR2-AP-derived pseudopeptides as
inhibitors of kallikrein 5 and 6
Abstract: A series of protease activated receptor 2 activat- mechanism. All but four kallikrein-related peptidases
ing peptide (PAR2-AP) derivatives (1–15) were designed present an Asp residue homologous to that of trypsin,
and synthesized. The obtained compounds were tested located at the base of the substrate-binding pocket,
on a panel of human kallikreins (hKLK1, hKLK2, hKLK5, addressing the specificity toward positively charged Lys
hKLK6, and hKLK7) and were found completely inactive and Arg residues. Human tissue kallikrein 5 (hKLK5), first
toward hKLK1, hKLK2, and hKLK7. Aiming to investigate isolated from human plantar stratum corneum (Brattsand
the mode of interaction between the most interesting and Egelrud, 1999), is found in many tissues and it appears
compounds and the selected hKLKs, docking studies were to be most abundantly expressed in human skin (Shaw
performed. The described compounds distinguish the dif- and Diamandis, 2007). hKLK5 was reported to play role
ferent human tissue kallikreins with compounds 1 and 5 in a kallikrein proteolytic cascade pathway activating the
as the best hKLK5 and hKLK6 inhibitors, respectively.
prostate kallikreins hKLK2 and hKLK3 that participate in
seminal clot liquefaction and possibly in prostate cancer
Keywords: human kallikrein 5; human kallikrein 6; inhib- progression (Michael et al., 2006). Moreover, hKLK5 can
itor; molecular modeling; serine protease.
play a major role in the pathobiology of Netherton syn-
drome (NS), a rare genetic skin disease (Hovnanian,
2
012). hKLK5 has a trypsin-like specificity, with a strong
DOI 10.1515/hsz-2014-0190
Received April 22, 2014; accepted July 25, 2014; previously published
online July 31, 2014
preference for Arg over Lys residues at P1 position of the
substrates. The specificities of the other subsites were not
determined in detail; however, the S2 subsite seems to
accept hydrophobic amino acids as Pro and Phe (Michael
et al., 2005; Debela et al., 2006).
Human kallikreins (hKs) are a multigene family of 15
secreted serine-type proteases (Yousef et al., 2000; Yousef
and Diamandis, 2001). Kallikrein-related peptidases show
a high degree of homology with trypsin; this similarity is
related both to the primary structure and to the catalytic
A number of hKLKs are expressed in the central
nervous system (CNS); among them, hKLK6 is the most
studied and its involvement in several neurodegenerative
diseases such as Alzheimer’s disease, Parkinson’s disease,
and multiple sclerosis has been documented (Diamandis
et al., 2000; Ogawa et al., 2000). A detailed analysis of
hKLK6 substrate specificity showed that it is a restricted
arginyl hydrolase and has high preference for hydropho-
bic amino acids at P2 and P1′ positions (Angelo et al.,
2006). On the basis of this study, the best substrate thus
far reported for hKLK6 was synthesized, Abz-AFR↓FSQ-
EDDnp, which was cleaved at the R-F bond (as indicated
by ↓). It is noteworthy that the AFRFS sequence was found
as a motif in the amino-terminal domain of seven human
ionotropic glutamate receptor subunits. It has been sug-
gested that hKLK1 and hKLK6 are directly implicated in
*
Corresponding author: Beatrice Severino, Dipartimento di
Farmacia, Università degli Studi di Napoli ‘Federico II’, Via D.
Montesano, 49, I-80131 Napoli, Italy, e-mail: bseverin@unina.it
Ferdinando Fiorino, Angela Corvino, Giuseppe Caliendo, Vincenzo
Santagada, Francesco Frecentese and Elisa Perissutti: Dipartimento
di Farmacia, Università degli Studi di Napoli ‘Federico II’, Via D.
Montesano, 49, I-80131 Napoli, Italy
Diego Magno Assis, Juliana R. Oliveira, Luiz Juliano and Maria
Aparecida Juliano: Departamento de Biofisica, Escola Paulista de
Medicina, Universidade Federal de São Paulo, Rua Três de Maio 100,
São Paulo 04044-020, Brasil
Serena Manganelli and Emilio Benfenati: IRCCS, Istituto di Ricerche
Farmacologiche Mario Negri, Via La Masa 19, I-20156 Milano, Italy
Brought to you by | Brigham Young University
Authenticated
Download Date | 5/19/15 9:48 AM

4
6ꢀꢀꢀꢀꢁꢀB. Severino et al.: Inhibitors of hKLK5 and hKLK6
neurodegeneration and could be useful tools to follow Table 1ꢂChemical data for compounds 1–4.
multiple sclerosis progression (Scarisbrick et al., 2008).
NH -Ser-Leu --- SPACER --- Arg-Leu-NH
Aside from their involvement in pathologies of the
ꢁ
ꢁ
skin and CNS, overexpression of hKLK5 and hKLK6 Compdꢂ
occurs in several malignancies, including ovarian, breast
and testicular cancer. Accumulating evidence suggests
SPACER
ꢂ
t (min)ꢂ
MW
R
ꢂ
Calcdꢂ
Found
ꢃ
that hKLKs are activators of protease activated receptors
HN
ꢃ
ꢄꢅ.ꢄ7ꢃ
ꢄ7.ꢊꢊꢃ
ꢆꢄꢄ.ꢇꢃ
ꢆꢄꢈ.ꢉ
ꢆꢊꢉ.ꢆ
(
(
PARs) that comprise four G protein-coupled receptors
PAR1, PAR2, PAR3, and PAR4) (Caliendo et al., 2012). It
ꢃ
O
ꢃ
ꢃ
ꢁ
3
ꢆꢄꢅ.ꢇꢃ
ꢆꢌ7.ꢇꢃ
has been reported that hKLK5, through PAR2 signaling,
increases the expression of proinflammatory cytokines
and chemokines in human primary keratinocytes (Briot
et al., 2009); hKLK6 evoked intracellular Ca2 flux medi-
ated by PAR1 in neurons and both PAR1 and PAR2 in
astrocytes (Vandell et al., 2008). Moreover, a study on the
substrate specificity of hKLK6, performed using fluores-
cence resonance energy transfer peptides derived from
sequences that span the cleavage sites for activation of
NH
O
ꢃ
O
ꢊꢋ.ꢊꢊꢃ
ꢊꢄ.ꢇꢌꢃ
ꢆꢌꢇ.7
ꢆꢉꢆ.ꢆ
HN
3
+
ꢃ
ꢃ
O
ꢃ
NH
ꢄ
ꢆꢉꢋ.7ꢃ
Compounds 1–15 were synthesized by SPPS using the standard
Fmoc method in a manual reaction vessel and in a stepwise fashion.
PAR1–4, has shown that only the substrate with the PAR2 For each compound, 0.5 g (0.35 mequiv) of Rink amide MBHA resin
sequence was hydrolyzed (Angelo et al., 2006).
was used. Each Fmoc-amino acid was coupled in a 3-fold excess,
through TBTU/HOBt/DIPEA, except for the first amino acid that was
used in a 5-fold excess and coupled through DIPCDI/HOBt. The Fmoc
protecting groups were removed by treatment with a 25% piperidine
in DMF (45 min). The commercially available difunctional scaffolds
cis-4-aminocyclohexanecarboxylic acid, 4-(aminomethyl)benzoic
hKLK5 and hKLK6 inhibitors would be very useful
pharmacological tools for elucidating the physio-path-
ological processes where the mentioned kallikrein iso-
forms are involved; moreover, these inhibitors might serve
as lead compounds for developing new drugs to treat acid, 4-(4-aminophenyl)butanoic acid, trans-4-(aminomethyl)
cyclohexanecarboxylic acid, 2-aminobenzoic acid, and 4-ami-
diseases that result from their uncontrolled activities.
nobenzoic acid were dissolved in 9% Na CO and converted to the
2
3
The development of hKLK6 inhibitors is thus far scarce,
although hKLK6 is involved in neurodegenerative dis-
eases and cancer. A series of N-(4-aminomethylphenyl)-
corresponding Fmoc derivatives by treatment with a solution of
N-(9-fluorenylmethoxycarboniloxy)succinimide (Fmoc-OSu) in DMF.
Each protected pseudopeptide was cleaved from the resin, and
2
-hydroxy-benzamide derivatives have been recently the amino acid side chains were simultaneously deprotected by
reported, and the most potent inhibitor presented a K of treatment with a mixture of TFA/DCM/TIS/anisole (90:5:3:2) for 2 h
i
at room temperature under nitrogen atmosphere. The conditions
0
.3 μꢀ (Liang et al., 2012a). Some natural isocoumarins,
employed for purification and characterization of the pseudopep-
tides are reported in Table 2.
instead, have shown their ability in inhibiting hKLK5 and
hKLK7, with vioxanthin being the most potent inhibitor
toward hKLK5 with K ꢀ= ꢀ2 2.9 μꢀ (Teixeira et al., 2011).
i
Starting from the consideration that PAR2 is a hKLK5 were synthesized by Fmoc solid-phase peptide synthesis
and hKLK6 substrate, we hypothesize that compounds (SPPS). Mature hKLK1, hKLK2, hKLK5, hKLK6, and hKLK7
derived from the PAR2 activating peptide (PAR2-AP), were expressed and purified from baculovirus/insect cell
namely the sequence SLIGRL, could be prepared and line system as previously described (Bernett et al., 2002).
assayed for activity against hKLK5 and hKLK6. Moreover, The compounds presented in Tables 1 and 2 were com-
to verify their selectivity toward other hKLKs, the com- pletely inactive toward hKLK1 and hKLK2, which also have
pounds were tested also toward hKLK1, hKLK2, and hKLK7. a high preference for Arg at the cleavage site, and were
Compounds 1–4 were already described in a previ- also inactive on hKLK7, which has a high specificity for
ous article reporting a SAR study on PAR2-AP derivatives Phe and Tyr. As shown in Table 3, several of the described
3
4
obtained replacing Ile -Gly dipeptide backbone with compounds display interesting K values toward hKLK5
i
more rigid difunctional scaffolds (Santagada et al., 2002). and/or hKLK6. Therefore, these compounds distinguish
Compounds 5–15 are lower molecular weight (MW) deriva- the different human tissue kallikreins. The compounds
tives in which an aliphatic or aromatic nucleus, variously of the series NH -S-L-spacer-R-L-NH (1–4) were more
2
2
substituted, has been linked to the PAR2-AP carboxy- effective inhibitors for hKLK5 (K values ranging from
i
terminal dipeptide Arg-Leu. The structures of the com- 1.0 ꢀ± ꢀ0 .05 to 31 ꢀ± ꢀ2 μꢀ) than for hKLK6 (K values ranging
i
pounds are reported in Tables 1 and 2. Compounds 1–15 from 2.9 ꢀ± ꢀ0 .2 to 254 ꢀ± ꢀ2 0 μꢀ). In contrast, the compounds
Brought to you by | Brigham Young University
Authenticated
Download Date | 5/19/15 9:48 AM

B. Severino et al.: Inhibitors of hKLK5 and hKLK6ꢂꢁꢀꢀꢀꢂ47
Table 2ꢂChemical data for compounds 5–15.
Table ꢈꢂK values toward hKLK5 and hKLK6, respectively, of
i
compounds 1–15 and the lowest ΔG docking values of the most
Spacer --- Arg-Leu-NHꢁ
significant clusters of hKLK5 and hKLK6 complexed with compounds
1, 2, 5, and 7.
Compdꢃ
Spacer
ꢃ
tR (min)ꢂ
MW
ꢂ
Calcdꢂ
Found
Compdꢂ hKLKꢅ
ꢂ
ꢂ
ꢂ
hKLKꢆ ꢂ
ꢂ
hKLKꢅ ꢂ
ꢂ
hKLKꢆ
ꢅ
ꢃ
ꢃ
K (μꢉ)
K (μꢉ) ΔG (kcal/mol) ΔG (kcal/mol)
i
i
CI
ꢄꢇ.ꢋꢌꢃ
ꢄꢆ.ꢉꢌꢃ
ꢄꢅ.ꢊꢆꢃ
ꢌꢊꢌ.ꢅꢃ
ꢌꢊꢋ.ꢇ
ꢌꢊꢆ.ꢋ
ꢈꢅ7.7
O
ꢃꢃ ꢄ.ꢉ ꢀ± ꢀꢉ .ꢉꢋ
ꢃ
ꢃ
ꢃ
ꢃ
ꢊ.ꢅ ꢀ± ꢀꢉ .ꢊ ꢃ
ꢌ.ꢌ ꢀ± ꢀꢉ .ꢈ ꢃ
ꢌꢋ ꢀ± ꢀꢈ ꢃ
-7.ꢊꢅ ꢃ
-ꢆ.ꢇ; -ꢆ.ꢊ ꢃ
– ꢃ
-ꢇ.ꢄꢈ
-ꢅ.ꢋꢌ
ꢃ
ꢁ
3
ꢃ
ꢃ
ꢄ.ꢊ ꢀ± ꢀꢉ .ꢉ7
ꢊ.ꢅ ꢀ± ꢀꢉ .ꢄ
–
–
ꢆ
ꢃ
ꢌꢊꢋ.ꢆꢃ
ꢈꢅꢆ.ꢋꢃ
H N
O
2
ꢄꢃ ꢈꢄ ꢀ± ꢀꢊ
ꢊꢋꢌ ꢀ± ꢀꢊ ꢉ ꢃ
– ꢃ
a
a
ꢃ
ꢃ
ꢅꢃ Not inhibitedꢃ
ꢆꢃ ꢅ.ꢉ ꢀ± ꢀꢉ .ꢆ
ꢇꢃ Not inhibitedꢃ
ꢄ.ꢋ ꢀ± ꢀꢉ .ꢉꢇ ꢃ -ꢇ.ꢌꢌ ; -7.ꢈꢇ
ꢃ
-ꢇ.ꢆꢊ; -ꢇ.ꢈꢆ
ꢇ
ꢃ
ꢈ.ꢅ ꢀ± ꢀꢉ .ꢊ ꢃ
ꢋ.7 ꢀ± ꢀꢉ .ꢈ ꢃ
ꢇ.ꢌ ꢀ± ꢀꢉ .ꢆ ꢃ
ꢄꢈ.ꢉ ꢀ± ꢀꢉ .ꢇ ꢃ
ꢊꢄ.ꢇ ꢀ± ꢀꢊ ꢃ
ꢊꢆ.ꢈ ꢀ± ꢀꢈ ꢃ
ꢊꢆ.ꢈ ꢀ± ꢀꢊ ꢃ
77 ꢀ± ꢀꢆ ꢃ
– ꢃ
–
-ꢇ.ꢈꢉ
O
-ꢇ.ꢅ7a
ꢃ
ꢃ
– ꢃ
– ꢃ
– ꢃ
– ꢃ
– ꢃ
– ꢃ
– ꢃ
– ꢃ
F CO
3
8ꢃ ꢄꢋ ꢀ± ꢀꢄ
ꢊꢄ ꢀ± ꢀꢊ
ꢃ
ꢃ
ꢃ
ꢃ
ꢃ
ꢃ
ꢃ
–
–
–
–
–
–
–
–
9ꢃ
8
ꢊꢊ.ꢇꢉꢃ
ꢄꢇ.ꢈꢇꢃ
ꢌ7ꢌ.ꢋꢃ
ꢌꢊꢆ.ꢋꢃ
ꢌ7ꢋ.ꢈ
ꢌꢊ7.ꢇ
ꢃ0ꢃ ꢌꢈ.ꢉ ꢀ± ꢀꢋ
ꢃꢃꢃ ꢈꢇ ꢀ± ꢀꢊ
O
ꢃ
O
S
ꢃ
ꢃꢁꢃ ꢌ.ꢅ ꢀ± ꢀꢉ .ꢌ
ꢃ3ꢃ ꢌ.ꢈ ꢀ± ꢀꢉ .ꢈ
ꢃꢄꢃ ꢄꢄ.ꢅ ꢀ± ꢀꢉ .ꢅ
9
ꢃ
O
ꢄꢌꢇ ꢀ± ꢀꢇ ꢃ
ꢄ7ꢇ ꢀ± ꢀꢄ ꢈ ꢃ
ꢃꢅꢃ Not inhibitedꢃ
ꢃ
ꢃ
ꢃ
F C
ꢃ0
3
ꢊꢊ.ꢈꢉꢃ
ꢄꢌ.ꢉꢅꢃ
ꢌꢋꢇ.ꢋꢃ
ꢌꢉꢋ.ꢋꢃ
ꢌꢋꢅ.ꢆ
ꢌꢉꢆ.ꢈ
Assays were carried out at 3ꢍ°C in 50 mꢎ Tris/HCl and 1 mꢎ EDTA
pH 9.0). The substrate used was 5.3 μꢎ Abz-KLRSSKQ-EDDnp
O
ꢃ
(
-
1
H N
(K ꢀ= ꢀ2 .9 μꢎ and K ꢀ= ꢀ1 4.ꢍ s ; enzyme concentration, 2 nꢎ). The K
ꢃꢃ
2
m
c
a
t
i
O
values were calculated using the relations v /v ꢀ= ꢀ1 +[I]/K (app), where
0 i i
v is the hydrolysis velocity in the absence of inhibitors and v is the
ꢃ
O
0
i
velocity in the presence of different concentrations of the inhibitors.
values were obtained using the relation K ꢀ= ꢀK (app)/(1+[S]/K ).
ꢃꢁ
ꢄꢆ.ꢅ7ꢃ
ꢌꢌ7.ꢆꢃ
ꢌꢌꢇ.ꢋ
H N
K
i
i
i
m
2
3
ꢃ
Mature hKLK1, hKLK2, hKLK5, hKLK6, and hKLKꢍ were expressed
and purified from baculovirus/insect cell line system, as previously
described (Bernett et al., 2002). The final yield of purified mature
hKLKs was typically 20–25 mg/l of culture. The purity of mature
hKLKs was assessed by SDS-PAGE (w/v 15% polyacrylamide gel)
and Coomassie brilliant blue staining, N-terminal sequencing, and
MALDI-TOF MS. The protein samples were about 98% homogene-
ous, with no visible degradation products and had the expected
ꢃ
ꢃ
ꢃ
ꢃ3
ꢊꢌ.ꢉ7ꢃ
ꢄꢌ.ꢊꢄꢃ
ꢈꢅꢉ.ꢋꢃ
ꢌꢄꢅ.ꢋꢃ
ꢈꢅꢄ.ꢌ
ꢌꢊꢉ.ꢇ
O
ꢃ
ꢃꢄ
H N
O
2
ꢃ
O
ꢃꢅ
ꢊꢉ.ꢇꢇꢃ
ꢈꢅꢆ.ꢋꢃ
ꢈꢅ7.ꢌ mature N-terminus corresponding to the mature proteases. Molar
concentrations of active hKLK6 and hKLK1 were determined by titra-
tion with 4-methylumbelliferyl p-guanidinobenzoate hydrochloride
by spectrofluorimetric titration as described previously (Jameson
et al., 19ꢍ3).
ꢃ
The crude compounds were purified by reversed phase (RP)-HPLC
on a preparative Vydac C18 column (15–20 μm, 50 ꢀ× ꢀ2 50 mm; The
Separations Group, Inc., Hesperia, CA, USA) using a gradient of
a
These ΔG values refer to the different binding modes supposed to
CH CN in 0.1% aqueous TFA (from 0% to 45% in 35 min) at a flow
be responsible for the lack of inhibition of these compounds.
3
rate of 30 ml/min. Analytical RP-HPLC was done using the following
conditions: RP Vydac C18 column (5 μm, 4.6 ꢀ× ꢀ2 50 mm) and a two-
solvent system: eluent A, 0.05% TFA (v/v) in water; eluent B, 0.05%
TFA (v/v) in acetonitrile; eluted in a linear gradient from 0% to 50%
B during 35 min, UV detection at 220 nm, and flow rate at 1 ml/
min. Analytical RP-HPLC indicated a purity of ꢀ> ꢀ9 8%, and the correct
MWs were confirmed by electrospray ionization-MS, performed on
a ThermoFinnigan LCQ Ion-Trap mass spectrometer (Thermo Fisher
Scientific Inc., Waltham, MA, USA).
NH -S-L-spacer-R-L-NH incorporated both flexible (1) as
2
2
well as rigid spacers (2). However, the most rigid struc-
ture, as compound 4, reduced the inhibition by two orders
of magnitude. The lower inhibition activity of compound
3
compared with that of 2 indicated that the length of the
spacer has significant influence on the interaction with
hKLK6.
of the series spacer-R-L-NH (5–15) were less effective
Aiming to investigate the mode of interaction between
2
inhibitors for hKLK5, with 12 and 13 as the best inhibitors the most interesting compounds and the selected hKLKs,
of the series, which, in contrast, are very poorly effective docking studies were performed. To verify the presence of
toward hKLK6. The best inhibitors for hKLK6 in the series direct and frequent interactions between some PAR2-AP
Brought to you by | Brigham Young University
Authenticated
Download Date | 5/19/15 9:48 AM

4
ꢊꢀꢀꢀꢀꢁꢀB. Severino et al.: Inhibitors of hKLK5 and hKLK6
derivatives and, respectively, hKLK5 and hKLK6, molecular the coordinates of the atoms. One way to measure the
docking with Autodock v4.2 was performed. Compounds 1, reliability of a result is to compare the rmsd of the lowest
2
, 5, and 7 were considered the most significative, as they energy conformations and their rmsd to one another, to
all inhibited hKLK6, while, tested on hKLK5, the first two group them into families of similar conformations (clus-
chemicals showed the lowest experimental K values and ters). In this process, the lowest energy conformation is
i
the last two had no inhibition activity. The two proteases used as the seed for the first cluster. Next, the second
and the assayed compounds were considered as rigid conformation is compared with the first. If it is within
receptors and flexible ligands, respectively. At first, the the rmsd tolerance, it is added to the first cluster. If not,
flexible ligands were designed with Marvin Sketch v5.10, it becomes the first member of a new cluster. This process
followed by the investigation of the lowest energy con- is repeated with the rest of the docked results, grouping
former within the force field MMFF94. At the same time, them into families of similar conformations. In this work,
the crystal-bound structures of hKLK5 and hKLK6 com- different clusters were made and examined at different
plexed with their inhibitors, leupeptin and 4-{[(3R)-3-{[(7- rms tolerance values until 3.5 Å was chosen as the most
methoxynaphthalen-2-yl)sulfonyl](thiophen- 3-ylmethyl) proper value for grouping. Before making this choice for
amino}-2-oxopyrrolidin-1-yl]methyl}thiophene-2-carbox- the synthesized compounds, the originally bound struc-
imidamide, respectively (Liang et al., 2012b), were down- tures were redocked for the calibration. Using an rmsd
loaded as.pdb files (2PSX and 3VFE) from the Protein Data tolerance of 2.0 Å, a cluster analysis is automatically per-
Bank and visualized to define their specific binding sites. formed by Autodock. The lowest binding energy cluster
Two grids of proper sizes centered on these identified of redocked leupeptin and 4-{[(3R)-3-{[(7-methoxynaph-
portions were built. Their associated parameter values thalen-2-yl)sulfonyl](thiophen- 3-ylmethyl)amino}-2-ox-
were specific for each hKLK. Coordinates (x, y, and z) of opyrrolidin-1-yl]methyl}thiophene-2-carboximidamide)
the center of the grid of the two hKLKs were chosen on was consistent with their crystallographic poses, and they
the basis of the crystal structure of the protease-ligand had minimum free binding energy values of -7.93 and -9.68
complex. This choice was made considering the protein kcal/mol, respectively.
binding sites, keeping the fundamental amino acids for
interactions and also in line with the dimensions of the binding values. Outputting structurally similar clusters are
ligands to be docked.
ranked by Autodock in order of increasing energy. Thus,
The docking scores in Autodock are provided by ΔG
Moreover, genetic algorithm (GA) was selected as the the lowest binding energy value cluster is assigned to the
docking mode. GAs are a class of optimization methods cluster rank 1. The ranking of the ligands was correct for
that are based on various computational models of Dar- all the docked compounds, and their free binding energy
winian evolution (Goldberg, 1989) (the term ‘evolutionary range values were consistent with those of the original
algorithms’ is also used to describe such methods). GAs inhibitors.
involve the creation of a population of potential solu-
Compound 1 had the highest torsional degree of
tions that gradually evolves toward better solutions, and freedom owing to the coexistence of four amino acid free
the use of a chromosome to encode each member of the side chains and a cyclohexane moiety in its two lowest
population. In these methods, each chromosome in a pop- energy chair conformations. These two main structural
ulation encodes one conformation of the ligand together features let these derivatives accommodate in differ-
with its orientation within the binding site. A scoring ent ways. Thus, clusters of similar docked poses were
function is used to calculate the fitness of each member poorly populated, and the output was slightly ambigu-
of the population and to select individuals for each itera- ous. However, in the complexes with the lowest binding
tion. As with the Monte Carlo search methods, the under- energy, the Arg side chain extended to the S1 pocket,
lying random nature of the GA means that it is usual to and its carbonyl group was in close contact with Ser195
perform a number of runs and to select the structures with Oγ or Gly193, thus resembling the P1-Arg3i of leupeptin
the highest scores (Leach and Gillet, 2007). GA was com- with its terminal carbonyl group and the amidinothio-
bined with a method of local search, providing a hybrid phene group of 4-{[(3R)-3-{[(7-methoxynaphthalen-2-yl)
algorithm, named Lamarckian GA.
sulfonyl](thiophen-3-ylmethyl)amino}-2-oxopyrrolidin-
The reliability of a docking result depends on the 1-yl]methyl}thiophene-2-carboximidamide, respectively
energies of the docked structures found at the end of each (Figure 1). Moreover, a high degree of superposition
run and on their similarities to each other. In Autodock, turned out for these Arg residues in the lowest energy
the similarity of docked structures is measured by com- docked structures of 1 with the positively charged moieties
puting the root-mean-square deviation (rmsd) between current in the relative cocrystallized ligands (Figure 2).
Brought to you by | Brigham Young University
Authenticated
Download Date | 5/19/15 9:48 AM

B. Severino et al.: Inhibitors of hKLK5 and hKLK6ꢂꢁꢀꢀꢀꢂ4ꢋ
Figure 1ꢂ(A) Compound 1 in the hKLK5 substrate binding site, where the carbon atom of compound 1 are in yellow and those of hKLK5 are
in light gray. (B) Compound 1 in the hKLK6 substrate binding sites, where the carbon atoms of compound 1 are in yellow and those of hKLK6
are in light gray. Docking studies: compounds 1–15 were designed with Marvin Sketch v5.10, followed by the investigation of the lowest
energy conformer within the force field MMFF94. The visualization was performed with UCSF Chimera v1.ꢍ. Molecular docking was done with
Autodock v4.2, using the following protocol for both hKLK5 and hKLK6: genetic algorithm (GA); number of GA runs: 100; population size:
1
50; maximum number of evaluations: long (25 000 000); other parameters set with their default values. To address the ligand on the sup-
posed binding site, Autogrid 4 was performed, selecting the following parameters for hKLK5: number of points in x-dimension: 50; number
of points in y-dimension: 48; number of points in z-dimension: 40; spacing: 0.3ꢍ5 (default); center grid box (x, y, z): 6.566, -2.509, 20.503.
For hKLK6, Autogrid 4 was performed, keeping the same number of points in x, y, and z dimensions and the same spacing, but the following
center grid box (x, y, z): -3.08ꢍ, -10.988, -16.0ꢍ4.
In general, all the supposed inhibitors kept this binding cluster and in the largest cluster complexes with hKLK5
mode. Thus, it may be considered responsible for the (Figure 3). The relative free binding energy ranges were
lack of proteolytic activity or at least strictly related to it. 1.05 kcal/mol (minimum ꢀ= ꢀ- 6.8 kcal/mol) and 3.81 kcal/mol
The lowest free binding energy values for compound 1 (minimum ꢀ= ꢀ- 6.2 kcal/mol), respectively.
complexed with hKLK5 and hKLK6 were -7.29 and -8.13
kcal/mol, respectively.
In the first mentioned complex, the CO of the aromatic
2
spacer was in close contact with the Nε atom of Gln192 and
Compound 2 showed this conservative orienta- its hydrophobic benzene ring was partially aligned with
tion of Arg residue both in the lowest binding energy the P2-Leu2i side chain of the leupeptin crystallographic
Figure 2ꢂSuperposition of compound 1 with crystallographic poses of leupeptin (A) and 4-{[(3R)-3-{[(ꢍ-methoxynaphthalen-2-yl) sulfonyl]
(
thiophen-3-ylmethyl) amino}-2-oxopyrrolidin-1-yl] methyl} thiophene-2-carboximidamide (B).
The carbon atoms of leupeptin and amidinothiophene derivative are in light gray, while those of compound 1 are in purple (A) and cyan (B),
respectively.
Brought to you by | Brigham Young University
Authenticated
Download Date | 5/19/15 9:48 AM

5
ꢌꢀꢀꢀꢀꢁꢀB. Severino et al.: Inhibitors of hKLK5 and hKLK6
Figure ꢈꢂThe lowest energy complex of the lowest binding energy cluster (A) and of the largest cluster (B) complexes of compound 2 in the
hKLK5 substrate binding site.
The carbon atoms of compound 2 are in sea green and those of hKLK5 are in light gray.
pose. In the second one, the spacer was superposed on (minimum ꢀ= ꢀ- 8.36 kcal/mol). The Arg side chain extends
the P3-Leu1 residue, while the C-terminal Leu was on the into S1 resembling the amidinothiophene group, in close
P2-Leu2i residue of leupeptin. For hKLK6, 25 analogous contact with the carboxylic side chain of Asp189 and the
complexes with compound 2 were obtained in the lowest backbone carbonyl oxygen of Asn217, while the Leu side
binding energy cluster that was also the most populated chain is situated adjacent to Ser195, simulating the pyrro-
one, with a ΔG range of 4.94 kcal/mol (minimum ꢀ= ꢀ- 9.54 lidinone group. On the other hand, frequently p-chloro-
kcal/mol).
phenyl group slots into the S4 cleft, delimited by Trp215,
Two similar main clusters were fed by 22 and 19 which is free in the crystallographic pose (Figure 4).
docked structures of compound 5 complexed with However, the specificity in S4 for hydrophobic residues
hKLK6, respectively. The ΔG range values were 2.27 follows the order hKLK4 ꢀ> ꢀh KLK5 ꢀ> ꢀh KLK6 ꢀ> ꢀh KLK7 (non-
kcal/mol (minimum ꢀ= ꢀ- 8.62 kcal/mol) and 1.23 kcal/mol specific), which means that this pocket in hKLK6 can
Figure 4ꢂThe lowest energy cluster (A) and the largest cluster (B) complexes of compound 5 in the hKLK6 substrate binding site.
The carbon atoms of compound 5 are in orange and those of hKLK6 are in light gray.
Brought to you by | Brigham Young University
Authenticated
Download Date | 5/19/15 9:48 AM

B. Severino et al.: Inhibitors of hKLK5 and hKLK6ꢂꢁꢀꢀꢀꢂ51
Figure 5ꢂSuperposition of crystallographic poses of leupeptin with compounds 5 (A and B) and 7 (C and D).
The carbon atoms of leupeptin are in light gray and those of compounds 5 and 7 are in magenta and yellow, respectively. The hydrophobic
spacers of the two non-inhibitors could be aligned with the opposite moiety in leupeptin.
accommodate any non-polar group (Debela et al. 2008). suggested a correlation between specific conformations
These conformations were visible in the results of the and activities.
alignment process with the crystallized inhibitor. In
In conclusion, the PAR2-AP derivatives 1–15 have been
contrast, a completely opposite orientation with the Arg prepared and tested for their ability to inhibit a panel of
side chain extending into the catalytic triad was kept hKs. The compounds were able to distinguish the different
by compound 5 docked with hKLK5 in the most signifi- hKLKs acting as inhibitors only toward hKLK5 and hKLK6.
cant cluster. This chain could not be aligned with the Compound 1 (K ꢀ= ꢀ1 .0 ꢀ± ꢀ0 .05 μꢀ) was the most effective
i
leupeptin correspondent one, but only matches with the toward hKLK5, while 5 (K ꢀ= ꢀ1 .5 ꢀ± ꢀ0 .08 μꢀ), besides being
i
hydrophobic Leu residues were allowed (Figure 5A,B). the best hKLK6 inhibitor, was also very selective, being
Thus, this different binding mode and an incorrect completely inactive toward all the other hKLKs tested in
alignment with cocrystallized leupeptin were sup- our study. The promising results here obtained indicate
posed to be responsible for the lack of inhibition in the these new compounds as valid leads for further optimi-
docking output with hKLK5 for both compounds 5 and zation studies aiming to improve the potency of this new
7
(Figure 5C,D). For these two compounds, the interac- class of hKLK5 and hKLK6 inhibitors.
tions with Asp189 were replaced by that with Asp102
of the catalytic triad. Even if compound 7 docked with Acknowledgments: This work was supported by Fundação
hKLK6 showed two different main poses and not only de Amparo à Pesquisa do Estado de São Paulo (FAPESP
that seen thus far to be related to the inhibition activ- – project-12/50191-4R), Conselho Nacional de Desen-
ity, the overall results supported the hypothesis of an volvimento Científico e Tecnológico (CNPq – projects
inhibition mechanism based on the binding mode and 471340/2011-1 and 470388/2010-2).
Brought to you by | Brigham Young University
Authenticated
Download Date | 5/19/15 9:48 AM

5
2ꢀꢀꢀꢀꢁꢀB. Severino et al.: Inhibitors of hKLK5 and hKLK6
para-amidobenzylanmine P1 group identified through virtual
screening. Bioorg. Med. Chem. Lett. 22, 2450–2455.
Liang, G., Chen, X., Aldous, S., Pu, S.F., Mehdi, S., Powers, E.,
Giovanni, A., Kongsamut, S., Xia, T., Zhang, Y., et al. (2012b).
Virtual screening and X-ray crystallography for human kal-
likrein 6 inhibitors with an amidinothiophene P1 group. ACS
Med. Chem. Lett. 3, 159–164.
Michael, I.P., Sotiropoulou, G., Pampalakis, G., Magklara, A., Ghosh,
M., Wasney, G., and Diamandis, E.P. (2005). Biochemical and
enzymatic characterization of human kallikrein 5 (hK5), a novel
serine protease potentially involved in cancer progression.
J. Biol. Chem. 280, 14628–14635.
References
Angelo, P.F., Lima, A.R., Alves, F.M., Blaber, S.I., Scarisbrick I.A.,
Blaber, M., Juliano, L., and Juliano, M.A. (2006). Substrate
specificity of human kallikrein 6: salt and glycosaminoglycan
activation effects. J. Biol. Chem. 281, 3116–3126.
Bernett, M.J., Blaber, S.I., Scarisbrick, I.A., Dhanarajan, P.,
Thompson, S.M., and Blaber, M. (2002). Crystal structure and
biochemical characterization of human kallikrein 6 reveals that
a trypsin-like kallikrein is expressed in the central nervous
system. J. Biol. Chem. 277, 24562–245ꢍ0.
Brattsand, M. and Egelrud, T. (1999). Purification, molecular clon-
ing, and expression of a human stratum corneum trypsin-like
serine protease with possible function in desquamation.
J. Biol. Chem. 274, 30033–30040.
Briot, A., Deraison, C., Lacroix, M., Bonnart, C., Robin, A., Besson,
C., Dubus, P., and Hovnanian, A. (2009). Kallikrein 5 induces
atopic dermatitis-like lesions through PAR2-mediated thymic
stromal lymphopoietin expression in Netherton syndrome.
J. Exp. Med. 206, 1135–114ꢍ.
Caliendo, G., Santagada, V., Perissutti, E., Severino, B., Fiorino,
F., Frecentese, F., and Juliano, L. (2012). Kallikrein protease
activated receptor (PAR) axis: an attractive target for drug
development. J. Med. Chem. 55, 6669–6686.
Michael, I.P., Pampalakis, G., Mikolajczyk, S.D., Malm, J., Sotiropou-
lou, G., and Diamandis, E.P. (2006). Human tissue kallikrein
5
is a member of a proteolytic cascade pathway involved in
seminal clot liquefaction and potentially in prostate cancer
progression. J. Biol. Chem. 281, 12ꢍ43–12ꢍ50.
Ogawa, K., Yamada, T., Tsujioka, Y., Taguchi, J., Takahashi, M.,
Tsuboi, Y., Fujino, Y., Nakajima, M., Yamamoto, T., Akatsu, H.,
et al. (2000). Localization of a novel type trypsin-like serine
protease, neurosin, in brain tissues of Alzheimer’s disease and
Parkinson’s disease. Psychiatry Clin. Neurosci. 54, 419–426.
Santagada, V., Caliendo, G., Severino, B., Perissutti, E., Fiorino,
F., Cicala, C., De Filippis, V., and Cirino, G. (2002). Minimal
structural requirements for agonist activity of PAR-2 activating
peptides. Bioorg. Med. Chem. Lett. 12, 21–24.
Scarisbrick, I.A., Linbo, R., Vandell, A.G., Keegan, M., Blaber, S.I.,
Blaber, M., Sneve, D., Lucchinetti, C.F., Rodriguez, M., and Dia-
mandis, E.P. (2008). Kallikreins are associated with secondary
progressive multiple sclerosis and promote neurodegenera-
tion. Biol. Chem. 389, ꢍ39–ꢍ45.
Debela, M., Magdolen, V., Schechter, N., Valachova, M.,
Lottspeich, F., Craik, C.S., Choe, Y., Bode, W., and Goettig, P.
(
2006). Specificity profiling of seven human tissue kallikreins
reveals individual subsite preferences. J. Biol. Chem. 281,
56ꢍ8–25688.
2
Debela, M., Beaufort, N., Magdolen, V., Schechter, N.M., Craik, C.S.,
Schmitt, M., Bode, W., and Goettig, P. (2008). Structures and
specificity of the human kallikrein-related peptidases KLK 4, 5,
Shaw, J.L.V. and Diamandis, E.P. (200ꢍ). Distribution of 15 human
kallikreins in tissues and biological fluids. Clin. Chem. 53,
1423–1432.
6, and ꢍ. Biol. Chem. 389, 623–632.
Diamandis, E.P., Yousef, G.M., Petraki, C., and Soosaipillai, A.R.
Teixeira, T.S.P., Freitas, R.F., Abrahao, Jr., A., Devienne, K.F., de
Souza, L.R., Blaber, S.I., Blaber, M., Kondo, M.J., Juliano, M.A.,
Juliano, L., et al. (2011). Biological evaluation and docking
studies of natural isocoumarins as inhibitors for human kal-
likrein 5 and ꢍ. Bioorg. Med. Chem. Lett. 21, 6112–6115.
Vandell, A.G., Larson, N., Laxmikanthan, G., Panos, M., Blaber, S.I.,
Blaber, M., and Scarisbrick, I.A. (2008). Protease-activated
receptor dependent and independent signaling by kallikreins
1 and 6 in CNS neuron and astroglial cell lines. J. Neurochem.
107, 855–8ꢍ0.
Yousef, G.M. and Diamandis, E.P. (2001). The new human tissue
kallikrein gene family: structure, function, and association to
disease. Endocr. Rev. 22, 184–204.
Yousef, G.M., Chang, A., Scorilas, A., and Diamandis, E.P. (2000).
Genomic organization of the human kallikrein gene family on
chromosome 19q13.3-q13.4. Biochem. Biophys. Res. Commun.
276, 125–133.
(
2000). Human kallikrein 6 as a biomarker of Alzheimer’s
disease. Clin. Biochem. 33, 663–66ꢍ.
Goldberg D.E. (1989). Genetic Algorithms in Search, Optimization,
and Machine Learning (Reading, MA: Addison-Wesley).
Hovnanian, A. (2012). Netherton syndrome: new advances in the
clinic, disease mechanism and treatment. Exp. Rev. Dermatol.
7, 81–92.
Jameson, G.W., Roberts, D.V., Adams, R.W., Kyle, W.S.A., and
Elmore, D.T. (19ꢍ3). Determination of the operational molar-
ity of solutions of bovine α-chymotrypsin, trypsin, thrombin
and factor Xa by spectrofluorimetric titration. Biochem. J. 131,
1
0ꢍ–11ꢍ.
Leach, A.R. and Gillet, V.J. (200ꢍ). An Introduction to Chemoinfor-
matics (Dordrecht, the Netherlands: Springer).
Liang, G., Chen, X., Aldous, S., Pu, S.F., Mehdi, S., Powers, E., Xia,
T., and Wang, R. (2012a). Human kallikrein 6 inhibitors with a
Brought to you by | Brigham Young University
Authenticated
Download Date | 5/19/15 9:48 AM