Selective catalysis with peptide dendrimers

Article abstract of DOI:10.1021/ja049276n

Peptide dendrimers incorporating 3,5-diaminobenzoic acid 1 as a branching unit (B) were prepared by solid-phase synthesis of ((Ac-A³) ₂B-A²)₂B-Cys-A¹-NH₂ followed by disulfide bridge formation. Twenty-one homo- and heterodimeric dendrimers were obtained by permutations of aspartate, histidine, and serine at positions A¹, A², and A³. Two dendrimers catalyzed the hydrolysis of 7-hydroxy-N-methyl-quinolinium esters (2-5), and two other dendrimers catalyzed the hydrolysis of 8-hydroxy-pyrene-1,3,6- trisulfonate esters (10-12). Enzyme-like kinetics was observed in aqueous buffer pH 6.0 with multiple turnover, substrate binding (K_M = 0.1 -0.5 mM), rate acceleration (k_cat/k_uncat > 10³), and chiral discrimination (E = 2.8 for 2-phenylpropionate ester 5). The role of individual amino acids in catalysis was investigated by amino acid exchanges, highlighting the key role of histidine as a catalytic residue, and the importance of electrostatic and hydrophobic interactions in modulating substrate binding. These experiments demonstrate for the first time selective catalysis in peptide dendrimers.

Full text of DOI:10.1021/ja049276n

Published on Web 05/28/2004
Selective Catalysis with Peptide Dendrimers
Ce´line Douat-Casassus, Tamis Darbre, and Jean-Louis Reymond*
Contribution from the Department of Chemistry & Biochemistry, UniVersity of Berne,
Freiestrasse 3, CH-3012 Berne, Switzerland
Received February 10, 2004; E-mail: jean-louis.reymond@ioc.unibe.ch
Abstract: Peptide dendrimers incorporating 3,5-diaminobenzoic acid 1 as a branching unit (B) were prepared
by solid-phase synthesis of ((Ac-A3)2B-A2)2B-Cys-A1-NH2 followed by disulfide bridge formation. Twenty-
one homo- and heterodimeric dendrimers were obtained by permutations of aspartate, histidine, and serine
at positions A1, A2, and A3. Two dendrimers catalyzed the hydrolysis of 7-hydroxy-N-methyl-quinolinium
esters (2-5), and two other dendrimers catalyzed the hydrolysis of 8-hydroxy-pyrene-1,3,6-trisulfonate
esters (10-12). Enzyme-like kinetics was observed in aqueous buffer pH 6.0 with multiple turnover, substrate
binding (KM ) 0.1-0.5 mM), rate acceleration (kcat/kuncat > 103), and chiral discrimination (E ) 2.8 for
2-phenylpropionate ester 5). The role of individual amino acids in catalysis was investigated by amino acid
exchanges, highlighting the key role of histidine as a catalytic residue, and the importance of electrostatic
and hydrophobic interactions in modulating substrate binding. These experiments demonstrate for the first
time selective catalysis in peptide dendrimers.
Introduction
been obtained in dendrimers by incorporating catalytically active
subunits such as metal complexes and cofactors either at the
Enzyme catalysis is based on the productive encounter of
substrates with amino acid side-chains and cofactors within
catalytic sites.1 The ultimate understanding of enzyme catalysis
should result in the ability to prepare artificial enzymes. This
goal has been actively pursued with catalytic antibodies,2
catalytic polymers,3 catalytic peptides and protein models,4-6
and designed metalloproteins and metallopeptides.7 Recently,
we reported a new approach to artificial enzymes by showing
that amino acids can be assembled into catalytically functional
peptide dendrimers.8 Dendrimers are ramified structures that
adopt a globular or disk-shaped structure as a consequence of
topology rather than folding. Dendrimers are being extensively
explored in several areas of chemistry and display a range of
special properties and functions.9 Catalysis, for instance, has
surface or at the core of the dendrimer. The dendrimer is used
to provide a particular microenvironment or to increase molec-
ular size and facilitate catalyst separation and recovery.10 By
contrast, our catalytic peptide dendrimer approach aims at
studying catalysis and selectivity arising from the interplay
between amino acids within the dendrimer structure. While
peptide dendrimers have been investigated as protein mimics,
antiviral and anticancer agents, vaccines and drug and gene
delivery systems,11 the construction of such catalytic peptide
dendrimers had not been studied previously.
Herein, we report a combinatorial series of 21 different
peptide dendrimers based on the Fmoc-protected 3,5-diami-
(7) (a) Xing, G.; DeRose, V. J. Curr. Opin. Chem. Biol. 2001, 5, 196-200.
(b) Rossi, P.; Felluga, F.; Tecilla, P.; Formaggio, F.; Crisma, M.; Toniolo,
C.; Scrimin, P. J. Am. Chem. Soc. 1999, 121, 6948-6949. (c) Sissi, C.;
Rossi, P.; Felluga, F.; Formaggio, F.; Palumba, M.; Tecilla, P.; Toniolo,
C.; Scrimin, P. J. Am. Chem. Soc. 2001, 123, 3169-3173. (d) Moffet, D.
A.; Certain, L. K.; Smith, A. J.; Kessel, A. J.; Beckwith, K. A.; Hecht, M.
H. J. Am. Chem. Soc. 2000, 122, 7612-7613. (e) Berkessel, A.; He´rault,
D. A. Angew. Chem., Int. Ed. 1999, 32, 102-105. (f) Di Costanzo, L.;
Wade, H.; Geremia, S.; Randaccio, L.; Pavone, V.; DeGrado, W. F.;
Lombardi, A. J. Am. Chem. Soc. 2001, 123, 12749-12757. (g) Benson,
D. E.; Haddy, A. E.; Hellinga, H. W. Biochemistry 2002, 41, 3262-3269.
(h) Looger, L. L.; Dwyer, M. W.; Smith, J. J.; Hellinga, H. W. Nature
2003, 423, 185-190. (i) Gibney, B. R.; Rabanal, F.; Skalicky, J. J.; Wand,
A. J.; Dutton, P. L. J. Am. Chem. Soc. 1998, 120, 4952-4960. (j) Gibney,
B. R.; Isogai, Y.; Rabanal, F.; Reddy, K. S.; Grosset, A. M.; Moser, C. C.;
Dutton, P. L. Biochemistry 2000, 39, 11041-11049. (k) Dai, Q.-H.;
Tommos, C.; Fuentes, E. J.; Blomberg, M. R. A.; Dutton, P. L.; Wand, A.
J. J. Am. Chem. Soc. 2002, 124, 10952-10953.
(8) (a) Esposito, A.; Delort, E.; Lagnoux, D.; Djojo, F.; Reymond, J. L. Angew.
Chem., Int. Ed. 2003, 43, 1381-1383. (b) Lagnoux, D.; Delort, E.; Douat-
Casassus C.; Esposito, A.; Reymond, J. L. Chem.-Eur. J. 2004, 10, 1215-
1226.
(9) (a) Smith, D. K. Tetrahedron 2003, 59, 3797-3798. (b) Hecht, S.; Fre´chet,
J. M. J. Angew. Chem. 2001, 113, 76-94; Angew. Chem., Int. Ed. 2001,
40, 74-91. (c) Grayson, S. M.; Fre´chet, J. M. J. Chem. ReV. 2001, 101,
3819-3868. (d) Zeng, F.; Zimmerman, S. C. Chem. ReV. 1997, 97, 1681-
1712. (e) Newkome, G. R.; Moorefield, C. N.; Vo¨gtle, F. Dendritic
Molecules: Concepts, Synthesis, PerspectiVes; VCH: Weinheim, 1996.
(1) (a) Stryer, L. Biochemistry; WH Freeman and Co.: New York, 1999. (b)
Fersht, A. Structure and Mechanism in Protein Science; WH Freeman and
Co.: New York, 1999. (c) Kirby, A. J. Angew. Chem., Int. Ed Engl. 1996,
35, 707-724.
(2) (a) Lerner, R. A.; Benkovic, S. J.; Schultz, P. G. Science 1991, 252, 659-
667. (b) Schultz, P. G.; Lerner, R. A. Acc. Chem. Res. 1993, 26, 391-395.
(c) Schultz, P. G.; Lerner, R. A. Science 1995, 269, 1835-1842. (d)
MacBeath, G.; Hilvert, D. Chem. Biol. 1996, 3, 433-445.
(3) (a) Wulff, G. Angew. Chem., Int. Ed. Engl. 1995, 34, 1812-1832. (b)
Hodge, P. Chem. Soc. ReV. 1997, 26, 417-424. (c) Suh, J. AdV. Supramol.
Chem. 2000, 6, 245-286. (d) Suh, J.; Sung-Ju, M. Inorg. Chem. 2001, 40,
4890-4895. (e) Menger, F. M.; Eliseev, A. V.; Migulin, V. A. J. Org.
Chem. 1995, 60, 6666-6667.
(4) (a) Johnsson, K.; Allemann, R. K.; Widmer, H.; Benner, S. A. Nature 1993,
365, 530-532. (b) Zouhair Atassi, M.; Manshouri, T. Proc. Natl. Acad.
Sci. U.S.A. 1993, 90, 8282-8286. (c) Perreault, D. M.; Anslyn, E. V.
Angew. Chem., Int. Ed. Engl. 1997, 36, 432-451.
(5) (a) Broo, K. S.; Brive, L.; Ahlberg, P.; Baltzer, L. J. Am. Chem. Soc. 1997,
119, 11362-11372. (b) Kerstin, S.; Nilsson, H.; Nilsson, J.; Baltzer, L. J.
Am. Chem. Soc. 1998, 120, 10287-10295. (c) Baltzer, L. K.; Broo, S.;
Nilsson, H.; Nilsson, J. Bioorg. Med. Chem. 1999, 7, 83-91. (d) Nilsson,
J.; Baltzer, L. Chem.-Eur. J. 2000, 6, 2214-2220. (e) Andersson, L. K.;
Caspersson, M.; Baltzer, L. Chem.-Eur. J. 2002, 8, 3687-3689.
(6) (a) Bolon, D. L.; Mayo, S. L. Proc. Natl. Acad. Sci. U.S.A. 2001, 98,
14274-14279. (b) Wei, Y.; Hecht, M. H. Protein Eng. Des. 2004, 17,
67-75.
9
10.1021/ja049276n CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 7817-7826
7817

A R T I C L E S
Douat-Casassus et al.
Scheme 1. Solid-Phase Synthesis of Peptide Dendrimers
nobenzoic acid 1 as a building block for the branching unit and
bearing the catalytic triad amino acids serine, histidine, and
aspartate at variable positions on the dendrimer branches. Our
peptide dendrimers are obtained by dimerization of two second-
generation peptide dendrimers by disulfide bond formation
(Scheme 1). Two dendrimers were found to catalyze the
hydrolysis of 7-hydroxy-N-methyl quinolinium esters 2-5 in
water with significant rate enhancement and chiral discrimina-
tion. Two other dendrimers in the series catalyzed the hydrolysis
of hydropyrene-trisulfonate esters 10-12. Amino acid ex-
changes establish the role of individual amino acids in catalysis.
These experiments show the first example of chemo- and
stereoselective catalysis with peptide dendrimers.
Results and Discussion
The challenge in supramolecular catalysis is to harness
productive interactions between functional groups within the
supramolecular structure to produce a system with selectivity
and catalysis properties not encountered at the level of the
building blocks. We chose to investigate the hydrolysis of
fluorogenic 7-hydroxy-N-methyl quinolinium esters 2-8 and
8-hydroxypyrene-1,3,6-trisulfonate esters 10-12 for the study
of catalytic peptide dendrimers as esterase models (Scheme 2).
Both types of esters are well-known substrates for lipases and
esterases, and their charged fluorogenic leaving groups offer
multiple opportunities to achieve molecular recognition by
hydrophobic, π-stacking, or electrostatic interactions. In addition,
the hydrolysis of these esters is weakly catalyzed by 4-methyl
imidazole in aqueous medium. We reasoned that peptide
dendrimers incorporating histidine as one of the building blocks
might therefore display catalytic properties for the hydrolysis
of these substrates, which might be combined with molecular
recognition through interactions with the dendrimer backbone.
We would also use serine and aspartate as building blocks to
probe the possibility of catalytic-triad-type catalysis.
(10) (a) Tomalia, D. A.; Dvornic, P. R. Nature 1994, 372, 617-618. (b) Knapen,
J. W. J.; van der Made, A. W.; de Wilde, J. C.; van Leeuwen, P. W. N.
M.; Wijkens, P.; Grove, D. M.; van Koten, G. Nature 1994, 372, 659-
662. (c) Balogh, L.; Tomalia, D. A. J. Am. Chem. Soc. 1998, 120, 7355-
7356. (d) Zhao, M.; Crooks, R. M. Angew. Chem. 1999, 111, 375-377;
Angew. Chem., Int. Ed. 1999, 38, 364-366. (e) Bhyrappa, B.; Vaijayan-
thimala, G. J. Am. Chem. Soc. 1999, 121, 262-263. (f) Nlate, S.; Ruiz, J.;
Sartor, V.; Navarro, R.; Blais, J. C.; Astruc, D. Chem.-Eur. J. 2000, 6,
2544-2553. (g) Weyermann, P.; Diederich, F. HelV. Chim. Acta 2002,
85, 599-617. (h) Sellner, H.; Rheiner, P. B.; Seebach, D. HelV. Chim.
Acta 2002, 85, 352-387. (i) Twyman, L. J.; King, A. S. H.; Martin, I. K.
Chem. Soc. ReV. 2002, 31, 69-82. (j) Astruc, D.; Chardac, F. Chem. ReV.
2001, 101, 2991-3023. (k) Diederich, F.; Felber, B. Proc. Natl. Acad. Sci.
U.S.A. 2002, 99, 4778-4781.
(11) Peptide dendrimers: (a) Cloninger, M. J. Curr. Opin. Chem. Biol. 2002,
6, 742-748. (b) Crespo, L.; Sanclimens, G.; Montaner, B.; Pe´rez-Toma´s,
R.; Royo, M.; Pons, M.; Albericio, F.; Giralt, E. J. Am. Chem. Soc. 2002,
124, 8876-8863. (c) Sadler, K.; Tam, J. P. ReV. Mol. Biotechnol. 2002,
90, 195-229. (d) Tam, J. P.; Lu, Y.-A.; Yang, J.-L. Eur. J. Biochem. 2002,
269, 923-932. (e) Higashi, N.; Koga, T.; Niwa, M. ChemBioChem 2002,
3, 448-454. (f) Boas, U.; Sontjens, S. H. M.; Jensen, K. J.; Christensen,
J. B.; Meijer, E. W. ChemBioChem 2002, 3, 433-439. (g) Kinberger, G.
A.; Welbo, C.; Goodman, M. J. Am. Chem. Soc. 2002, 124, 15162-15163.
(h) Muhanna, A. M. A.; Ortiz-Salmeron, E.; Garcia-Fuentes, L.; Gimenez-
Martinez, J. J.; Vargas-Berenguel, A. Tetrahedron Lett. 2003, 44, 6125-
6128. (i) Tung, C.-H.; Mueller, S.; Weissleder, R. Bioorg. Med. Chem.
2002, 10, 3609-3614. (j) Wimmer, N.; Marano, R. J.; Kearns, P. S.;
Rakoczy, E. P.; Toth, I. Bioorg. Med. Chem. 2002, 12, 2635-2637.
9
7818 J. AM. CHEM. SOC. VOL. 126, NO. 25, 2004

Selective Catalysis with Peptide Dendrimers
A R T I C L E S
Scheme 2. Peptide Dendrimer-Catalyzed Ester Hydrolysis
Reactions
Figure 1. Bis-(Fmoc)-protected diamino acid branching units for peptide
dendrimers.
zotriazol-1-yloxy)-tris-(dimethylamino)-phosphonium-hexafluo-
rophosphate)) as the in situ coupling reagent.13 However,
acylation of the solid-phase bound 3,5-diaminobenzoyl group
only gave a mono-acylated product under these conditions,
preventing dendrimer growth. The second acylation most likely
did not take place due to the inductive deactivation induced by
the first acylation. The problem had been circumvented in a
previous report of dendrimer synthesis by using a prefunction-
alized 3,5-diaminobenzoic acid bis-phenylalanine amide as the
dendron.14 However, this strategy was not practical for the
planned synthesis of diverse dendrimer structures. We therefore
decided to look for a viable coupling protocol that would avoid
the preparation of prefunctionalized units and yield the bis-
acylated diaminobenzoic branching as a solid-supported step
in the sequence. The use of mixed anhydrides,15 strong coupling
In a previous approach to this problem, we developed the
synthesis of peptide dendrimers based on the small aliphatic
diamino acids 15-17 (Figure 1) as branching units and the
dimeric architecture shown in Scheme 1.8 The dendrimers were
functionalized with serine, histidine, and aspartate as variable
amino acids at positions A1, A2, and A3. These peptide
dendrimers indeed showed catalytic properties; however, cata-
lytic activity was limited to the fluorogenic quinolinium ester
substrates 2-5 and only occurred when the amino acid histidine
occupied the outermost position at the dendrimer surface.
Furthermore, no chiral discrimination was observed with chiral
substrates. The dendrimers derived from the aliphatic branching
units 15-17 probably adopted a compact structure barring
access of the substrate to the dendrimer core. To circumvent
these difficulties, we set out to investigate catalytic peptide
dendrimers based on a different branching unit that might result
in more open structures allowing substrate access to the
dendrimer core. We selected the 3,5-diaminobenzoic acid (as
Fmoc-derivative 1) as the branching unit because it provided a
rigid, achiral, and symmetrical structure. Simple modeling
considerations indicated that the resulting peptide dendrimers
might adopt a relatively open structure, allowing for interaction
of inner-shell residues with the substrates.
19
reagents (pyBop,16 TFFH17), BTC/collidine,18 or POCl3 was
unsuccessful. Coupling proceeded with excellent yields using
the symmetrical anhydrides of N-Fmoc amino acids in dichlo-
romethane in the absence of base.20 The introduction of histidine
after the branching unit required the Fmoc-His(Boc)-OH
because the symmetrical anhydride of Fmoc-His(trt)-OH (N-
protected amino acid commonly used in SPPS) did not undergo
coupling with the branching unit.
Six monomeric dendrimers incorporating all possible per-
mutations of aspartate, histidine, and serine were prepared by
SPPS and obtained in 10-30% yields after purification on
preparative HPLC (A-F in Scheme 1). The monomeric
dendrimers were then dimerized using Aldrithiol,21,22 allowing
(12) (a) Lloyd-Williams, P.; Albericio, F.; Giralt, E. Chemical Approaches to
the Synthesis of Peptides and Proteins; CRC Press: Boca Raton, FL, New
York, 1997. (b) Chan, C.; White, P. D. Fmoc Solid-Phase Peptide
Synthesis: a practical approach; Oxford Press: New York, 2000.
(13) Carpino, L. A.; El-Faham, A. J. Am. Chem. Soc. 1995, 117, 5401-5402.
(14) (a) Prabhat, A.; Venugopal Rao, N.; Singkhonrat, J.; Alper, H.; Bourque,
S. C.; Manzer, L. E. J. Org. Chem. 2000, 65, 1881-1885. (b) Lu, S.-M.;
Alper, H. J. Am. Chem. Soc. 2003, 125, 13126-13131.
(15) (a) Orlowski, M.; Meister, A. Biochim. Biophys. Acta 1963, 73, 679-681.
(b) Pozdnev, V. F. Int. J. Pept. Protein Res. 1994, 44, 36-48. (c)
Schutkowski, M.; Mrestani-Klaus, C.; Neubert, K. Int. J. Pept. Protein Res.
1995, 45, 257-265.
Synthesis of Peptide Dendrimers. The introduction of 3,5-
diaminobenzoic acid as a linker posed a synthetic challenge due
to the low reactivity of the aromatic amines for coupling,
requiring the development of a new protocol for the solid-phase
synthesis. Dendrimer synthesis was carried out on a Rink amide
resin by the Fmoc strategy (Scheme 1).12 The coupling of the
first amino acid (A1) to the resin and the subsequent introduction
of cysteine and the Fmoc-protected 3,5-diaminobenzoic acid 1
followed the standard coupling methodology with BOP ((ben-
(16) Martinez, J.; Bali, J. P.; Rodriguez, M.; Castro, B.; Magous, R.; Laur, J.;
Lignon, M. F. J. Med. Chem. 1985, 28, 1874-1879.
(17) Carpino, L. A.; El-Faham, A. J. Am. Chem. Soc. 1995, 117, 5401-5402.
(18) Gilon, C.; Falb, E.; Yechezkel, T.; Salitra, Y. J. Pept. Res. 1999, 53, 507-
517.
(19) Tesser, G. I.; Rijkers, D. T. S.; Adams, H. P. H. M.; Hemker, H. C.
Tetrahedron 1995, 51, 11235-11250.
(20) (a) Fu, Y.; Hammer, R. P. Org Lett. 2002, 4, 237-40. (b) Jensen, K. J.;
Alsina, J.; Songster, M. F.; Vagner, J.; Albericio, F.; Barany, G. J. Am.
Chem. Soc. 1998, 120, 5441-5452.
9
J. AM. CHEM. SOC. VOL. 126, NO. 25, 2004 7819

A R T I C L E S
Douat-Casassus et al.
Figure 2. Time-plot of dendrimer-catalyzed hydrolysis of ester 3. Formation
of 7-hydroxy-N-methyl quinolinium 9 from isobutyrate 3 with catalytic
peptide dendrimer DD (b), GG (O), and background reaction (+).
Conditions: 200 µM substrate 3, 5 µM dendrimer, 26 °C, in 20 mM aqueous
Bis-Tris buffer pH 6.0. The reaction was followed using a SpectraMAX
fluorescence plate reader (λexc ) 350 nm, λem ) 505 nm). Fluorescence
was converted to product concentration using a calibration curve, which
was linear in the concentration range used.
Table 1. Hydrolysis of Quinolinium Ester 2 in the Presence of
Peptide Dendrimersa
V
net/V
A
B
C
D
E
F
uncat
A
B
C
D
E
F
0.3
1.4
0.3
2.7
1.3
1.2
0.9
3.5
0.4
0.3
0.5
0.4
0.3
1.0
0.2
1.5
1.1
0.7
0.8
1.5
0.6
7.8
1.3
1.2
1.2
1.0
2.2
a Conditions: 200 µM 2, 5 µM dendrimer, 25 °C, 20 mM aqueous Bis-
Tris pH 6.0. The ratio Vnet/Vuncat is reported for monomeric dendrimers A-F
(first column) and all combinations of dimers. Vuncat is the apparent
spontaneous rate of formation of 9 in buffer alone, and Vnet ) Vapp - Vuncat
,
where Vapp is the apparent rate of formation of 9 in the presence of
dendrimer. The reactions were run in 96-well polystyrene half area microtiter
plates and followed using a SpectraMAX fluorescence detector with λexc
) 350 nm, λem ) 505 nm. Fluorescence was converted to product
concentration using a calibration curve, which was linear in the concentration
range used.
the preparation of all 21 possible combinations of disulfide-
bridged homo- and heterodimeric dendrimers in yields of 40-
60% (AA-FF). All dendrimers were purified by semiprepar-
ative RP-HPLC and characterized by ESI MS.
Figure 3. Double reciprocal plot of kinetic data. (A) Hydrolysis of 3
catalyzed by dendrimers AB (b) and DD (4). (B) Hydrolysis of (R)-5 (2)
and (S)-5 (O) catalyzed by dendrimer DD. For conditions, see legend of
Table 2.
Dendrimer-Catalyzed Hydrolysis of Quinolinium Esters.
The esterolytic activity of the peptide dendrimers derived from
diaminobenzoic acid 1 was first investigated using 7-acetoxy-
N-methyl-quinolinium ester 2 as fluorogenic substrate (200 µM)
in aqueous buffer pH 6.0 in the presence of 2.5 mol % catalyst
(5 µM dendrimer). Among the six monomers and 21 dimers
assayed, dimers AB and DD showed strong catalytic activity,
while all of the other dendrimers showed almost no activity
(Figure 2, Table 1). Catalysis was further confirmed by the
observation of multiple turnovers. Products were further identi-
fied by HPLC analysis of the reaction mixture.
Peptide dendrimers AB and DD also catalyzed the hydrolysis
of isobutyryl ester 3, hexanoyl ester 4, and the chiral 2-phe-
nylpropionate ester 5. In contrast, there was no reaction with
pivalate 6, benzoate 7, or the isomeric 6-acetoxy-1-meth-
ylquinolinium 8. The kinetics of hydrolysis followed the
Michaelis-Menten model, with Michaelis-Menten constants
KM in the order of 0.1-0.55 mM and kcat/kuncat values in the
range of 5 × 102 to 4 × 103 (Table 2, Figure 3). Catalysis was
proportional to dendrimer concentration up to 50 µM dendrimer,
suggesting that the dendrimers did not aggregate in the buffer
at pH 6. Isobutyryl ester 3 was the best substrate for both
dendrimers, with a maximum rate enhancement of kcat/kuncat of
4000 for dendrimer DD. The most active dendrimer DD also
(21) Akaji, K.; Fujino, K.; Tatsumi, T.; Kiso, Y. Tetrahedron Lett. 1992, 33,
1073-1075.
(22) Ruiz-Gayo, M.; Royo, M.; Fernandez, I.; Albericio, F.; Giralt, E.; Pons,
M. J. Org. Chem. 1993, 58, 6319.
9
7820 J. AM. CHEM. SOC. VOL. 126, NO. 25, 2004

Selective Catalysis with Peptide Dendrimers
A R T I C L E S
Table 2. Kinetic Parameters for Dendrimer-Catalyzed Hydrolysis of N-Methyl Quinolinium Esters 2-5a
b
2
3
4
(S)-5
(R)-5
E
4-MeIm
k2 (mM-1 min -1
)
6.7 × 10-3
4.3 × 10-3
5.6 × 10-3
5.0 × 10-3
4.8 × 10-3
Kuncat (min-1
KM (mM)
)
7.7 × 10-4
1.9 × 10-4
3.2 × 10-4
2.9 × 10-4
2.9 × 10-4
DD
(DHS)2
0.13
0.41
0.25
0.77
0.14
0.45
0.09
0.50
0.23
0.46
2.8
1.0
1.2
kcat (min-1
)
k
cat/kuncat
540
470
4000
720
1400
570
1740
1100
1600
420
(kcat/KM)/k2
AB
SDH-SHD
K
k
k
M (mM)
0.17
0.26
340
230
0.54
0.61
3200
260
0.12
0.27
840
400
0.10
0.33
1150
660
0.07
0.22
760
660
cat (min-1
)
cat/kuncat
(kcat/KM)/k2
M (mM)
kcat (min-1
cat/kuncat
(kcat/KM)/k2
GG
(DHA)2
K
0.17
0.23
300
200
0.44
0.83
4300
440
0.30
0.72
2200
430
0.08
0.21
740
530
0.11
0.24
830
450
)
k
a Conditions: 50-800 µM ester substrate, 5 µM dendrimer in 20 mM aqueous Bis-Tris pH 6.0, 25 °C. The kinetic constants given are derived from the
linear double reciprocal plots of 1/Vnet versus 1/[S] (Figure 3). b E ) (kcat/KM ((S)-5))/(kcat/KM ((R)-5)).
Scheme 3. Solid-Phase Synthesis of Peptide Dendrimers
exhibited significant chiral discrimination between both enan-
Incorporating Alanine Residues
tiomers of the 2-phenylpropionate ester (S)-5 and (R)-5, with
an enantiomeric ratio E ) 2.8 in favor of the (S)-enantiomer.
Interestingly, the catalytic rate constants of both enantiomers
were similar, and chiral discrimination was caused mainly by
the lower KM value for (S)-5 (KM ) 0.090 mM) as compared
to (R)-5 (KM ) 0.50 mM).
The catalytic mechanism involved in the dendrimer-catalyzed
hydrolysis reaction was investigated next. Because 4-methyl-
imidazole weakly catalyzes the reaction, a conservative mecha-
nistic hypothesis would attribute dendrimer catalysis to the
histidine side-chains alone, with the imidazole group acting
either as a nucleophilic or as a general-base catalyst and the
remainder of the dendrimer structure assisting catalysis by
molecular recognition of the substrate. Along these lines,
comparison of the specificity constant kcat/KM for dendrimer
catalysis with the second-order rate constant for the reaction
with 4-methyl-imidazole k2 shows that imidazole catalysis is
increased by approximately 700-fold within dendrimer DD. The
rate acceleration is partly due to molecular recognition (KM),
but might also indicate a cooperative effect between two or more
histidines, one acting as a nucleophile or general base and
another, in protonated form, stabilizing an oxyanion intermedi-
ate. This mechanism has been proposed by Baltzer et al. to
rationalize a similar enhancement of histidine reactivity in a
helical peptide bearing six histidine residues as catalytic side-
chains.5 Alternatively, the large reactivity increase observed
could indicate assistance of histidine catalysis by the other two
amino acids aspartate and serine, possibly by covalent catalysis
in the sense of an enzyme-like catalytic triad.
A series of amino acid exchange experiments were conducted
to test these mechanistic hypotheses for the more active
homodimeric dendrimer DD. An alanine scan study was carried
out by sequential alanine replacement of each of the four
histidine side-chains at position A2. Dendrimers with a serine-
to-alanine exchange at position A1 were also prepared. The
synthesis of dendrimers incorporating single His-Ala exchanges
in position A2 was realized by exploiting the chemoselectivity
observed during synthesis in the acylation of the diaminobenzoic
acid branching unit 1 (Scheme 3). Thus, the branching unit
following A1 and cysteine was first coupled under standard
conditions with the amino acid histidine, which resulted in a
monofunctionalized branch. Coupling with the symmetrical
anhydride of alanine resulted in coupling of alanine to the second
amino-group. Further elaboration then led to the complete
monomeric unit bearing one histidine and one alanine at position
A2. It should be noted that this synthetic strategy was free of
any stereochemical complexity because no chiral center was
created by the monofunctionalization of the diaminobenzoic acid
branching unit, this in contrast to the situation which would
9
J. AM. CHEM. SOC. VOL. 126, NO. 25, 2004 7821

A R T I C L E S
Douat-Casassus et al.
Table 3. Hydrolysis of Pyrenesulfonate Ester 10 in the Presence
of Peptide Dendrimersa
V
net/V
A
B
C
D
E
F
uncat
A
B
C
D
E
F
0.3
0.0
-0.1
0.4
1.9
1.6
0.2
0.2
0.1
0.1
0.3
0.2
0.4
0.6
0.1
1. 4
3.6
2.0
0.3
0.6
0.4
2.5
0.3
0.4
4.5
1.4
6
a Conditions: 200 µM 10, 5 µM dendrimer, 25 °C, 20 mM aqueous
Bis-Tris pH 6.0. The ratio Vnet/Vuncat is reported for monomeric dendrimers
A-F (first column) and all combinations of dimers. Vuncat is the apparent
spontaneous rate of hydrolysis of 10 in buffer alone, and Vnet ) Vapp
-
V
uncat, where Vapp is the apparent rate of hydrolysis of 10 in the presence of
dendrimer. The reactions were run in 96-well polystyrene half area microtiter
plates and followed using a SpectraMAX fluorescence detector with λexc
) 460 nm, λem ) 530 nm. Fluorescence was converted to product
concentration using a calibration curve, which was linear in the concentration
range used.
Figure 4. Dendrimer-catalyzed hydrolysis of quinolinium isobutyryl ester
3 in Ala-scan series. The apparent hydrolysis rate of 3 over background
(Vnet/Vuncat) is reported for each dendrimer. Conditions: 200 µM 3, 5 µM
dendrimer, 25 °C, 20 mM aqueous Bis-Tris pH 6.0. Vuncat is the hydrolysis
rate of 3 in buffer alone, and Vnet ) Vapp - Vuncat, where Vapp is the apparent
hydrolysis rate of 3 in the presence of dendrimer. See also legend of Table
1.
site that involves a cooperative effect between different amino
acid residues.
A direct role of the serine residue as a possible nucleophile
was ruled out from the data with the Ser-Ala exchange. Indeed,
the Ser-Ala dendrimer (DHA)2 (GG) showed slightly better
catalytic activity than DD with quinolinium esters 3 and 4,
whereby both dendrimers showed overall very comparable
catalytic efficiencies (Table 2). The absence of a direct role for
serine was further supported by the absence of burst-kinetics
in the initial phase of the reaction. In addition, the catalytic
dendrimers DD and GG appeared unchanged upon HPLC-
analysis of the reaction mixture, ruling out the existence of a
long-lived covalent intermediate. The much lower activity of
DG with only one Ser-Ala replacement was surprising and
suggested that additional effects were at play in catalysis.
Given the nonessential role of serine at position A1, the
absence of any catalytic effect in dendrimer BB bearing four
catalytic residues at position A2 like dendrimer DD, but having
aspartate at A1 and serine at A3, suggests that the aspartate
residues at position A3 also play a decisive role. These anionic
residues might be necessary for molecular recognition of the
cationic quinolinium ester substrates 2-5 by dendrimer DD.
Such an effect, however, does not explain the activity of peptide
dendrimer AB, which has eight serine residues at the outer
position A3.
have arisen by an asymmetric functionalization of the aliphatic
branching units 15-17.
Modified dendrimers with one, two, three, and four histidine-
alanine replacements were obtained by preparing all possible
dimeric combinations involving the single-exchange monomeric
dendrimer K, monomer D (no alanine replacement), and the
double histidine-alanine dendrimer H. Monomer G with Ser-
Ala exchange at position A1 was homodimerized and het-
erodimerized with monomer D to give a dendrimer with one
(DHA-DHS) or two (DHA-DHA) Ser-Ala exchanges at
position A1. All resulting dimeric peptide dendrimers were
obtained in good yields and purified by semipreparative HPLC
as before.
The catalytic efficiency of each Ala-dendrimer was assessed
by measuring the apparent hydrolysis rate of the isobutyryl ester
3 at pH 6.0 (Figure 4). Dendrimer Ala-4 with a 4-fold His-
Ala exchange did not show any measurable esterolytic activity,
confirming the role of histidine as an essential catalytic residue.
Most interestingly, catalytic efficiency was already reduced to
25% by a single His-Ala exchange (Ala-1). Moreover, den-
drimers with one (Ala-3), two (Ala-2 and Ala-2′), or three
histidine residues showed a comparable catalytic efficiency. The
activity observed in the dendrimer with three histidines ex-
changed by alanine (Ala-3) might be explained by simple
nucleophilic catalysis by the single histidine residue. The
absence of reactivity increase upon addition of the second and
third histidine suggests that these additional histidines cannot
operate independently within the dendrimer. The sharp increase
in catalytic efficiency upon introduction of the fourth histidine
suggests a cooperative effect, for example, in the sense of one
histidine residue acting as a nucleophile and the other three
histidines providing electrostatic assistance in stabilizing the
oxyanion in the tetrahedral intermediate. The rate enhancement
effect might also be due to proximity-induced modulation of
pKa values by electrostatic or H-bonding interactions between
the different histidines. In any case, the data suggest that
catalytic peptide dendrimer DD possesses only a single catalytic
Dendrimer-Catalyzed Hydrolysis of Hydroxypyrene-
trisulfonate Esters. Screening of the peptide dendrimers series
using the anionic fluorogenic substrate 8-acetoxypyrene-1,3,6-
trisulfonate 10 identified the two homodimers EE and FF as
being catalytically active for this substrate (Table 3). Both
dendrimers displayed histidine as the outer position A3. An even
stronger activity was observed with the corresponding butyrate
ester 11. Catalysis followed the Michaelis-Menten model with
KM in the order of 0.1 mM and kcat/kuncat values from 400 to
900 (Table 4). The corresponding nonanoyl ester 12 was also
hydrolyzed by the dendrimers; however, the KM value was very
small (KM , 10 µM) and could not be determined precisely.
The lower KM values with longer acyl chain in the series 10-
11-12 suggest that substrate binding involves hydrophobic
interactions between the dendrimer and the aliphatic chain of
the acyl group.
The catalytic dendrimers EE and FF were between 210-fold
and 670-fold more reactive toward the pyrene-trisulfonate
9
7822 J. AM. CHEM. SOC. VOL. 126, NO. 25, 2004

Selective Catalysis with Peptide Dendrimers
A R T I C L E S
Table 4. Michaelis-Menten Parameters for the Most Active Dendrimers on Pyrene Substrates 10-12a
10
11
12b
4-MeIm
kuncat (min-1
)
8.5 × 10-5
4.4 × 10-5
4.9 × 10-5
k2 (mM-1 min-1
)
1.1 × 10-3
8.1 × 10-4
8.2 × 10-4
EE
(HDS)2
KM (mM)
0.20
0.036
410
170
0.23
0.039
880
210
<0.01
0.011
kcat (min-1
)
)
)
kcat/kuncat
kcat/KM/k2
FF
(HSD)2
K
M (mM)
0.10
0.055
620
520
0.07
0.038
853
670
<0.01
0.0098
kcat (min-1
kcat/kuncat
kcat/KM/k2
JJ
(HAA) 2
K
M (mM)
kcat (min-1
cat/kuncat
cat/KM/k2
0.07
0.08
900
1100
0.03
0.037
840
1500
<0.01
0.0096
K
k
a Conditions: 50-800 µM ester substrate 10-12, 5 µM dendrimer in 20 mM aqueous Bis-Tris pH 6.0, 25 °C. The kinetic constants given are derived
from the linear double reciprocal plots of 1/Vnet versus 1/[S]. See also legend of Table 3. b Only Vmax was observed with this substrate.
substrate 11 as compared to free 4-methyl imidazole, as judged
by the ratio of the specificity constant for dendrimer catalysis
kcat/KM to the second-order rate constant of 4-imidazole catalysis
k2. The presence of histidine residues at the surface position A3
of both catalytic dendrimers suggested that catalysis was largely
due to the imidazole side-chains. The protonated histidine side-
chains probably assist substrate binding by electrostatic binding
to the sulfonate groups, while the free base form acts as a general
base or nucleophilic catalyst for hydrolysis. Hydrolysis of these
pyrene-sulfonate esters is also catalyzed by self-assembled pores
similarly bearing multiple histidine side-chains.23
The role of histidine residues in catalysis was confirmed by
exchanging all other residues to alanine to give a similarly
catalytically active dendrimer. In fact, the (HAA)2 dendrimer
JJ catalyzed the reaction even more efficiently than the parent
dendrimers EE or FF in terms of both rate acceleration kcat/
kuncat and specificity constant kcat/KM (Table 4). Interestingly,
substrate binding as measured by KM decreased in the series
EE-FF-JJ. The trend is consistent with significant interactions
between the substrates and the dendrimer core (residues A1 and
A2) in terms of either electrostatic substrate repulsion (lower
KM with decreasing number of negatively charged aspartates)
or hydrobic substrate binding (lower KM with increasing number
of hydrophobic residues).
The electrostatic component in binding was supported by the
observation of competitive inhibition of catalysis by the
quadruply negatively charged 1,3,6,8-pyrene-tetrasulfonic acid
14 (Scheme 2). This compound inhibited dendrimer catalysis
with KI ) 130 µM for EE and KI ) 30 µM for FF. These
inhibition constants are too large to be compatible with a
transition state analogue-type inhibition. Indeed, the calculated
transition state dissociation constants of dendrimers EE, FF,
and JJ for the pyrene-trisulfonate hydrolysis are in the range
of KTS ) KM/(kcat/kuncat) ) 0.08-0.25 µM. Furthermore, 1,3,6,8-
pyrene-tetrasulfonic acid 14 did not show any inhibition for the
dendrimer JJ, suggesting that hydrophobicity-driven substrate
binding was more important than electrostatic interactions in
this case.
literature. A four-helix bundle esterase catalyst with six histidine
residues described by Baltzer et al. was reported to hydrolyze
mono-p-nitrophenyl fumarate (KM ) 1 mM), a substrate similar
in reactivity to the ester substrates above, with a catalytic
proficiency of kcat/KM/k2 ) 230 relative to 4-methyl imidazole
at pH 5.1.5b This four-helix bundle catalyst exhibited a chiral
discrimination of E ) 2 for the corresponding D- and L-
norleucine nitrophenyl ester, which is similar to the selectivity
observed with dendrimer DD and the chiral esters (R)-5 and
(S)-5. Another four-helix bundle selected for folding from a
library by Hecht et al. and bearing 12 histidine and 8 surface
lysine residues was recently reported to catalyze the hydrolysis
of nitrophenyl acetate with a catalytic proficiency of kcat/KM/k2
) 100 over 4-methyl imidazole at pH 7.0 and KM ) 3 mM,
corresponding to a rate acceleration of kcat/kuncat ) 8700 above
background at that pH.6b The high KM values of the nitrophenyl
ester substrates observed with these four-helix bundles were
attributed to the fact that substrate binding was not designed
specifically.6b As discussed above, substrate binding by our
catalytic peptide dendrimers (0.03 mM < KM < 0.54 mM) is
reinforced by electrostatic complementarity between the charged
leaving group and the amino acid side-chains. An esterase active
site computationally engineered to bind and catalyze the
hydrolysis of nitrophenyl acetate at the surface of thioredoxin
with a single histidine residue exhibited catalysis at pH 7.0 (kcat/
KM/k2 ) 25, kcat/kuncat ) 180) with a tighter substrate binding
(KM ) 0.17 mM).6a Catalysis and chiral discrimination with
relatively tight and specific substrate binding have been amply
demonstrated by various catalytic antibodies raised against
transition state analogues.2 These antibodies operate with rate
accelerations in the range kcat/kuncat ) 102-104 and 0.006 mM
< KM < 0.2 mM. In contrast to artificial systems exhibiting
nucleophilic histidine catalysis, catalytic antibodies may also
display other reaction mechanisms involving direct water attack
or formation of an acyl-antibody intermediate at a catalytic
serine side-chain.
Experimental Section
Comparison with Other Esterase Models. The catalytic
peptide dendrimers above can be compared in efficiency and
selectivity to other esterase models recently reported in the
Materials and Methods. All reagents were either purchased from
Aldrich or Fluka Chemica (Switzerland) or synthesized following
literature procedures. Amino acids and their derivatives were purchased
from Senn Chemicals or Novabiochem (Switzerland). BOP and DIC
(23) Baumeister, B.; Sakai, N.; Matile, S. Org. Lett. 2001, 3, 4229-4232.
9
J. AM. CHEM. SOC. VOL. 126, NO. 25, 2004 7823

A R T I C L E S
Douat-Casassus et al.
were from Fluka. Rink amide resin was purchased from Novabiochem
(Switzerland). All solvents used were analytical grade. Peptide syntheses
were performed manually in a glass reactor. Analytical RP-HPLC was
performed in a Waters (996 Photodiode array detector) chromatography
system using a chromolith performance RP-18e, 4.6 × 100 mm, flow
rate 3 mL min-1 column. Compounds were detected by UV absorption
at 214 and 234 nm. Preparative RP-HPLC was performed with HPLC-
grade acetonitrile and MilliQ deionized water in a Waters prepak
cartridge 500 g (RP-C18 20 mm, 300 Å pore size) installed on a Waters
Prep LC4000 system from Millipore (flow rate 100 mL min-1, gradient
TFA Cleavage. The cleavage was carried out using a TFA
(trifluoroacetic acid)/EDT (1,2-ethanedithiol)/H2O/TIS (triisopropylsi-
lane) 94.5/2.5/2.5/1 solution for 4 h. The peptide was precipitated with
methyl tert-butyl ether and then dissolved in a water/acetonitrile
mixture. All of the dendrimers were purified by preparative HPLC.
General Procedure of Homodimerization. The monomeric den-
drimer X (1 mg) was dissolved in water (25 µL), and a methanolic
solution of Aldrithiol was added (45.4 mM, 0.45 equiv). The pH of
the solution was adjusted to 8 with a (NH4)HCO3 buffer solution (20
mM). The solution was stirred for 30 min and then acidified with one
drop of TFA. The homodimer XX was purified by semipreparative
RP-HPLC.
1% min-1 CH3CN). Semipreparative RP-HPLC (flow rate, 4 mL min-1
;
eluant A, water and 0.1% TFA; eluant B, acetonitrile, water, and TFA
(3/2/0.1%)); column, Vydac 218 TP (1.0 cm×25 cm, 300 Å pore size).
Compounds were detected by UV absorption at 220 nm. Chromatog-
raphy (flash) was performed with Merck silica gel 60 (0.040 ( 0.063
mm). TLC was performed with fluorescent F254 glass plates. Fluo-
rescence measurements were carried out with a spectraMAX fluores-
cence detector. NMR spectra were recorded on a Bruker-AC-300 (1H,
300 MHz; 13C, 75 MHz); MS was provided by Dr. Thomas Schnee-
berger (University of Bern, Switzerland).
General Procedure of Heterodimerization. To a solution of
dendrimer X (1 mg/25 µL) in water was added a methanolic solution
of Aldrithiol (45.4 mM, 5 equiv). The thiol-activation was followed
by HPLC, and, after the evaporation of methanol, the Aldrithiol excess
was removed by extraction (3×) with DCM. To this activated dendrimer
solution was added a solution of dendrimer Y (1 mg/25 µL, 1 equiv)
in water, and the pH of the solution was adjusted to 8 with a (NH4)-
HCO3 buffer solution (20 mM). The solution was stirred during 30
min and then acidified with one drop of TFA. The heterodimer XY
was purified by semipreparative RP-HPLC.
3,5-Bis-(9H-fluoren-9-yloxycarbonylamino)benzoic Acid (1). 3,5-
Diaminobenzoic acid (0.5 g, 3.28 mmol) was dissolved in a 10%
solution of NaHCO3 (25 mL, 23.6 mmol). Dioxane (7.5 mL) was added,
and the mixture was stirred in an ice-water bath. 9-Fluoromethyl
chlorocarbonate (1.87 g, 7.23 mmol) dissolved in dioxane (25 mL) was
added dropwise, and stirring was continued at 0 °C for 1 h and then at
room temperature for 8 h. The reaction mixture was then poured into
water (100 mL) and extracted (2×) with diethyl ether. The aqueous
solution was cooled in an ice-water bath and acidified with 3 N HCl
solution until pH 2. The aqueous phase was extracted (3×) with ethyl
acetate. The organic layers were combined, washed (1×) with a
saturated solution of NaCl, dried under sodium sulfate, and concentrated
under vacuum. The product was obtained as a white powder by
precipitation of the crude material in a mixture of diethyl ether/hexane
with a yield of 72%. mp: 244-246 °C. 1H NMR (300 MHz, DMSO-
d6) δ (ppm): 4.33 (t, J ) 7.0 Hz, 2H), 4.47 (d, J ) 7.0 Hz, 4H), 7.3-
7.5 (m, 8H), 7.73-7.75 (m, 6H), 7.92 (d, 4H), 8.05 (s, 1H). 13C NMR
(75 MHz, DMSO-d6) δ (ppm): 51.76, 70.90, 118.60, 125.40, 130.42,
132.34, 132.93, 144.94, 145.98, 148.93, 158.56. MS (+TOF): 597.2157.
General Procedure. Coupling of the First Amino Acids on Rink
Amide Resin. The Rink amide resin (0.61 mmol/g) was acylated with
3 equiv of N-Fmoc amino acid or bis-Fmoc diaminobenzoic acid in
the presence of 3 equiv of BOP (benzotriazol-1-yl-oxy-tris-(dimethyl-
amino)-phosphoniumhexafluorophosphate and 5 equiv of DIEA (N,N′-
diisopropylethylamine). After 45 min, the resin was washed (2× each)
with DMF, MeOH, and DMC (dichloromethane) and then controlled
with the TNBS (trinitrobenzenesulfonic acid) test.
Yields and characterizations of monomers and the corresponding
active dimer dendrimers are described below. Yields, purification
methods, and characterizations of all dendrimers are found in the
Supporting Information. When not indicated, conditions for the
analytical RP-HPLC are 100% A to 50% A in 10 min.
Dendrimer A. Starting with 100 mg of Rink amide resin (0.61
mmol/g), the sequence ((Ac-Ser)2-B-Asp)2-B-Cys-His-NH2 was
obtained as a colorless foamy solid after preparative HPLC purification
(8.7 mg, 10%). RP-HPLC: tR ) 5.95 MS (ES+): calcd for
C58H71N17O23S, 1405.75; found, 1405.46.
Dendrimer B. Starting with 100 mg of Rink amide resin (0.61 mmol/
g), the sequence ((Ac-Ser)2-B-His)2-B-Cys-Asp-NH2 was ob-
tained as a colorless foamy solid after preparative HPLC purification
(10.7 mg, 12%). RP-HPLC: tR ) 6.09 MS (ES+): calcd for
C60H72N18O22S, 1427.88; found, 1428.48.
Dendrimer C. Starting with 100 mg of Rink amide resin (0.61
mmol/g), the sequence ((Ac-Asp)2-B-Ser)2-B-Cys-His-NH2 was
obtained as a colorless foamy solid after preparative HPLC purification
(17.0 mg, 19%). RP-HPLC: tR ) 5.84 MS (ES+): calcd for
C60H71N17O25S, 1461.45; found, 1461.63.
Dendrimer D. Starting with 150 mg of Rink amide resin (0.61
mmol/g), the sequence ((Ac-Asp)2-B-His)2-B-Cys-Ser-NH2 was
obtained as a colorless foamy solid after preparative HPLC purification
(38.2 mg, 29%). RP-HPLC: tR ) 6.17 MS (ES+): calcd for
C63H73N19O24S, 1511.48; found, 1512.00.
Fmoc Removal. The Fmoc protecting group was removed with a
solution of piperidine in DMF (1:4) for 3 min. After filtration, the
procedure was repeated for 7 min and then washed (2× each) with
DMF, MeOH, and DCM.
Dendrimer E. Starting with 100 mg of Rink amide resin (0.61 mmol/
g), the sequence ((Ac-His)2-B-Asp)2-B-Cys-Ser-NH2 was ob-
tained as a colorless foamy solid after preparative HPLC purification
(10.0 mg, 11%). RP-HPLC: tR ) 5.87 MS (ES+): calcd for
C60H71N17O25S, 1555.54; found, 1555.88.
Incorporation of the N-Fmoc Amino Acid on the Free Amines
of the Branching Unit. To a solution of N-Fmoc amino acid (12 equiv
relative to resin loading for the first generation and 24 equiv for the
second one) in dry DCM (a few drops of DMF may be required to aid
complete dissolution) under nitrogen atmosphere was added diisopro-
pylcarbodiimide (6 or 12 equiv relative to resin loading). The reaction
mixture was stirred for 20 min at 0 °C. At the same time, the resin
was suspended in dry DCM. After filtration of the resin, the symmetrical
anhydride solution was added with a few drops of DMF to the resin,
and the resin was shaken 4 h for the first generation and overnight for
the second one. The resin was then filtered and washed as before. The
coupling efficiency was checked with the chloranil test.
Dendrimer F. Starting with 200 mg of Rink amide resin (0.61 mmol/
g), the sequence ((Ac-His)2-B-Ser)2-B-Cys-Asp-NH2 was ob-
tained as a colorless foamy solid after preparative HPLC purification
(30.0 mg, 16%). RP-HPLC: tR ) 5.76 MS (ES+): calcd for
C66H77N23O19S, 1527.55; found, 1528.13.
Dendrimer G. Starting with 150 mg of Rink amide resin (0.61
mmol/g), the sequence ((Ac-Asp)2-B-His)2-B-Cys-Ala-NH2 was
obtained as a colorless foamy solid after preparative HPLC purification
(19 mg, 14%). RP-HPLC: tR ) 6.23 MS (ES+): calcd for
C63H73N19O23S, 1495.48; found, 1495.63.
Capping of the N-Terminus. At the end of the synthesis, the resin
was acetylated with a solution of acetic anhydride in DCM (1:1) for
30 min. The resin was then dried under vacuum and stored at -4 °C.
Dendrimer H. Starting with 100 mg of Rink amide resin (0.61
mmol/g), the sequence ((Ac-Asp)2-B-Ala)2-B-Cys-Ser-NH2 was
obtained as a colorless foamy solid after preparative HPLC purification
9
7824 J. AM. CHEM. SOC. VOL. 126, NO. 25, 2004

Selective Catalysis with Peptide Dendrimers
A R T I C L E S
(8.5 mg, 10%). RP-HPLC: tR ) 7.36 MS (ES+): calcd for
Conclusion
C57H69N15O24S, 1379.44; found, 1381.38.
Two pairs of catalytic peptide dendrimers based on catalytic
triad amino acids and the diaminobenzoic acid branching unit
1 have been identified. Dendrimers AB and DD catalyze the
hydrolysis of 7-hydroxyquinolinium esters 2-5, while den-
drimers EE and FF catalyze the hydrolysis of pyrene-trisulfonate
esters 10-12. The orthogonal reactivities of these dendrimers
are noteworthy and highlight the selectivity of catalysis. This
selectivity is made possible by molecular recognition of the
substrates, which is probably mediated by electrostatic interac-
tions between surface aspartates and the cationic quinolinium
group in the case of dendrimer DD, and between surface
histidines and the anionic trisulfonate group with dendrimers
EE and FF. In the latter case, a hydrophobic component of
substrate binding is also evident in the lower KM values observed
for substrates with longer acyl chains. On the other hand, the
observation of significant chiral discrimination in the hydrolysis
of chiral substrate 5 by dendrimer DD highlights the existence
of precise molecular interactions between substrate and den-
drimers. The observation of catalysis with the heterodimeric
dendrimers AB also suggests that additional interactions leading
to catalysis are possible within the dendrimer framework. For
dendrimer DD, the loss of 75% of activity upon exchange of a
single histidine residue (Ala-1) clearly shows that catalysis
involves cooperativity between different amino acids within the
dendrimer.
Our goal was to produce a dendritic catalyst that would be
active in aqueous medium, the most relevant medium for the
study of enzyme mimics. This stands in contrast to many
supramolecular systems based on hydrogen-bonding interactions
for both structure and molecular recognition, which are generally
investigated in organic solvent.24,25 Our design was based on
topology by providing a dendrimeric framework enforcing
spatial proximity between amino acid side-chains irrespective
of the actual nature of each amino acid and therefore allowing
for the preparation of a series of peptide dendrimers. The peptide
dendrimer framework was found to be very robust to amino
acid substitution and well-suited for aqueous catalysis. Indeed,
the peptide dendrimers were all water soluble and did not form
aggregates, as evidenced by the proportionality of catalysis to
dendrimer concentration. This is also supported by the absence
of any signals in light scattering measurements (data not shown).
Dendrimer synthesis combined divergent solid-phase synthesis
of the monomeric dendrimers followed by convergent disulfide-
bridge formation to the dimers and provided a versatile and rapid
access to dendrimers in excellent yields and purity. This factor
not only enabled the assembly of a dendrimer family from which
active members could be identified, but also the preparation of
modified dendrimers as a means to understand the role of
individual amino acids in catalysis.
Dendrimer K. Starting with 150 mg of Rink amide resin (0.61
mmol/g), the sequence ((Ac-Asp)2-B-His)-((Ac-Asp)2-B-Ala)-
B-Cys-Ala-NH2 was obtained as a colorless foamy solid after
preparative HPLC purification (12.6 mg, 9.5%). RP-HPLC: tR ) 6.68
MS (ES+): calcd for C60H71N17O24S, 1445.46; found, 1445.63.
Dendrimer J. Starting with 50 mg of Rink amide resin (0.61 mmol/
g), the sequence ((Ac-His)2-B-Ala)2-B-Cys-Ala-NH2 was ob-
tained as a colorless foamy solid after preparative HPLC purification
(5 mg, 11.3%). RP-HPLC: tR ) 6.61 MS (ES+): calcd for
C65H77N23O15S, 1451.57; found, 1452.25.
Dendrimer AB. 0.3 mg, 15%, RP-HPLC: tR ) 6.12 MS (ES+):
calcd for C119H146N36O44S2, 2832.83; found, 2831.88.
Dendrimer DD. 0.6 mg, 60%, RP-HPLC: tR ) 5.76 MS (ES+):
calcd for C136H142N38O48S2, 3020.96; found, 3022.25.
Dendrimer EE. 0.6 mg, 60%, RP-HPLC: tR ) 6.13 MS (ES+):
calcd for C134H150N46O40S2, 3109.08; found, 3109.88.
Dendrimer FF. 0.4 mg, 40%, RP-HPLC: tR ) 6.16 MS (ES+):
calcd for C134H150N46O40S2, 3053.10; found, 3054.25.
Dendrimer GG. 0.5 mg, 50%, RP-HPLC: tR ) 6.24 MS (ES+):
calcd for C126H144N38O46S2, 2988.96; found, 2989.88.
Dendrimer JJ. 0.5 mg, 50%, RP-HPLC: tR ) 7.17 MS (ES+):
calcd for C130H152N46O30S2, 2901.14; found, 2902.50.
Kinetic Measurements. The kinetic measurements were carried out
by using a SPECTRAMax fluorescence detector with preset values of
the excitation and emission wavelengths corresponding to the measured
substrate (pyrene substrates λex ) 460 nm, λem ) 530 nm, N-methyl
quinolinium substrates λex ) 350 nm, λem ) 505 nm) at 25 °C. Assays
were followed in individual wells of round-bottom polystyrene 96-
well plates (Costar). Kinetic experiments were followed for 2 h. The
dendrimers were kept at -20 °C in 1 mM stock solution in B
(acetonitrile/water: 1/1). Dendrimer stock solutions were freshly diluted
to 0.05 mM solution in 20 mM aqueous Bis-Tris pH 6.0. The Bis-Tris
buffer pH 6.0 was prepared using MilliQ deionized water, the pH being
adjusted with aqueous NaOH and aqueous HCl solutions.
Fluorescence data were converted to product concentration by means
of a calibration curve. Initial reaction rates were calculated from the
steepest part observed during the first 2000 s of each curve. In a typical
experiment, 20 µL of aqueous Bis-Tris pH 6.0 (20 mM) was first added
in a well, and then 2.5 µL of a dendrimer solution (0.05 mM in aqueous
Bis-Tris pH 6.0, final concentration in the well: 5 µM), and last 2.5
µL of substrate solution (2 mM in B, final concentration in the well:
200 µM). The rate observed under these conditions is the apparent rate
Vapp. Vuncat is the rate observed with 22.5 µL of aqueous Bis-Tris pH
6.0 (20 mM) and 2.5 µL of substrate solution (2 mM in B, final
concentration in the well: 200 µM). The observed rate enhancement
V
net/Vuncat is defined as (Vapp/Vuncat) - 1.
Michaelis-Menten parameters were obtained from the linear double
reciprocal plot of 1/Vnet versus 1/[S] measured similarly with (final
concentrations) 5 µM dendrimer (Vapp) or no dendrimer (Vuncat) and
50, 75, 100, 150, 200, 300, 400, 500, 600, 700, and 800 µM substrate,
20 mM Bis-Tris, 25 °C. Kinetic parameters such as maximum velocity
(Vmax) and the Michaelis constant (KM) were determined by the least-
squares method from a Lineweaver-Burk plot. The reaction rate with
4-methylimidazole (4-MeIm) was obtained under the same conditions
with 50, 75, 100, 150, 200, 300, 400, 500, 600, 700, and 800 µM
4-MeIm and 200 µM substrate. The second-order rate constants k2 were
calculated from linear regression of the experimentally measured pseudo
first-order rate constants k′ as a function of 4-MeIm concentrations.
The inhibition constant KI was obtained from the Dixon-plot 1/Vnet
versus [I] measured similarly with 5 µM dendrimer, 200 µM substrate,
and 0, 10, 20, 40, 60, 80, 100, 200, 400, 600, 800, and 1000 µM
inhibitor, 20 mM Bis-Tris, 25 °C. Experiments were repeated twice
and gave reproducible values with (10% error.
The use of 3,5-diaminobenzoic acid 1 as a branching unit
was thought to provide dendrimers with open structures,
allowing substrate access to the inner residues at positions A1
and A2. This hypothesis seemed to be confirmed by the
observation of catalysis and chiral discrimination with dendrimer
DD bearing catalytic histidines at position A2. In any case, this
is the first example of a catalytic peptide dendrimer capable of
(24) Rebek, J. Acc. Chem. Res. 1999, 32, 278-286.
(25) Wennemers, H.; Conza, M.; Nold, M.; Krattiger, P. Chem.-Eur. J. 2001,
7, 3342-3347.
9
J. AM. CHEM. SOC. VOL. 126, NO. 25, 2004 7825

A R T I C L E S
Douat-Casassus et al.
selectivity in terms of both substrate type and chiral discrimina-
tion. The rate enhancements obtained are up to 4000-fold, which
compares well with catalytic peptides and most catalytic
antibodies reported for similar esterolytic reactions. The straight-
forward synthetic access can be used to prepare a large variety
of such dendrimers by variation of the amino acids at positions
A1, A2, and A3, which corresponds to millions of different
dendrimers counting only proteinogenic amino acids. A sys-
tematic investigation of these dendrimers for binding and
catalysis using a combinatorial protocol should be possible and
should allow for the discovery of further catalytic peptide
dendrimers with higher rate acceleration and selectivity.
Acknowledgment. This work was supported by the Swiss
National Science Foundation, the Novartis foundation, and the
University of Berne.
Supporting Information Available: HPLC and MS data and
spectra for all monomeric and dimeric peptide dendrimers. This
material is available free of charge via the Internet at
http://pubs.acs.org.
JA049276N
9
7826 J. AM. CHEM. SOC. VOL. 126, NO. 25, 2004