Combinatorial discovery of peptide dendrimer enzyme models hydrolyzing isobutyryl fluorescein

Article abstract of DOI:10.1021/co200006z

Two 6750-membered one-bead-one-compound peptide dendrimer combinatorial libraries L (X⁴)₈(LysX³)₄(LysX ²)₂LysX¹ (X^1-4 = 14 different amino acids or deletion, Lys = branching lysine residue) and AcL (with N-terminal acetylation) were prepared by split-and-mix solid phase peptide synthesis. Screening toward fluorogenic substrates for esterase and aldolase activities using the in silica off-bead assay (N. Maillard et al., J. Comb. Chem.2009, 11, 667-675) and bead decoding by amino acid analysis revealed histidine containing sequences active against fluorescein diacetate. Isobutyryl fluorescein, a related hydrophobic fluorogenic substrate, was preferentially hydrolyzed by dendrimers from library AcL containing hydrophobic residues such as AcH3 (AcHis)₈(LysLeu)₄(LysVal)₂LysLysOH, compared to simple oligohistidine peptides as reference catalysts. Polycationic dendrimers from library L with multiple free N-termini such as H8 (His) ₈(LysβAla)₄(LysThr)₂LysaProNH₂ (aPro = (2S,4S)-4-aminoproline) showed stronger reactivity toward 8-acetoxypyrene-1,3,6-trisulfonate with partial acylation of N-termini. These experiments highlight the role of noncatalytic amino acids to determine substrate selectivity in peptide dendrimer esterase models.

Full text of DOI:10.1021/co200006z

RESEARCH ARTICLE
pubs.acs.org/acscombsci
Combinatorial Discovery of Peptide Dendrimer Enzyme Models
Hydrolyzing Isobutyryl Fluorescein
Noꢀelie Maillard, Rasomoy Biswas, Tamis Darbre, and Jean-Louis Reymond*
Department of Chemistry and Biochemistry, University of Berne, Freiestrasse 3, CH-3012, Berne, Switzerland
S Supporting Information
b
ABSTRACT: Two 6750-membered one-bead-one-compound pep-
tide dendrimer combinatorial libraries L (X4)8(LysX3)4(LysX2)2LysX1
(X1ꢀ4 = 14 different amino acids or deletion, Lys = branching lysine
residue) and AcL (with N-terminal acetylation) were prepared by
split-and-mix solid phase peptide synthesis. Screening toward fluoro-
genic substrates for esterase and aldolase activities using the in silica
off-bead assay (N. Maillard et al., J. Comb. Chem. 2009, 11, 667ꢀ675)
and bead decoding by amino acid analysis revealed histidine contain-
ing sequences active against fluorescein diacetate. Isobutyryl fluor-
escein, a related hydrophobic fluorogenic substrate, was preferentially
hydrolyzed by dendrimers from library AcL containing hydrophobic
residues such as AcH3 (AcHis)8(LysLeu)4(LysVal)2LysLysOH, compared to simple oligohistidine peptides as reference catalysts.
Polycationic dendrimers from library L with multiple free N-termini such as H8 (His)8(LysβAla)4(LysThr)2LysaProNH2 (aPro =
(2S,4S)-4-aminoproline) showed stronger reactivity toward 8-acetoxypyrene-1,3,6-trisulfonate with partial acylation of N-termini.
These experiments highlight the role of noncatalytic amino acids to determine substrate selectivity in peptide dendrimer esterase
models.
KEYWORDS: combinatorial libraries, peptide dendrimer, hydrolysis, isobutyryl fluorescein, enzyme models
’ INTRODUCTION
this approach only a small fraction of the theoretically possible
peptide dendrimer sequence space was actually surveyed. Indeed
in a typical third generation peptide dendrimer library with eight
variable amino acid positions with sequence (X8X7)8(BX6X5)4-
(BX4X3)2BX2X1 (X1ꢀ8 = variable amino acid, B = branching
diamino acid), a choice of 15 amino acids implied that over 2
billion possible sequences were possible, while we used only four
different amino acids per variable position and thus prepared
65,536 sequences, corresponding to only 0.0026% of the theo-
retically possible sequence space. This left open the possibility
that the difficulty to identify esterase catalysts toward hydro-
phobic substrates was caused by incomplete screening of
sequence space.
In our efforts to realize a more thorough coverage of peptide
dendrimer sequence space, we herein report the screening of
the 6750-membered combinatorial libraries L (X4)8(LysX3)4-
(LysX2)2LysX1 (X1ꢀ4 = variable amino acid, Lys = branching
lysine residue) and AcL (with N-terminal acetylation, Figure 1)
with more extensive survey of sequence space for three of its four
variable positions.The library was screened for conversion of
various fluorogenic substrates for aldolase and esterase activities
using our recently reported off-bead catalysis assay, which allows
to assay the libraries in solution on the surface of the silica gel
Dendrimers are regularly branched tree-like synthetic macro-
molecules displaying a range of interesting properties.1 Most
dendrimers consist of predefined organic cores to which variable
end groups are attached for function.2 Recently we reported that
peptide dendrimers can be obtained by solid-phase peptide
synthesis alternating branching diamino acids and standard R-
amino acids in the sequence.3 Peptide dendrimers resemble
proteins in their composition and adopt a globular shape because
of their topology rather than by folding. In approaches similar to
those reported by others for libraries of linear peptides,4,5 our
screening of one-bead-one-compound (OBOC) combinatorial
libraries6 identified peptide dendrimers with various functions
such as esterase7 and aldolase8 catalysis and selective binding to
metal ions,9 cofactors,10 and proteins.11 In the case of enzyme
model studies for esterase activities, our experiments best
succeeded when using charged fluorogenic substrates such as
8-acetoxypyrene-1,3,6-trisulfonate 4 (Scheme 1). On the other
hand screening with hydrophobic, non charged substrates gen-
erally failed to identify significant catalysis in any of the libraries.
The peptide dendrimer combinatorial libraries screened for
esterase catalysis were based on sequence designs in which each
amino acid appeared only twice in the synthesis in two successive
branches. This design produced library sizes manageable by
SPPS, which is limited by the number of solid support beads in
the experiment, and enabled decoding by the inexpensive and
efficient amino acid analysis (AAA) of the beads. However, with
Received: January 13, 2011
Revised:
March 18, 2011
Published: March 25, 2011
r
2011 American Chemical Society
310
dx.doi.org/10.1021/co200006z ACS Comb. Sci. 2011, 13, 310–320
|

ACS Combinatorial Science
RESEARCH ARTICLE
Scheme 1. Peptide Dendrimer Catalyzed Fluorogenic Ester
Hydrolysis Reactions
plate, a method which is less prone to bead-surface artifacts than
simple on-bead activity assays.12 The screening led to the
identification of several sequences catalyzing the hydrolysis of
fluorescein diacetate 1. A comparative kinetic study with iso-
butyryl fluorescein 2 as a model hydrophobic substrate and
8-acetoxypyrene-1,3,6-trisulfonate 4 as a typical polyanionic
substrate reveals unprecedented aspects of structure-activity
relationships in peptide dendrimer enzyme models, in particular
the role of noncatalytic amino acids to determine substrate
selectivity.
’ RESULTS AND DISCUSSION
Library Design. AAA is applicable for decoding SPPS synth-
esis beads because the amount of peptide per bead is sufficient for
the sensitivity of AAA, which requires at least 90 pmols. While
our previous examples of AAA decodable libraries were limited to
cases where each amino acid is only used in two of the variable
positions during library synthesis, the method is theoretically
suitable for decoding an exhaustive combinatorial library of third
generation peptide dendrimers if each branch bears only a single
variable amino acid position, as shown in Table 1.
Figure 1. Synthesis of third generation peptide dendrimer combinatorial
libraries AcL and L. Conditions: (a) MSNT (1-(2-mesitylenesulfonyl)-
3-nitro-1H-1,2,4-triazole), FmocX1OH, 1-methylimidazole, CH2Cl2, 1 h,
(2ꢁ);(b) piperidine20%inDMF(2ꢁ 10 min); (c) FmocLys(Fmoc)OH,
PyBop, iPrNEt2, NMP; (d) split in 15 portions, then FmocXiOH, PyBop,
iPrNEt2, NMP; (e) for AcL: Ac2O/CH2Cl2 (1:1), 1 h; (f) TFA/TIS/H2O
(94:5:1), 4 h. aPro = (2S,4S)-4-aminoproline, introduced with (N-Boc-trans-
4-N-Fmoc-amino-L-proline).
Taking into account the possible difficulties in uniquely
resolving the copy number above 10 copies by AAA, we set
out to explore this exhaustive library design in a focused example
considering the variable positions X2, X3, and X4 (Figure 1). The
core position X1 would be limited to 4-aminoproline or lysine
only to selectively explore single-site enamine-type organocata-
lytic aldolase dendrimers, a type of activity previously identified
in peptides13 and peptide dendrimers.8 Fourteen amino acids
were included for the variable positions including aromatic
(phenylalanine and tyrosine), hydrophobic (leucine, proline,
and valine), small and polar (alanine, glutamine, glycine, serine,
and threonine), charged and catalytic residues (arginine, aspar-
tate, histidine). β-alanine and a deletion (which can be identified
by AAA) were also included to allow size variation in the
dendrimer. We excluded tryptophan (not detectable by AAA),
glutamate and asparagine (indistinguishable in AAA from gluta-
mine respectively aspartate), isoleucine (too similar to leucine),
lysine (already used as branching unit), and sulfur containing
amino acids.
Library Synthesis and Screening. The library was prepared
by split-and-mix Fmoc-SPPS on a 1.5 g batch of hydroxymethyl-
photolinker NovaSyn TG resin (0.24 mmol/g, bead diameter
130 μm, ∼8 ꢁ 105 bead/g, 200ꢀ300 pmol/bead). After the final
Fmoc deprotection, library L was either used directly, thus
311
dx.doi.org/10.1021/co200006z |ACS Comb. Sci. 2011, 13, 310–320

ACS Combinatorial Science
RESEARCH ARTICLE
Table 1. Amino Acid Copy Numbers from AAA and Position
Attribution in a 3rd Generation Peptide Dendrimer Library
(X4)8(LysX3)4(LysX2)2LysX1
Scheme 2. Reactions of Fluorogenic Substrates That Gave
No Hits in the off-Bead Assays with Libraries L and AcL
copy number from AAA
positions occupied
1
X1
2 X2
2
3
2 X2 þ X1
4
4 X3
5
4 X3 þ X1
6
4 X3 þ 2 X2
4 X3 þ 2 X2 þ X1
8 X4
7
8
9
8 X4 þ X1
10
11
12
13
14
15
8 X4þ 2 X2
8 X4 þ2 X2 þ X1
8 X4þ 4 X3
8 X4þ 4 X3 þ X1
8 X4þ 4 X3 þ 2 X2
8 X4 þ 4 X3 þ 2 X2 þ X1
featuring multiple free N-termini with possible aldolase type
catalytic properties, or acetylated to provide library AcL in which
only the lysine or aminoproline residue at position X1 displayed an
enamine reactive amino acid. The library quality was checked by
submitting 72 beads to AAA. A total of 56 beads (78%) returned a
readable sequence, showing the expected even distribution of amino
acids (Supporting Information, Table S1). The remaining beads
contained 10 incomplete sequences (14%) and 6 empty beads (8%)
reflecting the yield of library synthesis.
Screening was performed by off-bead screening following our
recently reported protocol.12 In this assay the library beads are
partially photolyzed in the absence of solvent, and subsequently
placed on the surface of a silicagel plate freshly impregnated with
a solution of a fluorogenic substrate for the reaction of interest.
The photolyzed portion of the dendrimer dissolves and diffuses
into the surrounding wet silicagel. Synthesis beads carrying an
active sequence are detected by the formation of a fluorescent
halo indicating product formation. The beads are then retrieved
with a pipet, washed, and subjected to decoding by AAA of the
nonphotolyzed portion of the dendrimer which has remained
attached to the bead. Decoding requires only total hydrolysis
under acidic conditions, amino acid derivatization, and a single
HPLC analysis per bead to give the amino acid composition. This
method is much simpler than Edman sequencing, does not
require a free N-terminus, and can even be adapted to the
analysis of linear and cyclic peptide libraries.14
The procedure was applied by photolyzing 10 mg portions of
libraries L and AcL in the absence of solvent and applying the
beads to silica gel plates freshly impregnated with aqueous
buffered solutions of the fluorogenic substrates at 200 μM in
the case of the water-soluble substrates 6 and 8. For the water-
insoluble substrate 1 and 9 the silicagel plates were first im-
pregnated with a 100 μM solution of the fluorogenic substrate in
CH2Cl2, dried, and impregnated with aqueous buffer prior to
applying the library beads. Substrates 6a/b, 8, and 9 release
hydroxycoumarin by retroaldol reaction (6a/b), enediol forma-
tion, and tautomerization (8) or ester hydrolysis (9).
Figure 2. Fluorescent active halo from an off-bead screening with
fluorescein diacetate 1. The photolyzed beads (here AcL) were poured
on an impregnated TLC with PB buffer 60 mM pH 7.6 and 100 μM 1.
The picture is taken under illumination with a 366 nm TLC-UV lamp.
was expected to involve with either the primary or the secondary
R-amino group of the C-terminal amino acid at the core position
X1 in the case of library AcL or the multiple N-terminal amino
group in the case of library L, did not occur at a detectable level in
this assay (Scheme 2). While the hydrophobic esterase substrate
9 was also unreactive, active beads were observed in the case of
the esterase substrate 1 for both libraries L and AcL (Figure 2). In
each assay with 1 a fluorescent halo appeared after 24 h at room
temperature (RT) around 7ꢀ9 beads (0.1% of approximately
8,000 beads). These active beads as well as a selection of control
beads without halo were transferred to vials and subjected to
AAA. From 50 beads picked in the assays with AcL and L, 37
returned a readable sequence (74%, Table 2), 8 contained
incomplete sequences (16%), and 5 were empty (10%). The
analysis of the 60 non-hits provided 43 readable sequences
(72%) (Supporting Information, Table S2), 6 incomplete se-
quences (10%), and 11 empty beads (18%). It should be noted
that approximately 5ꢀ10% of the positive hit beads were
aggregated in groups of 2ꢀ5 beads, in which case the beads
were not analyzed.
The tests with the aldol-type fluorogenic substrates 6, 8, and 9
were all negative, suggesting that enamine-type reactivity, which
312
dx.doi.org/10.1021/co200006z |ACS Comb. Sci. 2011, 13, 310–320

ACS Combinatorial Science
RESEARCH ARTICLE
Table 2. Sequences for 37 Positive Hits in the off-Bead Assay
with Fluorescein Diacetate 1
library
no.
X4
X3
X2
X1
AcL
1
His
His
His
His
His
His
His
His
His
His
His
His
Ser
His
Gly
His
Arg
Leu
Phe
Gln
Gln
Val
Val
Arg
Ala
His
Leu
Arg
Ser
Val
His
Thr
Arg
Val
His
Ala
His
His
His
Lys
2
AcH1
AcH2
Lys
3
Lys
4
aPro
Lys
5
6
Tyr
βAla
Leu
Val
Lys
7
aPro
Lys
8
AcH3
AcH4
9
aPro
aPro
aPro
aPro
aPro
aPro
aPro
aPro
aPro
10
11
12
13
14
15
16
17
βAla
βAla
Gly
Arg
Ala
Figure 3. Composition of the dendrimer sequences. Control: beads
picked randomly from library L (56 beads). AcLþ: positive hits from
AcL with fluorescein diacetate 1 (17 beads); AcL-: inactive beads from
AcL with 1 (19 beads); Lþ: positive hits from L with 1 (20 beads); L-:
inactive hits from L with 1 (24 beads). The total number of occurrences
of each amino acid is calculated as 8m þ 4n þ 2p þ q with m, n, p, and q
denoting the number of sequences with that amino acid at positions X4,
X3, X2, and X1, respectively.
AcH5
His
Leu
L
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
His
Phe
Gln
His
His
Phe
His
Thr
His
Pro
Phe
His
Phe
His
Leu
His
His
Phe
His
His
Leu
His
Phe
Leu
Phe
His
bAla
His
Leu
Arg
His
Ala
Pro
Phe
Thr
His
Gly
Arg
His
Val
Ala
Ser
His
Ala
Pro
Ser
Arg
Ala
Lys
Table 3. Synthesis of Hits and Non-Hit Peptide Dendrimers
Identified by off-Bead Screening with Fluorescein Diacetate 1
with OBOC Libraries AcL and L
aPro
aPro
Lys
H6
H7
MS calc/ MS
aPro
aPro
aPro
aPro
Lys
no.
sequence
yield, mg (%)
obs
obsd
H8
AcH1 (AcH)8(KF)4(KR)2 KKOH
AcH2 (AcH)8(KQ)4 (KH)2KaPNH2 21.5 mg (8%) 3245.7
AcH3 (AcH)8(KL)4(KV)2KKOH 50.3 mg (16%) 3126.8
66.2 mg (18%) 3376.8
3377
3246
3127
3056
2642
2735
2906
2609
2254
3043
2385
3455
2239
2410
H9
His
His
Phe
His
Leu
His
Lys
AcH4 (AcH)8 (KβA)4(KR)2KaPNH2 2.4 mg (1%)
AcH5 (AcG)8(KH)4(KH)2 KaPNH2 2.0 mg (1%)
3055.7
2641.4
H10
Lys
Lys
H6
H7
H8
H9
(H)8(KL)4(KA)2KKOH
(H)8(KF)4(KP)2KaPNH2
(H)8(KβA)4(KT)2KaPNH2
(H)8(K)4(KG)2KKOH
62.0 mg (22%) 2734.7
15.8 mg (7%) 2905.6
aPro
aPro
aPro
aPro
aPro
aPro
aPro
Lys
3.3 mg (2%)
2609.5
73.8 mg (29%) 2254.3
64.0 mg (21%) 3042.7
H10 (F)8(KH)4(KH)2KKOH
AcN1 (AcS)8(K)4(KY)2KaPNH2
AcN2 (AcF)8(KR)4(KH)2KKOH
Leu
His
Arg
Phe
7.9 mg (5%)
28.8 mg (9%) 3454.9
1.3 mg (1%) 2239.4
46.5 mg (22%) 2410.4
2385.2
N3
N4
(A)8(KP)4(KQ)2KaPNH2
(S)8(KV)4(KH)2KKOH
Control and inactive beads showed approximately random
amino acid composition. By contrast the sequences on active
beads showed a strong consensus for histidine (Figure 3). All
positive hits in libraries AcL and L had at least one histidine in the
sequence, most frequently at the outer position X4 (8 copies),
such that the average amino acid composition comprised
approximately 50% histidine. Histidine at position X4 was
generally combined with hydrophobic amino acids at the inner
positions X3 and X2, in particular phenylalanine and leucine. The
cationic arginine occurred in 9 of the 37 positive hits, but the
anionic aspartate was totally absent from positive hits.
C-terminal carboxyl or carboxamide (the nature of the C-termi-
nus turned out to have negligible influence on catalysis). The
dendrimers were obtained in good yields and purity after
purification by RP-HPLC. The hit sequences carried 6 to 10
histidine residues. Two of the non-hits had no histidine residues
and two had a pair of histidines at position X2. A series of linear
oligohistidine peptides from 1 to 15 residues (His1-His15) and
the histidine-lysine oligomer P25 (AcHKHKHKHKHKHNH2),
which were known to exhibit good esterolytic activity toward 4,15
were also included in the study. Kinetics were measured in 96-
well microtiter plates following product formation by fluores-
cence in aqueous buffered solutions containing the peptide
dendrimer at 5 μM and the substrates 1, 2 or 4, including 17%
acetonitrile cosolvent with 1 and 11% acetonitrile cosolvent with
2 (Table 4 and Supporting Information, Table S3, Figure 4).
Hit Resynthesis and Kinetic Studies. Ten hits and four non-
hits from screening with fluorecein diacetate 1 were selected
equally from libraries AcL and L and resynthesized by SPPS
(Table 3), using either Wang or Rink amide resin yielding a
313
dx.doi.org/10.1021/co200006z |ACS Comb. Sci. 2011, 13, 310–320

ACS Combinatorial Science
RESEARCH ARTICLE
Table 4. Kinetic Parameters for Dendrimer and Peptide Catalyzed Ester Hydrolysis
no.
sequence
NHis
kcat/KM (1)a Mꢀ1 minꢀ1
kcat/KM (2)b Mꢀ1 minꢀ1
kcat/KM (4)c Mꢀ1 minꢀ1
AcH1
AcH2
AcH3
AcH4
AcH5
H6
(AcH)8(KF)4(KR)2 KKOH
(AcH)8(KQ)4 (KH)2KaPNH2
(AcH)8(KL)4(KV)2KKOH
(AcH)8 (KβA)4(KR)2KaPNH2
(AcG)8(KH)4(KH)2 KaPNH2
(H)8(KL)4(KA)2KKOH
(H)8(KF)4(KP)2KaPNH2
(H)8(KβA)4(KT)2KaPNH2
(H)8(K)4(KG)2KKOH
(F)8(KH)4(KH)2KKOH
(AcS)8(K)4(KY)2KaPNH2
(AcF)8(KR)4(KH)2KKOH
(A)8(KP)4(KQ)2KaPNH2
(S)8(KV)4(KH)2KKOH
AcHNH2
8
10
8
24 ( 2
34 ( 8
25 ( 9
19 ( 0.6
112 ( 7
160 ( 23
260 ( 70
170 ( 20
47 ( 5
745 ( 30
1100 ( 47
910 ( 30
8
1170 ( 230
1570 ( 110
1770 ( 65
3820 ( 260
3400 ( 80
2770 ( 26
1690 ( 40
6
8
20 ( 5
33 ( 8
H7
8
47 ( 7
310 ( 70
36 ( 8
H8
8
35 ( 12
21 ( 10
43 ( 19
H9
8
30 ( 3
H10
6
130 ( 6
AcN1
AcN2
N3
2
68 ( 9
670 ( 33
137 ( 12
660 ( 13
3 ( 0.2
N4
2
18 ( 5
5 ( 0.5
10 ( 1
19 ( 1
25 ( 2
36 ( 5
35 ( 3
42 ( 3
60 ( 7
65 ( 8
81 ( 5
59 ( 2
70 ( 7
75 ( 4
79 ( 3
114 ( 13
33 ( 3
His1
His2
His3
His4
His5
His6
His7
His8
His9
His10
His11
His12
His13
His14
His15
P25
1
nd
AcHHNH2
2
nd
27 ( 0.4
AcHHHNH2
3
nd
132 ( 4
AcHHHHNH2
4
nd
220 ( 6
AcHHHHHNH2
5
42 ( 14
39 ( 12
36 ( 13
43 ( 16
41 ( 20
56 ( 27
49 ( 32
nd
270 ( 7
AcHHHHHHNH2
6
380 ( 2
AcHHHHHHHNH2
7
796 ( 130
960 ( 140
1010 ( 130
1240 ( 110
1200 ( 130
1540 ( 180
1570 ( 250
1690 ( 220
2210 ( 190
2380 ( 200
AcHHHHHHHHNH2
AcHHHHHHHHHNH2
AcHHHHHHHHHHNH2
AcHHHHHHHHHHHNH2
AcHHHHHHHHHHHHNH2
AcHHHHHHHHHHHHHNH2
AcHHHHHHHHHHHHHHNH2
AcHHHHHHHHHHHHHHHNH2
AcHKHKHKHKHKHNH2
8
9
10
11
12
13
14
15
6
77 ( 46
83 ( 41
nd
nd
a Conditions: 3.75ꢀ15 μM peptide or dendrimer, 3ꢀ100 μM 1, 20 mM phosphate pH 7.6 with 17% acetonitrile, 34 °C. b Conditions: 3.75ꢀ20 μM
peptide or dendrimer, 3ꢀ100 μM 2, 5 mM citrate pH 5.5, with 11% acetonitrile, 34 °C. c Conditions: 3.75- 20 μM peptide or dendrimer, 58.5ꢀ1000 μM
substrate 4, 5 mM citrate pH 5.5, 34 °C. Spontaneous background reaction kuncat = 3.22 ꢁ 10ꢀ4 minꢀ1 for 1, 3.37 ꢁ 10ꢀ5 minꢀ1 for 2, 3.58 ꢁ 10ꢀ5
minꢀ1 for 4. k2 is the 2nd order rate constant with 4-methyl imidazole, k2 =2.26 Mꢀ1 minꢀ1 for 2, k2 = 1.0 Mꢀ1 minꢀ1 for 4. 120 μL assays in
microtiterplate wells were followed by fluorescence (λexc = 450 ( 25 nm, λem = 530 ( 12 nm). All kinetics were run in triplicate. See Experimental
Section and Supporting Information, Table S3 for complete data.
The hydrolysis of fluorescein diacetate 1 to fluorescein was
investigated in phosphate buffer pH 7.6 in the presence of 17%
acetonitrile cosolvent to ensure substrate solubility. Unfortu-
nately experiments showed poor reproducibility and only pro-
vided an estimation of the apparent specificity constant kcat/KM.
The data confirmed the screening results with all but one positive
hit showing the expected activity and the four non-hit sequences
showing no activity with this substrate. The apparent activity was
comparable to that observed with linear oligohistidine peptides
when counted per catalytic histidine residues.
To better characterize the unprecedented esterolytic activity
of dendrimers with hydrophobic substrates, we prepared iso-
butyryl fluorescein 2 (Scheme 1). This substrate requires only a
single ester bond cleavage to release the fluorescent product
fluorescein, and was expected to show higher aqueous solubility
because of the presence of a free carboxyl group. We also
anticipated a better reproducibility of the kinetic data from our
previous experiences with related straight-chain aliphatic fluor-
escein monoesters as lipase and esterase substrates.16 Indeed
substrate 2 gave better reproducible data than 1, and showed
activity both at pH 7.6 and at pH 5.5. The best kinetic data were
collected at pH 5.5; however, the substrate concentration was
held at 100 μM or below because of limited solubility, which
precluded a determination of the MichaelisꢀMenten constant
KM. The data were therefore interpreted in terms of the
specificity constant kcat/KM. The kinetic study was also carried
out with the trianionic substrate 4 at the same pH because of its
strong reactivity with histidine containing catalysts.
Structure-Activity Relationships. A comparative analysis of
the catalytic proficiencies of the different peptide dendrimers
with the hydrophobic ester substrate 2 and the polyanionic
substrate 4 provided a useful insight into the structure-activity
relationships. For each substrate the relative specificity constants
compared to the reference small molecule catalyst 4-methyl-
imidazole were computed as (kcat/KM)/k2(2) and (kcat/KM)/
k2(4), respectively ((kcat/KM)/k2)/NHis(2) and ((kcat/KM)/
k2)/NHis(4) for the effect counted per histidine residue
(Figure 5A and 5B).
314
dx.doi.org/10.1021/co200006z |ACS Comb. Sci. 2011, 13, 310–320

ACS Combinatorial Science
RESEARCH ARTICLE
Figure 4. Catalytic reaction rates with peptide dendrimers H7, H8, and AcH3. Conditions: aq. 5 mM citrate pH 5.5 with 5 μM dendrimer.
Figure 5. Relative catalytic proficiencies of histidine containing peptides and peptide dendrimers with substrates 2 and 4 in aqueous 5 mM citrate buffer
pH 5.5. (A) Relative catalytic proficiencies (kcat/KM)/k2. k2 is the 2nd order rate constant with 4-methyl-imidazole as reference catalyst. (B) Relative
catalytic proficiencies per histidine residue ((kcat/KM)/k2)/NHis. NHis is the number of histidine residue in the peptide or peptide dendrimer catalyst.
(red solid triangles): peptide dendrimers from library L; (blue solid triangles): peptide dendrimers from library AcL; (black solid circles): oligohistidine
peptide from 1 to 15 amino acid residues, with N-terminal acetylation. (black open circles): octapeptide AcHKHKHKHKNH2. See Table 4 for detailed
kinetic data and conditions.
The reference oligohistidine peptides His1-His15 showed
increasing relative catalytic proficiencies (kcat/KM)/k2 as a
function of length for both substrates. The relative catalytic
proficiencies for substrate 4 were on average 40-fold larger
than for substrate 2. The catalytic effect per histidine residue for
the trianionic substrate 4 ((kcat/KM)/k2)/NHis(4) strongly
increased up to seven histidine residues before leveling off. By
contrast, the catalytic proficiency per histidine residue for the
isobutyryl fluorescein substrate 2 ((kcat/KM)/k2)/NHis(2) was
essentially independent of peptide length. Both the higher
catalytic activity with 4 compared to 2 and the length depen-
dence of ((kcat/KM)/k2)/NHis(4) may be explained by signifi-
cant electrostatic binding with substrate 4 requiring cooperative
action of several histidine residues, while the more hydrophobic
substrate 2 does not bind and is therefore insensitive to peptide
length.
substrate 2. The most selective catalyst for substrate 2 was peptide
dendrimer AcH3 (AcHis)8(LysLeu)4(LysVal)2LysLysOH, which
combined eight histidine residues in the third generation
branches with hydrophobic residues in the first and second
generation branches. On the other hand peptide dendrimer
AcH5 (AcGly)8(LysHis)4(LysHis)2 LysaProNH2 with six his-
tidine residues in the inner first and second generation branches
and only glycine residue in the third generation branch was
more selective toward the trianionic substrate 4. Interestingly
the negative control dendrimer AcN2 (AcPhe)8(LysArg)4-
(LysHis)2LysLysOH, which was not active with fluorescein
diacetate 1, showed significant activity with the isobutyrate 2
as well as with the trisulfonate 4. This reactivity stood out in
particular when considering the specific catalytic effect per
histidine residue, and probably reflects substrate binding inter-
actions with the arginine side chains.
The peptide dendrimers showed a differentiated catalytic
behavior toward 2 and 4, with several dendrimers standing out
by their reactivities, in particular relative to the linear peptides.
Dendrimers with acetylated N-termini from library AcL were
generally more active toward the hydrophobic fluorescein
The peptide dendrimers from library L with free N-termini
were more reactive toward the trianionic substrate 4, in particular
for dendrimer H8 (His)8(LysβAla)4(LysThr)2LysaProNH2, re-
flecting electrostatic substrate binding. This dendrimer showed
only poor reactivity with 2 attributable to the absence of
315
dx.doi.org/10.1021/co200006z |ACS Comb. Sci. 2011, 13, 310–320

ACS Combinatorial Science
RESEARCH ARTICLE
hydrophobic residues. Peptide dendrimer H7 (His)8(LysPhe)4-
(LysPro)2LysaProNH2 on the other hand had the strongest
reactivity of all catalysts toward both substrates. A particularly
favorable combination of eight catalytic histidine residues at the
N-terminus with multiple positive charges from the free N-ter-
mini favored activity with 4. In addition the hydrophobic
interior of H7 constituted of four phenylalanines in the second
generation branch and two proline residues in the first genera-
tion branch favored activity with 2. Peptide dendrimer H10
(Phe)8(LysHis)4(LysHis)2LysLysOH featuring an inverted
combination with eight hydrophobic phenylalanines in the third
generation branches and six catalytic histidines in the inner
branches showed a weaker but similarly balanced reactivity between
2 and 4. By comparison the linear peptide P25 AcHKHK-
HKHKHKHNH2 which combines six histidine residues with five
ammonium group from lysine side chains showed only enhanced
catalysis with the trisulfonate 4 but not with isobutyryl fluorescein
2, because of the absence of hydrophobic residues.
Mass Spectrometric Analysis. As previously observed with
P25, the peptide dendrimers from library L underwent signifi-
cant acylation of the N-termini in contact with substrate 4, as
detected by MS for the reaction of H9 and H10 under high
substrate and dendrimer concentration (see Supporting In-
formation). In the case of dendrimer H9 MS analysis also showed
the formation of substrate-dendrimer complexes with one, two,
and three substrates per dendrimer, in agreement with the
multivalent nature of these catalysts. On the other hand acylation
with isobutyrate 2 occurred only to a very low level, probably
because of the slower reaction rates and the bulkier acyl group in
that substrate.
derivatives were purchased from Advanced ChemTech (U.S.A.)
or Novabiochem (Switzerland). Amino acids were used as the
following derivatives: Fmoc-Ala-OH, Fmoc-βAla-OH, Fmoc-
Arg(Pbf)-OH, Fmoc-Asp(OtBu)-OH, Fmoc-Gln(Trt)-OH, Fmoc-
Gly-OH, Fmoc-His(Boc)-OH, Fmoc-Ile-OH, Fmoc-Leu-OH,
Fmoc-Lys(Boc)-OH, Fmoc-Lys(Fmoc)-OH, Fmoc-Phe-OH,
Fmoc-Pro-OH, Fmoc-Ser(tBu)-OH, Fmoc-Thr(tBu)-OH, Fmoc-
Tyr(tBu)-OH, Fmoc-Val-OH. Fmoc-aPro(Boc)-OH was purchased
from Aldrich. TG S RAM (loading: 0.23 or 0.24 mmol gꢀ1) resin
3
was purchased from Rapp Polymere (Germany) while hydroxy-
methyl-photolinker NovaSyn TG (loading: 0.24 mmol gꢀ1) and
3
TGA (loading: 0.26 mmol gꢀ1) were purchased from Novabio-
3
chem. Peptide dendrimer syntheses were performed manually in
polypropylene syringes fitted with a polyethylene frit, a Teflon
stopcock and stopper. Analytical RP-HPLC was performed in
Waters (996 Photo diode array detector) chromatography
system using Atlantis column (dC18, 5 μm, 4.6 ꢁ 100 mm, flow
rate 1.4 mL minꢀ1). Analytical RP-UPLC was performed by
3
Ultimate 3000 Rapid Separation LC System using diode array
detector (DAD-3000RS) and Acclaim RSLC 120 C18 column
(2.2 μm, 120 Å, 3.0 ꢁ 50 mm, flow 1.2 mL minꢀ1) from Dionex.
Preparative RP-HPLC was performed in Waters Prep LC4000
chromatography system using a Delta-Pak column (C18, 15 μm,
pore size 300 Å, flow rate 80 mL minꢀ1) or an Atlantis OBD
3
column (dC18, 5 μm, 19 ꢁ 100 mm, flow rate 20 mL minꢀ1).
3
Compounds were detected by UV absorption at 214 nm. All RP-
HPLC were using HPLC-grade acetonitrile and mQ-deionized
water. The elution solutions were A H2O with 0.1% TFA; B
H2O/MeCN (50:50); C H2O/MeCN (10:90) with 0.1%
TFA; D H2O/MeCN (40:60) with 0.1% TFA. MS spectra
and amino acid analyses measurements were provided by
Mass Spectrometry, and Protein Analysis services respectively
of the Department of Chemistry and Biochemistry at the
University of Bern. Kinetic measurements were carried out
using a CytoFluor Series 4000 multiwell plate reader from
PerSeptive Biosystems.
’ CONCLUSION
The experiments above showed the first example of a combi-
natorial peptide dendrimer library extensively covering its se-
quence space by using 14 amino acids and a deletion at each
variable position, exploiting the fact that such exhaustive sam-
pling is compatible with AAA for bead decoding. Screening of
catalysis with fluorogenic substrates in the off-bead in silica assay
uncovered specific reactivities with fluorescein diacetate 1. While
this substrate proved unsuitable for kinetic studies, the related
isobutyryl fluorescein 2, which requires only a single ester bond
cleavage to release its fluorescent product, allowed the character-
ization of this reactivity. Structure-activity relationship studies
comparing the catalyzed ester hydrolysis reaction rates of 2 with
those of the trianionic substrate 4 showed that efficient hydro-
lysis of the hydrophobic substrate 2 requires a combination of
multiple catalytic histidine residues with hydrophobic amino
acids such as phenylalanine and leucine. These hydrophobic
residues presumably enable substrate binding, although pre-
equilibrium binding was not visible in the accessible concentra-
tion range. By contrast peptide dendrimers with mostly cationic
residues and no hydrophobic groups showed stronger activity
with the trisulfonate 4. These experiments highlight the power of
combinatorial approaches to identify peptidic enzyme models
and the role of noncatalytic amino acids to determine substrate
selectivity in these systems.
Synthesis of Split-and-Mix Library. The peptide dendrimer
library was prepared from 1.5 gꢀ100 mg per split-resin batch of
hydroxymethyl-photolinker NovaSyn TG resin (0.24 mmol
g
3
ꢀ1) divided equally in 15 reactors. The attachment of the first
amino acid to hydroxymethyl resin was done by using MSNT
with 1-MeIm.17 The resin was placed in a syringe and swelled
with dry CH2Cl2 under nitrogen. In a round-bottom flask, the
appropriate amino acid (5 equiv) was dissolved in dry CH2Cl2
(3 mL per mmol) and a few drops of tetrahydrofuran (THF).
MSNT (5 equiv) and 1-methylimidazole (3.75 equiv) were
added, dissolved, and the mixture was added to the syringe.
Then the syringe was agitated at RT for 1 h. The resin was
washed with CH2Cl2 (5 ꢁ 5 mL). The coupling was repeated
two times. A sensitive color test (alizarin-cyanuric chloride
test)18 was used to detect the presence of hydroxyl groups on
solid support; the presence of unreacted hydroxyl groups was
indicated by red-colored beads. The Fmoc-protecting groups of
the resin were removed with a solution of 20% piperidine in
dimethylformamide (DMF, 2 ꢁ 10 min). For further couplings,
the resin was acylated with one of the amino acid (3 eq/G, G =
generation) in the presence of PyBOP (3 eq/G) and DIEA (5
eq/G) in NMP.3c Amino acids, derivatives or diamino acids were
coupled for 30 min (G0), 1 h (G1), 2 h (G2), 4 h (G3). The
completion of the reaction was checked using 2,4,6-trinitroben-
zenesulfonic acid solution (TNBS) or chloranil test.19 If the
beads were red (brown for proline), there were some free amino
’ EXPERIMENTAL SECTION
General Procedures. All reagents were either purchased from
Aldrich, Fluka, or Acors Organics. PyBOP, amino acids, and their
316
dx.doi.org/10.1021/co200006z |ACS Comb. Sci. 2011, 13, 310–320

ACS Combinatorial Science
RESEARCH ARTICLE
groups and the resin test was positive. If they were colorless, there
were no more free amino groups, and the resin test was negative.
The coupling was repeated after a positive test. After each
coupling, the resin in each syringe was deprotected (20%
piperidine in DMF, 2 ꢁ 10 min) followed by TNBS or chloranil
test (test must be positive). Then the resin batches were mixed
together, vortexed during 1 min, and split equally. These split-
and-mix steps were repeated after each amino acid coupling.
After each coupling or deprotection, the resin was successively
washed with NMP, MeOH, and CH2Cl2 (3 ꢁ 5 mL, with each
solvent). At the end of the synthesis, the Fmoc protected library
was dried and stored at ꢀ18 °C. For all the steps, the syringes
were covered with aluminum to prevent exposure to light.
Off-Bead Screening. Before the screening of the library, the
Fmoc protecting groups were removed and the last amino acid
was acetylated with Ac2O/CH2Cl2 (1:1) for 1 h (library AcL) or
not acetylated (library L). The side-chain protecting groups were
removed with TFA/TIS (triisopropylsilane)/H2O (94:5:1) dur-
ing 4 h, resulting in a functional dendrimer library on-beads.
Screening was performed by washing the beads (5 ꢁ 3 mL) with
the buffer and swelling those (10 mg, 2 ꢁ 1 h) in the same
aqueous buffered solution. After solution filtration and drying
under vacuum, the resin was poured on a Petri dish (diameter:
3 cm) and was partially irradiated under an Hg lamp (100 W)
with a 366 nm filter for 30 min. A glass TLC plate (10 ꢁ 10 cm,
SIL G25 without fluorescent indicator) was impregnated with
100 μM fluorescein diacetate 1 dissolved in CH2Cl2, and dried.
The TLC was sprayed with PB buffer (60 mM, pH 7.6) and the
photolyzed beads were poured on the slightly wet TLC and incubated
in a close Petri dish for 12 h. Under UV irradiation (366 nm) a green
fluorescent halo appeared on the TLC around some of the beads.
These active beads were picked and subjected to AAA.
(AcHis)8(LysGln)4(LysHis)2LysaProNH2 (AcH2). Starting with
250 mg of TG S RAM (0.24 mmol gꢀ1), dendrimer AcH2 was
3
obtained as a white foamy solid after cleavage from the resin and
preparative RP-HPLC purification (21.5 mg, 8%). Anal. RP-
HPLC: tR = 7.0 min. MS (ESþ) calcd for C143H214N55O34
[MþH]þ: 3245.7, found: 3246.0; [MþK]þ: 3287.7, found:
3285.0.
(AcHis)8(LysLeu)4(LysVal)2LysLysOH (AcH3). Starting with
300 mg of TGA (0.26 mmol gꢀ1), dendrimer AcH3 was
3
obtained as a white foamy solid after cleavage from the resin
and preparative RP-HPLC purification (50.3 mg, 16%). Anal.
RP-HPLC: tR = 8.3 min. MS (ESþ) calcd for C146H233N46O31
[MþH]þ: 3126.8, found: 3127.0; [MþNa]þ: 3149.8, found:
3149.0; [MþK]þ: 3165.8, found: 3167.0.
(AcHis)8(LysβAla)4(LysArg)2LysaProNH2 (AcH4). Starting
with 250 mg of TG S RAM (0.24 mmol gꢀ1), dendrimer
3
AcH4 was obtained as a white foamy solid after cleavage from
the resin and preparative RP-HPLC purification (2.4 mg, 1%).
Anal. RP-HPLC: tR = 6.8 min. MS (ESþ) calcd for
C135H212N53O30 [MþH]þ: 3055.7, found: 3056.0; [MþK]þ:
3093.7, found: 3097.0.
(AcGly)8(LysHis)4(LysHis)2LysaProNH2 (AcH5). Starting with
250 mg of TG S RAM (0.24 mmol gꢀ1), dendrimer AcH5 was
3
obtained as a white foamy solid after cleavage from the resin and
preparative RP-HPLC purification (2.0 mg, 1%). Anal. RP-
HPLC: tR = 7.0 min. MS (ESþ) calcd for C115H178N43O30
[MþH]þ: 2641.4, found: 2642.0; [MþK]þ: 2679.4, found:
2682.0; [MþTFA]þ: 2753.4, found: 2756.0.
(His)8(LysLeu)4 (LysAla)2LysLysOH (H6). Starting with 300 mg
of TGA (0.26 mmol gꢀ1), dendrimer H6 was obtained as a
3
white foamy solid after cleavage from the resin and preparative
RP-HPLC purification (62.0 mg, 22%). Anal. RP-UPLC:
tR = 1.312 min (A/D = 100/0 to 0/100 in 2.2 min, 100 D in
1.3 min, A/D = 0/100 to 100/0 in 0.5 min, 100 A in 1 min
Sequence Determination. Single dendrimer-containing re-
sin beads were hydrolyzed with 6 M aqueous HCl at 110 °C for
22 h. The amino acids were derivatized with phenylisothiocya-
nate (PITC) and the phenylthiocarbamyl (PTC) derivatives
analyzed on a RP-C18 Novapack column.
1.2 mL minꢀ1). MS (ESþ) calcd for C126H209N46O23
þ3
[MþH] : 2734.7, found: 2735.0; [MþK]þ: 2773.7, found:
2775.0.
Dendrimer Synthesis. The resin (TG S RAM, or TGA) was
swelled in CH2Cl2, and the Fmoc-protecting groups of the resin
were removed with a solution of 20% piperidine in DMF (2 ꢁ 10
min). Then the resin was acylated with each amino acid,
derivative, or diamino acid (3 eq/G) using PyBOP (3 eq/G)
and DIEA (5 eq/G) in NMP. The first coupling on the TGA
resin used the described protocol (MSNT, 1-MeIm). The
completion of the reaction was checked using the TNBS or
chloranil test (alizarin-cyanuric chloride test for the first coupling
on TGA). At the end of the synthesis, the resin was acetylated
with Ac2O/CH2Cl2 (1:1) for 1 h for library AcL. The cleavage
was carried out with TFA/TIS/H2O (94:5:1) during 4 h. The
peptide was precipitated with tert-butylmethylether (MTBE)
then dissolved in a H2O/MeCN mixture. Peptides were purified
by preparative RP-HPLC and obtained as TFA salts after
lyophilization. Unless otherwise mentioned, the gradient used
for analytical HPLC is A/D = 100/0 to 0/100 in 10 min, 100 D in
(His)8(LysPhe)4(LysPro)2LysaProNH2 (H7). Starting with 250 mg
of TG S RAM (0.24 mmol gꢀ1), dendrimer H7 was obtained as a
3
white foamy solid after cleavage from the resin and preparative RP-
HPLC purification (15.8 mg, 7%). Anal. RP-HPLC: tR = 7.7 min.
MS (ESþ) calcd for C141H202N47O22 [MþH]þ: 2905.6, found:
2906.0; [MþK]þ: 2943.6, found: 2946.0.
(His)8(LysβAla)4(LysThr)2LysaProNH2 (H8). Starting with 250
mg of TG S RAM (0.24 mmol gꢀ1), dendrimer H8 was obtained
3
as a white foamy solid after cleavage from the resin and
preparative RP-HPLC purification (3.3 mg, 2%). Anal. RP-
HPLC: tR = 7.0 min. MS (ESþ) calcd for C115H186N47O24
[MþH]þ: 2609.5, found: 2609.0.
(His)8(Lys)4(LysGly)2LysLysOH (H9). Starting with 300 mg of
TGA (0.26 mmol gꢀ1), dendrimer H9 was obtained as a white
3
foamy solid after cleavage from the resin and preparative RP-
HPLC purification (73.8 mg, 29%). Anal. RP-UPLC: tR = 1.212
min (A/D = 100/0 to 0/100 in 2.2 min, 100 D in 1.3 min, A/D =
5 min, 1.4 mL minꢀ1
.
0/100 to 100/0 in 0.5 min, 100 A in 1 min 1.2 mL minꢀ1). MS
3
3
(AcHis)8(LysPhe)4(LysArg)2LysLysOH (AcH1). Starting with
(ESþ) calcd for C100H161N42O19 [MþH]þ: 2254.3, found:
2254.0; [MþNa]þ: 2276.3, found: 2277.0; [MþK]þ: 2293.3,
found: 2295.0.
300 mg of TGA (0.26 mmol gꢀ1), dendrimer AcH1 was
3
obtained as a white foamy solid after cleavage from the resin
and preparative RP-HPLC purification (66.2 mg, 18%). Anal.
RP-HPLC: tR = 8.1 min. MS (ESþ) calcd for C160H231N52O31
[MþH]þ: 3376.8, found: 3377.0; [MþK]þ: 3414.8, found:
3416.0; [MþTFA]þ: 3488.8, found: 3491.0.
(Phe)8(LysHis)4(LysHis)2LysLysOH (H10). Starting with 300
mg of TGA (0.26 mmol gꢀ1), dendrimer H10 was obtained as a
3
white foamy solid after cleavage from the resin and preparative
RP-HPLC purification (64.0 mg, 21%). Anal. RP-HPLC:
317
dx.doi.org/10.1021/co200006z |ACS Comb. Sci. 2011, 13, 310–320

ACS Combinatorial Science
RESEARCH ARTICLE
tR = 7.3 min (A/D = 100/0 to 0/100 in 7 min, 100 D in 3 min,
pH 5.5 were used as buffer, and the pH was adjusted to the
desired value with HCl 1.0 M or NaOH 1.0 M using a Metrohm
692 pH/ion meter. Eight final substrate concentrations ranging
from 5.8 to 100 μM in MeCN/buffer (17:83) for substrate 1,
5.8 to 100 μM in MeCN/buffer (11:89) for substrate 2, 58.5 to
1000 μM in buffer for substrate 4 (dilutions by 2/3) were
measured for each MichaelisꢀMenten plot. Eight solutions of
fluorescein sodium salt (FNa) ranging from 0 to 10 μM in in
MeCN/buffer (17:83) for substrate 1, in MeCN/buffer (11:89)
for substrate 2, 8-hydroxypyrene-1,3,6-trisulfonic acid, trisodium
salt (HPTS) from 0 to 100 μM in buffer for substrate 4 were used
for the calibration curve. A 40 μL portion of dendrimer was
mixed with 40 μL of buffer and 40 μL of substrate in a 96-well half
area flat bottom plate (190 μL). The formation of FNa or HPTS
was followed by fluorescence emission using absorbance filter
450/50 and emission filter 530/25. The calibration curve and the
blank (40 μL substrate, 40 μL buffer and 40 μL mQ-H2O) were
recorded for every experiment in parallel. Prior to every experi-
ment the gain was adjusted to a signal of ∼50000 for the
maximum concentration of FNa (for 1 nd 2) or HPTS (for 4)
well. The temperature inside the instrument was set to 34 °C.
Kinetic experiments were followed for typically 1ꢀ3 h, and the
data points were measured every 90 s. Fluorescence data were
converted into product concentration by means of the calibration
curve. Initial reaction rates were calculated from the steepest
linear part observed in the curve that gives fluorescence
versus time.
1.4 mL minꢀ1). MS (ESþ) calcd for C156H213N42O23
þ3
[MþH] : 3042.7, found: 3043.0; [MþK]þ: 3082.7, found:
3083.0.
(AcSer)8(Lys)4(LysTyr)2LysaProNH2 (AcN1). Starting with 250
mg of TG S RAM (0.24 mmol gꢀ1), dendrimer AcN1 was
3
obtained as a white foamy solid after cleavage from the resin and
preparative RP-HPLC purification (7.9 mg, 5%). Anal. RP-
HPLC: tR = 7.7 min. MS (ESþ) calcd for C105H170N27O36
[MþH]þ: 2385.2, found: 2385.0; [MþNa]þ: 2407.2, found:
2407.0; [MþK]þ: 2423.2, found: 2423.0.
(AcPhe)8(LysArg)4(LysHis)2LysLysOH (AcN2). Starting with
300 mg of TGA (0.26 mmol gꢀ1), dendrimer AcN2 was
3
obtained as a white foamy solid after cleavage from the resin
and preparative RP-HPLC purification (28.8 mg, 9%). Anal. RP-
UPLC: tR = 1.777 min (A/D = 100/0 to 0/100 in 2.2 min, 100 D
in 1.3 min, A/D = 0/100 to 100/0 in 0.5 min, 100 A in 1 min,
1.2 mL minꢀ1). MS (ESþ) calcd for C172H249N46O31
þ3
[MþH] : 3454.9, found: 3455.0; [MþTFA]þ: 3567.9, found:
3570.0.
(Ala)8(LysPro)4(LysGln)2LysaProNH2 (N3). Starting with 250
mg of TG S RAM (0.24 mmol gꢀ1), dendrimer N3 was obtained
3
as a white foamy solid after cleavage from the resin and
preparative RP-HPLC purification (1.3 mg, 1%). Anal. RP-
HPLC: tR = 6.6 min. MS (ESþ) calcd for C101H180N33O24
[MþH]þ: 2239.4, found: 2239.0; [MþK]þ: 2277.4, found:
2278.0.
(Ser)8(LysVal)4(LysHis)2LysLysOH (N4). Starting with 300 mg
Kinetic Parameters kcat and KM. Vcat is the apparent rate in
the presence of dendrimer catalyst, and Vuncat is the rate in buffer
alone. The observed rate enhancement is defined as Vnet/Vuncat
with Vnet = Vcat ꢀ Vuncat. MichaelisꢀMenten parameters kcat
(rate constant) and KM (Michaelis constant) were obtained by
fitting the data to the MichaelisꢀMenten model. With substrate
2 where saturation was not reached no determination of kcat and
KM was possible and only the specificity constant kcat /KM is
given. The rate constant kuncat without catalyst was calculated
from the slope of the linear curve that gives Vuncat (as product
concentration per time) versus substrate concentration [S].
k2 (4-Methylimidazole). The solutions of 4-methylimidazole
were prepared by serial dilution from a stock solution (3 mM)
adjusted to the desired pH value using HCl 1 M. The reaction
rate with 4-methylimidazole was obtained under the same
conditions as described above. The final concentrations in
the plate were 0, 88, 132, 198, 296, 444, 667 and 1000 μM of
4-MeIm, 200 μM of substrate, and 60 mM of buffer. The
second order rate constants k2 were calculated from linear
regression of the0 experimentally measured pseudo first order
rate constants k2 as a function of 4-methylimidazole concen-
trations. The second order rate constants k2 is given by k2 =
of TGA (0.26 mmol gꢀ1), dendrimer N4 was obtained as a
3
white foamy solid after cleavage from the resin and preparative
RP-HPLC purification (46.5 mg, 22%). Anal. RP-HPLC: tR = 6.9
min. MS (ESþ) calcd for C104H189N34O31 [MþH]þ: 2410.4,
found: 2410.0; [MþNa]þ: 2432.4, found: 2432.0; [MþK]þ:
2449.4, found: 2450.0.
Isobutyryl Fluorescein (2). Fluorescein (500 mg. 1.5 mmol)
was added to a suspension of 43 mg NaH (1.2 equiv.) and 50 mg
ZnCl2 (0.25 equiv.) in 10 mL of anhydrous DMF in the presence
of molecular sieves. Isobutyryl chloride (160 μL, 1 equiv.) in
0.25 mL of dry DMF was added dropwise to the first solution for
5 min. After 18 h the reaction mixture was poured into 100 mL of
ethyl acetate, washed with brine (3 ꢁ 100 mL), and dried over
Na2SO4. The solvents were removed, and the crude was purified
with flash chromatography (hexane-ethyl acetate 6:4)) to yield
233 mg of pure 2. Yield 38%.1H NMR (CDCl3, 400 MHz) δ =
8.04ꢀ7.99 (m, 1H), 7.69ꢀ7.56 (m, 2H), 7.11ꢀ7.04 (m, 2H),
6.83ꢀ6.72 (m, 2H), 6.69 (d, J = 2.3 Hz, 1H), 6.55 (dt, J = 8.7, 5.5
Hz, 2H), 2.83 (dt, J = 17.7, 7.0 Hz, 1H), 1.33 (d, J = 1.2 Hz, 3H),
1.31 (d, J = 1.2 Hz, 3H). 13C NMR (CDCl3, 400 MHz) δ = 175.8,
170.1, 158.7, 152.9, 152.6, 152.5, 152.2, 135.4, 130.1, 129.5,
126.7, 125.5, 124.4, 117.6, 116.8, 113.1, 110.9, 110.6, 103.4, 34.5,
19.1. Hrms (ESIþ) Calculated for C24H19O6 [MþH]þ,
403.1181, observed 403.1176, [MþNa]þ 426.39, observed
425.10.
Kinetic Measurements. Peptides solutions were used as 11.5
μM (His10-His14), 15 μM (all dendrimers), 22.5 μM (His8-
His9), 45 μM (His5-His7) for substrate 1, 11.5 μM (His11-
His15, P25), 15 μM (all dendrimers), 22.5 μM (His8-His10), 45
μM (His5-His7), 60 μM (His1-His4) for substrate 2, 11.5 μM
(His15, P25), 15 μM (all dendrimers, His7-His14), 22.5 μM
(His8-His10), 45 μM (His5-His6), 60 μM (His1-His4) for
substrate 4, as freshly prepared solutions in mQ-H2O. Phosphate
buffer 60 mM pH 7.6, citric acid-sodium citrate buffer 15 mM at
0
k2 /[S].
’ ASSOCIATED CONTENT
S
Supporting Information. Tables S1ꢀS3, catalytic den-
b
drimers (HPLC traces, MS and kinetics spectra). This material is
available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
*Fax:þ 41 31 631 80 57. E-mail: jean-louis.reymond@ioc.unibe.ch.
318
dx.doi.org/10.1021/co200006z |ACS Comb. Sci. 2011, 13, 310–320

ACS Combinatorial Science
RESEARCH ARTICLE
’ ACKNOWLEDGMENT
Nature 1991, 354, 84–86. (d) Lam, K. S.; Lebl, M.; Krchnak, V. The
“One-Bead-One-Compound” Combinatorial Library Method. Chem.
Rev. 1997, 97, 411–448.
This work was supported financially by the University of
Berne, the Swiss National Science Foundation, and the Marie
Curie Training Network IBAAC.
(7) (a) Clouet, A.; Darbre, T.; Reymond, J.-L. A Combinatorial
Approach to Catalytic Peptide Dendrimers. Angew. Chem., Int. Ed. 2004,
43, 4612–4615. (b) Maillard, N.; Clouet, A.; Darbre, T.; Reymond, J.-L.
Combinatorial Libraries of Peptide Dendrimers: Design, Synthesis, On-
Bead High-Throughput Screening, Bead Decoding and Characteriza-
tion. Nat. Protoc. 2009, 4, 132–142. (c) Javor, S.; Delort, E.; Darbre, T.;
Reymond, J.-L. Peptide Dendrimer Enzyme Model with a Single
Catalytic Site at the Core. J. Am. Chem. Soc. 2007, 129, 13238–13246.
(8) Kofoed, J.; Darbre, T.; Reymond, J.-L. Artificial aldolases from
peptide dendrimer combinatorial libraries. Org. Biomol. Chem.
2006, 3268–3281.
(9) Uhlich, N. A.; Sommer, P.; B€uhr, C.; Sch€urch, S.; Reymond, J.-L.;
Darbre, T. Remote Control of Bipyridine-Metal Coordination with a
Peptide Dendrimer. Chem. Commun. 2009, 6237–6239.
(10) (a) Sommer, P.; Uhlich, N. A.; Reymond, J.-L.; Darbre, T. A.
Peptide Dendrimer Model for Vitamin B12 Transport Proteins. Chem-
BioChem 2008, 9, 689–693. (b) Uhlich, N.; Natalello, A.; Kadam, R. U.;
Doglia, S. M.; Reymond, J.-L.; Darbre, T. Structure and Binding of
Peptide Dendrimer Ligands to Vitamin B12. ChemBioChem 2010,
11, 358–365.
(11) (a) Kolomiets, E.; Johansson, E. M. V.; Renaudet, O.; Darbre,
T.; Reymond, J.-L. Neoglycopeptide Dendrimer Libraries as a Source of
Lectin Binding Ligands. Org. Lett. 2007, 9, 1465–1468. (b) Kolomiets,
E.; Swiderska, M. A.; Kadam., R. U.; Johansson, E. M. V.; Jaeger, K.-E.;
Darbre, T.; Reymond, J.-L. Glycopeptide dendrimers with high affinity
for the fucose binding lectin PA-IIL from Pseudomonas Aeruginosa.
ChemMedChem 2009, 4, 562–569.
(12) Maillard, N.; Darbre, T.; Reymond, J.-L. Identification of
Catalytic Peptide Dendrimers by “Off-Bead” in Silica High-Throughput
Screening of Combinatorial Libraries. J. Comb. Chem. 2009,
11, 667–675.
’ REFERENCES
(1) (a) Newkome, G. R.; Moorefield, C. N.; V€ogtle, F. Dendritic
Molecules: Concepts, Synthesis, Applications; VCH: Weinheim, 2001;
(b) Kofoed, J.; Reymond, J.-L. Dendrimers as artificial enzymes. Curr.
Opin. Chem. Biol. 2005, 9, 656–664. (c) Lee, C. C.; MacKay, J. A.; Frechet,
J. M. J.; Szoka, F. C. Designing dendrimers for biological applications. Nat.
Biotechnol. 2005, 23, 1517–1526. (d) Astruc, D.; Boisselier, E.; Ornelas, C.
Dendrimers Designed for Functions: From Physical, Photophysical, and
Supramolecular Properties to Applications in Sensing, Catalysis, Molecular
Electronics, Photonics, and Nanomedicine. Chem. Rev. 2010, 110,
1857–1959.
(2) (a) Kreiter, R.; Kleij, A. W.; Klein Gebbink, R. J. M.; van Koten,
G. Dendritic Catalysts. In Topics in Current Chemistry; Springer-Verlag:
Berlin, 2001; Vol. 217, pp 163ꢀ199. (b) Garcia-Martinez, J. C.;
Lezutekong, R.; Crooks, R. M. Dendrimer-Encapsulated Pd Nanopar-
ticles as Aqueous, Room-Temperature Catalysts for the Stille Reaction.
J. Am. Chem. Soc. 2005, 127, 5097–5103. (c) Lu, S. M.; Alper, H.
Intramolecular Carbonylation Reactions with Recyclable Palladium-
Complexed Dendrimers on Silica: Synthesis of Oxygen, Nitrogen, or
Sulfur-Containing Medium Ring Fused Heterocycles. J. Am. Chem. Soc.
2005, 127, 14776–14784. (d) Tang, W.-J.; Yang, N.-F.; Yi, B.; Deng, G.-
J.; Huang, Y.-Y.; Fan, Q.-H. Phase selectively soluble dendrimer-bound
osmium complex: a highly effective and easily recyclable catalyst for
olefin dihydroxylation. Chem. Commun. 2004, 1378–1379. (e) Liang,
C. O.; Helms, B.; Craig, J.; Frechet, J. M. J. Dendronized cyclocopoly-
mers with a radial gradient of polarity and their use to catalyze a difficult
esterification. Chem. Commun. 2003, 2524–2525.
(13) (a) Kofoed, J.; Nielsen, J.; Reymond, J.-L. Discovery of New
Peptide-based Catalysts for the Direct Asymmetric Aldol Reaction.
Bioorg. Med. Chem. Lett. 2003, 13, 2445–2447. (b) Martin, H. J.; List,
B. Mining sequence space for asymmetric aminocatalysis. N-Terminal
prolylpeptides efficiently catalyze enantioselective aldol and Michael
reactions. Synlett 2003, 1901–1902. (c) Tang, Z.; Yang, Z.-H.; Cun, L.-
F.; Gong, L.-Z.; Mi, A.-Q.; Jiang, Y.-Z. Small Peptides Catalyze Highly
Enantioselective Direct Aldol Reactions of Aldehydes with Hydroxya-
cetone: Unprecedented Regiocontrol in Aqueous Media. Org. Lett.
2004, 6, 2285–2287. (d) Notz, W.; List, B. Catalytic Asymmetric
Synthesis of anti-1,2-Diols. J. Am. Chem. Soc. 2000, 122, 7386–7387.
(e) Shi, L.-X.; Sun, Q.; Ge, Krattiger, P.; Kovasy, R.; Revell, J. D.; Ivan, S.;
Wennemers, H. Increased structural complexity leads to higher activity:
peptides as efficient and versatile catalysts for asymmetric aldol reac-
tions. Org. Lett. 2005, 7, 1101–1103.
(14) (a) Kofoed, J.; Reymond, J. L. A general method for designing
combinatorial peptide libraries decodable by amino acid analysis.
J. Comb. Chem. 2007, 9, 1046–52. (b) Kofoed, J.; Reymond, J. L.
Identification of protease substrates by combinatorial profiling on
TentaGel beads. Chem. Commun. 2007, 4453–4455. (c) Fluxa, V. S.;
Reymond, J. L. On-bead cyclization in a combinatorial library of 15,625
octapeptides. Bioorg. Med. Chem. 2009, 17, 1018–1025. (d) Dulery, V.;
Uhlich, N. A.; Maillard, N.; Fluxa, V. S.; Garcia, J.; Dumy, P.; Renaudet,
O.; Reymond, J. L.; Darbre, T. A cyclodecapeptide ligand to vitamin B12.
Org. Biomol. Chem. 2008, 6, 4134–4141. (e) Fluxa, V. S.; Maillard, N.;
Page, M. G.; Reymond, J. L. Bead diffusion assay for discovering
antimicrobial cyclic peptides. Chem. Commun. 2011, 1434–1436.
(15) Biswas, R.; Maillard, N.; Kofoed, J.; Reymond, J.-L. Comparing
dendritic with linear esterase peptides by screening SPOT arrays for
catalysis. Chem. Commun. 2010, 8746–8748.
(3) (a) Esposito, A.; Delort, E.; Lagnoux, D.; Djojo, F.; Reymond, J.-
L. Catalytic Peptide Dendrimers. Angew. Chem., Int. Ed. 2003,
42, 1381–1383. (b) Douat-Casassus, C.; Darbre, T.; Reymond, J.-L.
Selective Catalysis with Peptide Dendrimers. J. Am. Chem. Soc. 2004,
126, 7817–7826. (c) Delort, E.; Darbre, T.; Reymond, J.-L. A Strong
Positive Dendritic Effect in a Peptide Dendrimer-Catalyzed Ester
Hydrolysis Reaction. J. Am. Chem. Soc. 2004, 126, 15642–15643.
(d) Delort, E.; Nguyen-Trung, N.-Q.; Darbre, T.; Reymond, J.-L. Synthesis
and Activity of Histidine Containing Catalytic Peptide Dendrimers.
J. Org. Chem. 2006, 71, 4468–4480. (e) Darbre, T.; Reymond, J.-L.
Peptide Dendrimers as Artificial Enzymes, Receptors, and Drug-Delivery
Agents. Acc. Chem. Res. 2006, 39, 925–934.
(4) (a) Berkessel, A.; Herault, D. A. Discovery of peptide-zirconium
complexes that mediate phosphate hydrolysis by batch screening of a
combinatorial undecapeptide library. Angew. Chem., Int. Ed. 1999,
38, 102–105. (b) Copeland, G. T.; Miller, S. J. Selection of enantiose-
lective acyl transfer catalysts from a pooled peptide library through a
fluorescence-based activity assay. An approach to kinetic resolution of
secondary alcohols of broad structural scope. J. Am. Chem. Soc. 2001,
113, 6496–6502.
(5) (a) Krattiger, P.; McCarthy, C.; Pfaltz, A.; Wennemers, H.
Catalyst-Substrate Coimmobilization. A Strategy for Catalysts Discov-
ery in Split.and.Mix Libraries. Angew. Chem., Int. Ed. 2003,
42, 1722–1724. (b) Bauke-Albada, H.; Liskamp, R. M. J. TAC-Scaf-
folded Tripeptides as Artificial Hydrolytic Receptors. A Combinatorial
Approach Toward Esterase Mimics. J. Comb. Chem. 2008, 10, 814–824.
(6) (a) Furka, A.; Sebestyꢀen, F.; Asgedom, M.; Dibꢀo, G. General
method for rapid synthesis of multicomponent peptide mixtures. Int. J.
Pept. Protein Res. 1991, 37, 487–493. (b) Lam, K. S.; Salmon, S. E.;
Hersh, E. M.; Hruby, V. J.; Kazmierski, W. M.; Knapp, R. J. A new type of
synthetic peptide library for identifying ligand-binding activity. Nature
1991, 354, 82–84. (c) Houghten, R. A.; Pinilla, C.; Blondelle, S. E.;
Appel, J. R.; Dooley, C. T.; Cuervo, J. H. Generation and use of synthetic
peptide combinatorial libraries for basic research and drug discovery.
(16) Yang, Y.; Babiak, P.; Reymond, J.-L. New Monofunctionalized
Fluorescein Derivatives for the Efficient High-Throughput Screening of
Lipases and Esterases in Aqueous Media. Helv. Chim. Acta 2006,
89, 404–415.
319
dx.doi.org/10.1021/co200006z |ACS Comb. Sci. 2011, 13, 310–320

ACS Combinatorial Science
RESEARCH ARTICLE
(17) Blankemeyer-Menge, B.; Nimtz, M.; Frank, R. An efficient
method for anchoring Fmoc-amino acids to hydroxyl-functionalized
solid supports. Tetrahedron Lett. 1990, 31, 1701–1704.
(18) Attardi, M. E.; Falchi, A.; Taddei, M. A sensitive visual test for
detection of OH groups on resin. Tetrahedron Lett. 2000,
41, 7395–7399.
(19) (a) Hancock, W. S.; Battersby, J. E. A new micro-test for the
detection of incomplete coupling reactions in solid-phase peptide
synthesis using 2,4,6-trinitrobenzenesulfonic acid. Anal. Biochem.
1976, 71, 260–264. (b) Vojkovsky, T. Detection of secondary amines
on solid phase. Pept. Res. 1995, 8, 236–237.
320
dx.doi.org/10.1021/co200006z |ACS Comb. Sci. 2011, 13, 310–320