Peptide-cleaving catalyst selective for peptide deformylase

Article abstract of DOI:10.1021/ja044043h

A peptide-cleaving catalyst selective for peptide deformylase (PDF) was obtained from a library containing about 15000 catalyst candidates. The catalyst cleaved the polypeptide backbone of PDF at Gln(152)-Arg(153). Docking simulations suggested multiple m

Full text of DOI:10.1021/ja044043h

Published on Web 02/02/2005
Peptide-Cleaving Catalyst Selective for Peptide Deformylase
Pil Seok Chae, Myoung-soon Kim, Chul-Seung Jeung, Seong Du Lee, Hwangseo Park,
Sangyoub Lee, and Junghun Suh*
Department of Chemistry, Seoul National UniVersity, Seoul 151-747, Korea
Received September 30, 2004; E-mail: jhsuh@snu.ac.kr
Previously, we reported1 the first artificial protease selective for
a target protein by designing synthetic homogeneous catalysts that
recognized myoglobin (Mb) and cleaved the polypeptide backbone
of Mb by hydrolysis using metal complexes2-4 as the catalytic
center. If the target protein for a peptide-cleaving catalyst is an
enzyme related to cancers or is a viral or bacterial enzyme,
destruction of the enzyme by cleavage of the polypeptide backbone
may cure the disease. Since a catalytic amount may be sufficient
for a peptide-cleaving catalytic drug, the dosage and the side effects
of the drug may be reduced substantially. In addition, harmful
proteins lacking active sites, such as prion related to mad-cow
diseases or â-amyloid related to Alzheimer’s disease, can be
destroyed by the peptide-cleaving catalysts. Although the Mb-
cleaving catalysts were the first target-selective artificial proteases,
their molecular weights (ca. 3000) were too large to use the catalysts
as drugs and to analyze the mechanism of the catalytic action. In
addition, Mb is not related to a disease. To establish drug discovery
exploiting peptide-cleaving catalysts, it is desirable to design arti-
ficial proteases possessing considerably smaller molecular weights
as well as high selectivity for proteins directly related to diseases.
Here, we report the first peptide-cleaving catalyst meeting those
criteria designed by using peptide deformylase (PDF) as the target.
PDF is involved in deformylation of the formyl-methionyl
derivative of proteins formed in the prokaryotic translational
systems, and thus, its inhibitors are searched as candidates for new
antibiotic drugs.5 The active PDF has an Fe(II) ion in the active
site, which reacts readily with oxygen. To obtain a stable variant,
the Fe(II) ion is often substituted with Zn(II), although Zn(II)-
PDF has reduced activity by 2-3 orders of magnitude. Escherichia
coli Zn(II)-PDF provided by LG Life Sciences, Ltd. was used as
the target enzyme in the present study.
two enantiomers, the total number of catalyst candidates prepared
in the present study was about 15 000. The number of compounds
to be placed in one reaction vessel was adjusted to produce nine
different combinations and, consequently, 18 catalyst candidates
in each vessel. As checked randomly by MALDI-TOF MS before
incorporating Co(III) ion into the Cyc moieties, formation of more
than 80% of the condensation products was confirmed.
A mixture of 18 catalyst candidates generated in one vessel was
screened together for the peptide-cleaving activity toward PDF. In
a typical screening condition, the concentration of PDF was 5 µM,
whereas that of each catalyst candidate was 1.5-3 µM assuming
that the overall yield for the synthetic steps was 100%. After the
solution was incubated overnight at pH 7.5 and 37 °C, whether
new protein fragments was formed by the cleavage of PDF was
examined by MALDI-TOF MS. When the formation of some
protein fragments was indicated by a batch of the catalyst
candidates, each of the Co(III)Cyc derivatives contained in the batch
was individually generated and then incubated with PDF to identify
the active compound. The only compound that manifested PDF-
cleaving activity positively was the Co(III) complex of 1 (Co(III)-
1), which was subsequently synthesized on a large scale as described
in Supporting Information.
The PDF-cleaving catalyst was searched with a library of catalyst
candidates synthesized by the Ugi reaction6 (eq 1). The catalyst
candidates are N-acylamino acid amides with various polar and
nonpolar pendants as well as the Co(III) complex of cyclen (Cyc).
The Co(III)Cyc moiety was chosen as the proteolytic center in view
of the results of the previous study.1c Cyc with three secondary
amines protected with t-boc groups was incorporated in either the
carboxyl or the amine component of the Ugi reaction. Later, the
t-boc groups were removed, and Co(III) ion was inserted to the
Cyc portion (see Supporting Information).
MALDI-TOF MS (Figure 1) of a reaction mixture obtained by
incubation of PDF (peak a; m/z ) 19 198) with Co(III)1 disclosed
that PDF was dissected to produce a new peak (peak b) with m/z
value of 17 236. Peak b may be formed by cleavage of PDF at
either Gln(152)-Arg(153) or Ala(17)-Lys(18), which produces a
protein fragment with m/z of 17 236 or 17 243, respectively.
The cleavage site was identified as Gln(152)-Arg(153) by
treating the cleavage product with carboxypeptidase A, which
releases amino acid residues from the C-terminus in a sequential
manner. When treated with carboxypeptidase A, peak b produced
new MALDI-TOF MS peaks3c with m/z reduced by 128, 256, and
385. These are consistent with the theoretical values of 128, 256,
and 384 for the C-terminal amino acid residues of Lys(150)-Gln-
(151)-Gln(152).7
The MALDI-TOF MS data indicated that the optimum pH was
7.5.8 The relative amount of proteins, such as a and b of Figure 1,
can be estimated fairly accurately by taking the MALDI-TOF MS
several times.9 On the basis of 10 different MALDI-TOF MS
measurements, it appears that about one-half of PDF was cleaved
By using various compounds listed in Supporting Information,
about 7500 combinations of the four components were employed
for the Ugi reaction. As each Ugi condensation reaction produces
9
2396
J. AM. CHEM. SOC. 2005, 127, 2396-2397
10.1021/ja044043h CCC: $30.25 © 2005 American Chemical Society

C O M M U N I C A T I O N S
expected that new catalysts with therapeutic potential can be
designed on the basis of peptide-cleaving catalysts, especially with
regard to cancer-related, viral, and bacterial proteins as well as toxic
proteins lacking active sites.
Acknowledgment. This work was supported by a grant (R01-
2004-000-10354-0) from the Basic Research Program of the Korea
Science and Engineering Foundation, and by a grant (M104KH-
010015-04K0801-01510) from the Center for Biological Modulators
of the 21st Century Frontier R&D Program, the Ministry of Science
and Technology, Korea.
Figure 1. MALDI-TOF MS of PDF (5.2 µM; peaks a and a/2) before and
72 h after incubation with Co(III)1 (1.0 µM; calculated by assuming that
only one enantiomer of 1 is active) at pH 7.5 (0.05 M Hepes) and 37 °C.
Supporting Information Available: Structures of compounds used
for construction of the library, experimental procedures for construction
of the library, synthesis of Co(III)1, details of theoretical analysis, and
a list of proteins tested with Co(III)1. This material is available free of
charge via the Internet at http://pubs.acs.org.
References
(1) (a) Jeon, J. W.; Son, S. J.; Yoo, C. E.; Hong, I. S.; Song, J. B.; Suh, J.
Org. Lett. 2002, 4, 4155-4158. (b) Suh, J. Acc. Chem. Res. 2003, 36,
562-570. (c) Jeon, J. W.; Son, S. J.; Yoo, C. E.; Hong, I. S.; Suh, J.
Bioorg. Med. Chem. 2003, 11, 2901-2910.
(2) For general discussion of hydrolytic peptide cleavage by metal complexes
either tethered or untethered to the substrates, see: (a) Sutton, P. A.;
Buckingham, D. A. Acc. Chem. Res. 1987, 20, 357-364. (b) Chin, J.;
Jubian, V.; Mrejen, K. J. Chem. Soc., Chem. Commun. 1990, 1326-1328.
(c) Suh, J. Acc. Chem. Res. 1992, 25, 273-279. (d) Zhu, L.; Qin, L.;
Parac, T. N.; Kostic, N. M J. Am. Chem. Soc. 1994, 116, 5218-5224. (e)
Hegg, E. L.; Burstyn, J. N. J. Am. Chem. Soc. 1995, 117, 7015-7016. (f)
Jang, B.-B.; Lee, K.-P.; Min, D.-H.; Suh, J. J. Am. Chem. Soc. 1998,
120, 12008-12016. (g) Kasai, M.; Ravi, R. G.; Shealy, S. J.; Grant, K.
B. Inorg. Chem. 2004, 43, 6130-6132.
(3) For site-selective hydrolytic peptide cleavage by metal complexes un-
tethered to the substrates, see: (a) Moon, S.-J.; Jeon, J. W.; Kim, H.;
Suh, M. P.; Suh, J. J. Am. Chem. Soc. 2000, 122, 7742-7749. (b) Suh,
J.; Moon, S.-J. Inorg. Chem. 2001, 40, 4890-4895. (c) Yoo, C. E.; Chae,
P. S.; Kim, J. E.; Jeong, E. J.; Suh, J. J. Am. Chem. Soc. 2003, 125,
14580-14589. (d) Milovic, N. M.; Badjic, J. D.; Kostic, N. M. J. Am.
Chem. Soc. 2004, 126, 696-697.
(4) For site-selective peptide cleavage by metal complexes in the presence
of oxidoreductive additives, see: (a) Shepartz, A.; Cuenoud, B. J. Am.
Chem. Soc. 1990, 112, 3247-3249. (b) Hoyer, D.; Cho, H.; Schultz, P.
G. J. Am. Chem. Soc. 1990, 112, 3249-3250. (c) Rana, T. M.; Meares,
C. F. Proc. Natl. Acad. Sci. U.S.A. 1991, 88, 10578-10582.
(5) Ravi Rajagopalan, P. T.; Grimme, S.; Pei, D. Biochemistry 2000, 39, 779-
790.
Figure 2. The lowest-energy conformation of the Co(III)1-PDF complex
predicted by the docking simulations. Full view, (a) arrow indicates the
cleavage site, and expanded view, (b) P148-E159 come from the helix
and M143 and D144 from the loop; the catalyst is shown in pink.
on incubation with 0.19 equiv of the catalyst for 72 h (Figure 1),
which corresponds to ko of 0.01 h-1. Here, the initial fraction of
PDF complexed with the catalyst cannot exceed 20%. The ko
observed when PDF is fully bound to the catalyst is kcat. The lower
limit of kcat is, therefore, estimated to be 0.05 h-1. For the Mb-
cleaving catalyst, kcat was 0.02 h-1 under the same conditions.1a,c
To gain insights into the mechanism of the PDF cleavage by
Co(III)1, docking experiments were performed for the complex
formed between PDF and Co(III)1.10 The docking simulations
indicated that the S isomer of Co(III)1 produces a complex with
PDF that is more stable than that with the R isomer.12 Figure 2a
shows the lowest-energy conformation of (S)-Co(III)1 in the PDF
surface thus predicted. In the complex, the catalytic head of the
Co(III)Cyc and the central acyclic chain of the catalyst interact
with the C-terminal R-helix, while the three aromatic tails make
contact with the helical and the loop structures residing above the
active site. Thus, the catalyst would not recognize other proteins
that do not have these helical and loop structures.13 An expanded
view (Figure 2b) disclosed several modes of interactions between
the catalyst and the side chains of PDF; the hydroxo ligand of the
Co(III) ion, the putative nucleophile attacking the peptide group
of Gln(152)-Arg(153), is situated in proximity of Gln(152) and
Gln(156). Each of the three phenyl rings of the catalyst forms
independent van der Waals contact with the side chain of Met-
(143), Pro(148), or Leu(149). Hydrogen bonds are formed between
Cyc N-H of Co(III)1 and the carboxylate group of Glu(159) and
between the ammonium group of the catalyst and the side chains
of both Asp(144) and Gln(151).14
(6) Domling, A.; Ugi, I. Angew. Chem., Int. Ed. 2000, 39, 3168-3210.
(7) Expected values are 71, 227, and 298 if the cleavage site is Ala(17)-
Lys(18).
(8) In the study on Mb-cleaving catalysts,1a,c kinetic data were collected by
electrophoresis, which requires a much larger amount of the protein than
the total amount (1.5 mg) of PDF used in this work. Kinetic analysis
was, therefore, mainly performed with MALDI-TOF MS data in this study.
(9) Hollemeyer, K.; Altmeyer, W.; Heinzle, E. Anal. Chem. 2002, 74, 5960-
5968.
(10) The geometry of Co(III)1 was first optimized using the JAGUAR 4.1
program. The optimized structure was then docked onto the surface of
PDF around the C-terminal R-helix including the cleavage site. The
docking simulation was first carried out with the AutoDock 3.0.5 program.
Here, the protein structure was fixed, but the program allowed torsional
flexibility of Co(III)1. The coordinates of the protein atoms were taken
from the X-ray structure reported in the literature.11 To examine the
dynamic stability of the PDF-Co(III)1 complex, we carried out 1.0 ns
molecular dynamic simulation in aqueous solution with the SANDER
module of AMBER 7. Used in this calculation was an HP GS320 SMP
machine in Supercomputing Center of Korea Institute of Science and
Technology Information. The results showed that the time evolutions of
root-mean-square deviation from the starting structure (RMSDinit) remained
<2.0 Å for all CR atoms of PDF and 0.7 Å for all heavy atoms of Co-
(III)1. Furthermore, the RMSDinit values of Co(III)1 were maintained lower
than those of PDF CR atoms during the entire course of simulation,
indicating that the movement of Co(III)1 on the protein surface is highly
restricted compared to the conformational change of the protein. See
Supporting Information for details of theoretical analysis.
The high selectivity observed between Co(III)1 and PDF can be
attributed to the multiple modes of interaction. The Co(III) center,
an effective Lewis acid catalyst2c for peptide hydrolysis, can be
located in a highly productive position in the PDF-catalyst
complex. The high effective molarity of the Co(III) center thus
achieved would lead to effective peptide cleavage.
(11) Becker, A.; Schlichting, I.; Kabsch, W.; Groche, D.; Schultz, S.; Wagner,
A. F. Nat. Struct. Biol. 1998, 5, 1053-1058.
(12) Attempts to isolate (R)-1 or (S)-1 by resolution of 1 or enantioselective
stepwise synthesis were unsuccessful.
(13) When tested with 15 other proteins listed in the Supporting Information,
Co(III)1 did not cleave the proteins.
The present study reports the first peptide-cleaving catalyst
selective for a protein related to a disease. Moreover, mechanistic
analysis of the protein cleavage was performed for the first time
owing to the moderate size (MW of 1 ) 644) of the catalyst. It is
(14) Docking simulation of Co(III)1 in the active site of PDF indicated that
the active site was too narrow to accommodate Co(III)1 properly, and
thus, the binding in the active site causes bad van der Waals contacts
between protein and catalyst atoms.
JA044043H
9
J. AM. CHEM. SOC. VOL. 127, NO. 8, 2005 2397