Artificial aldolases from peptide dendrimer combinatorial libraries

Article abstract of DOI:10.1039/b607342e

Peptide dendrimers were investigated as synthetic models for aldolase enzymes. Combinatorial libraries were prepared with aldolase active residues such as lysine and proline placed at the dendrimer core or near the surface. On-bead selection for aldolase activity was carried out using the dye-labelled 1,3-diketone 1a, suitable for covalent trapping of enamine-reactive side-chains, and the fluorogenic enolization probe 6. Aldolase dendrimers catalyzed the aldol reaction of acetone, dihydroxyacetone and cyclohexanone with nitrobenzaldehyde. Much like enzymes, the dendrimers exhibited strong aldolase activity in aqueous medium, but were also active in organic solvent. Dendrimer-catalyzed aldol reactions reached complete conversion in 3 h at 25 °C with 1 mol% catalyst and gave aldol products with up to 65% ee. A positive dendritic effect in catalysis was observed with both lysine and proline based aldolase dendrimer catalysts. The Royal Society of Chemistry 2006.

Full text of DOI:10.1039/b607342e

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Artificial aldolases from peptide dendrimer combinatorial libraries†
Jacob Kofoed, Tamis Darbre and Jean-Louis Reymond*
Received 23rd May 2006, Accepted 5th July 2006
First published as an Advance Article on the web 26th July 2006
DOI: 10.1039/b607342e
Peptide dendrimers were investigated as synthetic models for aldolase enzymes. Combinatorial libraries
were prepared with aldolase active residues such as lysine and proline placed at the dendrimer core or
near the surface. On-bead selection for aldolase activity was carried out using the dye-labelled
1,3-diketone 1a, suitable for covalent trapping of enamine-reactive side-chains, and the fluorogenic
enolization probe 6. Aldolase dendrimers catalyzed the aldol reaction of acetone, dihydroxyacetone and
cyclohexanone with nitrobenzaldehyde. Much like enzymes, the dendrimers exhibited strong aldolase
activity in aqueous medium, but were also active in organic solvent. Dendrimer-catalyzed aldol
reactions reached complete conversion in 3 h at 25 ◦C with 1 mol% catalyst and gave aldol products
with up to 65% ee. A positive dendritic effect in catalysis was observed with both lysine and proline
based aldolase dendrimer catalysts.
Introduction
The aldol reaction is one of the most important C–C bond forming
reactions in organic synthesis. The reaction can be catalyzed in
water by enzymes, which operate by an enamine (type I) or an
enolate (type II) mechanism.1 Type I aldolase mimics are known
based on catalytic antibodies acting primarily in water,2 and small
molecule organocatalysts such as proline, which are usually more
active in organic solvent.3,4 Zn–proline is a type II aldolase mimic
acting in water.5
We asked the question whether a synthetic macromolecular
enzyme model could be obtained that would reproduce the type I
aldolase activity in an aqueous environment, working on the basis
of peptide dendrimers as a framework. Dendrimers are tree-like
macromolecules under investigation for various uses in technology
and medicine.6,7 We recently reported artificial esterases on a
peptide dendrimer basis,8 showing that these macromolecules
are suitable as catalysts in an aqueous environment.9 A recent
report showed that multivalent prolines at the surface of a
poly(propylene) imine dendrimer exhibit comparable activities
to proline itself for aldolization in organic solvent, however
the study did not address the issue of aqueous catalysis as an
enzyme model.10 Herein, we report the discovery of aldolase
peptide dendrimers by functional selection from dendrimer
combinatorial libraries using probes specific for aldolase active
residues. The aldolase dendrimers are shown to operate by a
type I mechanism in water. The most active catalysts, such as
L2D1, display multiple N-terminal prolines or primary amines as
catalytic groups (Scheme 1). The activity of proline residues in the
dendritic multivalent display is enhanced compared to monovalent
catalysts.
Department of Chemistry and Biochemistry, University of Berne, Freiestrasse
3, CH-3012, Berne, Switzerland. E-mail: jean-louis.reymond@ioc.unibe.ch
† Electronic supplementary information (ESI) available: HPLC chro-
matograms, NMR and MS spectra, and kinetic data. See DOI:
10.1039/b607342e
Scheme 1 Peptide dendrimer catalyzed aldol reaction. Conditions: 1 mM
dendrimer in DMSO–acetone (4 : 1, v/v), 36 h, rt.
3268 | Org. Biomol. Chem., 2006, 4, 3268–3281
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
dendrimers with multiple histidines.8d Thus, lysine was used as
one of the variable amino acids for positions A5 and A7, and
proline as one of the variable amino acids for positions A6 and
A8. The other amino acids were distributed evenly in the available
variable positions. In this case, the library was left with a free N-
terminus after removal of the last Fmoc protecting groups and the
side chains deprotected, giving library L2.
Four different probes were prepared for functional screening
of the dendrimer libraries. The first probe was the dye-labelled
1,3-diketone 1a. Such 1,3-diketones react rapidly with lysine side-
chains to form a stable enaminone, and the method has been
used to select catalytic antibodies by reactive immunization with
hapten 1b (Scheme 2).2 The probe was obtained by alkylation of
acetylacetone with nitrobenzyl bromide at the terminal position
via the dienolate to intermediate 2, hydrogenation of the nitro
group to the unstable aniline 3, and coupling with monoester
5, obtained by acylation of Disperse Red 1 (4) with succinic
anhydride.
Results and discussion
Combinatorial discovery of aldolase dendrimers
We recently reported a combinatorial approach to peptide den-
drimers based on functional screening of split-and-mix11 libraries
on a solid support.12 The peptide dendrimers in these libraries
contain eight variable amino acid positions along three successive
branches. Using four different amino acids per variable position
results in a combinatorial library of 48 = 65 536 members (Fig. 1).12
We reasoned that focused dendrimer libraries incorporating the
essential features of known aldolase enzymes and catalysts might
contain catalytic aldolase dendrimers, and that these functional
dendrimers could be discovered using appropriate probes for
screening.
Fig. 1 Combinatorial split-and-mix synthesis of peptide dendrimers.
᭹ = L-2,3-diaminopropanoic acid (Dap). Library L1 is acetylated at
the N-terminus. Using 4 amino acids per variable position Ai (i = 1–8)
gives 48 = 65 536 members. Conditions: a) suspend the whole resin batch
in DMF–DCM (2 : 1, v/v), mix via nitrogen bubbling for 15 min, and
split the batch in four equal portions 1–4; b) in each portion x = 1–4:
i
2.5 eq. Fmoc–A –OH, 2.5 eq. PyBOP, 6.0 eq. DIEA, DMF, 2g × 60 min
x
(where g = generation number), then DMF–piperidine (4 : 1, v/v), 2 ×
10 min; c) same as b) with Fmoc–Dap(Fmoc)–OH.
A first library (L1) was designed to mimic type I aldolase
enzymes and catalytic antibodies by using lysine residues as one
of the variable residues for the core positions A1 and A3 (Fig. 1).13
The core also featured hydrophobic and aromatic residues to
create a hydrophobic microenvironment, while most other charged
and polar residues were placed near the surface. The library was
acetylated at the N-terminus and the side chains deprotected to
yield library L1. A second library (L2) featured multiple catalytic
residues at the surface (free N-termini, N-terminal proline, or
lysine), with the aim of exploiting a possible dendritic effect in
aldolase catalysis similar to that observed with esterase peptide
Scheme 2 Aldolization mechanisms and enaminone formation with
diketone probe 1a. Synthesis: (a) 2.1 eq. LDA, THF, HMPA, −78 ◦C,
2 h, 4-nitrobenzyl bromide (40%); (b) H2, Pd/C (10 mol%), DCM, rt,
18 h, not isolated; (c) succinic anhydride, DMAP, Et3N, DCM, rt, 5 h,
(33%); (d) 3, EDC, HOBt, DCM, 0 ◦C to rt, 18 h, (95%).
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 3268–3281 | 3269
©

View Article Online
The dihydroxyacetone derivative 6 was used as a second
screening probe (Scheme 3). Ketone 6 undergoes enolization in
the presence of aldolase catalysts by rate-limiting deprotonation
of the a-carbon atom, followed by b-elimination of the fluorescent
product umbelliferone 7. The probe detects both enamine (type I)
aldolase catalysts such as catalytic antibody 38C2 and enolate
(type II) aldolase catalyst such as Zn(Pro)2.14 Ketone 6 was
prepared from umbelliferone 7 by alkylation with allylbromide
to form allyl ether 8, dihydroxylation to form glycerol ether 9,
protection of the primary alcohol to silyl ether 10, oxidation of
the secondary alcohol to ketone 11, and acidic deprotection and
purification by reverse-phase HPLC.
undergo a fluorogenic retro-aldolization reaction and are suitable
for detecting aldolase antibodies by fluorescence (Scheme 4).
Indeed, the aldol reaction between a ketone and an aldehyde is
a near-equilibrium process, and the direction of the reaction is
given by the concentration of the reactants,17 allowing the use of
retro-aldolization as a test for aldol catalysis.
Scheme 4 Fluorogenic probes for retro-aldolization.
Activity screening was carried out several times for each of
the probes. Each screening used a 50 mg batch of the library
corresponding to approximately 62 500 beads,18 ensuring 60%
coverage of the library.19 Since the beads were acidic after removal
of the protecting group with trifluoroacetic acid, they were
equilibrated several times with aqueous phosphate buffer saline
pH 7.4 (PBS) and DMF for neutralization before each assay.
The binding assay with diketone 1a was carried out in a 1 :
1 mixture of PBS and dimethylsulfoxide (DMSO), followed by
washing. The cosolvent was necessary to solubilize the probe,
and the conditions were also favorable for aldol catalysis. The
concentration of 1a was adjusted to 50 lM and the incubation
time to 30 min to produce only very few beads (ca. 50, 0.1%) per
screening batch (Fig. 2). There was no detectable staining with up
to 10 mM Disperse Red (3) alone under the screening conditions,
showing the specificity of the staining reaction with the diketone
probe 1a. The beads stained by diketone 1a in library L1 contained
dendrimerswithoneorthreelysineresidues, withastrongselection
for lysine at position A1, consistent with enaminone formation
Scheme
3 Fluorogenic enolization probe 6. Synthesis conditions:
(a) allyl bromide, K2CO3, acetone, reflux, 18 h, (85%); (b) NMO, OsO4,
t-BuOH–H2O (2 : 1, v/v), rt, 18 h, (95%); c) TBDMSCl, imidazole,
DCM–DMF (3 : 1, v/v), rt, 18 h, (58%); d) oxalyl chloride, DMSO, Et3N,
1 h, −78 ◦C, 78% yield; e) TFA–H2O (9 : 1, v/v), rt, 1 h, (33%).
Two additional aldolase screening assays were carried out
using the known retro-aldolization probes 1215 and 14,16 which
were prepared using the published procedures. Theses substrates
Fig. 2 High throughput screening of the libraries. Left: diketone probe 1b, the beads containing binding sequences are deep red, beads with inactive
sequences are colorless. Right: enolization probe 6, the beads containing catalytic sequences appear in light blue, beads with inactive sequences are black.
3270 | Org. Biomol. Chem., 2006, 4, 3268–3281
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
Table 1 Peptide dendrimer sequencesa identified by amino acid analysis of active beads from the combinatorial libraries L1 and L2. Sequences L1D1–
L1D7 and L2D1–L2D9; hits for the diketone 1a with library L1 and L2. Sequences L2K1–L2K8; hits for the ketone 6 with library L2
Assay
Dendrimer
A8
A7
A6
A5
A4
A4
A2
A1
Enaminone formation with 1ab
L1D1
L1D2
L1D3
L1D4
L1D5
L1D6
L1D7
L2D1
L2D2
L2D3
L2D4
L2D5
L2D6
L2D7
L2D8
L2D9
L1
L2K1
L2K2
L2K3
L2K4
L2K5
L2K6
L2K7
L2K8
Leu
Glu
Glu
Leu
Glu
Arg
Arg
Pro
Pro
Glu
Thr
Glu
Ser
Ile
Glu
Glu
Arg
Arg
Arg
Glu
Leu
Pro
Pro
Thr
Glu
Ser
Thr
Tyr
Ile
Ser
Phe
Phe
Trp
Phe
Phe
Phe
Tyr
Ile
Tyr
Ile
Tyr
Tyr
b-Ala
Ile
Val
Gly
Ala
Gly
Lys
Lys
Val
Leu
Ala
Leu
Ala
Ala
Gly
Gly
Gly
Gly
Ser
Ser
Asp
Phe
Ser
Ser
Asp
Ile
Tyr
Phe
Tyr
Phe
Phe
Phe
Tyr
Phe
Lys
Lys
Gly
Ala
Lys
Lys
Lys
Gly
Ala
Val
Gly
Val
Gly
Val
Gly
Gly
Thr
Thr
Ile
Tyr
Tyr
Ile
Lys
Lys
Lys
Lys
Lys
Lys
Lys
Lys
Lys
Thr
Thr
Tyr
Thr
Lys
Lys
Asp
Lys
Lys
Lys
Arg
Arg
Asp
Ser
Pro
Ser
Glu
Glu
Pro
Ser
Ile
Catalysis of direct b-elimination of ketone 6c
No hits with this library were observed
Pro
Thr
Pro
Pro
Pro
Ser
Lys
Arg
Lys
Lys
Lys
Asp
Lys
Lys
Ser
Ser
Thr
Thr
Ser
Pro
Ser
Glu
His
Arg
Lys
His
Asp
His
Arg
Asp
b-Ala
b-Ala
b-Ala
Tyr
Ala
Val
Ala
Gly
Gly
Gly
Val
Gly
Ile
Val
Gly
Gly
Val
Leu
Gly
Leu
Ala
Tyr
Phe
Phe
Ile
Tyr
Tyr
Tyr
Ile
b-Ala
b-Ala
Ile
Pro
Glu
a Dendrimer structure according to Fig. 1. Dendrimers marked bold were resynthesized by Fmoc–SPPS and purified. b In library L1 N-termini were
acetylated; in library L2 N-termini were free amine. Screening conditions: beads shaken in DMSO–PBS buffer (pH = 7.4) (1 : 1, v/v) with 50 lM diketone
1a for 30 min, 25 ◦C. Hits are deep red (see Fig. 2). c In library L1 N-termini were acetylated; in library L2 N-termini were free amine. Beads soaken in
aqueous bicine buffer pH = 8.50 with 1% acetonitrile with 100 lM probe 6 for 40 min, 25 ◦C. Hits are light blue under radiation with a TLC UV-lamp
set at 356 nm (see Fig. 2). Statistically significant selections for L1: Lys at A1, Phe(Trp) at A4, Arg and Glu at A6/A8. L2 with 1a: Lys at A7. L2 with 6:
Pro at A8.
(Table 1). The positive selection of lysine residues by probe 1a
at position A1 was confirmed by sequencing of 7 beads from
library L1 picked at random. In this case residues at positions
A1 corresponded to statistical distribution (Lys: 21%, Ala: 21%,
Gly: 43%, Val: 15%). The situation was similar for the hits from
library L2 featuring surface amine residues. The selection of lysine
residues at position A5 and A7 was surprising in this case since N-
terminal amino groups might also engage in enaminone formation.
As for L1, random picking of beads from L2 gave a statistically
even representation of amino acids at positions A5 and A7 (6 beads,
Lys: 25%, His 21%, Arg: 31%, Asp: 23%).
residues, in agreement with the known aldolase reactivity of N-
terminal prolyl peptides.4d Lysine was also often present.
The retro-aldolization probes 12 and 14 were tested at 100 lM
concentration in aqueous borate buffer pH 9 with 5% acetonitrile,
which are the conditions under which catalytic antibodies15 show
activity. The screening protocol for probe 6 above was used. None
of the beads gave any detectable reaction with these probes in
either of the libraries.
Aldol catalysis with peptide dendrimers
Hits from the combinatorial experiment were synthesized on Tent-
agel at 0.25 mmol g−1, deprotected and cleaved, and purified by
preparative HPLC (Table 2). The dendrimers were investigated for
catalysis of the aldol reaction of acetone with nitrobenzaldehyde 16
to give aldol 17 (Scheme 5).4 The reaction was first tested in organic
solvent (DMSO–acetone 4 : 1, v/v), typical for organocatalysis of
this reaction using 0.1 M aldehyde and 1 mM (1 mol%) peptide
dendrimer. While dendrimers from library L1 with core lysines
were not active, all the peptide dendrimers from library L2 were
catalytically active (Table 3). The four dendrimers containing N-
terminal proline residues (L2D1, L2D7, L2K4, L2K7) were clearly
more active than the others, and showed moderate enantioselec-
tivity (up to 61% ee for (R)-17 with dendrimer L2D1).
The enolization probe 6 was assayed in aqueous bicine buffer
pH = 8.50 with 1% acetonitrile at 100 lM, under which conditions
it was fully soluble. The beads were first equilibrated with bicine
buffer, washed, and suspended in a freshly prepared solution
of probe 6. The beads were then plated out onto a silica gel
plate, which absorbed most of the solution and left each bead
well separated from the others as a free-standing microreactor.
Under these conditions the fluorescent product umbelliferone
(7) cannot diffuse away and remains inside the polymer bead,
ensuring that it can be detected by fluorescence. Library L1
gave no detectable fluorescent beads with this assay. This was
surprising, since ketone 6 reacts well with catalytic antibody 38C2
in an active site lysine specific reaction.14 Screening of library L2
featuring surface catalytic residues produced fluorescent beads,
which appeared slowly over the course of the first 40 min (Fig. 2).
The beads were picked and sequenced as above (Table 1). The
majority of active beads with probe 6 featured N-terminal proline
The dendrimers were also tested with acetone and aldehyde 17
under aqueous conditions (acetone–water 1 : 1, pH 8.5) mimicking
those of an enzyme-catalyzed reaction. While the dendrimers from
library L1 were again not active, the dendrimers from library L2
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 3268–3281 | 3271
©

View Article Online
Table 2 Yield and molecular weight of the synthesized peptide dendrimers from L1 and L2a
Yield as TFA-salts/mg (%) Expected mass [M + H]+ Observed mass [M + H]+
Origin
Dendrimer
Enaminone formation with 1a AcEY·RT·FK·SK L1D5 18 mg, 9%
AcRY·EY·FK·SK L1D6 21 mg, 10%
5088.35
5444.65
4046.00
3942.30
3846.15
4064.36
4060.28
4010.42
4228.05
5088.38
5444.25
4045.88
3942.12
3846.13
4064.38
4060.38
4010.75
4228.50
PK·PK·YL·IG
EK·SKYA·FV
SK·SK·YG·FG
PK·ER·bAG·FV
PK·TH·TG·FV
PK·SR· bAV.YL
EK·ED·IG·YA
L2D1 86 mg, 25%
L2D5 20 mg, 8%
L2D6 33 mg, 13%
L2D7 19 mg, 7%
L2K4 11 mg, 4%
L2K7 28 mg, 10%
L2K8 17 mg, 6%
b-Eliminaton of ketone 2
a The peptide dendrimers were synthesized on NovasynTGR resin (200 mg, 0.25 mmol g−1) using Fmoc SPPS and were purified by RP-HPLC.
Table 3 Activities of the peptide dendrimers for the aldol reaction of acetone with nitrobenzaldehyde 16
Dendrimer
Conditionsa
Time/h
Conversion (%)b
ee (%)c
PK·PK·YL·IG
EK·SKYA·FV
SK·SK·YG·FG
PK·ER·bAG·FV
PK·TH·TG·FV
PK·SR· bAV.YL
EK·ED·IG·YA
L2D1
L2D5
L2D6
L2D7
L2K4
L2K7
L2K8
DMSO–acetone (4 : 1, v/v)
aq. buffer–acetone (1 : 1, v/v)
DMSO–acetone (4 : 1, v/v)
aq. buffer–acetone (1 : 1, v/v)
DMSO–acetone (4 : 1, v/v)
aq. buffer–acetone (1 : 1, v/v)
DMSO–acetone (4 : 1, v/v)
aq. buffer–acetone (1 : 1, v/v)
DMSO–acetone (4 : 1, v/v)
aq. buffer–acetone (1 : 1, v/v)
DMSO–acetone (4 : 1, v/v)
aq. buffer–acetone (1 : 1, v/v)
DMSO–acetone (4 : 1, v/v)
aq. buffer–acetone (1 : 1, v/v)
36
3
36
3
36
3
72
3
36
3
36
3
36
3
69
>99
<5
14
<5
11
37
>99
68
92
95
>99
16
34
61
<5
N.a.
N.a.
N.a.
N.a.
41
<5
35
<5
46
<5
N.a.
<5
a Conditions: 100 mM aldehyde, 1 mM peptide dendrimer in the indicated solvent mixture. The aqueous buffer is bicine pH 8.5. The DMSO was buffered
x eq. of NMM (where x = number of TFA–amine salts for the dendrimer tested). b Conversion was measured by RP-HPLC at 254 nm. c Enantiomeric
excess of (R)-17 was determined by HPLC at 254 nm on a Daicel Chiralpak AS column.
aldol reaction with 1 mol% catalyst was complete in 60 min at room
temperature (>99% conversion), compared to 40–95% conversion
after 36 h in organic solvent (Table 3). Unfortunately, the reactions
were not enantioselective under these aqueous conditions.
The aldol reaction of dihydroxyacetone with 2-bromobenzal-
dehyde 18 to give aldol 19 was investigated next. Hydroxyke-
tones are typical substrates for enzyme-catalyzed transaldolase
processes.1 The reaction was carried out under aqueous conditions
(MeOH–water 1 : 1, pH 8.5) using 0.1 M aldehyde, 5 M dihydro-
xyacetone and 1 mM (1 mol%) peptide dendrimer.20 The L1
dendrimers were once more inactive. However, the L2 dendrimers
showed good to moderate activity for this aldol reaction (Table 4).
In fact, only the dendrimers derived from screening with enoliza-
tion probe 2 showed significant activity, while those identified with
the 1,3-diketone probe 1a were only weakly active. There was no
significant diastereoselectivity.
The dihydroxyacetone aldolase reactivity of the selected den-
drimers L2K4, L2K7 and L2K8 might be a general-base type
reactivity which is much weaker in the other dendrimers. Indeed,
we showed recently that aldol reactions of dihydroxyacetone in
water are catalyzed by general bases not capable of enamine
formation such as N-methylmorpholine, while typical enamine-
forming catalysts such as L-proline are poor catalysts for this
substrate.14
Scheme 5 Direct aldol reactions investigated with aldolase peptide
dendrimers.
were active and showed much stronger activity than in organic
The selection of dihydroxyacetone aldolase reactive dendrimers
during library screening with probe 6 could not be traced back to a
solvent, particularly for dendrimers with N-terminal proline
residues. With three dendrimers (L2D1, L2D7 and L2K7), the
3272 | Org. Biomol. Chem., 2006, 4, 3268–3281
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
Table 4 Activities of the peptide dendrimers for the aldol reaction of
dihydroxyacetone with bromobenzaldehyde 18.a Dendrimers L1D5 and
L1D6 were not active for this reaction
lectivity. Very high conversions were also observed with dendrimer
L2K8 (EK·ED·IG·YA) featuring only primary amines as possible
catalytic groups, namely the lysine side chains and the N-terminal
amino groups. The reactions showed moderate diastereo- and
enantioselectivities. These results were surprising to us because
of the lack of activity with acetone with these dendrimers under
the same conditions, suggesting a substrate specific interaction
with the ketone, which might originate from hydrophobic binding
of the cyclohexanone as compared to acetone.
Cyclohexanone is not miscible with water, allowing this aldol
reaction to be run as an emulsion, conditions which seem
to be favorable for stereoselective organocatalysis as suggested
by recent reports.4h The peptide dendrimers might form mi-
celles under these conditions in which the reaction takes place.
The most promising dendrimer L1D5, L2D1 and L2K8 were
studied in water–cyclohexanone emulsion. Indeed, we observed
enantiomeric excess in the aldol product under these aqueous
conditions with dendrimer L2K8 displaying multiple primary
amines. On the other hand, dendrimer L1D5 with core lysines
showed negligible stereoselectivity.
Dendrimer
Conversion (%)b (anti : syn)ratiob
PK·PK·YL·IG
EK·SK·YA·FV
SK·SK·YG·FG
PK·ER·bAG·FV
PK·TH·TG·FV
PK·SR· bAV.YL
EK·ED·IG·YA
L2D1
L2D5
L2D6
L2D7
L2K4
L2K7
L2K8
7 (1.0 : 1.0)
8 (1.0 : 1.5)
16 (1.0 : 1.3)
<5 (n.a. : n.a.)
40 (1.0 : 1.4)
74 (1.0 : 1.5)
36 (1.0 : 1.8)
a Conditions: 5 M dihydroxyacetone, 100 mM aldehyde, 1 mM peptide
dendrimer in methanol–aq. bicine buffer pH 8.5 (1 : 1, v/v) 66 h,
room temperature. b Conversion and diastereomeric excess of syn-19 was
measured by RP-HPLC at 254 nm.
specific reactivity with this probe. Indeed, all dendrimers catalyzed
the fluorogenic reaction of probe 6 (see the ESI). Although the
three dendrimers derived from the screening with 6 are the most
reactive, the difference is less than 2-fold and is not sufficient
to explain their reactivity with dihydroxyacetone. Probe 6 shows
relatively low KM values with the dendrimers, which probably
indicates hydrophobic substrate–dendrimer interactions. These
interactions should be much weaker or even absent with the more
polardihydroxyacetone, resultinginamoredifferentiatedbehavior
of the dendrimers with this substrate.
Cyclohexanone, which was recently shown to be an excellent
substrate for organocatalyzed aldolization,4f was investigated
as a third substrate (Table 5). In this case dendrimer L1D5
(AcEY·RT·FK·SK), equipped with catalytic lysine residues at the
dendrimer core, was catalytically active, but showed no stereose-
Mechanistic investigations
The mechanism of the aldol reaction of acetone and cyclohex-
anone with aldehyde 16 was investigated. Four of the most active
aldolase dendrimers (L2D1, L2D7, L2K4 and L2K7) featured an
N-terminal Pro–Lys dyad. This catalytic dyad was therefore used
to prepare a regular dendritic series R2 of increasing generation
number to investigate a possible dendritic effect in catalysis.
A regular dendrimer series R1 featuring the known4l catalytic
dipeptide Pro–Thr in its branches was also prepared (Table 6).
All dendritic peptides of the regular series R1 and R2 were
Table 5 Activities of the peptide dendrimers for the aldol reaction of cyclohexanone with nitrobenzaldehyde 16 in DMSO and aqueous buffera
DMSO
Aqueous buffer (pH 8.5)
Catalyst/dendrimer
Conversiona,b (anti : syn ratio)
eeb (anti/syn, %)
Conversionc,b (anti : syn ratio)
eeb (anti/syn, %)
L-Prolined
95 (1.4 : 1.0)
32 (1.0 : 2.2)
18 (1.0 : 1.3)
96 (1.0 : 2.0)
97 (1.5 : 1.0)
26 (1.0 : 1.9)
88 (1.0 : 1.5)
98 (1.0 : 2.4)
70/80
14/18
<5/<5
40/12
22/16
64/24
14/16
50/28
AcEY·RT·FK·SK
AcRY·EY·FK·SK
PK·PK·YL·IG
SK·SK·YG·FG
PK·TH·TG·FV
PK·SR· bAV.YL
EK·ED·IG·YA
L1D5
L1D6
L2D1
L2D6
L2K4
L2K7
L2K8
32 (2.2 : 1.0)
94 (1.6 : 1.0)e
5/8
12/12
54 (1.9 : 1.0)
65/45
a Conditions: 100 mM aldehyde, 1 mM peptide dendrimer in DMSO–cyclohexanone (1 : 1, v/v) buffered with x eq. of NMM (where x = number of
TFA–amine salts for the dendrimer tested), 18 h, room temperature. b Conversion and diastereomeric excess of 20 was measured by RP-HPLC at 268 nm.
Enantiomeric excess of anti-20 and syn-20 respectively were measured by chiral phase HPLC at 268 nm on a ChiralPak OD-H column. c Conditions:
100 mM aldehyde, 1 mM peptide dendrimer in cyclohexanone–aq. bicine buffer pH 8.5 (1 : 1, v/v) 4 d, room temperature. d 30 mol% catalyst was used.
e Reaction time 18 h.
Table 6 Yield and molecular weight for synthesis of the regular series R1 and R2a
Origin
Dendrimer
Yield (as TFA-salts) (mg, %) Expected mass [M + H]+ Observed mass [M + H]+
Regular series Pro–Thr R1 PT·PT
PT·PT·PT
PT·PT·PT·PT
Regular series Pro–Lys R2 PK·PK
PK·PK·PK
PK·PK·PK·PK R2G3 18 mg, 20%
R1G1
R1G2 15 mg, 78%
R1G3 14 mg, 26%
R2G1
R2G2 17 mg, 58%
7 mg, 79%
696.78
1661.84
3591.97
777.99
1851.32
3997.99
697.38
1661.88
3591.84
778.50
1851.33
3998.00
9 mg, 85%
a The peptide dendrimers were synthesized on NovasynTGR resin (200 mg, 0.25 mmol g−1) using Fmoc SPPS and were purified by RP-HPLC.
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 3268–3281 | 3273
©

View Article Online
Table 7 Activities of the regular series of peptide dendrimers for the aldol reaction of acetone and cyclohexanone with nitrobenzaldehyde 16
Dendrimer
PT
Conditionsa
Time/h
Conversion (%)b
ee (%)c
R1G0
R1G1
R1G2
DMSO–acetone (4 : 1, v/v)
aq. buffer–acetone (1 : 1, v/v)
DMSO–acetone (4 : 1, v/v)
aq. buffer–acetone (1 : 1, v/v)
DMSO–acetone (4 : 1, v/v)
aq. buffer–acetone (1 : 1, v/v)
DMSO–cyclohexanone (1 : 1, v/v)
DMSO–acetone (4 : 1, v/v)
aq. buffer–acetone (1 : 1, v/v)
DMSO–cyclohexanone (1 : 1, v/v)
DMSO–acetone (4 : 1, v/v)
aq. buffer–acetone (1 : 1, v/v)
DMSO–cyclohexanone (1 : 1, v/v)
aq. buffer–cyclohexanone (1 : 1, v/v)
DMSO–acetone (4 : 1, v/v)
aq. buffer–acetone (1 : 1, v/v)
DMSO–cyclohexanone (1 : 1, v/v)
aq. buffer–cyclohexanone (1 : 1, v/v)
DMSO–acetone (4 : 1, v/v)
aq. buffer–acetone (1 : 1, v/v)
DMSO–cyclohexanone (1 : 1, v/v)
aq. buffer–cyclohexanone (1 : 1, v/v)
DMSO–acetone (4 : 1, v/v)
36
3
36
3
36
3
18
36
3
18
36
3
18
2
36
3
18
2
36
3
18
2
36
3
18
2
20
15
80
68
N.a.
<5
44
<5
20
<5
54/26
34
PT·PT
PT·PT·PT
>99
>99
16 (1.3 : 1.0)
>99
>99
26 (1.4 : 1.0)
8
14
<5
<5
27
41
<5
<5
87
93
PT·PT·PT·PT
R1G3
R2G0
<5
59/20
N.a.
<5
N.a.
N.a.
52
PK
PK·PK
R2G1
R2G2
R2G3
<5
N.a.
N.a.
49
PK·PK·PK
PK·PK·PK·PK
<5
32 (1.0 : 2.2)
43 (2.1 : 1.0)
81
>99
56 (1.0 : 2.5)
96 (1.2 : 1.0)
46/8
2/16
60
<5
40/22
30/22
aq. buffer–acetone (1 : 1, v/v)
DMSO–cyclohexanone (1 : 1, v/v)
aq. buffer–cyclohexanone (1 : 1, v/v)
a Conditions: 100 mM aldehyde, 1 mM peptide dendrimer in the indicated solvent mixture. The aqueous buffer is bicine pH 8.5, room temperature. The
DMSO was buffered x eq. of NMM (where x = number of TFA–amine salts for the dendrimer tested). b For aldol 17: conversion was measured by
RP-HPLC at 254 nm; for aldol 20: conversion and anti : syn ratio were measured by RP-HPLC at 268 nm. c For aldol 17: enantiomeric excess of (R)-17
was determined by HPLC at 254 nm on a Daicel Chiralpak AS column; for aldol 20: enantiomeric excess of anti-20 and syn-20 were determined by HPLC
at 268 nm on a Daicel Chiralpak OD–H column.
catalytically active in both organic solvent and aqueous conditions
with diastereo- and enantioselectivities comparable to those of the
library-derived dendrimers (Table 7). The R1 (Pro–Thr) series was
more active with acetone, while in the case of cyclohexanone the
R2 (Pro–Lys) series was more active, in agreement with the fact
that dendrimers from library L1 with core active site lysine residues
showed activity with this ketone.
The catalytic mechanism of the peptide dendrimer-catalyzed
aldol reactions with acetone and cyclohexanone should involve
enamine formation. Reductive alkylation of amino groups was
investigated with dendrimer R2G3 featuring both N-terminal
proline residues and core lysine residues. Incubation with acetone
and NaBH4 under the reaction conditions produced N-isopropyl
derivatives of the dendrimer. 1H NMR analysis showed that four
N-terminal prolines were alkylated in R2G3, as indicated by the
integral ratio between the two different a-H (11 : 4, proline–N-
isopropyl proline) and the appearance of a signal at ca. 1.2 ppm
corresponding to the isopropyl group. The formation of the im-
minium intermediate between acetone and the N-terminal proline
under the reaction conditions supports an enamine mechanism
for catalysis.14 A small extent of alkylation at lysine residues was
also observed. Nevertheless, an enamine at the e-amino group of
lysine is probably not significant with acetone because lysine itself
is not an efficient aldolase catalyst for acetone.
L-lysine with cyclohexanone in water and subsequent reduction
with NaBH4 gave approximately 50% alkylation of the e-amino
group. The same experiment with L-proline gave approximately
75% alkylation of the pyrrolidine nitrogen. A similar alkylation
experiment was attempted with dendrimer L1D5, however due to
1
the complexity of the H NMR spectrum it was not possible to
estimate the extent of alkylation. Nevertheless, N-alklyation of
both lysine and proline with cyclohexanone is consistent with an
enamine mechanism for both primary-amine and proline-based
catalysts, including our peptide dendrimers.
The reactions under aqueous conditions were followed by
HPLC to obtain a precise measure of catalytic efficiency (Fig. 3).
In the case of acetone, the activity per N-terminal proline residue
increased strongly with dendrimer size in the regular series R1 and
R2 up to the second generation dendrimers with four N-terminal
residues, with a stronger activity in the R1 (Pro–Thr) series
(Fig. 4A). Although the G3 dendrimers were slightly less active per
N-terminal residue than the G2 dendrimers, there was a further
increase in activity per N-terminal residue in the library derived G3
dendrimers L2D1 (PK·PK·YL·IG) and L2K7 (PK·SR·bAV·YL)
relative to the related regular series R2 (Pro–Lys). Dendrimer
L2K7 was even more active than R1G2 of the Pro–Thr series.
The higher activity of these library derived dendrimers suggests
that the hydrophobic core residues present in both dendrimers
play a role in enhancing catalysis.
The activity data shows that acetone aldol catalysis is most
efficient with N-terminal proline residues, while cyclohexanone
reacts preferentially with primary amino groups, as in the case
of dendrimers L1D5, L2D6 and L2K8. Incubation of N-a-acetyl-
The rate data for the dendritic series R2 was also investigated
with cyclohexanone since lysine containing dendrimers were
particularly active with this substrate (Fig. 3B and 4B). In this
3274 | Org. Biomol. Chem., 2006, 4, 3268–3281
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
Fig. 3 (A) Plot of concentration of 17 versus time in the peptide dendrimer catalyzed direct aldol reaction of 16 with acetone. Conditions: 100 mM
aldehyde 16, 1 mM dendrimer in acetone–aq. bicine buffer pH = 8.5 (1 : 1, v/v). Reaction was followed by RP-HPLC at 254 nm. (B) Plot of concentration
of 20 versus time in the peptide dendrimer catalyzed direct aldol reaction of 16 with cyclohexanone. Conditions: 100 mM aldehyde 16, 1 mM dendrimer
in aq. bicine buffer pH = 8.5–cyclohexanone (1 : 1. v/v). Reaction was followed by RP-HPLC at 268 nm.
Fig. 4 (A) Rate of aldol product 17 formation per N-terminal proline versus number of N-terminal prolines in the regular series R1 and R2 and catalytic
peptide dendrimers L2K7 and L2D1. Conditions are as in Fig. 3A. (B) Rate of aldol product 20 formation per N-terminal proline versus number of
N-terminal prolines in the regular series R1 and R2 and catalytic peptide dendrimers L2K7 and L2D1. Conditions are as in Fig. 3B.
case a positive dendritic effect in catalysis occurred up to the
third generation dendrimer R2G3, which had similar activity to
L2D1, the most active dendrimer under water emulsion conditions
with cyclohexanone. It should be noted that with cyclohexanone,
dendrimer R2G1 (featuring three core lysine residues and two N-
terminal prolines) was less active (<5% conversion, Table 7) than
the library derived dendrimer L1D5 with three core lysine residues
(32% conversion, Table 5), highlighting the particular activity of
the library-derived dendrimer.
Modulation of the pKa of the N-terminal proline in the
dendrimers might influence catalytic activity and thus explain the
positive dendritic effect observed in aldol catalysis. For example,
hydroxyproline, with a pKa of 9.73 for the pyrrolidine, is more effi-
cient than proline (pKa = 10.60) for aldol catalysis under aqueous
conditions.14 However, the apparent pKa of the N-terminal proline
residues was found to be largely independent of dendrimer size,
Fig. 5 Acid–base titration of peptide dendrimers R1G0–R1G3 at equimo-
and decreased from pKa = 8.43 in the Pro–Thr–NH2 dipeptide
lar concentration with respect to N-terminal prolines. pKa values obtained:
R1G0 to pKa = 8.13 in dendrimer R1G3 (Fig. 5). The much lower
R1G0: 8.43, R1G1: 8.39, R1G2: 8.21 and R1G3: 8.13.
pKa value for the N-terminal prolinamide compared to free proline
stabilizing the protonated pyrrolidine in free proline. However,
considering that R1G0 (Pro–Thr–NH2) catalyzes the aldol only
(pKa = 10.60) can be explained by the change from carboxylate
to carboxamide, which removes the electrostatic component
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 3268–3281 | 3275
©

View Article Online
slightly better than L-Proline, the small downward shift in pKa
between R1G0 (Pro–Thr) and R1G3 cannot explain the increased
aldolase activity in higher generation peptide dendrimers.
R2G3). The effect might be caused by the macromolecular nature
of the peptide dendrimer as enzyme model, such as a hydrophobic
microenvironment in the larger dendrimers favoring substrate
binding and/or enamine formation, or by bifunctional catalysis.
The present study clearly points to the pyrrolidine ring as the
residue of choice for catalyzing acetone aldolizations, and primary
amines (either N-termini or lysine side-chains) for catalyzing
cyclohexanone aldolizations in the peptide dendrimer enzyme
model. The multivalent dendritic display is advantageous for
catalytic efficiency, as evidenced by the positive dendritic effect
on reaction rates observed in regular peptide dendrimer series of
increasing generation number. On the other hand, the presence
of multiple, stereochemically non-equivalent catalytic sites might
be deleterious for obtaining stereoselectivity. Future experiments
towards aldolase peptide dendrimer enzyme models will use
analogs of library L1 with a single core catalytic residue containing
either a primary amine for cyclohexanone aldolization, or a
pyrroldine for reactions with acetone.
We propose that the positive dendritic effect observed in al-
dolase catalysis is caused by a hydrophobic effect either increasing
substrate binding or shifting the imine formation equilibrium
towards the reactive enamine, which both would be favorable for
catalysis. Such an effect might also explain the enhanced reactivity
of the library derived dendrimers L2D1 and L2K7 which both
feature a strongly hydrophobic core in comparison to the regular
series dendrimers R1 and R2. The results obtained could also
indicate bifunctional catalysis, with a free amino group involved
in enamine formation being assisted by another protonated amino
group for activation of the aldehyde towards aldolization. The fact
that only four alkylated prolines were observed in the iminium
trapping experiment corroborates this assumption. Bifunctional
catalysis is an accepted hypothesis for the proline catalyzed aldol
reaction supported by computational studies indicating that H-
bonding activation of the aldehyde electrophile may lower the
activation barrier for aldol addition by as much as 17 kcal mol−1.21
Experimental section
Conclusion
General
Peptide dendrimer libraries were screened for aldolase reactivity
using four different probes, including the 1,3-diketone 1a suitable
for enaminone formation with enamine reactive side chains,
the fluorogenic coumarin ether of dihydroxyacetone 6 selective
for enolization, and the two known retro-aldolase fluorogenic
substrates 12 and 14. Library L1, with core active-site lysines gave
positive hits with diketone 1a. Library L2 with catalytic residues
at the surface gave positive hits with diketone 1a and with probe
6. Most hit sequences showed N-terminal proline residues.
Peptide dendrimers selected from library L1 showed aldolase
activity with nitrobenzaldehyde and cyclohexanone as substrates.
The activity must be attributed to primary amines from lysine
side-chains at the dendrimer core since these are the only
aldolase catalytic groups available. Peptide dendrimers selected
from library L2 catalyzed the aldol reaction of nitrobenzaldehyde
with both acetone and cyclohexanone, under organic or aqueous
conditions. Aldol (S)-17 was formed from acetone with 61% ee
with dendrimer L2D1 (PK·PK·YL·IG) in DMSO, and anti-20
was formed from cyclohexanone with 65% ee with dendrimer
L2K8 (EK·ED·IG·YA) in water–cyclohexanone mixture. In the
case of acetone, the reaction was much faster in water (complete
conversion in 3 h at 25 ◦C with 1 mol% catalyst) but not
enantioselective. Dendrimers selected with the enolization probe 6
also showed good activity with dihydroxyacetone as the substrate,
which was not found in any of the other dendrimers in the study.
The reaction with dihydroxyacetone presumably involves general-
base reactivity via an enolate intermediate.
Reagents were purchased in the highest quality available from
Fluka, Sigma, Bachem, Novabiochem, NeoMPS or Aldrich.
All solvents used in reactions were bought in p.a. quality or
distilled and dried prior to use. Solvents for extractions were
distilled from technical quality. Sensitive reactions were carried
out under nitrogen or argon, the glassware being heated under
HV. Chromatographic purifications (flash) were performed with
silica gel 60 from Merck or Fluka (0.04
0.063 nm; 230
400 mesh ASTM). Preparative RP-HPLC (flow rate 100 mL min−1)
was performed with a Waters Delta Prep 4000 system with a
Waters Prepak Cartridge (500 g) as column and Waters 486
Tunable Absorbance Detector. Semi-preparative RP-HPLC (flow
rate 4 mL min−1) was performed with a Water 510 Pump operated
with a Waters Automated Gradient Controller and Jasco VU-
2075 Plus Detector on a Vydac 218 TP (1.0 × 25 cm) column.
Analytical RP-HPLC was performed on Waters 600E systems with
a Waters Atlantis (4.6 mm × 100 mm, dC18, 5 lm) column, UV
detection with Waters 996 photodiode array detector). Eluents
for all systems were: A: water and 0.1% TFA; D: acetonitrile,
water and TFA (3/2/0.1%). TLC monitoring was performed
with Alugram SIL G/UV254 silica gel sheets (Macherey-Nagel),
followed by coloration with cerium solution (10.5 g Ce(IV)sulfate,
21 g phosphomolybdic acid, 60 mL conc. H2SO4 in 900 mL
water), anisaldehyde stain and heating or observation under
a UV lamp. MS and HRMS analyses were provided by the
mass spectrometry service of the Department of Chemistry and
Biochemistry, University of Berne. 1H and 13C NMR spectra were
recorded on Bruker AC 300 (300 MHz) and DRX 500 or Avance
500 (500 MHz) instruments. Chemical shifts d are given in ppm,
coupling constants (J) in Hertz (Hz).
The reactions with acetone and cyclohexanone involve an
enamine intermediate, as evidenced by reductive trapping of
the imminium with NaBH4. Dendritic effects in catalysis were
investigated using regular series dendrimers featuring the catalytic
dyad Pro–Lys (R2) and Pro–Thr (R1) with increasing generation
number. A positive dendritic effect was observed with both
acetone (maximum activity per N-terminal proline residue with
2nd generation dendrimer R1G2) and cyclohexanone (maximum
activity per free amino-group with 3rd generation dendrimer
6-(4-Nitro-phenyl)hexane-2,4-dione (2)2a
To a solution of acetyl acetone (750 lL, 7.30 mmol) (freshly
distilled) in dried THF (5 mL) was added under N2 at room
temperature a commercial solution of LDA (2N, 7.5 mL) The
3276 | Org. Biomol. Chem., 2006, 4, 3268–3281
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
◦
◦
solution was warmed to 40 C for 1 h. After cooling to −78 C
4-nitrobenzyl bromide (1.50 g, 7.30 mmol) in HMPA (5 mL)
was added dropwise via a syringe. The reaction was stirred for
at −78 ◦C and followed by TLC. The solution was quenched
carefully with water and evaporated to dryness, taken up in
CH2Cl2, extracted with 1 N HCl, brine and water, dried (MgSO4)
and flash chromatographed using hexane–ethyl acetate (2 : 1) to
give the title compound (673 mg, 40%). Rf = 0.26 (n-hexane–ethyl
7-(Allyloxy)-2H-chromen-2-one (8)
A solution of umbelliferone 7 (810 mg, 5.00 mmol), allyl bromide
(360 lL, 4.2 mmol) and potassium carbonate (690 mg, 5.00 mmol)
in acetone (50 mL) was stirred overnight at reflux. The reaction
mixture was evaporated to dryness and the residue taken up in
ethyl acetate (50 mL). The organic phase was extracted with 1 N
NaOH (50 mL) and brine (50 mL), dried (Na2SO4) and evaporated
to yield a white solid (yield 950 mg, 85%). M.p. = 78–82 ◦C. Rf =
0.42 (n-hexane–ethyl acetate (2 : 1)). IR (neat) m˜ = 3081, 1712, 1710,
1603, 1600, 1557, 1511, 1424, 1398, 1367, 1354, 1283, 1225, 1208,
1
acetate (2 : 1)). H NMR (300 MHz, CDCl3) enol d = 15.25 (s,
1H), 8.13 (d, 2H, J = 8.67 Hz), 7.33 (s, 2H, J = 8.67 Hz), 5.49 (s,
1H), 3.05 (t, 2H, J = 7.63 Hz), 2.68 (t, 2H, J = 7.63 Hz), 2.08 (s,
3H) ppm. FAB MS(+): m/z 236 (M+ + 1).
1124, 1012, 994, 940, 892, 856, 838 cm−1. H NMR (300 MHz,
1
CDCl3) d = 7.59 (d, J = 9.4 Hz, 1H), 7.39 (d, J = 8.5 Hz, 1H),
6.74 (dd, J = 2.45, 8.48 Hz, 2H), 6.76 (d, J = 2.5 Hz, 1H), 6.17
(d, J = 9.4 Hz, 1H), 5.96 (m, 1H), 5.35 (ddd, J = 1.3, 2.6 Hz,
17.2, 1H), 5.30 (ddd, J = 1.5, 3.0, 10.6 Hz, 1H), 4.51 (dt, J = 1.4,
5.3 Hz, 2H) ppm. 13C NMR (300 MHz, CDCl3) d = 162.5, 161.8,
156.6, 144.0, 132.8, 129.4, 119.2, 113.9, 113.8, 102.4, 70.0 ppm.
Succinic acid mono-(2-{ethyl-[4-(4-
nitrophenylazo)phenyl]amino}ethyl) ester (5)22
To a solution of Disperse Red 1 4 (200 mg, 0.636 mmol), Et3N
(117 lL, 0.837 mmol) and DMAP (8 mg, 0.0646 mmol) in CH2Cl2
(8 mL) was added succinic acid anhydride (76 mg, 0.761 mg).
The reaction was followed by TLC. After stirring overnight the
solution was evaporated to dryness and the residue was purified
by flash chromatography (CH2Cl2–MeOH (20 : 1), 0.1% AcOH)
yielding 5 as a red solid (86 mg, 33%). Rf = 0.18 (CH2Cl2–MeOH
(20 : 1), 0.1% AcOH). 1H NMR (300 MHz, CDCl3) d = 8.33 (d,
2H, J = 9.05 Hz), 7.93 (dd, 4H, J = 6.60, 8.95 Hz), 6.81 (d, 2H, J =
9.24 Hz), 4.34 (t, 2H, J = 6.40 Hz), 3.71 (t, 2H, J = 6.22 Hz), 3.54
(q, 2H, 7.16 Hz), 2.71–2.55 (m, 4H), 1.32 (t, 3H, J = 6.97 Hz) ppm.
ESI MS(+): calcd for [M + H]+ C20H23N4O6+ 415.42, found 415.20.
ESI MS(+): calcd for [M + H]+ C12H11O3 203.07, found 203.03.
+
7-(2,3-Dihydroxypropoxy)-2H-chromen-2-one (9)
To a solution of 8 (800 mg, 4.00 mmol) in tert-butanol–water
mixture (40 mL, 2 : 1), was added N-methylmorpholine-N-oxide
(703 mg, 6.00 mmol) and a solution of osmium tetroxide (2.5%
in tert-butanol, 0.3 mL). After stirring overnight, sodium sulfite
solution 10% (30 mL) was added and the aqueous layer was
extracted with EtOAc. The combined organic layers were washed
with brine and dried over MgSO4. After purification on silica gel
(n-hexane–ethyl acetate (2 : 3)), the desired diol 9 was obtained
as a white solid (890 mg, 95%), m.p. = 118–122 ◦C. Rf = 0.17 (n-
hexane–ethyl acetate (2 : 3)). IR (neat) m˜ = 3309, 3065, 2939, 1698,
1618, 1606, 1550, 1508, 1398, 1294, 1236, 1138, 1118, 1102, 1054,
1036, 1002, 944, 885, 837 cm−1. 1H NMR (300 MHz, CD3OD) d =
7.79 (d, J = 9.4 Hz, 1H), 7.58 (d, J = 8.7 Hz, 1H), 7.02 (dd, J =
2.5, 8.7 Hz, 1H), 6.96 (d, J = 2.7 Hz, 1H), 6.26 (d, J = 9.4 Hz,
1H), 3.96 (dd, J = 6.26, 11.1 Hz, 1H), 3.40 (m, 1H), 2.95 (m,
1H), 2.80 (m, 1H) ppm. 13C NMR (300 MHz, CD3OD) d = 162.0,
161.5, 156.1, 143.9, 129.4, 113.7, 113.3, 113.2, 102.0, 69.7, 50.2,
44.9 ppm. ESI MS(+): calcd for [M + Na]+ C12H10O5Na+ 259.06,
found 259.00.
N-[4-(3,5-Dioxohexyl)phenyl]succinamic acid
2-{ethyl-[4-(4-nitrophenylazo)phenyl]amino}ethyl ester (1a)
The diketone 2 was dissolved in CH2Cl2 and Pd/C (10 mole%) was
added and the solution was degassed with N2 three times. H2 was
bubbled through the solution for 1 min and the solution was stirred
under H2 (atm) for 2 h. The reaction was followed by TLC. The
solution was filtered through celite and evaporated to give the
crude amine 3 which was used directly for the next coupling step.
A solution of amine 3 was taken up in CH2Cl2 and EDC, HOBt
and 5 was added at 0 ◦C. The solution was stirred for 2 h at 0 ◦C and
then it was allowed to warm up to room temperature overnight.
The reaction was followed by TLC. An aqueous workup (1 N
HCl, 10% NaHCO3 and brine successively), following by drying
with MgSO4, evaporation and flash chromatography (CH2Cl2–
MeOH (40 : 1 v/v)) yielded the title compound as a red solid
(39 mg, 95%). M.p. = 125–127 ◦C. Rf = 0.21 (CH2Cl2–MeOH
(20 : 1). IR (neat) m˜ = 3320, 2899, 2360, 1727, 1667, 1589, 1515,
1411, 1387, 1335, 1313, 1160, 1139, 1106, 998, 858, 821 cm−1. 1H
NMR (300 MHz, CDCl3) enol d = 15.35 (s, 1H), 8.33 (d. 2H,
J = 9.04 Hz), 7.81 (dd, 4H, J = 7.15, 9.04 Hz), 7.35 (d, 2H, J =
8.10 Hz), 7.04 (d, 2H, J = 8.1 Hz), 6.67 (d, 2H, J = 9.24 Hz), 5.39
(s, 1H), 4.22 (t, 2H, J = 6.21 Hz), 3.62 (t, 2H, J = 6.21 Hz), 3.47
(m, 2H), 2.85–2.78 (m, 2H), 2.71–2.62 (m, 2H), 2.59–2.52 (m, 2H),
2.51–2.44 (m, 2H), 1.99 (s, 3H), 1.23 (t, 3H, J = 6.74 Hz) ppm.
13C NMR (300 MHz, CDCl3) enol d = 193.4, 172.9, 169.8, 169.3,
156.7, 151.3, 147.5, 147.5, 143.8, 137.1, 128.8, 126.3, 124.7, 122.7,
119.9, 111.5, 100.1, 61.7, 55.6, 48.7, 45.7, 39.9, 32.5, 31.8, 29.3,
7-(2-Hydroxypropoxy-3-tert-butyldimethylsilyloxy)-2H-chromen-
2-one (10)
The diol 9 (890 mg, 3.9 mmol) was dissolved in dichloromethane–
dimethylformamide (3 : 1) (50 mL), tert-butyldimethyl silyl
chloride (570 mg, 3.7 mmol) and imidazole (385 mg, 5.7 mmol)
were added and the mixture was stirred overnight. After addition
of dichloromethane (50 mL) and water (100 mL) the organic phase
was dried (Na2SO4) and evaporated to dryness. The residue was
purified by flash chromatography n-hexane–ethyl acetate (2 : 1) to
give 10 (yield 533 mg, 58%) as an oil. Rf = 0.67 (n-hexane–ethyl
acetate (2 : 1)). IR (neat) m˜ = 3320, 2928, 2856, 1725, 1712, 1612,
1556, 1507, 1471, 1404, 1349, 1292, 1279, 1249, 1238, 1197, 1105,
1027, 937, 824, 775 cm−1. 1H NMR (500 MHz, CDCl3) d = 7.59
(d, J = 9.5 Hz, 1H), 7.32 (d, J = 8.5 Hz, 1H), 6.79 (dd, J = 2.40,
8.50 Hz, 1H), 6.75 (d, J = 2.4 Hz, 1H), 6.12 (d, J = 9.50 Hz, 1H),
3.96 (m, 3H), 3.68 (m, 3H), 0.80 (s, 9H), −0.03 (s, 6H) ppm. 13C
NMR (500 MHz, CDCl3) d = 143.7, 129.2, 113.8, 113.2, 102.1,
12.3 ppm. ESI MS(+): calcd for [M + H]+ C32H36N5O7 602.26,
+
found 602.46.
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 3268–3281 | 3277
©

View Article Online
70.3, 69.4, 63.9 ppm. ESI MS(+): calcd for [M + H]+ C18H27O5Si+
351.16, found 351.11.
A and injected on analytical RP-HPLC. The aldol product was
purified by diluting the reaction mixture with A and injecting
on semi-preparative RP-HPLC and the pure aldol product was
injected on chiral phase HPLC separating the enantiomers.
1H NMR (300 MHz, CDCl3) d = 8.20 (d, J = 8.5 Hz, 1 H), 7.55
(d, J = 8.5 Hz, 1 H), 5.25 (m, 1 H), 3.56 (d, J = 3.2 Hz, 1 H),
2.85 (m, 2 H), 2.25 (s, 3 H) ppm. 13C NMR (300 MHz, CDCl3)
d = 208.51, 150.04, 147.22, 126.37, 123.71, 123.57, 68.94, 51.43,
30.67 ppm. Anal. chiral HPLC [Daicel Chiralpak AS, iPrOH–
hexane (20 : 80), UV 254 nm, flow rate 2.0 mL min−1]: tR (major) =
8.47 min; tR (minor) = 10.02 min.
7-(2-Oxopropoxy-3-tert-butyldimethylsilyloxy)-2H-chromen-2-
one (11)
To oxalyl chloride (240 lL, 3.2 mmol) in dry CH2Cl2 (5 mL) at
−78 ◦C was added dry dimethyl sulfoxide (410 lL, 5.8 mmol).
After 15 min 10 (450 mg, 1.29 mmol) was added and the mixture
was stirred for 1 h at −78 ◦C. Triethylamine (1170 lL, 8.39 mmol)
was added and the mixture was stirred for 1.5 h. The reaction
mixture was warmed to rt and quenched with sat. NH4Cl (20 mL).
The mixture was extracted with CH2Cl2 (3 × 20 mL) and the
organic phase was dried (Na2SO4), concentrated in vacuo and flash
chromatographed (n-hexane–ethyl acetate, 2 : 1) to yield 11 as an
oil (350 mg, 78%). Rf = 0.60 (n-hexane–ethyl acetate (2 : 1)). IR
(neat) m˜ = 2928, 2856, 1725, 1608, 1508, 1424, 1402, 1348, 1277,
4-(2-Bromophenyl)-1,3,4-trihydroxybutan-2-one (19)10d
Catalysis in water. 50 lL of a 200 mM solution of 2-bromo-
benzaldehyde in methanol, dihydroxyacetone (0.5 mmol, 45 mg)
and 50 lL a 2 mM dendrimer catalyst solution in 100 mM aqueous
R
bicine buffer pH = 8.5 was shaken in an Eppendorfꢀ PP-tube
1
1232, 1109, 1055, 995, 893, 833, 775 cm−1. H NMR (500 MHz,
for 66 h. Final concentrations: 5 M dihydroxyacetone, 100 mM
aldehyde and 1 mM catalyst. 10 lL aliquots of the reaction mixture
were diluted with 100 lL A and injected on analytical RP-HPLC
separating the diastereomers. Anal. HPLC (Waters Atlantis, UV
254 nm, gradient 90% A, 10% D to 50% A, 50% D in 10 min), tR
(major, syn) = 4.96 min and tR = (minor, anti) 4.37 min.
CDCl3) d = 7.50 (d, J = 9.4 Hz, 1H), 7.26 (d, J = 8.7 Hz, 1H), 6.74
(dd, J = 2.50, 8.70 Hz, 1H), 6.59 (d, J = 2.50 Hz, 1H), 6.14 (d, J =
9.60 Hz, 1H), 4.85 (s, 2H), 4.26 (s, 2H), 0.82 (s, 9H), −0.01 (s,6H)
ppm. 13C NMR (500 MHz, CDCl3) d = 204.9, 161.3, 161.2, 156.1,
143.6, 129.4, 114.1, 113.3, 102.0, 26.1, 18.6, 14.6, −5.2 ppm. ESI
MS(+): calcd for C18H25O5Si+ [M + H]+ 349.15 found 349.08.
syn-19. 1H NMR (300 MHz, CD3OD) d = 7.70 (d, J = 8.3 Hz,
1H), 7.56 (d, J = 8.3 Hz, 1H), 7.39 (app. t, 1H), 7.16–7.21 (m, 1H),
5.45 (d, J = 2.07 Hz, 1H), 4.59 (s, 2 H), 4.41 (d, J = 2.07 Hz, 1H)
ppm. 13C NMR (CD3OD) d = 213.31, 142.09, 133.79, 131.35,
130.54, 128.72, 122.67, 79.00, 74.72, 68.44 ppm. HR ESI MS(+):
calcd for C10H11BrNaO4 296.9738; found 296.9749. Anal. HPLC
(254 nm; gradient 90% A, 10% D to 50% A, 50% D in 10 min):
tR = 5.0 min.
anti-19. 1H NMR (300 MHz, CD3OD) d = 7.50–7.77 (m,
2H), 7.35 (app. t, 1H), 7.1–7.2 (m, 1H), 5.22 (d, J = 4.9 Hz,
1H), 4.35–4.50 (m, 3H) ppm. 13C NMR (300 MHz, CD3OD)
d = 212.17, 141.69, 133.91, 130.66, 130.60, 128.86, 124.16, 79.47,
75.89, 68.79 ppm. HR ESI MS(+): calcd. For C10H11BrNaO4
296.9738; found 296.9732. Anal. HPLC (254 nm; gradient 90%
A, 10% D to 50% A, 50% D in 10 min): tR = 4.4 min.
7-(3-Hydroxy-2-oxopropoxy)-2H-chromen-2-one (6)
A solution of 11 (40 mg, 0.115 mmol) in TFA–H2O (9 : 1) (10 mL)
was stirred for 1 h. The TFA–H2O mixture was evaporated under
high vacuum to yield an oily film which was taken up in H2O–
CH3CN (85 : 15) (50 mL) and purified directly using preparative
RP-HPLC (RP-HPLC conditions: H2O–CH3CN (85 : 15) with
0.1% TFA to H2O–CH3CN (50 : 50) in 70 min, k = 254 nm and
flow rate = 100 mL min−1) to yield 6 (9 mg, 33%). M.p. 138–142 ◦C.
Rf = 0.31 (n-hexane–ethyl acetate (2 : 1)). IR (neat) m˜ = 3422, 2922,
1699, 1615, 1558, 1507, 1418, 1400, 1353, 1283, 1230, 1158, 1123,
1
1010, 985, 836, 815 cm−1. H NMR (500 MHz, d6-acetone) d =
7.78 (d, 1H, J = 9.60 Hz), 7.49 (d, 1H, J = 8.67 Hz), 6.84 (dd, 1H,
J = 2.45, 8.67 Hz), 6.09 (d, 1H, J = 9.42 Hz), 4.99 (s, 2H), 4.31
(s, 2H) ppm. 13C NMR (500 MHz, d6-acetone) d = 206.2, 161.4,
160.6, 155.6, 144.6, 129.8, 113.1, 113.0, 101.7, 80.9, 66.1 ppm. ESI
+
MS(+): calcd for C12H11O5 [M + H]+ 235.0606 found 235.0605.
2-[Hydroxy-(4-nitrophenyl)methyl]cyclohexanone (20)
4-(4-Nitrophenyl)-4-hydroxybutan-2-one (17)4a
Catalysis in DMSO. 100 lL of a 200 mM solution of 4-nitro-
benzaldehyde in cyclohexanone and 100 lL of a 2 mM dendrimer
solution in DMSO buffered with x eq. of NMM (where x =
number of TFA–amine salts for the dendrimer tested) was
Catalysis in water. 50 lL of a 200 mM solution of 4-nitro-
benzaldehyde in acetone and 50 lL of a 2 mM dendrimer catalyst
solution in 100 mM aqueous bicine buffer pH = 8.5 was shaken in
R
shaken in an Eppendorfꢀ PP-tube for 18 h. Final concentrations:
R
an Eppendorfꢀ PP-tube for 2–3 h. Final concentrations: 100 mM
Aldehyde 100 mM, catalyst 1 mM. 10 lL aliquots of the
reaction mixture were diluted with 100 lL A and injected on
analytical RP-HPLC. The aldol product was purified by diluting
the reaction mixture with A and injecting on semi-preparative RP-
HPLC separating the diastereomers. The pure aldol products were
injected on chiral phase HPLC separating the enantiomers.
aldehyde and 1 mM catalyst. 10 lL aliquots of the reaction mixture
were diluted with 100 lL A and injected on analytical RP-HPLC
running isocratic 66% A, 34% D.
Catalysis in DMSO. 50 lL of a 200 mM solution of 4-nitro-
benzaldehyde in DMSO–acetone (4 : 1, v/v) and 50 lL of a 2 mM
dendrimer solution in DMSO–acetone (4 : 1, v/v) buffered with
x eq. of NMM (where x = number of TFA–amine salts for the
Catalysis in water. 50 lL of a 200 mM solution of 4-nitro-
benzaldehyde in cyclohexanone and 50 lL a 2 mM dendrimer
catalyst solution in 100 mM aqueous bicine buffer pH = 8.5 was
R
dendrimer tested) was shaken in an Eppendorfꢀ PP-tube for 36–
R
72 h. Final concentrations: aldehyde 100 mM, catalyst 1 mM.
shaken in an Eppendorfꢀ PP-tube for 2 h. Final concentrations:
10 lL aliquots of the reaction mixture were diluted with 100 lL
100 mM aldehyde and 1 mM catalyst. 10 lL aliquots of the
3278 | Org. Biomol. Chem., 2006, 4, 3268–3281
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
reaction mixture were diluted with 100 lL A and injected on
analytical RP-HPLC. The aldol product was purified by diluting
the reaction mixture with A and injecting on semi-preparative RP-
HPLC separating the diastereomers. The pure aldol products were
injected on chiral phase HPLC separating the enantiomers.
mixture was removed by filtration and 1 mL of 100 lM solution
of 6 in 20 mM bicine buffer pH = 8.5 with 1% acetonitrile was
added. The bead suspension was then plated out onto a silica gel
plate and incubated for 40 min. The beads were observed under
a microscope with a UV lamp irradiating at 365 nm. Single blue
fluorescent beads were transferred via a syringe needle to amino
acid analysis vials.
anti-20. IR (neat) m˜ = 3504, 1688, 1604, 1531, 1508, 1342,
1131, 1044, 855, 842, 800, 702 cm−1. 1H NMR (300 MHz, CDCl3)
d = 8.22 (d, J = 8.8 Hz, 2H), 7.51 (d, J = 8.7 Hz, 2H), 4.95 (d, J =
8.3 Hz, 1H), 4.09 (br, 1H), 2.32–2.71 (m, 3H), 2.12–2.29 (m, 1H),
1.31–1.87 (m, 5H) ppm. 13C NMR (300 MHz, CDCl3) d = 214.7,
145.6, 145.2, 127.8, 123.5, 74.0, 57.2, 42.7, 30.8, 27.6, 24.7 ppm.
ESI MS(+): calcd for C13H15NNaO4 272.3; found 272.5. Anal.
HPLC (268 nm; gradient 90% A, 10% D to 10% A, 90% D in 15
min): tR = 11.3 min. Anal. chiral HPLC [Daicel Chiralpak OD–H,
iPrOH–hexane (10 : 90), UV 268 nm, flow rate 1.5 mL min−1]: tR
(major); = 16.1 min. tR (minor) = 22.5.
On bead assay with retro-aldolase substrates 12 and 14. 50 mg
library resin was swollen overnight in 20 mM borate buffer pH =
8.8. The swelling mixture was removed by filtration and 1 mL of
100 lM solution of 12 or 14 in 20 mM borate buffer pH = 8.8 with
5% acetonitrile was added. The bead suspension was then plated
out onto a silica gel plate and incubated for up to 2 h. The beads
were observed under a microscope with a UV lamp irradiating at
365 nm. No hits were observed.
Bead analysis. Single dendrimer-containing resin beads were
hydrolyzed with aqueous HCl (6 M) at 110 ◦C for 22 h, and
their amino acid composition was determined quantitatively by
HPLC after derivatization with phenyl isothiocyanate (PITC).
Such amino acid analyses are routine for protein composition
analysis. The analysis detects as little as 5 pmol per amino acid,
which is sufficient for single resin bead analysis, as these contain
50–200 pmol of dendrimers. False positives are most likely due to
manipulation errors during bead picking. Sometimes more than
one bead is transferred to the pipette, and the fluorescence staining
is diluted when the beads are washed down the pipette used for
picking, which does not allow staining to be rechecked afterwards.
syn-20. IR (neat) m˜ = 3491, 1693, 1602, 1508, 1447, 1343, 1186,
1131, 1091, 852, 796, 703 cm−1. 1H NMR (300 MHz, CDCl3) d =
8.22 (d, J = 8.7 Hz, 2H), 7.49 (d, J = 8.9 Hz, 2H), 5.45 (d, J =
2.1 Hz, 1H), 3.05 (br, 1H), 2.51–2.61 (m, 1H), 2.29–2.49 (m, 2H),
1.95–2.11 (m, 1H), 1.41–1.82 (m, 5H) ppm. 13C NMR (300 MHz,
CDCl3) d = 214.0, 144.6, 143.8, 126.6, 123.5, 70.2, 56.8, 42.6, 27.8,
25.9, 24.8 ppm. ESI MS(+): calcd for C13H15NNaO4 272.3; found
272.4. Anal. HPLC (268 nm; gradient 90% A, 10% D to 10% A,
90% D in 15 min): tR = 11.7 min. Anal. chiral HPLC [Daicel
Chiralpak OD–H, iPrOH–hexane (10 : 90), UV 268 nm, flow rate
1.5 mL min−1]: tR (minor) = 13.5 min; tR (major) = 15.1 min.
Library synthesis and screening
Cleavage of the Fmoc protecting group. The Fmoc protecting
group was removed with 5 mL of a solution of DMF–piperidine
(1 : 4, v/v) for 10 min. After filtration, the procedure was repeated
and then washed (3 × each) with DMF, DCM and MeOH.
Coupling of the Fmoc-protected amino acids. The resin was
washed and swollen inside the reactor with DCM (2 × 5 mL)
and DMF (1 × 5 mL). The NovasynTGR (Tentagel with
Rink linker) (0.25 mmol g−1) was acylated with 2.5 equivalents
of N-Fmoc amino acid in the presence of 2.5 equivalents
of PyBOP ((benzotriazol-1-yloxy)tripyrrolidinophosphonium
N-Acetylation. The resin was acetylated with a solution of
acetic acid anhydride–DCM (1 : 1, v/v) for 10 min. After filtration,
the procedure was repeated and then washed (3 × each) with DMF,
DCM and MeOH.
hexafluorophosphate) and
6
equivalents of DIEA (N,Nꢀ-
diisopropylethylamine) in DMF. After 2g × 60 min (where g =
generation number) the resin was washed (3 × each) with DMF,
DCM and MeOH and controlled with the TNBS (trinitrobenzene-
sulfonic acid) or chloranil test followed by acetylation.
TFA cleavage. The cleavage was carried out using TFA–H2O–
TIS (triisopropylsilane) as a (95 : 2.5 : 2.5, v/v) solution for 6 h.
The peptide was precipitated with methyl tert-butyl ether then
dissolved in water–acetonitrile mixture. All the dendrimers were
purified by preparative RP-HPLC.
Resin mixing and splitting. The resin was suspended in DMF–
DCM (2 : 1, v/v), and mixed via nitrogen bubbling for 15 min, and
then distributed in four equal portions.
Yields, purification method and characterization of all dendrimers
On-bead assay with diketone 1a. 50 mg library resin was
swollen overnight in 20 mM DMSO–PBS buffer pH = 7.4 (1 :
1, v/v). The swelling mixture was removed by filtration and 1 mL
of 50 lM solution of 1a in 20 mM DMSO–PBS buffer pH =
7.4 (1 : 1, v/v) was added. The resin was shaken for 30 min and
washed extensively with PBS buffer, DMSO, DMF, MeOH, DCM,
MeOH, DMF and finally with PBS buffer again (3 × each). A
suspension of the resin in DMF was transferred to a silica gel
plate and the beads were observed under a microscope. Single red
colored beads were transferred via a syringe needle to amino acid
analysis vials.
The dendrimers were synthesized using the same conditions as for
the library synthesis.
Dendrimer
L1D5
(Ac–Glu–Tyr)8(Dap–Arg–Thr)4(Dap–
Phe–Lys)2(Dap–Ser–Lys): from NovasynTGR (200 mg,
0.25 mmol g−1), L1D5 was obtained as colorless foamy solid after
preparative HPLC purification (18 mg, 9%, as TFA-salt); anal.
RP-HPLC: tR = 4.75 min; ESI MS(+): calcd for C228H326N60O74:
5088.35, found: 5088.38.
Dendrimer
L1D6
(Ac–Arg–Tyr)8(Dap–Glu–Tyr)4(Dap–
Phe–Lys)2(Dap–Ser–Lys): from NovasynTGR (200 mg,
0.25 mmol g−1), L1D6 was obtained as colorless foamy solid
after preparative HPLC purification (21 mg, 10%, as TFA-salt);
anal. RP-HPLC (70% A, 30% D to 30% A, 70% D in 10 min):
On-bead assay with enolization probe 6. 50 mg library resin was
swollen overnight in 20 mM bicine buffer pH = 8.5. The swelling
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 3268–3281 | 3279
©

View Article Online
tR = 4.44 min; ESI MS(+): calcd for C252H352N72O66: 5444.65,
HPLC purification (14 mg, 26%, as TFA-salt); anal. RP-HPLC
(90% A, 10% D to 50% A, 50% D in 15 min): tR = 5.22 min; ESI
MS(+): calcd for C156H256N45O52: 3591.97, found: 3591.84.
Dendrimer R2G1 (Pro–Lys)2(Dap–Pro–Lys): from No-
vasynTGR (100 mg, 0.25 mmol g−1), R2G1 was obtained as
colorless foamy solid after preparative HPLC purification (9 mg,
85%, as TFA-salt); anal. RP-HPLC (100% A, 0% D to 10% A, 90%
D in 15 min): tR = 4.10 min; ESI MS(+): calcd for C36H67N12O7:
777.99, found: 778.50.
Dendrimer R2G2 (Pro–Lys)4(Dap–Pro–Lys)2(Dap–Pro–Lys):
from NovasynTGR (100 mg, 0.25 mmol g−1), R2G2 was obtained
as colorless foamy solid after preparative HPLC purification
(17 mg, 58%, as TFA-salt); anal. RP-HPLC (95% A, 5% D to
70% A, 30% D in 15 min): tR = 6.38 min; ESI MS(+): calcd for
C86H155N28O17: 1851.32, found: 1851.33.
found: 5444.25.
Dendrimer
L2D1
(Pro–Lys)8(Dap–Pro–Lys)4(Dap–Tyr–
Leu)2(Dap–Ile–Gly): from NovasynTGR (200 mg, 0.25 mmol g−1),
L2D1 was obtained as colorless foamy solid after preparative
HPLC purification (86 mg, 25%, as TFA-salt); anal. RP-HPLC
(90% A, 10% D to 50% A, 50% D in 10 min): tR = 6.74 min; ESI
MS(+): calcd for C191H327N56O40: 4046.00, found: 4045.88.
Dendrimer L2D5 (Ser–Lys)8(Dap–Pro–Arg)4(Dap–Ile–Gly)2-
(Dap–Tyr–Gly): from NovasynTGR (200 mg, 0.25 mmol g−1),
L2D5 was obtained as colorless foamy solid after preparative
HPLC purification (20 mg, 8%, as TFA-salt); anal. RP-HPLC
(90% A, 10% D to 50% A, 50% D in 10 min): tR = 7.87 min; ESI
MS(+): calcd for C166H302N64O47: 3942.00, found: 3942.12.
Dendrimer L2D6 (Ser–Lys)8(Dap–Ser–Lys)4(Dap–Tyr–Gly)2-
(Dap–Phe–Gly): from NovasynTGR (200 mg, 0.25 mmol g−1),
L2D6 was obtained as colorless foamy solid after preparative
HPLC purification (33 mg, 13%, as TFA-salt); anal. RP-HPLC
(90% A, 10% D to 50% A, 50% D in 10 min): tR = 6.83 min; ESI
MS(+): calcd for C162H284N56O52: 3846.15, found: 3846.13.
Dendrimer L2D7 (Pro–Lys)8(Dap–Glu–Arg)4(Dap–b-Ala–Gly)2-
(Dap–Phe–Val): from NovasynTGR (200 mg, 0.25 mmol g−1),
L2D7 was obtained as colorless foamy solid after preparative
HPLC purification (19 mg, 7%, as TFA-salt); anal. RP-HPLC
(90% A, 10% D to 10% A, 90% D in 15 min): tR = 5.13 min; ESI
MS(+): calcd for C177H306N64O46: 4064.36, found: 4064.38.
Dendrimer L2K4 (Pro–Lys)8(Dap–Thr–His)4(Dap–Tyr–Gly)2-
(Dap–Phe–Val): from NovasynTGR (200 mg, 0.25 mmol g−1),
L2K4 was obtained as colorless foamy solid after preparative
HPLC purification (11 mg, 4%, as TFA-salt); anal. RP-HPLC
(90% A, 10% D to 10% A, 90% D in 15 min): tR = 3.43 min; ESI
MS(+): calcd for C185H295N61O43: 4060.28, found: 4060.38.
Dendrimer L2K7 (Pro–Lys)8(Dap–Ser–Arg)4(Dap–b-Ala–Val)2-
(Dap–Tyr–Leu): from NovasynTGR (200 mg, 0.25 mmol g−1),
L2K7 was obtained as colorless foamy solid after preparative
HPLC purification (28 mg, 10%, as TFA-salt); anal. RP-HPLC
(90% A, 10% D to 10% A, 90% D in 15 min): tR = 5.34 min; ESI
MS(+): calcd for C176H313N65O42: 4010.42, found: 4010.75.
Dendrimer L2K8 (Glu–Lys)8(Dap–Glu–Asp)4(Dap–Ile–Gly)2-
(Dap–Tyr–Ala): from NovasynTGR (200 mg, 0.25 mmol g−1),
L2K8 was obtained as colorless foamy solid after preparative
HPLC purification (17 mg, 6%, as TFA-salt); anal. RP-HPLC
(90% A, 10% D to 10% A, 90% D in 15 min): tR = 4.10 min; ESI
MS(+): calcd for C173H388N53O70: 4228.05, found: 4228.50.
Dendrimer R1G1 (Pro–Thr)2(Dap–Pro–Thr): from No-
vasynTGR (100 mg, 0.25 mmol g−1), R1G1 was obtained as
colorless foamy solid after preparative HPLC purification (7 mg,
79%, as TFA-salt); anal. RP-HPLC (100% A, 0% D to 10% A, 90%
D in 15 min): tR = 4.92 min; ESI MS(+): calcd for C30H51N9O10:
696.78, found: 697.38.
Dendrimer R2G3 (Pro–Lys)8(Dap–Pro–Lys)4(Dap–Pro–Lys)2-
(Dap–Pro–Lys): from NovasynTGR (100 mg, 0.25 mmol g−1),
R2G3 was obtained as colorless foamy solid after preparative
HPLC purification (18 mg, 20%, as TFA-salt); anal. RP-HPLC
(90% A, 10% D to 50% A, 50% D in 15 min): tR = 5.794 min; ESI
MS(+): calcd for C186H331N60O37: 3997.97, found: 3998.00.
Reductive alkylation of R2G3
Dendrimer R2G3 (6.0 mg, 1 lmol) was stirred in 100 lL 100 mM
bicine buffer pH 8.5. Then, 100 lL of acetone was added and
the resulting mixture was stirred for 3 h at room temperature. A
solution of NaBH4 (4.0 mg, 100 lmol) in H2O (100 lL) was added
and the mixture was stirred overnight at room temperature. Then,
acetic acid was added (100 lL) and the mixture was lyophilized.
The 1H NMR spectrum was recorded.
Reductive alkylation of L1D5
Dendrimer L1D5 (5.4 mg, 1 lmol) was stirred in 100 lL 100 mM
bicine buffer pH 8.5. Then, 100 lL of cyclohexanone was added
and the resulting mixture was stirred for 3 h at room temperature. A
solution of NaBH4 (4.0 mg, 100 lmol) in H2O (100 lL) was added
and the mixture was stirred overnight at room temperature. Then,
acetic acid was added (100 lL) and the mixture was lyophilized.
The 1H NMR spectrum was recorded.
Acknowledgements
This work was supported by the University of Berne, the Swiss
National Science Foundation, the Marie Curie Training Network
IBAAC, and the COST Action D25. We thank Monique Lu¨thy
and Quirin B. Scheifele for assistance in the preparation of
compound 1a.
Dendrimer R1G2 (Pro–Thr)4(Dap–Pro–Thr)2(Dap–Pro–Thr):
from NovasynTGR (100 mg, 0.25 mmol g−1), R1G2 was obtained
as colorless foamy solid after preparative HPLC purification
(17 mg, 58%, as TFA-salt); anal. RP-HPLC (90% A, 10% D to
50% A, 50% D in 15 min): tR = 4.75 min; ESI MS(+): calcd for
C71H119N21O24: 1661.84, found: 1661.88.
References
1 T. D. Machajewski and C. H. Wong, Angew. Chem., Int. Ed., 2000, 39,
1352–1374.
2 (a) J. Wagner, R. A. Lerner and C. F. Barbas, III, Science, 1995, 270,
1797; (b) G. Zhong, R. A. Lerner and C. F. Barbas, III, Angew. Chem.,
Int. Ed., 1999, 38, 3738; (c) F. Tanaka, R. A. Lerner and C. F. Barbas,
III, J. Am. Chem. Soc., 2000, 122, 4835.
3 (a) Z. G. Hajos and D. R. Parish, J. Org. Chem., 1973, 38, 3239–3243;
(b) Z. G. Hajos and D. R. Parrish, J. Org. Chem., 1974, 39, 1615–21;
(c) U. Eder, G. Sauer and R. Wiecher, Angew. Chem., Int. Ed. Engl.,
Dendrimer
R1G3
(Pro–Thr)8(Dap–Pro–Thr)4(Dap–Pro–
Thr)2(Dap–Pro–Thr): from NovasynTGR (100 mg, 0.25 mmol
g−1), R1G3 was obtained as colorless foamy solid after preparative
3280 | Org. Biomol. Chem., 2006, 4, 3268–3281
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
1971, 10, 496–497; (d) S. Danishefsky and P. Cain, J. Am. Chem. Soc.,
1976, 98, 4975–4983.
Rev., 2002, 31, 69–82; (d) D. Astruc and F. Chardac, Chem. Rev., 2001,
101, 2991–3023; (e) A. Didier, K. Heuze, S. Gatard, D. Mery, S. Nlate
and L. Plault, Adv. Synth. Catal., 2005, 347, 329–338; (f) J. C. Garcia-
Martinez, R. Lezutekong and R. M. Crooks, J. Am. Chem. Soc., 2005,
127, 5097–5103; (g) S. M. Lu and H. Alper, J. Am. Chem. Soc., 2005,
127, 14776–14784; (h) W.-J. Tang, N.-F. Yang, B. Yi, G.-J. Deng, Y.-Y.
Huang and Q.-H. Fan, Chem. Commun., 2004, 1378–1379; (i) C. O.
Liang, B. Helms, J. Craig and J. M. J. Frechet, Chem. Commun., 2003,
2524–2525; (j) P. Weyermann and F. Diederich, Helv. Chim. Acta, 2002,
85, 599–617; (k) L. Liu and R. Breslow, J. Am. Chem. Soc., 2003, 125,
12110–12111.
8 For peptide dendrimers: (a) L. Crespo, G. Sanclimens, M. Pons, E.
Giralt, M. Royo and F. Albericio, Chem. Rev., 2005, 1663–1681; (b) M. J.
Cloninger, Curr. Opin. Chem. Biol., 2002, 6, 742–748; (c) L. Crespo,
G. Sanclimens, B. Montaner, R. Pe´rez-Toma´s, M. Royo, M. Pons, F.
Albericio and E. Giralt, J. Am. Chem. Soc., 2002, 124, 8876–8863; (d) K.
Sadler and J. P. Tam, Rev. Mol. Biotechnol., 2002, 90, 195–229; (e) J. P.
Tam, Y.-A. Lu and J.-L. Yang, Eur. J. Biochem., 2002, 269, 923–932;
(f) N. Higashi, T. Koga and M. Niwa, ChemBioChem, 2002, 3, 448–
454; (g) U. Boas, S. H. M. Sontjens, K. J. Jensen, J. B. Christensen and
E. W. Meijer, ChemBioChem, 2002, 3, 433–439; (h) G. A. Kinberger, C.
Welbo and M. Goodman, J. Am. Chem. Soc., 2002, 124, 15162–15163;
(i) C.-H. Tung, S. Mueller and R. Weissleder, Bioorg. Med. Chem.,
2002, 10, 3609–3614; (j) N. Wimmer, R. J. Marano, P. S. Kearns, E. P.
Rakoczy and I. Toth, Bioorg. Med. Chem., 2002, 12, 2635–2637; (k) D.
Lagnoux, T. Darbre, M. L. Schmitz and J.-L. Reymond, Chem.–Eur. J.,
2005, 11, 3941–3950.
4 For selected direct organocatalytic aldol reactions see: (a) B. List, R. A.
Lerner and C. F. Barbas, III, J. Am. Chem. Soc., 2000, 122, 2395;
(b) A. B. Northrup and D. W. C. MacMillan, J. Am. Chem. Soc.,
2002, 124, 6798–6799; (c) A. Bøgevig, N. Kumaragurubaran and K. A.
Jørgensen, Chem. Commun., 2002, 6, 620–621; (d) J. Kofoed, J. Nielsen
and J.-L. Reymond, Bioorg. Med. Chem. Lett., 2003, 13, 2445–2447;
(e) P. Krattiger, R. Kovasy, J. D. Revell, S. Ivan and H. Wennemers,
Org. Lett., 2005, 7, 1101–1103; (f) D. Enders and C. Grondal, Angew.
Chem., Int. Ed., 2005, 44, 1210–1212; (g) P. Dziedzic, W. B. Zou, J.
Hafren and A. Co´rdova, Org. Biomol. Chem., 2006, 4, 38–40; (h) A. B.
Northrup and D. W. C. MacMillan, Science, 2004, 305, 1752–1755;
(i) N. Mase, Y. Nakai, N. Ohara, H. Yoda, K. Takabe, F. Tanaka
and C. F. Barbas, III, J. Am. Chem. Soc., 2006, 128, 734–735; (j) Y.
Hayashi, T. Sumiya, J. Takahashi, H. Gotoh, T. Urushima and M.
Shoji, Angew. Chem., Int. Ed., 2006, 45, 958–961; (k) Z. Tang, Z.-H.
Yang, L.-F. Cun, L.-Z. Gong, A.-Q. Mi and Y.-Z. Jiang, Org. Lett.,
2004, 6, 2285–2287; (l) I. Ibrahem, W. Zou, Y. Xu and A. Co´rdova,
Adv. Synth. Catal., 2006, 348, 211–222; (m) Z. Tang, F. Jiang, L.-T.
Yu, X. Cui, L.-Z. Gong, A.-Q. Mi, Jiang and Y.-Z. Wu, J. Am. Chem.
Soc., 2003, 125, 5262; (n) A. Co´rdova, W. Notz and C. F. Barbas, III,
Chem. Commun., 2002, 67, 3034; (o) F. Tanaka, R. Fuller and C. F.
Barbas, III, Biochemistry, 2005, 44, 7583–7592; (p) A. Hartikka and
P. I. Arvidsson, Eur. J. Org. Chem., 2005, 4287–4295; (q) A. Berkessel,
B. Koch and J. Lex, Adv. Synth. Catal., 2004, 346, 1141–1146; (r) H. J.
Martin and B. List, SYNLETT, 2003, 12, 1901–1902; (s) Z. Tang, Z.-H.
Yang, X.-H. Chen, L.-F. Cun, L.-F., A.-Q. Mi, Y.-Z. Jiang, Y.-Z. and
L.-Z. Gong, J. Am. Chem. Soc., 2005, 127, 9285–9289; (t) Z. Tang, F.
Jiang, X. Cui, L.-Z. Gong, A.-Q. Mi, Y.-Z. Jiang and Y.-D. Wu, Proc.
Natl. Acad. Sci. U. S. A., 2004, 101, 5755–5760; (u) For a recent short
review on asymmetric organocatalysis see: J. Seayad and B. List, Org.
Biomol. Chem., 2005, 3, 719–724.
5 For selected Zn-catalyzed direct aldol reactions see: (a) T. Darbre and
M. Machuqueiro, Chem. Commun., 2003, 9, 1090–1091; (b) J. Kofoed,
M. Machuqueiro, J.-L. Reymond and T. Darbre, Chem. Commun.,
2004, 13, 1540–1541; (c) J. Kofoed, J.-L. Reymond and T. Darbre,
Org. Biomol. Chem., 2005, 3, 1850–1855; (d) R. Fernandez-Lopez, R. J.
Kofoed, M. Machuqueiro and T. Darbre, Eur. J. Org. Chem., 2005,
24, 5268–5276; (e) Y. M. A. Yamada, N. Yoshikawa, H. Sasai and M.
Shibasaki, Angew. Chem., Int. Ed. Engl., 1997, 36, 1871–1873; (f) N.
Yoshikawa, Y. M. A. Yamada, J. Das, M. Sasai and H. Shibasaki, J. Am.
Chem. Soc., 1999, 121, 4168–4178; (g) N. Kumagai, S. Matsunaga, N.
Yoshikawa, T. Ohshima and M. Shibasaki, Org. Lett., 2001, 3, 1539–
1542; (h) B. M. Trost and H. Ito, J. Am. Chem. Soc., 2004, 122, 12003–
12004; (i) B. M. Trost, E. R. Silcoff and H. Ito, Org. Lett., 2001, 3,
2497–2500; (j) B. M. Trost, H. Ito and E. R. Silcoff, J. Am. Chem. Soc.,
2001, 123, 3367–3368; (k) B. M. Trost, S. Shin and J. A. Sclafani, J. Am.
Chem. Soc., 2005, 127, 8602–8603.
6 (a) G. R. Newkome, C. N. Moorefield and F. Vo¨gtle, Dendritic
Molecules: Concepts, Synthesis, Perspectives, VCH, Weinheim, 1996;
(b) G. R. Newkome, C. N. Moorefield and F. Vo¨gtle, Dendritic
Molecules: Concepts, Synthesis, Applications, VCH: Weinheim, 2001;
(c) Dendrimers V: Functional and Hyperbranched Building Blocks,
Photophysical Properties and Applications in Materials and Life Science
(Top. Curr. Chem., vol. 228), ed. F. Vo¨gtle and C. A. Schalley, Springer-
Verlag, Berlin, 2003; (d) Dendrimers IV: Metal Coordination, Self
Assembly and Catalysis (Top. Curr. Chem., vol. 217), ed. F. Vo¨gtle
and C. A. Schalley, Springer-Verlag, Berlin, 2001; (e) D. A. Tomalia and
P. R. Dvornic, Nature, 1994, 372, 617–618; (f) J. W. J. Knapen, A. W. van
der Made, J. C. de Wilde, P. W. N. M. van Leeuwen, P. Wijkens, D. M.
Grove and G. van Koten, Nature, 1994, 372, 659–662; (g) C. Liang and
J. M. Fre´chet, Prog. Polym. Sci., 2005, 30, 385–402; (h) D. K. Smith,
Tetrahedron, 2003, 59, 3797–3798; (i) S. M. Grayson and J. M. J. Fre´chet,
Chem. Rev., 2001, 101, 3819–3868; (j) F. Zeng and S. C. Zimmerman,
Chem. Rev., 1997, 97, 1681–1712; (k) M. Witvrouw, V. Fikkert, W.
Pluymers, B. Matthews, K. Mardel, D. Schols, J. Raff, Z. Debyser, E.
DeClercq, G. Holan and C. Pannecouque, Mol. Pharmacol., 2000, 58,
1100–1108.
9 (a) A. Esposito, E. Delort, D. Lagnoux, F. Djojo and J.-L. Reymond,
Angew. Chem., Int. Ed., 2003, 42, 1381–1383; (b) D. Lagnoux, E.
Delort, C. Douat-Casassus, A. Esposito and J. L. Reymond, Chem.–
Eur. J., 2004, 10, 1215–1226; (c) C. Douat-Casassus, T. Darbre and J.-L.
Reymond, J. Am. Chem. Soc., 2004, 126, 7817–7826; (d) E. Delort, T.
Darbre and J.-L. Reymond, J. Am. Chem. Soc., 2004, 126, 15642–15643;
(e) A. Clouet, T. Darbre and J.-L. Reymond, Adv. Synth. Catal., 2004,
346, 1195–1204.
10 E. Bellis and G. Kokotos, J. Mol. Catal. A: Chem., 2005, 241, 166–174.
11 (a) A. Furka, A., F. Sebestye´n, M. Asgedom and G. Dibo´, Int. J. Pept.
Protein Res., 1991, 37, 487–493; (b) K. S. Lam, S. E. Salmon, E. M.
Hersh, V. J. Hruby, W. M. Kazmierski and R. J. Knapp, Nature, 1991,
354, 82–84; (c) R. A. Houghten, C. Pinilla, C., S. E. Blondelle, J. R.
Appel, C. T. Dooley and J. H. Cuervo, Nature, 1991, 354, 84–86;(d) K. S.
Lam, M. Lebl and V. Krchnak, Chem. Rev., 1997, 97, 411–448.
12 (a) A. Clouet, T. Darbre and J.-L. Reymond, Angew. Chem., Int. Ed.,
2004, 43, 4612–4615; (b) A. Clouet, T. Darbre and J.-L. Reymond,
Biopolymers, 2006, 84, 114–123.
13 18.75% of the library contains 1 lysine at A1, 18.75% 2 lysines at
position A3, and 6.25% three lysine residues at position A1 and A3,
giving a total of 43.75% of the library containing at least one lysine
residue.
14 J. Kofoed, T. Darbre and J.-L. Reymond, Chem. Commun., 2006, 1482–
1484.
15 (a) N. Jourdain, R. Pe´rez Carlo´n and J.-L. Reymond, Tetrahedron Lett.,
1998, 39, 9415–8; (b) R. Pe´rez Carlo´n, N. Jourdain and J.-L. Reymond,
Chem.–Eur. J., 2000, 6, 4154–4162; (c) E. M. Gonzalez-Garcia, V.
Helaine, G. Klein, M. Schuermann, G. Sprenger, W.-D. Fessner and
J.-L. Reymond, Chem.–Eur. J., 2003, 9, 893.
16 (a) B. List, C. F. Barbas and R. A. Lerner, Proc. Natl. Acad. Sci. U. S. A.,
1998, 95, 15351; (b) L. Hoang, S. Bahmanyar, K. N. Houk and B. List,
J. Am. Chem. Soc., 2003, 125, 16–17.
17 (a) J. P. Guthrie, J. Am. Chem. Soc., 1991, 113, 7249; (b) J. P. Guthrie,
Can. J. Chem., 1978, 56, 962.
18 Each bead has a diameter of 100 lm and weighs approximately 1 lg,
corresponding to 200 pmol per bead at 0.2 mmol g−1
.
19 Any sequence has a probability of (65 535/65 536)62 500 = 38.5% of never
being picked in the sample of 50 mg containing 62 500 beads picked
from a large excess (1.25 million beads per g) of library.
20 The reaction cannot be run in organic solvent due to the insolubility of
dihydroxyacetone.
21 (a) J. D. Bass, A. Solovyov, A. J. Pascall and A. Katz, J. Am. Chem.
Soc., 2006, 128, 3737–3747; (b) S. Saito and H. Yamamoto, Acc. Chem.
Res., 2004, 37, 570–579; (c) J. Hine and Y. Chou, J. Org. Chem., 1981,
46, 649–652.
7 For catalytic dendrimers: (a) J. Kofoed and J.-L. Reymond, Curr. Opin.
Chem. Biol., 2005, 9, 656–664; (b) R. Kreiter, A. W. Kleij, R. J. M.
Klein Gebbink and G. van Koten, Top. Curr. Chem., 2001, 217, 163–
199; (c) L. J. Twyman, A. S. H. M. King and I. K. Martin, Chem. Soc.
22 S. S. Yoon and W. C. Still, Tetrahedron, 1995, 51, 567–578.
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 3268–3281 | 3281
©