An approach for differentiating isozymes. Construction of libraries containing short aromatic peptides as part of a method to design selective inhibitors against lipases

Article abstract of DOI:10.1016/j.bmcl.2012.03.048

Through synthesis and assays of peptidyl substrates, we could select substrates having peptidyl complementary against lipases. The best substrate showed 20-fold improved K_m relative to non-peptidyl substrate. Using this information, we generated selective inhibitors. Lipase activities with peptidyl substrates were represented as fingerprints. Differences in fingerprints reflect different structures near active site of lipases, could be used for generating selective inhibitors.

Full text of DOI:10.1016/j.bmcl.2012.03.048

Bioorganic & Medicinal Chemistry Letters 22 (2012) 3147–3151
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
An approach for differentiating isozymes. Construction of libraries
containing short aromatic peptides as part of a method to design selective
inhibitors against lipases
⇑
Sangeun Park, Jaehoon Yu
Department of Chemistry & Education, Seoul National University, Seoul 151-742, Republic of Korea
a r t i c l e i n f o
a b s t r a c t
Article history:
Through synthesis and assays of peptidyl substrates, we could select substrates having peptidyl comple-
mentary against lipases. The best substrate showed 20-fold improved Km relative to non-peptidyl sub-
strate. Using this information, we generated selective inhibitors. Lipase activities with peptidyl
substrates were represented as fingerprints. Differences in fingerprints reflect different structures near
active site of lipases, could be used for generating selective inhibitors.
Received 18 February 2012
Revised 12 March 2012
Accepted 13 March 2012
Available online 21 March 2012
Ó 2012 Elsevier Ltd. All rights reserved.
Keywords:
Lipase
Peptidyl substrate
Selective inhibitor
Fingerprint
Isozymes
Although different types of enzymes can be easily discriminated
by using orthogonal substrates, differentiating isozymes by these
substances is a much more difficult task because their active site
structures are nearly identical. However, this formidable problem
must be overcome in order to develop selective inhibitors of ther-
apeutic target enzymes as side effect-free drugs. A general process
needed to uncover substances that discriminate between isozymes
requires acquisition of extensive data that lead to activity finger-
prints.1 However, the fingerprint method has seldom been used
to generate inhibitors for target enzymes, probably owing to rela-
tively small differences that exist in isozyme activities.
Lipases are a highly ubiquitous family of hydrolyzing enzymes
present in a variety of living systems ranging from bacteria to
mammals.2 These enzymes have served as therapeutic targets.
For example, mammalian pancreatic lipase (EC 3.1.1.3) is a key en-
zyme involved in digestion and lipid absorption. Because inhibition
of this enzyme reduces unwanted lipid intake, lipase inhibitors,
such as tetrahydrolipstatins,3 are currently marketed for the treat-
ment of obesity. However, these drugs generally lack selectivity
and, as a result, they can inhibit a variety of liver and intestinal li-
pases,4 mono-, di-acylglycerol lipases5 and liver esterases.6 Fur-
thermore, these types of drugs suffer from several adverse effects
such as abdominal pain, flatus with discharge, and oily stools.7
Substrate binding sites of lipases are composed of deep hydro-
phobic pockets that are tailored for their corresponding hydropho-
bic substrates.8 Lipase inhibitors bind to the enzymes with their
alkyl chains extending to the bottom of the pockets. Moreover, li-
pases are serine esterases that utilize typical catalytic triads and,
consequently, they have highly conserved active site amino acid
residues.9 In addition, hydrophobic surface around the active site
regions of lipases are present to bind ‘lipophilic interface’ or to sta-
bilize the enzymes.10 This hydrophobic surface contains different,
species- and tissue-specific sequence of amino acid residues11 that
may act to selectively recognize individual substrate. Thus, the
search for inhibitors that selectively bind to specific human lipases
needs to take into account differences that exist between the
hydrophobic surfaces around the active site regions.
Recently, we described a new systematic strategy for the design
of inhibitors of hydrolyzing enzymes that relies on information
gained about regions out side of the enzyme active sites.12 To un-
cover complementary groups that interact with residues in these
regions, we constructed and evaluated carbohydrate–peptide sub-
strate libraries whose members possess various peptide moieties
tethered to carbohydrates. Then, we were able to identify comple-
mentary structures that are responsible for recognition of regions
distant from the enzyme active sites by selecting the most active
substrates against each target
a-glucosidase. Information found
in the tethered-dipeptide moieties in the substrates was used to
design and synthesize selective inhibitors against cognate
⇑
Corresponding author.
E-mail address: jhoonyu@snu.ac.kr (J. Yu).
a-glucosidase.
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2012.03.048

3148
S. Park, J. Yu / Bioorg. Med. Chem. Lett. 22 (2012) 3147–3151
Table 1
We believed that the two compartment binding character of
Amino acids in 1st generation substrate (butyrate-sp-
X1X2)
lipases, consisting of deep pocket and hydrophobic surface, would
be useful in guiding a peptide-tethered strategy based selection of
substrates. In the effort described below, we first designed and
prepared a substrate library (Fig. 1) comprised of fatty acid linked
peptides and then screened members for their activities against a
series of lipases. On the basis of information gained from the
screening, we generated three types of selective inhibitors (Fig. 1),
in consequence. They display sub- to low micromolar affinities and
high selectivities for a bacterial lipase. Even though the initial at-
tempt to produce selective substrates against mammalian lipases
was not successful, the information obtained in studies with vari-
ous aromatic peptidyl substrates did shed light on strategies to de-
velop inhibitors that can differentiate isozymes of mammalian
lipases.
The 1st generation library (Fig. 1 and Table 1) was comprised of
substrates that are designed to contain tethered aromatic amino
acid groups (Trp, Phe, Tyr) on opposite sides of the hydrophobic
acyl chain in order to foster interactions with complementary ami-
no acid side chains on the hydrophobic surface of the enzymes.
Even though triacylglycerides are typical substrates for most
lipases,13 a short chain butyl ester group was used for hydrophobic
acyl part of the initial substrate library employed in this study. By
minimizing the size of the hydrophobic acyl group, we felt it would
be possible the linked peptide–ester library,14 the nitro group in
p-nitrophenyl butyrate (Fig. 1) was reduced to produce the corre-
sponding amine. This was followed by amide formation with
succinic acid to form a chain extended carboxylic acid that was
employed in solid phase peptide synthesis reactions. Fmoc-pro-
tected amino acids were used to make 2-AA peptides that were
employed in reactions with p-(succinyl)-anilino-butyrates to pro-
duce the corresponding tethered substrates. Cleavage from the
solid support afforded the peptidyl butyl esters butyrate-sp-X1X2
(Fig. 1 and Table 1) that were purified by using HPLC to give pure
substances in 30–40% isolated yields.14
Experiments to evaluate the activity of members of the library
employed p-nitrophenylbutyrate, non-peptidyl substrate, as the
competitive substrate and lipases from Thermomyces lanuginosus
(TLL), Chromobacterium viscosum (CVL), and porcine and human
pancreases (PPL and HPL). In the experiments, equimolar mixtures
of individual members of the substrate library along with the
p-nitrophenyl butyrate were incubated with each lipase while
monitoring the formation of the chromogenic hydrolysis product
of non-peptidyl substrate, utilizing HPLC.
Several interesting results arose from these studies. Firstly, the
rate of formation of p-nitrophenol was reduced in nearly every
sample containing a substrate relative to that of the non-peptidyl
substrate alone. Secondly, the peptidyl substrates caused a greater
reduction in the formation of p-nitrophenol when the lipases TLL
and CVL ( Fig. 2a,b) were used compared to HPL, or PPL
(Fig. 2c,d). These findings suggest that different complementary
No.
X1X2
1
2
3
4
5
6
7
8
9
FF
FW
FY
WF
WW
WY
YF
YW
YY
interactions exist in the substrate binding region of each lipase
and that this phenomenon leads to differences in substrate acces-
sibilities for all four of the lipases. For example, butyrate-sp-FF
caused the largest decrease in p-nitrophenol production for the
CVL catalyzed reaction, whereas the peptidyl substrates buty-
rate-sp-FW, butyrate-sp-YF and butyrate-sp-FY had the largest
effect on catalysis by TLL, PPL, and HPL, respectively. These obser-
vations indicate that the selected peptidyl substrates contain struc-
tural and functional features that are complementary with regions
of the lipases that are remote from their alkyl binding pockets. Ow-
ing to the fact that it displayed the largest effects on p-nitrophenol
formation and the greatest selectivity of the inhibitors probed, CVL
was chosen for further studies.
Among the first generation substrate explored, butyrate-sp-FF
displayed the most profound effect against CVL (Fig. 2b). The Km
value (0.18 mM) of this substrate was observed to be ca. 20-fold
less than the value (3.5 mM) of the non-peptidyl substrate.14 This
result demonstrates that the peptidyl moiety in butyrate-sp-FF
is well-recognized by the complementary region of the cognate en-
zyme. Two possible reasons exist for the activities of butyrate-sp-
FF and the other linked peptide esters. Firstly, these substances are
competitive substrates of the lipase catalyzed reaction as antici-
pated. Secondly, in the event that the linked peptidyl esters are
substrates, peptide product generated by hydrolysis of ester could
serve as an inhibitor of the lipase. Independent of the reasons, the
results show that the complementary of lipases is recognized by
peptidyl residues in the substrates or in the corresponding
products.
Three kinds of inhibitors were then prepared based on the infor-
mation gained in the studies described above (Fig. 1).14 Firstly,
spacer-FF (sp-FF), serving as the peptide portion of butyrate-sp-
FF, was prepared and its inhibitory constant against CVL was deter-
mined. This substance displayed
a mid-micromolar range Ki
(37 M, Fig. 3a) and it did not inhibit the other lipases (Table 3),
l
giving a discrimination ratio (DR) of more than 160. Encouraged
by this finding, we designed and synthesized the second genera-
tion inhibitors butyl-sp-FF ether and heptyl-sp-FF ether (Fig. 1),
O
R1
O
NO2
H
N
H
N
O
O
N
H
NH2
R1
O
O
O
R2
O
O
p-nitrophenyl butyrate
butyrate-sp-X1X2
O
O
H
H
N
R1
O
H
N
H
N
N
N
NH2
N
H
NH2
H
O
O
R2
O
O
n
O
R2
HO
butyl-sp-X1X2 ether (n=2)
heptyl-sp-X1X2 ether (n=5)
sp-X1X2
Figure 1. Structures of substrates (first line) and inhibitors (second line). X1 and X2 represent individual amino acids, R1 and R2represent side chains of X1 and X2, respectively.
Sprepresents the ‘spacer’ between lipid and peptidyl components.

S. Park, J. Yu / Bioorg. Med. Chem. Lett. 22 (2012) 3147–3151
3149
Figure 2. Initial screening of the 1st generation substrate library (butyrate-sp-X1X2) using p-nitrophenyl butyrate (1.0-equiv) as the competitive substrate for (a) TLL, (b)
CVL, (c) PPL and (d) HPL.14
Figure 3. Dixon plots corresponding to inhibition of CVL by (a) sp-FF, (b) butyl-sp-FF ether, and (c) heptyl-sp-FF ether.
in which sp-FF is directly tethered to hydrophobic chains through
ether linkages.15 Butyl-sp-FF ether was observed to be a potent
Introduction of a longer carbon chain into sp-FF as found in
heptyl-sp-FF ether did yield sub-micromolar inhibitor
inhibitor of the CVL (Ki = 2.6
lM, Figure 3b and Table 2). Even
(Ki = 0.37 lM, Fig. 3c and Table 2) and it did enhance selectivity
though incorporating a butyl group into sp-FF improved the
binding affinity to CVL20-fold, it did not lead to a large decrease
in lipase selectivity (DR = 84).
(DR = 140). It has been reported that CVL possesses a relatively
large hydrophobic pocket.16 The data accumulated in this effort
is consistent with this early observation that the length of alkyl

3150
S. Park, J. Yu / Bioorg. Med. Chem. Lett. 22 (2012) 3147–3151
Table 2
Ki (l
M) and DR values of inhibitors against various lipasesa,b
Lipase
sp-FF
Butyl-sp-FF ether
Heptyl-sp-FF ether
CVL
TLL
PPL
HPL
DR
37
NC
2.6
NC
230
210
84
0.37
NC
39
67
140
NCa
NCb
>160
a
NC: Values could not be calculated with 5 mM of the inhibitor.
b
Discrimination Ratio (DR) = Ki,averaged
lipases/Kiagainst
lipase from
against other
Chromobacteriumviscosum.
Table 3
Figure 4. Activity fingerprints derived from analysis of 19 peptidyl substrate of
various lipases. LPL, lipoprotein lipase from bovine milk (EC 3.1.1.34); PCL, lipase
from Pseudomonas cepacia; CRL, lipase from Candida rogosa; ROL, lipase from
Rhizopusoryzae.
Amino acids in 2st generation substrate (butyrate-sp-
X1X2)
No.
X1X2
10
11
12
13
14
15
16
17
18
19
Nal1Nal1
Nal1Nal2
Nal2Nal1
Nal2Nal2
FNal2
WNal2
YNal2
Nal2F
interactions associated with the naphthyl group(s) might lessen
selectivity.
Owing to the fact that the naphthyl ring containing substances
display inhibitory activities toward mammalian lipases that are
greater than those of the natural aromatic peptide analogs used
in the initial screen, we thought that peptidyl moieties might be
generally important in governing recognition of other source of li-
pases, too. Thus, activities of 8 lipases were assayed extensively
using total 19 substrates (9 from 1st and 10 from 2nd libraries).
The potencies of the 19 peptides were used to construct finger-
prints of activities ranging from the strongest (green in Fig. 4) to
the weakest (black in Fig. 4). Analysis of the fingerprints showed
that different complementarity exists between the active site and
near active site regions that are recognized by the peptidyl compo-
nents of the substrates. Differences between pairs of lipases from
different sources were then quantitatively evaluated by calculating
and normalizing the Euclidian distances19 in the fingerprints,
where 0% (black in Fig. 5) corresponds to identical activities and
100% to the largest activity differences (red in Fig. 5). 20 The calcu-
lated distances were then compared with amino acid sequence dif-
ferences that exist in pairs of lipases.14,21 Interestingly, the results
demonstrate that a few lipase-pairs exist in which the Euclidean
distances are much larger than the sequence differences. For exam-
ple, HPL/PPL and CVL/PCL display 47% and 56% differences in their
Euclidean distances while only a 15% differenceis present in their
amino acid sequences. Therefore, the activities of lipases with sim-
ilar phylogeny appear to be more distinguished than sequence
homologies.22 This is an encouraging observation considering the
fact that sequences in near active site regions of the two lipases
must be nearly identical. Even though our initial attempts to devel-
op peptidyl substrates that are capable of differentiating between
two lipases were not successful, the large differences observed be-
tween Euclidean distances in the fingerprints strongly suggest that
the peptide components of the substrates are complimentarily rec-
ognized by the two mammalian lipases.
Nal2W
Nal2Y
a
Nal1, Nal2 are (1-naphthyl and 2-naphthyl)alanine,
respectively.
chains in inhibitors is a critical determinant. The selected inhibi-
tors were observed to have excellent selectivities and reasonably
high inhibitory effect for CVL.
An investigation of the inhibition of mammalian lipases was
carried out next. The initial screening data (Fig. 2c,d) discussed
above suggested that butyrate-sp-YF and -FY might bethe best
substrates for PPL and HPL, respectively. However, the results of
experiments revealed that, in contrast to the best substrates of
CVL, these substrates bring about only similar Km values of PPL
and HPL catalyzed hydrolysis of p-nitrophenyl butyrate.14 In addi-
tion, synthesized inhibitors, butyl-sp-FY ether and butyl-sp-YF
ether are neither strong nor selective inhibitors of the mammalian
enzymes.14 These observations indicate that recognitions of the
peptidyl moiety by mammalian lipases are both weaker and less
selective than by other lipases. The reported results, exploring
phosphonates that serve as transition state analogs of esters partic-
ipating in hydrolysis reactions, show that two hydrophobic pockets
exist inactive sites of mammalian lipases.17 Therefore, the binding
affinities of inhibitors containing only one alkyl carbon chain might
be relatively weak in contrast to those, like phosphonate analogs,
which possess two alkyl chains.18
The observations made above do demonstrate that aromatic
peptidyl groups are somehow recognized by mammalian lipases.
In order to maximize these interactions, a 2nd generation substrate
library was constructed using the unnatural amino acid building
blocks, 1-naphthyl and 2-naphthylalanines (Table 3). It was antic-
ipated that naphthyl moieties would cause stronger hydrophobic
interactions with complimentary residues of the enzyme. As ex-
pected, the naphthyl alanine containing substances are more po-
tent substrates of both mammalian lipases than are the natural
aromatic amino acid containing substrates (Fig. S1). The inhibitor,
butylate-sp-Nal2Nal2 ether, designed based on the most active
unnatural peptidyl substrate, was found to display low micromolar
In summary, the results of the investigation described above, in
which peptidyl substrates were prepared and explored for their
activities against a range of lipases, have led to the identification
of substrates that have peptidyl complementary to residues pres-
ent on the hydrophobic surfaces of lipases. Butyrate-sp-FF, the
best substrate in the series against CVL has a 20-fold lower Km of
non-peptidyl substrate. Using information gained from analysis
of the substrate, we designed three kinds of inhibitors, sp-FF,
butyl-sp-FF ether and heptyl-sp-FF ether, which were found to
be selective inhibitors of the cognate enzyme with inhibitory con-
stants in the sub- to low micromolar ranges. Moreover, lipases
from other origins were not inhibited by these inhibitors
inhibition (Ki = 9.2 l
M).14 However, all of the substrates containing
naphthyl groups had very similar activities. No differences were
observed between those having one or two, 1-naphthyl and
2-naphthyl moieties (Fig. S1). Thus, the results suggest that strong

S. Park, J. Yu / Bioorg. Med. Chem. Lett. 22 (2012) 3147–3151
3151
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2012.03.048.
References and notes
1. Goddard, J. P.; Reymond, J. L. J. Am. Chem. Soc. 2004, 126, 11116.
2. Schmid, R. D.; Verger, R. Angew. Chem., Int. Ed. 1998, 37, 1608.
3. (a) Fernandez, E.; Borgström, B. Biochim. Biophys. Acta 1989, 1031, 249; (b)
Hvizdos, K. M.; Markham, A. Drugs 1999, 58, 743; (c) Hadváry, P.; Sidlerj, W.;
Meister, W.; Vetter, W.; Wolfer, H. J. Biol. Chem. 1991, 266, 2021; (d) Lüthi-Peng,
Q.; Märki, H. P.; Hadváry, P. FEBS Lett. 1992, 299, 111.
4. (a) Sternby, B.; Hartmann, D.; Borgström, B.; Nilsson, A. Clin. Nutr. 2002, 21,
395; (b) Lookene, A.; Skottova, N.; Olivecrona, G. Eur. J. Biochem. 1994, 222, 395.
5. Ortar, G.; Bisogno, T.; Ligresti, A.; Morera, E.; Nalli, M.; Di Marzo, V. J. Med.
Chem. 2008, 51, 6970.
6. Hadváry, P.; Lengsfeld, H.; Wolfer, H. Biochem. J. 1988, 256, 357.
7. (a) Ballinger, A. Expert Opin. Pharmacother. 2000, 1, 841; (b) Filippatos, T. D.;
Derdemezis, C. S.; Gazi, I. F.; Nakou, E. S.; Mikhailidis, D. P.; Elisaf, M. S. Drug Saf.
2008, 31, 53.
8. Bezzine, S.; Carrière, F.; De Caro, J.; Verger, R.; De Caro, A. Biochemistry 1998, 37,
11846.
9. Hide, W. A.; Chan, L.; Li, W. H. J. Lipid Res. 1992, 33, 167.
10. Fendri, A.; Frikha, F.; Miled, N.; Gargouri, Y. Biochimie 2006, 88, 1401.
11. Bodmer, M. W.; Angal, S.; Yarranton, G. T.; Harris, T. J.; Lyons, A.; King, D. J.;
Pieroni, G.; Riviere, C.; Verger, R.; Lowe, P. A. Biochim. Biophys. Acta 1987, 909,
237.
12. Park, S.; Hyun, S.; Do, S. I.; Yu, J. Bioorg. Med. Chem. Lett. 2011, 21, 2441.
13. Carrière, F.; Withers-Martinez, C.; van Tilbeurgh, H.; Roussel, A.; Cambillau, C.;
Verger, R. Biochim. Biophys. Acta 1998, 1376, 417.
Figure 5. A matrix of Euclidian distances calculated by using differences of
activities against various lipases.28 Black, zero distance; (0%); red, the largest
distance (100%).
(DR = 80–160), suggesting that the peptidyl moieties are essential
for promoting selectivity. Even though initial attempts to generate
selective inhibitors against mammalian lipases were unsuccessful,
the Euclidian distances, calculated using activity fingerprints,
showed that the substrates display a high degree of differentiation
especially for lipases that have similar amino acid sequences. It is
believed that the general approach developed in this effort might
be useful for generating selective inhibitors that discriminate be-
tween isozymes in other enzyme families.
14. See Supplementary data.
15. Butyl-sp-FF ether gave reasonable solubility in buffer conditions, while
heptyl-sp-FF and decyl-sp-FF ether showed very poor solubility.
16. Gascoyne, D. G.; Finkbeiner, H. L.; Chan, K. P.; Gordon, J. L.; Stewart, K. R.;
Kazlauskas, R. J. J. Org. Chem. 2001, 66, 3041.
17. Cavalier, J. F.; Buono, G.; Verger, R. Acc. Chem. Res. 2000, 33, 579.
18. (a) Cavalier, J. F.; Ransac, S.; Verger, R.; Buono, G. Chem. Phys. Lipids 1999, 100,
3; (b) Gargouri, Y.; Ransac, S.; Verger, R. Biochim. Biophys. Acta 1997, 1344, 6.
19. Lele, S.; Richtsmeier, J. T. Am. J. Phys. Anthropol. 1991, 86, 415.
20. (a) Grognux, J.; Reymond, J. L. ChemBioChem 2004, 5, 826; (b) Usui, K.;
Tomizaki, K. Y.; Ohyama, T.; Nokihara, K.; Mihara, H. Mol. BioSyst. 2006, 2, 113.
21. a) Corpet, F. Nucl. Acids. Res. 1988, 16, 10881; http://www.ncbi.nlm.nih.gov/
protein/ and http://www.ebi.ac.uk/Tools/msa/clustalw2/.
Acknowledgments
This work was financially supported by
a Grant (No.
20110002210) from National Research Foundation of Korea. We
thank Ms. Seulah Choi and Ms. Hee-jee Youn for assistance for
the synthesis of substances.
22. This trend was not observed where AA sequence difference between lipases is
larger (>90%) and where the activities of peptidyl substrates against lipase
(such as LPL) were too small.