Emergent Catalytic Behavior of Self-Assembled Low Molecular Weight Peptide-Based Aggregates and Hydrogels

Article abstract of DOI:10.1002/chem.201600344

We report a series of short peptides possessing the sequence (FE)_n or (EF)_n and bearing l-proline at their N-terminus that self-assemble into high aspect ratio aggregates and hydrogels. We show that these aggregates are able to catal

Full text of DOI:10.1002/chem.201600344

DOI: 10.1002/chem.201600344
Full Paper
&
Supramolecular Chemistry
Emergent Catalytic Behavior of Self-Assembled Low Molecular
Weight Peptide-Based Aggregates and Hydrogels
Marta Tena-Solsona,[a] Jayanta Nanda,[b] Santiago Díaz-Oltra,[a] Agata Chotera,[b]
Gonen Ashkenasy,*[b] and Beatriu Escuder*[a]
Abstract: We report a series of short peptides possessing
the sequence (FE)n or (EF)n and bearing l-proline at their N-
terminus that self-assemble into high aspect ratio aggre-
gates and hydrogels. We show that these aggregates are
able to catalyze the aldol reaction, whereas non-aggregated
analogues are catalytically inactive. We have undertaken an
analysis of the results, considering the accessibility of catalyt-
ic sites, pKa value shifts, and the presence of hydrophobic
pockets. We conclude that the presence of hydrophobic re-
gions is indeed relevant for substrate solubilization, but that
the active site accessibility is the key factor for the observed
differences in reaction rates. The results presented here pro-
vide an example of the emergence of a new chemical prop-
erty caused by self-assembly, and support the relevant role
played by self-assembled peptides in prebiotic scenarios. In
this sense, the reported systems can be seen as primitive al-
dolase I mimics, and have been successfully tested for the
synthesis of simple carbohydrate precursors.
Introduction
Fibrillar networks are also found in today’s cellular environ-
ments, either as structural frameworks within the cell (i.e.,
actin), or as components of the extracellular matrix. Remarka-
bly, these are also active components crucial for the cell life
cycle, as in the case of microtubules that participate in cell mo-
tility and division. Besides their structural roles in creating com-
partments and isolating chemicals and processes from the en-
vironment, fibrillar assemblies can help in the co-localization of
reactants and catalytic sites in a manner similar to enzymes,
thus leading to supramolecular catalytic effects. It has been
shown that the formation of arrays of functional groups on the
surface of self-assembled nanostructures, such as micelles[5]
and peptide hydrogel fibers,[6] may activate or enhance their
catalytic efficiency. Moreover, as specific examples relevant to
chemical evolution, short self-assembled peptides have been
used to display efficient self-replication.[7]
The emergence of new functional properties following the
self-assembly of small molecular components into supramolec-
ular structures is fundamental to many applications in the
fields of nanotechnology, catalysis, and nanomedicine. Such
functions evolve due to the transfer of physicochemical infor-
mation at the supramolecular level, and moreover as a result
of developing synergistic features, such as multivalent sub-
strate binding and cooperative reactivity.[1] In this sense, the
most significant emergent event on earth, the emergence of
life itself, has been postulated to have proceeded through the
self-assembly of simple prebiotic molecular components. Sev-
eral artificial protocell models have been consequently report-
ed, based on the self-assembly of amphiphilic components
into closed nanostructures (vesicles, capsules, polymersomes)
encapsulating reactants and catalysts.[2] Alternatively, peptide
membranes have also been proposed to play a role in chemi-
cal evolution, not only through the formation of closed vesi-
cles, but also through other morphologies, such as hollow
tubes, ribbons, and fibers.[3] In fact, in some recent reports,
self-assembled fibrillar networks were advocated as the most
applicable functional scaffolds for simple prebiotic systems.[4]
With the aim of developing supramolecular systems that
enable the emergence of additional catalytic properties so far
found only in the natural counterparts, we present here an ex-
ample of short peptides, each equipped with a catalytic func-
tional group (Scheme 1), which forms fibrillar networks and hy-
drogels. We show that these networks are catalytically active
for a CÀC bond-forming reaction, the direct aldol coupling,
whereas non-assembling analogues are inactive in solution, re-
vealing the emergence of enamine-based catalysis through
self-assembly. Moreover, the effect of subtle changes in the
amino acid sequence on the catalytic efficiency is observed,
suggesting the formation of tailored catalytic sites. Our net-
works represent a new example of a primitive system that
could have acted as a proto-enzyme assisting the production
of simple metabolite precursors. Indeed, CÀC bond-forming re-
actions are crucial for the biosynthesis of many known metab-
olites. Aldolases (transaldolase, transketolase), for example, cat-
[a] Dr. M. Tena-Solsona, Dr. S. Díaz-Oltra, Dr. B. Escuder
Departament de Química Inorgànica i Orgànica
Universitat Jaume I, 12071 Castelló (Spain)
E-mail: escuder@uji.es
[b] Dr. J. Nanda, A. Chotera, Prof. G. Ashkenasy
Department of Chemistry
Ben-Gurion University of the Negev, Be’er Sheva (Israel)
E-mail: gonenash@bgu.ac.il
Supporting information for this article can be found under
http://dx.doi.org/10.1002/chem.201600344.
Chem. Eur. J. 2016, 22, 6687 – 6694
6687
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
lytic studies. Compounds 2 and 4 having longer alkyl
chains were hardly soluble in aqueous solutions, and
their solubility in water could not be precisely deter-
mined because it was below the detection limit of
1H NMR spectroscopy. Compound 2 forms translu-
cent hydrogels after dissolution under heating and
cooling to RT at 1.5 mm, whereas dissolution of com-
pound 4 in such processes leads to a milky suspen-
sion (Figure 1). Self-assembly studies of peptides 5
and 6, possessing longer sequences but no alkyl tails
(in 0.5m phosphate buffer, pH 7), have shown that
both compounds form hydrogels at higher concen-
trations (ꢀ7.5 mm).
The microscopic morphology of these aggregates
was further studied by TEM, Cryo-TEM, and AFM
(Figure 2). The measurements revealed that the hy-
Scheme 1. Structures of peptides.
alyze the coupling of C2/C3 fragments leading to small carbo-
hydrates, such as tetroses and pentoses.[8] At the end of the
paper, we thus show the utility of these new fibrillar networks
in enhancing the synthesis of simple carbohydrate precursors.
Results and Discussion
To study the effect of fibril structure and gel formation on cat-
alysis, a new series of short lipophilic peptides equipped with
l-proline amino acids at their N-terminus was prepared (1–6;
Scheme 1).[9] Proline is known as a versatile organocatalyst for
CÀC forming reactions such as aldol and 1,4-conjugated addi-
tion reactions.[10] Additionally, the alternating sequence phenyl-
alanine (F)–glutamic acid (E) has been chosen to facilitate the
assembly of the peptides into b-sheet structures.[7a,11] We have
recently shown that closely related structures, including alter-
nating non-polar (aromatic)–polar (carboxylate) sequences, are
prone to assemble into b-sheet structures and also form ex-
tended non polar aromatic regions that are able to bind hydro-
phobic guests.[12] Two types of alternating sequences have
been designed, in which the amino acid neighboring proline is
either the F or E, to test their effect on the catalytic results.
Our previous experiments with related dipeptides have high-
lighted a hydrophobic residue located near the catalytic site to
be selective for hydrophobic substrates.[9b] The short peptides
in Scheme 1, compounds 1–4, present alkyl tails at the C-termi-
nus, which decrease their solubility in water. Compounds 5
and 6 are deca-peptides presenting longer alternating sequen-
ces but lacking the hydrophobic tails.
Figure 1. Macroscopic aspect of aggregates of compounds 2 (A, 1.5 mm), 4
(B, 1.5 mm), 5 (C, 7.5 mm) and 6 (D, 7.5 mm).
drogel of compound 2 was formed by the entanglement of
long rigid tubes of about 30 nm in diameter and 10 nm of wall
thickness (Figure 2B). AFM images suggest that these tubes
can be formed by the folding of flat nanostructures (nanobelts
or nanoribbons) into closed nanotubes (Figure 2D and 2E), as
previously reported for similar self-assembled peptide amphi-
philes.[13] Wide angle X-ray diffraction (WAXD) on freeze-dried
xerogel (the Supporting Information, Figure S56A) revealed
the presence of multiple diffraction peaks, following a periodici-
ty typical of a lamellar structure, with a low angle diffraction at
36.5 Š assignable to the width of an interdigitated bilayer. On
the other hand, aggregates of compound 4 formed a mesh of
helical fibers that was not able to percolate the solution into
a gel. These latter fibers of 50–70 nm in diameter present left-
handed twists with a pitch of about 100 nm (the Supporting
Information, Figure 2F–2J). In this case, WAXD of the freeze-
dried xerogel (the Supporting Information, Figure S56B)
showed a lower degree of crystallinity compared to that of the
previous compound and a shorter low-angle diffraction at
28.7 Š.
Self-assembly studies
The molecular arrangement within the different aggregates
was also probed by circular dichroism (CD) (Figures S2–S5 in
the Supporting Information,). The CD spectrum of aggregated
compound 2 revealed a Cotton effect typical to the b-sheet
secondary structure, with a positive lobe at 208 nm, zero-cross-
ing at 217 nm and a negative lobe minimum at 227 nm. These
values appear slightly redshifted relative to the usual amide
spectrum in b-sheet structures, probably indicating a twist
Supramolecular structures produced by the self-assembly of
peptides 1–6 were thoroughly characterized by various tech-
niques, with the main purpose to later on correlate their struc-
tural features with the observed catalysis. Compounds 1 and 3,
possessing the short C3 alkyl chains, were soluble under the
conditions used for catalysis (phosphate buffer 0.1m, pH 7)
and were taken as non-assembling model compounds for cata-
Chem. Eur. J. 2016, 22, 6687 – 6694
www.chemeurj.org
6688
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
again a fibrillar aspect. Compound 5 showed straight and long
fibers of about 30 nm in diameter (the Supporting Information,
Figure S1A,B). However, the resolution of the measurements
did not allow the identification of tubes, as with analogue 2.
Compound 6 presented longer and more flexible, about 20 nm
wide, fibers (the Supporting Information, Figure S1C,D). At
high magnification, some of these fibers displayed helical
structures, albeit less defined than compound 4, confirming
the tendency of the PEF sequence to adopt highly twisted
supramolecular arrangements. CD spectra of aggregated sam-
ples of 5 and 6 at 7.5 mm showed an intense negative band at
235 nm (the Supporting Information, Figure S5). In this case,
the effect of sequence changes on the supramolecular struc-
ture was not as evident as with the short analogues. This ob-
servation supports the hypothesis that strong supramolecular
effects result from competition between the hydrophobicity-
driven packing of the alkyl-tail regions and the H-bonding in
b-sheet regions. These two competing pathways modulate the
supramolecular arrangements in assemblies of 2 and 4, but are
not significant in the case of the decapeptide analogues. A
similar effect has been recently reported by Stupp et al. for
small PAs with alternating hydrophilic (E) and hydrophobic (V)
amino acids.[14] CD spectra of diluted samples of compounds 5
and 6 and of soluble analogues PFEC3 (1) and PEFC3 (3) dis-
played bisignate bands with negative lobes at about 210 nm
and positive lobes at about 220 nm, corresponding to the mo-
lecular CD associated with intramolecularly folded conforma-
tions (see the Supporting Information, Figures S2 and S4 for
CD spectra and S7 for calculated models of folding). In sum-
mary, it is remarkable that minor changes in the amino acid se-
quence, or elimination of the C-terminus alkyl chain, strongly
influence the aggregation tendency and furthermore result in
a significant difference in self-assembly packing and micro-
scopic architectures.
Figure 2. Morphology of aggregates formed by compound 2 (A–E) and com-
pound 4 (F–J) at 1.5 mm obtained by TEM (A,B,F,G), Cryo-TEM (C,H), and
AFM (D,E,I,J).
when the bilayer is closed into a nanotube.[14] Aggregates of
compound 4 showed a different CD spectrum with two less in-
tense negative peaks at 213 nm and 230 nm. This difference
may be attributed to a stronger degree of b-sheets twist
within the nanohelices (the Supporting Information, Figure S3).
FTIR of the xerogels affords additional information regarding
the H-bond networks present in the assemblies (the Support-
ing Information, Figure S19). The spectra of both compounds 2
and 4 showed a broad absorption band above 2500 cmÀ1, cor-
responding to the OÀH stretching vibrations of associated car-
boxylic acids. However, these spectra revealed differences in
bands related to the strength of H-bonding. The bands corre-
sponding to the amide NÀH and C=O stretching vibrations
and to the carboxylic acid C=O stretching vibration appear at
lower wavenumbers in the case of compound 2
Catalysis
The catalytic activity of peptides 1–6 was tested towards the
aldol coupling of cyclohexanone and 4-nitro-benzaldehyde as
a benchmark reaction (Scheme 2). This reaction has been previ-
ously studied, displaying an excellent test for the catalytic ac-
tivity of hydrophobic aggregates in water.[8] In the current
work, a 10% loading of catalyst and 10 equivalents of cyclo-
hexanone per aldehyde equivalent were employed, and the re-
action was run at 258C and buffered at pH 7 (phosphate
buffer; PB). Results appear collected in Table 1. As can be seen,
the use of soluble catalysts 1 and 3 did not lead to significant
aldol product yields even after 72 h (Table 1, entries 2 and 4),
and the outcomes of these reactions were comparable to that
(3284 and 1642 cmÀ1 (broad, amide, and carboxylic
acid C=O stretching)) than for compound 4 (3332,
1675, and 1646 cmÀ1). This difference reflects a lower
extent of b-sheet H-bonding for compound 4, in
agreement with the twisting observed in its fibrils.
In the case of the hydrogels formed by com-
pounds lacking the alkyl tail, TEM images showed Scheme 2. Direct aldol reaction.
Chem. Eur. J. 2016, 22, 6687 – 6694
www.chemeurj.org 6689
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Mechanistic studies
Table 1. Catalytic results for the aldol reaction using compounds 1–6 as
catalysts.[a]
Various examples have been previously published demonstrat-
ing that simple amino acids and peptides can serve as organo-
catalysts for direct “in water” and “on water” aldol reactions.[16]
However, although reports have appeared for such reactions
using emulsions and suspensions of catalysts in water, in most
of the cases the relationship between the supramolecular
structure of the aggregates and the catalysis was not investi-
gated, thus ignoring the paramount effects that may appear
after self-assembly.[17] However, it is important to fully under-
stand the structural characteristics of the catalytic sites and to
clarify the role of self-assembly on the catalytic properties of
the system. The catalytic mechanism for proline-based cata-
lysts has been shown to proceed by the formation of an enam-
ine intermediate between the catalyst and the ketone. After-
wards, the prolyl-enamine participates in the nucleophilic addi-
tion to the aldehyde. Therefore, it is important that the proline
amino group remains in its non-protonated form under the
catalytic conditions. For that reason, pH titration experiments
have been performed to identify the catalyst species at pH 7
(see Figures S15 and S16 in the Supporting Information). Titra-
tions of compounds 1–4 have been performed by starting
from basic solutions that showed no precipitation or gelation,
and then proceeding with subsequent slow addition of acid
aliquots until an acidic final pH was reached. A general
Scheme for the different steps along this process is shown in
Scheme 3, and the resulting protonation constant values are
collected in Table 2.
Entry
Catalyst
t
[h]
Yield[b]
[%]
d.r.[b]
(syn/anti)
e.r.[c]
anti
1
2
3
4
5
6
7
8
–
1
2
3
4
1
2
3
4
2
4
2
4
5
6
72
72
72
72
72
15
15
15
15
2
14
32
>99
17
>99
9
>99
6
>99
93
51
83
32
–
–
n.d.
12:88
n.d.
12:88
–
–
–
–
–
–
–
15:85
14:86
18:82
20:80
15:85
15:85
20:80
9:91
9
10
11
12
13
14
15
8:92
2
1
1
15:85
10:90
14:86
15:85
16:84
–
72
72
85
50
12:88
20:80
[a] Catalytic conditions for compounds 1–4: 6.5”10À3 mmol of catalyst
(10%) and 10 equiv of cyclohexanone in 2 mL of phosphate buffer 0.1m
at pH 7 and 258C. For compounds 5 and 6: 6.5”10À3 mmol of catalyst
(10%) and 10 equiv of cyclohexanone in 0.5 mL of phosphate buffer
0.5m at pH 7 and 258C. [b] Yields and diastereoisomer ratios were deter-
mined by 1H NMR spectroscopy. [c] Enantiomer ratios were determined
by chiral HPLC.
of reactions in control experiments without a catalyst (entry 1).
Remarkably, when self-assembled catalysts 2 and 4 were used,
a quantitative conversion of the aldehyde was obtained after
the same reaction period (Table 1, entries 3 and 5). Shortening
the reaction time revealed a difference in reaction rates be-
tween these two active assemblies (Figure 3). For example,
after 2 h, the aldol yield was 93% for reactions catalyzed by 2
(Table 1, entry 10), whereas for reactions with catalyst 4 it was
only 51% (entry 11).
Table 2. pKa values for compounds 1–4 determined by potentiometric ti-
trations.
Catalyst
pK1
pK2
Decapeptides 5 and 6 were also studied as catalysts, and it
was observed that reactions in the presence of these com-
pounds were significantly slower than in the presence of the
tripeptide analogues 2 and 4; even after 72 h, substrate con-
version was not yet complete (Table 1, entries 14 and 15).
1
2
3
4
8.22Æ0.02
9.72Æ0.03
8.27Æ0.02
7.60Æ0.01
4.34Æ0.02
6.85Æ0.04
4.21Æ0.04
6.92Æ0.01
In the case of compounds 1 and 3, which are soluble for the
entire pH range, the values for pK1 are 8.22 and 8.27, respec-
tively, meaning that at pH 7 the predominant species would
be a zwitterion (PH+À in Scheme 2). As a consequence, the
proline residue would be protonated and not available for the
enamine-based catalysis. On the other hand, compounds 2
and 4 bearing a C12 alkyl tail are much more hydrophobic,
and their molecular species at neutral pH tend to aggregate
into a hydrogel and a sticky suspension, respectively (PH in
Scheme 2). It is therefore reasonable to propose that in these
latter aggregates, which are formed by neutral non-charged
molecules as confirmed by the presence of ÀCOOH bands in
the FTIR spectra of freeze dried samples of aggregated 2 and 4
(the Supporting Information, Figure S19), the proline residues
are readily available to react with the substrates. The formation
of these aggregates has an effect on the pK1 values, which are
9.72 and 7.60 for 2 and 4, respectively.[18] However, both com-
Figure 3. Yield of the aldol reactions as a function of time in the presence of
~
*
self-assembled catalysts 2 ( ) and 4 ( ).
Chem. Eur. J. 2016, 22, 6687 – 6694
www.chemeurj.org
6690
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 3. pH-dependent speciation exemplified for compounds 1 and 2. Inset: detailed diagram including equilibrium constants valid for compounds 1–4.
pounds are strongly aggregated at pH 7 and, as can be seen in
the speciation diagram, at pH 7 the amount of the active spe-
cies PH is of 60% for compound 2 and 40% for compound 4
in pure water. Moreover, the catalytic experiments are per-
formed in phosphate buffer that causes a dramatic decrease of
the solubility due to the so-called Hofmeister effect,[19] as the
phosphate ion belongs to the group of ions that usually lower
the solubility of hydrophobic solutes in water, termed kosmo-
tropic ions.[20] For the studied case, this effect is strongly active
Kinetic studies were performed to analyze a possible
enzyme-like mechanism. Catalytic experiments were performed
by keeping the catalyst concentration constant and varying
the total concentration of substrate (at constant cyclohexa-
none/aldehyde ratio). The observed kinetics did not follow the
typical Michaelis–Menten velocity profile, and therefore rules
out the formation of a pre-organized catalyst–substrate com-
plex (Figures S20 and S21 in the Supporting Information).
On the other hand, p-nitrobenzaldehyde is poorly soluble in
water, and therefore should be incorporated into the hydro-
phobic regions of the gel phase to get access to the catalytic
center and react with the enamine intermediate. 1-Anilino-
naphthalene-8-sulphonate (ANS), a fluorescent probe exten-
sively used to track hydrophobic regions in proteins, was em-
1
for both compounds 2 and 4. Indeed, the H NMR spectra of
the two compounds in buffer at pH 7 did not show any signal
corresponding to free compounds, revealing that the soluble
fraction of both falls below the detection limit of this tech-
nique (<5%), and therefore >95% of the catalyst molecules
are in the non-protonated active form at that pH (the Support-
ing Information, Figures S17 and S18). All these data clearly
show that proline amino groups are non-protonated and ready
to react with electrophiles under the conditions used for the
catalytic experiments.
ployed to reveal the hydrophobicity of the aggregates.[21]
A
fluorescence blueshift is expected upon ANS binding, along
with an increase in the fluorescence intensity, due to an in-
crease in the rigidity of the probe within the hydrophobic
pockets. As can be seen in Figure S22 (the Supporting Informa-
tion), a fluorescence emission band appears at about 460 nm
in the spectrum of both compounds 2 and 4, blueshifted from
the weak emission of the blank solution at 490 nm. This band,
which is more intense for compound 4 than for compound 2,
corroborates the presence of hydrophobic regions. In the case
of compounds 5 and 6, the two compounds present a catalytic
residue similar to compounds 1–4, but they present longer se-
quences and do not bear alkyl tails at the C-terminus. The
ANS binding assays performed for compounds 5 and 6 in
comparison with those of compounds 2 and 4 revealed a sig-
nificantly lower hydrophobicity. On the other hand, the
amount of incorporated cyclohexanone was also quantified for
these compounds using 1H NMR spectroscopy, showing that
less than 0.1 equiv of cyclohexanone did react. These two re-
sults support the explanation that the slower catalysis rate ob-
served for 5 and 6, compared with 2 and 4, is due to a combi-
To understand the observed difference between the catalytic
activity of compounds 2 and 4, we then decided to study the
accessibility of the substrates to the catalytic sites. Therefore,
an experiment was set up in which samples of self-assembled
catalysts and a smaller excess of cyclohexanone (2.5 equiv)
were prepared, and the amount of cyclohexanone incorporat-
1
ed into the aggregated phase was quantified by H NMR spec-
troscopy (see the Supporting Information for details). Cyclo-
hexanone is significantly soluble in aqueous solution and
would only be incorporated into the aggregates if it reacts
with a proline residue forming an enamine. Interestingly, it was
observed that one equivalent of cyclohexanone was incorpo-
rated into the gel formed by catalyst 2, whereas aggregates of
catalyst 4 only entrapped 0.5 equiv of this ketone. This sug-
gests that catalytic sites of compound 2 are more accessible to
cyclohexanone than those of catalyst 4.
Chem. Eur. J. 2016, 22, 6687 – 6694
www.chemeurj.org
6691
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
nation of poor accessibility of both reagents to the catalytic
fibers.
Table 3. Catalytic results for the self-condensation of aldehydes 7a–
c and 9.[a]
Remarkably, a difference in catalytic performance related to
the peptide sequence could be observed again in favor of
compound 5, which contains a phenylalanine residue next to
proline. This general trend suggests that F plays a role in catal-
ysis, probably through secondary interactions at the transition
state of the reaction. It has to be mentioned that sequence
variations did not affect significantly the stereoselectivity of
the reactions. In all experiments, diastereomeric ratios syn/anti
were in the range 10:90 to 20:80, and enantiomer ratios for
the anti product were about 10:90. These facts support the hy-
pothesis that there is no specific substrate binding close to the
catalytic site but a preferential partitioning of the substrates
into the hydrophobic region of the gel fibers that is key to the
acceleration of the reaction, a trend that is also observed in
micellar catalysis.[5]
Entry
Catalyst
Aldehyde
t
[h]
Aldol
yield [%][b]
1
2
3
4
5
6
7
8
2
4
2
4
2
4
2
4
2
4
2
4
7a
7a
7b
7b
7c
7c
7c
7c
9
48
48
15
15
2
<5
<5
>95[c]
>95[c]
84[d]
50[d]
92
76
17
2
57
2
15
15
2
9
10
11
12
9
9
9
2
15
15
24
[a] Catalytic conditions: 6.5”10À3 mmol of catalyst (4.37 mg, 10%) and
6.5”10À2 mmol of the corresponding aldehyde in 2 mL of phosphate
buffer 0.1m at pH 7 and 258C. [b] Yields and diastereomeric ratio are de-
termined by 1H NMR spectroscopy (see the Supporting Information).
[c] Mixture of compounds, see text. [d] d.r. (syn/anti), 10:90, e.r. anti prod-
uct 91:9, determined as reported in ref. [22], see the Supporting Informa-
tion for details.
Biomimetic catalysis
At this point, the emergence of catalytic activity through pep-
tide self-assembly has been established, and a further step is
taken towards catalysis of biologically relevant aldol product
formation. Taking Nature as an inspiration, aldol reactions can
be used to obtain carbohydrate derivatives.[22] Aldolase I uses
catalytic enamine activation in the metabolic biosynthetic
pathways of carbohydrates. For instance, dihydroxyacetone
phosphate (DHP) is activated for the subsequent coupling with
C2 or C3 aldehydes, leading to either tetroses or pentoses, re-
spectively.
tives (7a–c) were studied as substrates leading to tetrose-
based products.[24] Results appear collected in Table 3.
Unprotected 2-hydroxyacetaldehyde (7a) did not react in
the presence of any of the self-assembled catalyst, even after
a long reaction time (Table 3, entries 1 and 2). However, the re-
action was accomplished quite fast when the a-hydroxyl
group was protected as a silyl ether (7b, 7c). The inactivity of
self-assembled catalysts against polar substrates has been ob-
served earlier for related systems, and supports the fact that
a hydrophobic environment of the fibers is required to dis-
solve/store the reagent near the catalytic site.[9b] Triisopropylsil-
yl (TIPS)-protected glycolaldehyde (7b) reacted as previously
described by Clarke et al., leading to mixtures of aldols togeth-
er with a trimeric acetal (see Figure S10 in the Supporting In-
formation).[17b] However, different results were ob-
tained for reactions of tert-butyldiphenylsilyl (TBDPS)
-protected analogue 7c. In this case, using self-as-
sembled catalysts 2 and 4, considerable yields of
aldol were already obtained after a few hours of reac-
tion. Again, the best results in terms of rate appeared
for catalyst 2, with a yield of 80% after 2 h of reac-
tion, compared with 50% obtained for catalyst 4.
Moreover, the aldehyde-containing products did not
participate in further aldol reactions for both catalysts
(see Figure S11 in the Supporting Information). It
seems that the presence of aromatic fragments in
substrate and F residues in the catalyst play a role
that favors the aldol product. The major product of
the reaction was determined to be the l-2-anti-eryth-
rose derivative, revealing a remarkable stereoselectiv-
ity (d.r. 90:10, enantiomeric ratio (e.r.) 91:9).
In this context, the relevance of organocatalysis in water,
and in particular proline catalysis, for the prebiotic origin of
carbohydrates, is evident and has been highlighted by several
authors.[23] Therefore, we decided to study the hydrogels and
aggregates formed by compounds 2 and 4 as catalysts for the
self-condensation of several a-oxy-aldehydes and phenylalkyl
aldehydes (Scheme 4). Firstly, 2-hydroxyacetaldehyde deriva-
In addition to the study of hydroxyacetaldehyde
derivatives, 2-phenyl acetaldehyde (8) and 3-phenyl
propionaldehyde (9) were used as substrates, testing
Scheme 4. Self-condensation of aldehydes.
Chem. Eur. J. 2016, 22, 6687 – 6694
www.chemeurj.org
6692
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
the role of the a-oxy group. When studying aldehyde 8, a com-
plex mixture of products was observed in H NMR analysis of
Acknowledgements
1
the crude, showing multiple aldehyde CÀH singlets between
d=9 and 10 ppm (not shown in Table 3; see Figure S12 in the
Supporting Information). Furthermore, an analysis of this crude
material by ESI-MS revealed the presence of different peaks
corresponding to poly-condensation products. In the case of
aldehyde 9, the main reaction product was the dehydrated
aldol (see Table 3, entries 9–12 and Figure S10 in the Support-
ing Information). For both aldehydes, using either 2 or 4 as
catalyst, the rates of reaction were slower than for the a-oxy-
analogues. Nevertheless, from a prebiotic chemistry point of
view, these condensation reactions resulted in the formation of
polyaldols, potential precursors of larger carbohydrates and
polyketide natural products.[24]
This work was supported by the Ministry of Economy and
Competitiveness of Spain (Grant CTQ2012-37735) and Universi-
tat Jaume I (Grant P1-1B2013-57). M.T.-S. thanks the Ministry of
Education, Culture and Sport of Spain for an FPU fellowship.
J.N. thanks the PBC program for post-doctoral fellowship. We
acknowledge the COST actions CM1005 and CM1304. We
thank Dr. N. Wagner (BGU) for help in editing the manuscript.
Keywords: aldol reaction · gels · peptides · self-assembly ·
supramolecular chemistry
[1] a) J. M. Lehn, Proc. Natl. Acad. Sci. USA 2002, 99, 4763; b) A. Mulder, J.
Huskens, D. N. Reinhoultd, Org. Biomol. Chem. 2004, 2, 3409; c) J. D.
Badjic, A. Nelson, S. J. Cantrill, W. B. Turnbull, J. F. Stoddart, Acc. Chem.
Res. 2005, 38, 723; d) L. C. Palmer, Y. S. Velichko, M. Olvera de La Cruz,
S. I. Stupp, Phil. Trans. RSC A: Math. Phys. Eng. Sci. 2007, 365, 1417; e) A.
Barnard, D. K. Smith, Angew. Chem. Int. Ed. 2012, 51, 6572; Angew.
Chem. 2012, 124, 6676; f) C. Fasting, C. A. Schalley, M. Weber, O. Seitz, S.
Hecht, B. Koksch, J. Dernedde, C. Graf, E.-W. Knapp, R. Haag, Angew.
Chem. Int. Ed. 2012, 51, 10472; Angew. Chem. 2012, 124, 10622; g) J. M.
Lehn, Angew. Chem. Int. Ed. 2013, 52, 2836; Angew. Chem. 2013, 125,
2906.
[2] a) D. Deamer, S. Singaram, S. Rajamani, V. Kompanichenko, S. Guggen-
heim, Phil. Trans. R. Soc. B 2006, 361, 1809; b) K. Kunnihara, M. Tamura,
K. Shohda, T. Toyota, K. Suzuki, T. Sugawara, Nat. Chem. 2011, 3, 775;
c) A. J. Dzieciol, S. Mann, Chem. Soc. Rev. 2012, 41, 79; d) K. Adamala,
J. W. Szostak, Nat. Chem. 2013, 5, 495; e) X. Huang, M. Li, D. C. Green,
D. S. Williams, A. J. Patil, S. Mann, Nat. Commun. 2013, 4, 2239; f) P.
Wade, H. Umakoshi, P. Stano, F. Mavelli, Chem. Commun. 2014, 50,
10177; g) X. Huang, A. J. Patil, M. Li, S. Mann, J. Am. Chem. Soc. 2014,
136, 9225.
[3] W. S. Childers, R. Ni, A. K. Mehta, D. G. Lynn, Curr. Opin. Chem. Biol.
2009, 13, 652.
[4] A. M. Brizard, J. H. van Esch, Soft Matter 2009, 5, 1320.
[5] For a review on catalytic micelles see: T. Dwars, E. Paetzold, G. Oehme,
Angew. Chem. Int. Ed. 2005, 44, 7174–7199; Angew. Chem. 2005, 117,
7338–7364.
[6] For recent examples of catalytic peptide fibers see: a) M. O. Guler, S. I.
Stupp, J. Am. Chem. Soc. 2007, 129, 12082; b) F. Rodríguez-Llansola, B.
Escuder, J. F. Miravet, J. Am. Chem. Soc. 2009, 131, 11478; c) Z. P. Huang,
S. W. Guan, Y. G. Wang, G. N. Shi, L. N. Cao, Y. Z. Gao, Z. Y. Dong, J. Y. Xu,
Q. Luo, J. Q. Liu, J. Mater. Chem. B 2013, 1, 2297–2304; d) C. M. Rufo,
Y. S. Moroz, O. V. Moroz, J. Stçhr, T. A. Smith, X. Hu, W. F. DeGrado, I. V.
Korendovych, Nat. Chem. 2014, 6, 303; e) C. Zhang, X. Xue, Q. Luo, Y. Li,
K. Yang, X. Zhuang, Y. Jiang, J. Zhang, J. Liu, G. Zou, X.-J. Liang, ACS
Nano 2014, 8, 11715; f) N. Singh, M. P. Conte, R. V. Ulijn, J. F. Miravet, B.
Escuder, Chem. Commun. 2015, 51, 13213; g) M. BØli›res, N. Chouini-La-
lanne, C. Dejugnat, RSC Adv. 2015, 5, 35830; h) K. S. Lee, J. R. Parquette,
Chem. Commun. 2015, 51, 15653; i) N. Singh, M. Tena-Solsona, J. F. Mira-
vet, B. Escuder, Isr. J. Chem. 2015, 55, 711; j) K. L. Duncan, R. V. Ulijn, Bio-
catalysis 2015, 1, 67.
Conclusion
We have presented here an example of the emergence of cata-
lytic activity through the self-assembly of short peptides in
water. Although the use of peptides in catalysis is widely re-
ported, in particular the so-called “on-water” reactions, we
were able to further rationalize here the catalytic activity in
terms of the supramolecular structure of the system, which is
not generally considered in the field of organocatalysis.
We have shown that the formation of aggregates is funda-
mental for the development of catalytic activity by comparing
the assembling compounds 2 and 4 with the soluble ana-
logues 1 and 3. Moreover, we have determined that there is
a relationship between the amino acid sequence, the supra-
molecular and microscopic organization of the material, and
the accessibility of the substrates to the reaction site. Addition-
ally, the presence of hydrophobic fragments and especially
alkyl tails has been shown to help in the solubilization of the
hydrophobic substrates and improve their catalytic per-
formance; compounds 5 and 6 lacking these tails had a re-
duced catalytic efficiency. It is noteworthy to remark that these
functional materials are formed by simple molecules with pre-
programmed information, which is in turn translated into com-
plex catalytic systems solely through noncovalent interactions.
In this sense, catalytic self-assembled peptides could be con-
sidered as supramolecular protein-like functional materials.
Finally, it is important to highlight that CÀC bond-forming
reactions such as aldol condensations are of great value be-
cause they increase molecular complexity and structural diver-
sity in one simple step. The results presented here are thus rel-
evant in the context of prebiotic chemistry. Probable prebiotic
scenarios for the synthesis of small peptides have thus been
proposed. In combination with the role of self-assembly, both
the formation of compartments and the emergence of cataly-
sis, they point to a plausible pathway towards the appearance
of chemical complexity.[25] In this context, catalytic self-assem-
bled peptides can be seen as intermediate evolutionary sys-
tems, between simple soluble molecular catalysts and complex
protein entities, namely, the enzymes.
[7] a) B. Rubinov, N. Wagner, H. Rapaport, G. Ashkenasy, Angew. Chem.
2009, 121, 6811; b) J. M. A. Carnall, C. A. Waudby, A. M. Belenguer,
M. C. A. Stuart, J. J. P. Peyralans, S. Otto, Science 2010, 327, 1502; c) B.
Rubinov, N. Wagner, M. Matmor, O. Regev, N. Ashkenasy, G. Ashkenasy,
ACS Nano 2012, 6, 7893.
[8] J. L. Tymoczko, J. M. Berg, L. Stryer, Biochemistry:
2nd ed., W. H. Freeman and Co., New York, 2013.
a Short Course,
[9] a) F. Rodríguez-Llansola, J. F. Miravet, B. Escuder, Chem. Commun. 2009,
7303; b) C. Berdugo, J. F. Miravet, B. Escuder, Chem. Commun. 2013, 49,
10608.
[10] B. List in Modern Aldol Reactions, Vol. 1 (Ed.: R. Mahrwald), Wiley-VCH,
Weinheim, 2004, p.161.
[11] a) H. Rapaport, K. Kjaer, T. R. Jensen, L. Leiserowitz, D. A. Tirrell, J. Am.
Chem. Soc. 2000, 122, 12523; b) N. R. Lee, C. J. Bowerman, B. J. Nilsson,
Chem. Eur. J. 2016, 22, 6687 – 6694
www.chemeurj.org
6693
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Pept. Sci. 2013, 100, 738; c) Y. Raz, B. Rubinov, M. Matmor, H. Rapaport,
G. Ashkenasy, Y. Miller, Chem. Commun. 2013, 49, 6561.
aggregation effects on pKa see, for instance: C. Tang, A. M. Smith, R. F.
Collins, R. V. Ulijn, A. Saiani, Langmuir 2009, 25, 9447.
[12] a) M. Tena-Solsona, J. F. Miravet, B. Escuder, Chem. Eur. J. 2014, 20, 1023;
b) M. Tena-Solsona, S. Alonso-de Castro, J. F. Miravet, B. Escuder, J.
Mater. Chem. B 2014, 2, 6192.
[13] See for instance: a) T. Shimizu, M. Masuda, H. Minamikawa, Chem. Rev.
2005, 105, 1401; b) L. Ziserman, H.-Y. Lee, S. R. Raghavan, A. Mor, D.
Danino, J. Am. Chem. Soc. 2011, 133, 2511; c) T. G. Barclay, K. Constanto-
poulos, J. Matisons, Chem. Rev. 2014, 114, 10217.
[14] a) E. T. Pashuck, H. Cui, S. I. Stupp, J. Am. Chem. Soc. 2010, 132, 6041;
b) E. T. Pashuck, S. I. Stupp, J. Am. Chem. Soc. 2010, 132, 8819.
[15] H. Cui, A. G. Cheetham, E. T. Pashuck, S. I. Stupp, J. Am. Chem. Soc.
2014, 136, 12461.
[16] E. A. C. Davie, S. M. Mennen, Y. Xu, S. J. Miller, Chem. Rev. 2007, 107,
5759.
[19] V. J. Nebot, J. J. Ojeda-Flores, J. Smets, S. Fernµndez Prieto, B. Escuder,
J. F. Miravet, Chem. Eur. J. 2014, 20, 14465.
[20] a) P. Lo Nostro, B. W. Ninham, Chem. Rev. 2012, 112, 2286; b) P. Lo Nos-
tro, A. Lo Nostro, B. W. Ninham, G. Pesavento, L. Fratoni, P. Baglioni,
Curr. Opin. Colloid Interface Sci. 2004, 9, 97.
[21] a) E. Schçnbrunn, S. Eschenburg, K. Luger, W. Kabsch, N. Amrhein, Proc.
Natl. Acad. Sci. USA 2000, 97, 6345; b) G. Chadha, Y. Zhao, J. Colloid In-
terface Sci. 2013, 390, 151.
[22] a) A. B. Northrup, D. W. C. MacMillan, Science 2004, 305, 1752; b) N.
Mase, C. F. Barbas III, Org. Biomol. Chem. 2010, 8, 4043.
[23] a) J. E. Hein, D. G. Blackmond, Acc. Chem. Res. 2012, 45, 2045; b) J. Mly-
narski, B. Gut, Chem. Soc. Rev. 2012, 41, 587.
[24] a) J. Rohr, Angew. Chem. Int. Ed. 2000, 39, 2847; Angew. Chem. 2000,
112, 2967; b) A. M. P. Koskinen, K. Karisalmi, Chem. Soc. Rev. 2005, 34,
677.
[17] a) N. Mase, Y. Nakai, N. Ohara, H. Yoda, K. Takabe, F. Tanaka, C. F. Barbas,
J. Am. Chem. Soc. 2006, 128, 734; b) L. Burroughs, P. A. Clarke, H. Forin-
tos, J. A. R. Gilks, C. J. Hayes, M. E. Vale, W. Wade, M. Zbytniewski, Org.
Biomol. Chem. 2012, 10, 1565; c) J. DuschmalØ, S. Kohrt, H. Wennemers,
Chem. Commun. 2014, 50, 8109.
[25] a) G. Danger, R. Plasson, R. Pascal, Chem. Soc. Rev. 2012, 41, 5416; b) K.
Ruiz-Mirazo, C. Briones, A. de La Escosura, Chem. Rev. 2014, 114, 285.
[18] There is an intriguing effect of aggregation on pKa values of com-
pounds 2 and 4. The relationship between the structure of aggregates
and pKa values is currently under study. For recent examples on peptide
Received: January 25, 2016
Published online on March 23, 2016
Chem. Eur. J. 2016, 22, 6687 – 6694
www.chemeurj.org
6694
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim