Simple Design of an Enzyme-Inspired Supported Catalyst Based on a Catalytic Triad

Article abstract of DOI:10.1016/j.chempr.2017.04.004

Enzyme active sites afford an intricate interplay of functional groups to mediate complex organic and inorganic reactions. Many hydrolytic enzymes use a catalytic triad comprising three different functional residues—(Ser(-OH), Hist(-imidazole), Asp(-CO₂H))—that catalyze the hydrolysis of numerous unique substrates. Inspired by this design, we have developed a simple one-step synthesis for preparing a new supported catalytic system in which the three reactive groups of the catalytic triad (alcohol, imidazole, and carboxylate) are incorporated into a single functional unit. These artificial active sites can be coupled to a solid-phase support (Merrifield resin) by copper(I)-catalyzed azide-alkyne cycloaddition “click chemistry,” and their effectiveness as esterolysis catalysts was demonstrated. Furthermore, tuning the local hydrophobicity of the resin particles with an approach analogous to the native enzyme hydrophobic pocket increased the catalytic efficiency. Quantum mechanics and molecular dynamics computational modeling were used to probe the catalytic effect and suggested a concerted two-step mechanism and hydrophobic nanoenvironment similar to that of hydrolytic enzymes.

Full text of DOI:10.1016/j.chempr.2017.04.004

Article
Simple Design of an Enzyme-Inspired
Supported Catalyst Based on a Catalytic Triad
Mitchell D. Nothling,
Aravindhan Ganesan, Karmen
Condic-Jurkic, ..., Megan L.
O’Mara, Michelle L. Coote,
Luke A. Connal
luke.connal@unimelb.edu.au
HIGHLIGHTS
Hydrolase-inspired catalyst based
on a ubiquitous active site: the
catalytic triad
A versatile synthesis enables
immobilization of catalyst onto
engineered supports
Inspired by enzyme hydrophobic
pockets, hydrophobic
modification increases catalysis
Computational modeling
supports a concerted mechanism
similar to that of native enzymes
The utility of enzymes as industrial catalysts is reduced by their low stability and
limited operating range. Connal and co-workers have developed a simple, one-
step procedure to afford an enzyme-inspired catalyst that contains the common
functional motif present at the active site of many hydrolases—the catalytic triad—
incorporated into a single unit. This enzyme-inspired catalyst is then easily
immobilized onto a resin support and hydrophobically tuned to perform enzyme-
like hydrolytic catalysis. Computational modeling provides clues to a concerted,
two-step mechanism analogous to native enzymes.
Nothling et al., Chem 2, 732–745
May 11, 2017 ª 2017 Elsevier Inc.
http://dx.doi.org/10.1016/j.chempr.2017.04.004

Article
Simple Design of an Enzyme-Inspired
Supported Catalyst
Based on a Catalytic Triad
Mitchell D. Nothling,1 Aravindhan Ganesan,2 Karmen Condic-Jurkic,3 Eric Pressly,4 Ashley Davalos,4
Michael R. Gotrik,4 Zeyun Xiao,1 Ezat Khoshdel,5 Craig J. Hawker,4,6 Megan L. O’Mara,3
Michelle L. Coote,2 and Luke A. Connal1,7,
*
SUMMARY
The Bigger Picture
The remarkable catalytic
Enzyme active sites afford an intricate interplay of functional groups to mediate
complex organic and inorganic reactions. Many hydrolytic enzymes use a
catalytic triad comprising three different functional residues—(Ser(-OH),
Hist(-imidazole), Asp(-CO2H))—that catalyze the hydrolysis of numerous unique
substrates. Inspired by this design, we have developed a simple one-step syn-
thesis for preparing a new supported catalytic system in which the three reactive
groups of the catalytic triad (alcohol, imidazole, and carboxylate) are incorpo-
rated into a single functional unit. These artificial active sites can be coupled
to a solid-phase support (Merrifield resin) by copper(I)-catalyzed azide-alkyne
cycloaddition ‘‘click chemistry,’’ and their effectiveness as esterolysis catalysts
was demonstrated. Furthermore, tuning the local hydrophobicity of the resin
particles with an approach analogous to the native enzyme hydrophobic pocket
increased the catalytic efficiency. Quantum mechanics and molecular dynamics
computational modeling were used to probe the catalytic effect and suggested
a concerted two-step mechanism and hydrophobic nanoenvironment similar to
that of hydrolytic enzymes.
performance of enzymes has
enabled their use in many
industrial applications. However,
the inherent instability and
restricted operating window of
enzymes limit their application.
The development of more robust
catalysts inspired by enzymes is a
significant step toward the United
Nations Sustainable
Developmental Goals for
responsible consumption and
production. The catalytic triad is a
textbook example of the complex
chemistry that enzymes control at
their active sites. We report an
approach to enzyme-inspired
catalysis by combining the three
groups of the catalytic triad into a
single molecule through a
INTRODUCTION
Nature has evolved efficient and highly orthogonal chemical strategies to control
complex multi-step chemical processes. This control has been optimized for many
different chemical transformations, and the specificity of enzymatic reactions is crit-
ical to many biological processes. In contrast, the design and performance of syn-
thetic catalysts that mimic enzymes in both range and efficiency of the chemical
transformations continues to be a major challenge.
versatile, one-step procedure. We
show that by coupling these active
sites onto hydrophobically
modified polymer resins, we can
tune the active site’s local
To address this challenge, we drew our attention to esterases, which are important
industrial and academic enzymes with broad applicability from food and paper
manufacturing to detergent formulations and pharmaceuticals.1 The mechanism
of many esterases is similar to the classic serine protease example of chymotrypsin,2
whereby a catalytic triad at the active site modulates the reaction. This textbook
example of an enzyme active site consists of serine, histidine, and aspartate residues
and has been studied extensively with the aspartate forming a hydrogen bond with
the imidazole ring of the histidine, which in turn deprotonates the alcohol of the
serine to form a potent nucleophile.3 A key enzymatic feature that is clearly demon-
strated in these triad systems is the presence of a ternary structure, which both gen-
erates a hydrophobic substrate-binding pocket and optimally positions the triad
environment to increase catalysis.
In addition to new industrial
catalysts, this study provides
insight into the mechanism of
enzymes, highlighting the
concerted action of multiple
functional groups to afford
catalysis.
732 Chem 2, 732–745, May 11, 2017 ª 2017 Elsevier Inc.

Figure 1. Multiple Catalytic Sites and Hydrophobic Tuning Afforded by Functionalized Resins
Comparison of the native catalytic active site—the catalytic triad—of esterases and the artificial
enzyme mimic loaded into polystyrene Merrifield resin.
residues in close proximity. Building on this bio-inspiration, a long-standing oppor-
tunity in catalyst design has been to develop simple synthetic systems that can
perform similar chemistry without the need for well-defined tertiary structures. Pio-
neering work by Breslow and Overman,4 Breslow,5 and Overberger et al.6 led to the
development of a range of strategies in the general area of enzyme mimicry. Exam-
ples of bio-inspired catalysis7 include metal-based active sties,8–12 polymeric,6,13–19
peptidic,20,21 and dendritic supports,22,23 and catalytic dyads.24,25
1Department of Chemical and Biomolecular
Engineering, University of Melbourne,
Melbourne, VIC 3010, Australia
In designing enzyme-inspired catalysts (EICs) based on esterases, our attention was
drawn to the challenge of introducing all three functional groups of the catalytic
triad—carboxylate, imidazole, and hydroxyl groups—into a single compound.
Importantly, we present a simple and versatile design strategy for the preparation
of a supported EIC with esterolytic activity from readily available starting materials.
The relative arrangement of functional groups can be further tuned via judicious se-
lection of the starting material. In addition to the functional units of the catalytic
triad, an acetylene group is also incorporated into the overall design for orthogonal
attachment to various solid supports. To illustrate the power and versatility of this
approach, we designed, synthesized, and evaluated the esterolytic activty of a range
of functionalized Merrifield resins26 that mimic the chemistry of the catalytic triad
(Figure 1). We also developed and evaluated control structures, where the three
groups of the catalytic triad were systematically eliminated, to confirm the impor-
tance of all three functional moieties (Figure 2D). We used computational modeling
of the reaction to support both the formation of a hydrophobic local environment
and the sequence of a two-step, acylation-deacylation catalytic mechanism, similar
to that performed by hydrolytic enzymes.
2ARC Centre of Excellence for Electromaterials
Science, Research School of Chemistry,
Australian National University, Canberra,
ACT 2601, Australia
3Research School of Chemistry, Australian
National University, Canberra, ACT 2601,
Australia
4Materials Research Laboratory and Materials
Department, University of California,
Santa Barbara, Santa Barbara, CA 93106, USA
5Unilever Research & Development Port Sunlight
Laboratory, Unilever, Bebington CH63 3JW, UK
6Department of Chemistry and Biochemistry,
University of California, Santa Barbara, Santa
Barbara, CA 93106, USA
7Lead Contact
*Correspondence: luke.connal@unimelb.edu.au
http://dx.doi.org/10.1016/j.chempr.2017.04.004
Chem 2, 732–745, May 11, 2017 733

A
B
C
D
Figure 2. Synthesis and Immobilization of an Artificial Catalytic Triad
(A) One-step synthesis of the protected artificial triad (M1) with a simple ring opening of a functional epoxide with a protected histidine analog.
(B) Immobilization of the artificial triad onto modified Merrifield resin by copper-catalyzed azide acetylene click chemistry and subsequent deprotection
to yield catalyst resins R1–R6.
(C) Inclusion of a C16 alkyl group to tune the resin’s local hydrophobicity with a second click reaction and subsequent deprotection to yield
hydrophobically modified resins R7–R11 (refer to Figures S1 and S2).
(D) Alternative mono-, di-, and tri-functionalized catalytic control structures used for confirming the concerted performance of the triad design during
catalysis (protecting groups: trt, trityl (triphenylmethyl); Boc, tert-butyl).
RESULTS AND DISCUSSION
Key to this work is the facile synthesis of multi-functional building blocks as the start-
ing motif. The utilization of a protected histidine derivative (H-Hist(1-Trt)-OtBu)
conveniently affords access to two of the functional groups of the catalytic triad:
the imidazole and carboxylate groups. The hydroxyl group is then introduced by
simple ring opening of an acetylene-functionalized epoxide (1) by the primary amine
of the protected histidine (2) (Figure 2A). Versatility is afforded here by the wide
range of available starting materials, giving access to a myriad of unique EIC dyad
and triad arrangements (Figure 2D). A single synthetic step leads to a multi-func-
tional building block containing the three groups of the catalytic triad together
734 Chem 2, 732–745, May 11, 2017

with an acetylene group for orthogonal conjugation to a wide variety of surfaces via
copper(I)-catalyzed azide-alkyne cycloaddition (CuAAC) chemistry.
Initially, commercial Merrifield resin (3.0 mmol gꢀ1 loading, polystyrene cross-linked
with 1% divinylbenzene, 100–200 mesh)27 was chosen as the support. Quantitative
reaction with sodium azide then afforded the desired azide functional resin particles
(3.0 mmol gꢀ1 of azide), which could be coupled with the triad functional molecule
(M1) under orthogonal CuAAC click chemistry conditions. This afforded access to a
range of resins with different loading of the catalytic unit (Figure 1). The combination
of monitoring the attenuation of the azide group (2,100 cmꢀ1) by Fourier transform
infrared measurements and secondary coupling with an acetylene-functionalized
coumarin dye followed by UV-visible (UV-vis) spectroscoopy analysis, allows the
level of M1 incorporation into the resin support to be quantified and confirmed (Fig-
ures S1 and S2). Treatment of the functional resins with trifluoroacetic acid (TFA) then
removes the tert-butyl and trityl protecting groups to yield the free acid, alcohol, and
imidazole functional groups immobilized on the polystyrene resin support. The ver-
satile nature of this process allows a range of particles with M1 loading varying from
0.1 to 3 mmol gꢀ1 to be prepared (R1–R6).
We could then screen esterase activity by observing the hydrolysis of p-nitrophenyl
butyrate (PNB) in the presence of the enzyme-inspired catalytic resins (R1–R6). In
contrast to the traditional p-nitrophenyl acetate model system, PNB is significantly
more stable under the assay conditions and has negligible background hydrolysis,
making PNB a more challenging and realistic target. Significantly, esterase-like
catalysis was observed for all EICs (R1–R6); however, it was not observed for the cat-
alytic control structures lacking an element of the catalytic triad (Figure 3A).
The stepwise nature of the resin functionalization procedure afforded access to a wide
series of candidate materials, which were then evaluated for esterolysis. Resins contain-
ing the artificial triad functionality—with and without hydrophobic modification—as well
as control structures lacking an element of the artificial triad, were all investigated. The
specific triad arrangement (M1) was selected for further analysis given that the alternate
triad and dyad candidates exhibited little esterolytic effect. We then implemented the
well-established Michaelis-Menten model for enzyme kinetics to compare the reaction
kinetics for each of the catalytic materials (Table 1). The first target was to vary the degree
of artificial catalytic triad functionalization on the resin support (0.6–3.0 mmol gꢀ1) in or-
der tounderstand the roleof increasing artificialactivesites for catalysis. Figure3B shows
the initial velocity (first 100 s) for the esterolysis of PNB with resins with varying levels of
incorporation of the artificial catalytic triad molecule only. Interestingly, a bell-shaped
curve was obtained with the fastest reaction rates for intermediate loading values of
1.5 mmol gꢀ1 (R3), which represents ꢁ50% functionalization of the available azides
with the triad molecule (M1). This was further supported by the finding that the material
exhibited the largest kcat value for substrate turnover (5.4 hrꢀ1) of all the triad-only resins.
Furthermore, the resin with 0.6 mmol gꢀ1 triad loading (R1) was a more effective catalyst
than the resin witha loading of 3 mmolgꢀ1 (R6) in terms ofboth turnover(kcat = 3.4 versus
0.07 hrꢀ1) and catalyst efficiency (kcat/KM 4.05 versus 0.09;valid because of the similar KM
parameters between materials).28 This suggests that simple catalyst concentration does
not dictatethe catalytic efficiency of these enzyme mimics. Ananalogy can bedrawnwith
the active site of the enzymes, where many factors, such as the hydrophobic nanoenvir-
onment around the active site, contribute to performance. In these initial examples,
increased loading of the triad unit significantly increases the hydrophilicity of the resin,
which could in turn interrupt the affinity of this material for the hydrophobic substrate
and decrease catalytic efficiency.
Chem 2, 732–745, May 11, 2017 735

Figure 3. Hydrolytically Active Synthetic Catalysts
(A) Time-course graph for the hydrolysis of nitrophenyl butyrate (100 mM concentration) by the triad
functional resin (R3) (blue) and without resin particles (gray).
(B) Initial rate (first 100 s, substrate concentration 400 mM) of esterolysis of nitrophenyl butyrate by
catalytic resin with different incorporations of artificial catalytic triad molecule (M1) only (triangles)
and both artificial catalytic triad and hexadecyne (circles). Total loading of artificial triad and
hexadecyne = 3 mmol gꢀ1
.
(C) We measured catalytic activity by monitoring the product formation over time for optimized
resins (R8 and R4) and resins functionalized with control structures (D2) at 540 mM substrate
concentration by UV-vis spectroscopy. The y axis is the product concentration divided by the
catalyst concentration (concentration of active sites), which corresponds to the number of substrate
turnovers.
736 Chem 2, 732–745, May 11, 2017

Table 1. Michaelis-Menten Rate Parameters of Resins with Different Functional Groups
Catalyst
Catalytic Triad
Hexadecyne
k
cat (hrꢀ1
)
KM (mM)
kcat/KM
Loading (mol gꢀ1
)
Loading (mol gꢀ1
)
(minꢀ1 Mꢀ1
)
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
0.6
0.9
1.5
2.1
2.7
3.0
2.4
2.1
1.5
0.9
0.6
–
3.4
14
34
45
17
15
13
40
12
30
53
10
4.05
1.23
2.00
1.86
0.11
0.09
2.41
136.1
4.65
0.16
0.67
–
2.5
–
5.4
–
1.9
–
0.10
0.07
5.8
–
0.6
0.9
1.5
2.1
2.4
98
8.37
0.52
0.40
For examination of this feature, the modularity of our synthetic approach allowed the
polarity of the supported resins to be tuned by conjugation of C16 alkyl chains
onto the resin, which effectively regulated the hydrophobicity of the catalytic local
environment. The same series of resins was used as described above, with one modi-
fication; the unreacted azide groups of the resin (R1–R6) were reacted in a second
step with hexadecyne under CuAAC conditions (R7–R11) (Figure 2C). Introducing
C16 alkyl chains into the resin particles allowed fine control of their hydrophobicity
and tuning of the catalytic environment (Figure 1). These hydrophobically modified
EIC materials were then screened for esterolytic catalysis with PNB; the velocity dur-
ing the burst phase of these hydrolysis reactions is displayed in Figure 3B. Signifi-
cantly, introduction of the unreactive C16-alkyl chain led to consistently increased
rates for the esterolysis reaction of PNB. For the functionalization levels studied
here, optimized conditions were observed for a 2.1 mmol gꢀ1 loading of M1 and
a 0.9 mmol gꢀ1 loading of hexadecyne (R8). This optimized material exhibited a cat-
alytic rate enhancement in excess of 50 times that of the same resin without hydro-
phobic tuning (R4) (kcat 98 versus 1.9 hrꢀ1). Figure 3C further highlights the impact of
hydrophobic tuning; the catalytic sites of resin R8 display enhanced substrate turn-
over as a result of the inclusion of the hexadecyl group. Once again, a bell-shaped
curve was observed for rate enhancements of the hydrophobically modified resins,
indicating that a trade-off between catalytic and hydrophobic groups is required for
optimized catalysis. These results suggest that hydrophobicity plays an important
role in the efficiency of the synthetic triad catalyst. The exact cause is under investi-
gation, but it is postulated that varying the hydrophobicity of the resin might result in
increased affinity for the hydrophobic substrate and regulate the local pKa of the
charged triad groups, leading to an increase in catalyst performance.
The importance of all three functional groups for catalytic function was further inves-
tigated synthetically. We applied chemical design to systematically remove the
three functional groups of the EIC to investigate their individual effects on catalysis
experimentally. Figure 2D shows some of the structures investigated. Significantly,
in all examples, including the fully protected (P1, which includes only N(a)H),
mono-functional (M1, which only includes the O(g) group), and di-functional (D1
and D2) resins, catalysis was essentially at background levels. Figure 3C further high-
lights the negligible esterolytic enhancement of the control structure D2, which has a
dyad-like structure including an imidazole and carboxylate but lacking the hydroxyl
Chem 2, 732–745, May 11, 2017 737

group. These control structures, paired with the catalytic enhancements of the hy-
drophobic resins, indicate that both inclusion of all three groups of the EIC and tun-
ing the local nanoenvironment play important roles in catalysis.
To further probe the functional group interactions responsible for catalysis, we were
motivated to conduct a thorough examination of the mechanism by which the resin-
supported EIC catalyzes the hydrolysis of PNB substrate through quantum me-
chanics (QM) calculations (refer to Supplemental Experimental Procedures, Tables
S1–S11, and Figures S5–S16). Because the synthetic design of the catalyst was
inspired by the native Ser-His-Asp active-site triad found in a number of enzymes,
such as the serine peptidase family,3,29,30 we initially examined a two-step mecha-
nistic pathway analogous to that exhibited by these enzymes, which involves succes-
sive acylation and deacylation steps (Figure 4, top). We discovered, however, that to
proceed via this pathway, the catalyst would have to adapt a very strained configu-
ration to facilitate the necessary hydroxyl-imidazole-carboxyl group interactions that
mimic the Ser-His-Asp interactions in enzymes. This is seen in Figure 5A, which
shows the so-called near-attack conformer (NAC), which is the conformation of the
substrate-enzyme complex that is poised for the subsequent reactions, in this
case, initial base catalysis by imidazole and nucleophilic attack by O(g) in a
concerted manner similar to that of enzymes. The unfavorability of this NAC high-
lights the important role of the enzyme in both binding the reacting groups in close
proximity and orientating them correctly for reaction; designing synthetic analogs
that can achieve this poses an additional challenge, and further work is underway
in this direction.
Nonetheless, although the NAC for the imidazole catalyzed process is very strained,
we discerned an alternative conformer that lies some 18 kcal molꢀ1 lower in energy.
In this case, the O(g)H group of the EIC forms a H bond with the secondary amine
N(a) of the catalyst rather than the N(d) of the imidazole ring (Figure 5B). The close
proximity of the O(g)H and N(a) groups on the EIC residue helps to drive the forma-
tion of a H bond between these groups over the more distant and sterically strained
imidazole nitrogen N(d). This NAC leads to an alternate reaction pathway (Figure 4,
bottom) in which the secondary amine N(a) of EIC acts as the specific base catalyst in
place of the imidazole ring, reminiscent of the Lys-Ser catalytic dyad seen in a range
of serine protease enzymes.31,32 To distinguish these two pathways, we refer to the
first as being imidazole mediated and the second as being N(a) mediated. We car-
ried out theoretical calculations on all of the model structures (Figure 4) to determine
the energy barriers along the different reaction pathways; we also studied the corre-
sponding uncatalyzed reaction under the same conditions for comparison. We per-
formed calculations in the gas phase because it is most representative of the EIC
environment, especially in the presence of hydrophobic chains. The potential energy
surfaces for each pathway are plotted in Figure 6; full details are provided in the
Supplemental Information.
From Figure 6, it is clear that the most energetically favorable process is the N(a)-me-
diated pathway, which leads to a substantial lowering of the barrier (25 kcal molꢀ1) in
comparison with the corresponding uncatalyzed process. In contrast, the imidazole
pathway is much higher in energy and only slightly more favorable (by 6.4 kcal molꢀ1
)
than the uncatalyzed process. The lack of significant catalysis by the imidazole
pathway can be explained by the absence of a H bond between the N( 3 ) atom of
the imidazole ring of EIC and its carboxylate group, which mimics the Asp-His inter-
actions that are found to play significant roles in enzyme catalysis.33–36 Moreover,
the anionic carbonyl oxygen atom (C–Oꢀ) in the intermediate structure of the EIC
738 Chem 2, 732–745, May 11, 2017

Figure 4. Proposed Catalysis Reaction Pathways
Postulated catalytic mechanisms for the enzyme-inspired catalyst (EIC) through the imidazole-mediated pathway (top) and the N(a)H-mediated
pathway (bottom). The wavy lines represent a unit of styrene, as shown in Figure 5 (refer to Tables S1–S11, Figures S5–S16, and Supplemental
Experimental Procedures).
(INT1 in Figure 4) is unstable and abstracts the proton back from the imidazole ring in
the catalyst, resulting in a neutral (i.e., C–OH) tetrahedral intermediate in the ground
state as opposed to the oxyanion intermediate found in the natural enzymatic sys-
tem. The absence of a favorable electrostatic environment, as seen in the enzymes,
can lead to instability of the oxyanion in the EIC-substrate intermediate (i.e., INT1),
which in turn is stabilized after H is extracted from the imidazole unit. As noted
Chem 2, 732–745, May 11, 2017 739

Figure 5. EIC Conformation Affording Catalysis and Hydrophobic Interactions
(A and B) Near-attack conformation of (A) the N(a) pathway and (B) the imidazole pathway. These are the conformations of the substrate-EIC complexes
that are poised for reaction. In the NAC for the imidazole pathway (relative conformer energy = 0 kcal molꢀ1), the O(g)H group of the EIC forms a strong
˚
intra-molecular hydrogen bond (H bond) with N(d) of the imidazole ring, and the carbonyl carbon (of the substrate) is only 2.69 A away from O(g), thus
making it convenient for the initial base catalysis by imidazole and nucleophilic attack by O(g). In the NAC of the N(a)-mediated scheme (relative
conformer energy = ꢀ18 kcal molꢀ1), the O(g)H group forms a H bond with N(a) of the catalyst (which then acts as the base), and the carbonyl carbon of
˚
the substrate is still conveniently located (within 3 A) in relation to O(g). These structures were identified from a full conformational search carried out for
the smaller catalyst-substrate model without polymer support (refer to Figures S5–S8). The color representations are as follows: carbon, gray; oxygen,
red; nitrogen, blue; and hydrogen, white.
(C) Representative conformations for the hydrophobically modified resins with zero, two, and four styrene units separating the EIC and alkyl chain
residues (EIC-0, EIC-2, and EIC-4, respectively). The C16 chain (green) can be seen to wrap closely around the EIC unit (violet) irrespectively of residue
spacing, highlighting the formation of a hydrophobic pocket around the EIC groups. The most populated cluster C1 for each fragment is represented
here (refer to Figures S3 and S4).
above, the steric strain incurred in accommodating the catalytic triad mimicking hy-
droxyl-imidazole-carboxyl interactions also inflates the reaction barriers by this
process.
Modeling of the hydrophobically modified resins provides support for the formation
of a hydrophobic pocket surrounding the EIC, analogous to the binding pocket of
many native proteases (Figures S3 and S4). Simulations showed that the C16 hydro-
phobic residues can form a number of different contacts with the EIC across all of the
examined fragments in which the modified residues were separated by zero to five
styrene units (Figure 5C). In general, the C16 hydrophobic chain adopted a folded
conformation, reflected in the consistent head-to-tail distance measured for this res-
idue (Figure S3). The distance between the center of mass of the EIC and C16 hydro-
˚
phobic residues on the resin support averaged approximately 10 A, and distances as
740 Chem 2, 732–745, May 11, 2017

Figure 6. Alternative Reaction Pathway Profiles
Comparison of the Gibbs free-energy profiles (298 K, kcal molꢀ1) of the entire catalytic processes
(from isolated reactants to isolated products) of the polymer-supported EIC via the N(a) pathway
and the imidazole pathway in the gas phase (see Figure 4 for mechanistic details). The uncatalyzed
pathway is also shown for purposes of comparison (refer to Tables S1–S5 and Figures S6–S16).
˚
low as 1 A were observed for some conformations where the C16 hydrophobic chain
enfolded the EIC. The strong interactions between the EIC and C16 hydrophobic
residues provide support for the formation of a local hydrophobic environment on
the resin that can both attract substrate and tune the electrostatic interactions
between EIC functional groups. These findings further support the experimental ob-
servations of balancing hydrophobic and catalytic residue inclusion on the resins for
optimal catalysis.
An interesting consequence of these studies is the indication that esterolysis by the
EIC is conducted by two structural groups, the O(g)H hydroxyl and the N(a)H sec-
ondary amine, which, as noted above, resemble the Ser-Lys catalytic dyads in
some native protease enzymes.32,33 However, the other functional groups in the
EIC, carboxylate and imidazole, are involved in a supporting role to stabilize the
transition state and intermediate structures. This result is supported experimentally
by the control structures (Figure 2D), which indicate that all three groups of the EIC
are required for catalysis. The consequence of specific functional-group placement
to best leverage the observed catalytic mechanism is of considerable interest, and
this strategy holds promise for future bio-inspired catalyst designs.
Conclusions
We have developed a simple one-step strategy for preparing multi-functional cata-
lytic building blocks with a triad of reactive groups—carboxylate, imidazole, and
hydroxyl—for CuAAC-mediated attachment to a variety of substrates. The robust
and quantitative nature of the CuAAC reaction also allows the stepwise introduction
of building blocks (C16 + triad) to be developed, leading to a series of polymer-sup-
ported EICs with finely tuned local hydrophobic environments. Significantly, the
incorporation of C16 alkyl chains results in a 50-fold faster initial rate of reaction
for the dual (C16 + triad) systems than for the parent triad functional resin. QM cal-
culations illustrate a modified mechanism in comparison with native serine prote-
ases, closely resembling the catalytic mechanism of Ser-Lys catalytic dyads. Both cal-
culations and experimental results indicate that all functional groups within the EIC
building block play an important role in the catalytic mechanism. Further work is
Chem 2, 732–745, May 11, 2017 741

underway to optimize the spatial arrangement of the functional moieties and nearby
chemical environment to better harness the observed catalytic mechanism.
EXPERIMENTAL PROCEDURES
Unless otherwise stated, all materials were purchased from Sigma-Aldrich.
H-His(1-Trt)-OtBu and H-His(1-Trt)-OH were purchased from Bachem. All materials
were used as received without further purification.
Synthesis of Protected Acetylene Functional Histidine M1
Glycidyl propargyl ether (185 mg, 1.65 mmol) and H-His(1-Trt)-OtBu (750 mg,
1.65 mmol) were dissolved in methanol (400 mL) and reacted at room temperature
for 48 hr. The product was concentrated by rotary evaporation and purified by flash
chromatography (dichloromethane [DCM]/MeOH 90:10) to yield M1 (tert-butyl
2-((2-hydroxy-3-(prop-2-yn-1-yloxy)propyl)amino)-3-(1-trityl-1H-imidazol-4-yl)propa-
noate) as a pale-yellow viscous oil (520 mg, 56%): 1H NMR (500 MHz, CDCl3): d = 7.38
(s, 1H), 7.32 (s, 9H), 7.18 (s, 6H), 6.15 (s, 1H), 4.20 (s, 2H), 3.80 (d, 1H), 3.55 (m, 3H),
2.90 (m, 2H), 2.75 (m, 2H), 2.40 (s, 1H), 1.40 (s, 9H). The product was analyzed with
electrospray ionization time-of flight (ESI-TOF) mass spectrometry: m/z: [M]+ calcd
565.29, C35H39N3O4; found 565.73.
Synthesis of Fully Protected Control Structure P1
Propargyl chloride (63.8 mL, 0.88 mmol), H-His(1-Trt)-OtBu (400 mg, 0.88 mmol), and
K2CO3 (183 mg, 1.30 mmol) were dissolved in acetonitrile (0.5 mL) and heated to
40ꢂC for 48 hr (refer to Figure 2D). The crude reaction mixture was concentrated
by rotary evaporation and purified by flash chromatography (DCM/MeOH 95:5)
to yield P1 as a pale-yellow viscous oil (35 mg, 33%): 1H NMR (500 MHz, CDCl3):
d = 7.38 (s, 1H), 7.32 (s, 9H), 7.18 (s, 6H), 6.70 (s, 1H), 3.65 (m, 1H), 3.40 (t, 2H), 3.05
(d, 1H), 2.85 (d, 1H) 2.15 (s, 1H), 1.40 (s, 9H). The product was analyzed with ESI-TOF
mass spectrometry: m/z: [M]+ calcd 491.26, C32H33N3O2; found 491.80. After conjuga-
tion to resin and deprotection with TFA, this yielded control structure D2.
Synthesis of Dyad Control Structure D1
Glycidyl propargyl ether (169 mg, 1.51 mmol) and H-His(1-Trt)-OH (600 mg,
1.51 mmol) were dissolved in methanol (400 mL) and reacted at room temperature
for 48 hr. The product was concentrated by rotary evaporation and purified by
flash chromatography (DCM/MeOH 90:10) to yield D1 as a pale-yellow viscous
oil (50 mg, 39%): 1H NMR (500 MHz, CDCl3): d = 7.38 (s, 1H), 7.32 (s, 9H),
7.18 (s, 6H), 6.15 (s, 1H), 4.20 (s, 2H), 3.80 (d, 1H), 3.55 (m, 3H), 2.90 (m, 2H),
2.75 (m, 2H), 2.40 (s, 1H). The product was analyzed with ESI-TOF mass spectro-
metry: m/z: [M]+ calcd 509.23 C31H31N3O4; found 509.61.
Synthesis of Protected Triad-Functionalized Polystyrene Resins R1–R6
Five particles were prepared with varying incorporations of M1 (R1–R6). For a typical
synthesis, azide functional polystyrene resin (50 mg, 0.15 mmol) was added to M1
(61 mg, 0.135 mmol) with 1 mL CuSO4 (5 mg/mL), 200 mL 1 M ascorbic acid, and
9.6 mL tetrahydrofuran. The solution was stirred for 2 hr, and another 200 mL 1 M as-
corbic acid was added and stirred for a further 2 hr. The particles were then vacuum
filtered and rinsed multiple times with dimethylformamide (DMF), DCM, and H2O.
The resin was then dispersed in 1 M N,N,N0,N00,N00-pentamethyldiethylenetriamine
solution in DMF and stirred overnight for the removal of any excess Cu2+. The resin
was then filtered under vacuum, rinsed sequentially with DMF, DCM, and H2O,
and dried under vacuum to yield a white solid (66 mg, 60%). The particles were
then suspended in a 1:1 mixture of DCM and TFA (1 mL) for the removal of protection
742 Chem 2, 732–745, May 11, 2017

groups. The particles were then rinsed multiple times with DCM and H2O, filtered
under vacuum, and finally dried under high vacuum.
Catalysis Experiments
The experiments were carried out on a Shimadzu UV3600 UV-VIS-NIR spectrometer.
The catalytic resins were dispersed in Tris buffer (pH 8.9). Nitrophenyl butyrate was
dissolved in DMSO at suitable concentrations and added to the dispersion immedi-
ately before analysis; the final concentration of DMSO was 3%. Each assay was per-
formed in triplicate with increasing substrate concentrations ranging from 40 to
400 mM nitrophenyl butyrate. The hydrolysis was monitored at 410 nm. Velocities
were determined as the slope of the progress curve of each reaction. The velocities
of uncatalyzed reactions were obtained directly from the control samples, and those
of catalyzed reactions were calculated by subtracting the uncatalyzed rate from the
total rate. Kinetic parameters were calculated by double reciprocal plots (Linewea-
ver-Burk method).
Computational Details
All QM calculations in this work were carried out with the hybrid density functional
M06-2X/6-31+G(d) level of theory in the Gaussian 09 computational chemistry
program.37 Initially, the molecular structures of the free species, styrene-free EIC,
styrene-attached EIC, PNB (substrate) The products (p-nitrophenol and butyrate)
were subjected to a thorough conformational search using the energy-directed
tree search (EDTS) algorithm,38 and their lowest-energy ground-state structures
were identified. Two structures of the EIC-substrate complexes (Figure 5), which
represent the NACs for the imidazole-mediated and N(a)H-mediated pathways,
were modeled on the structures identified from the QM calculations performed
for the smaller catalyst-substrate models without styrene support (refer to Supple-
mental Information). Subsequently, several intermediate structures and transition-
state-like structures along the two reaction schemes (as shown in Figure 4) were
modeled, again from the results obtained for the smaller models given in the Sup-
plemental Information. Geometry optimization and transition-state optimization,
followed by frequency calculations, were performed on all the structures for confirm-
ing the nature of the stationary points.
For the molecular dynamics simulations of the hydrophobic residues, polystyrene
fragments containing catalytic and hydrophobic residues were built and simulated
with the Amber16 package.39 Parameters for the styrene (STY), catalyst (CAT), and
hydrophobic (SPA) residues were assigned with Antechamber40 and the General
Amber Force Field (GAFF),41 whereas the charges were derived by a two-stage
RESP fitting procedure42 to the electrostatic potential calculated with Gaussian09
software37 at the HF/6-31G(d) level of theory. Full details are in the Supplemental
Information. Six fragments containing the catalyst and hydrophobic residue sepa-
rated by zero to five styrene units (EIC-0 to EIC-5) were built, and three additional
styrene units were added on each side of the catalyst and hydrophobic residue.
These fragments were then solvated with TIP3P waters in a truncated octahedron
˚
box with a 10 A buffer in each direction from the solute. The solvated fragments
were relaxed in four steps: (1) steepest descent minimization with a maximum of
10,000 cycles, (2) slow heating of the system from 0 to 300 K in 100 ps with harmonic
2
position restraints on the solute (2.5 kcal molꢀ1 Aꢀ ) under constant volume (NVT),
˚
(3) constant pressure (NPT) dynamics at 300 K for 1 ns with harmonic position re-
2
straints placed on the solute (2.5 kcal molꢀ1 Aꢀ ), and (4) constant pressure (NPT)
˚
dynamics at 300 K without restraints for another 1 ns. After relaxation (total of
2.1 ns), each system was simulated for 100 ns at 300 K under NVT. Periodic boundary
Chem 2, 732–745, May 11, 2017 743

conditions were used, and long-range electrostatic interactions were calculated with
˚
the Particle-Mesh Ewald technique with a cutoff of 8.0 A. The temperature in all sim-
ulations was set to 300 K and controlled by the Langevin thermostat combined with a
random seed generator.43 The SHAKE algorithm44 was used to constrain bonds
involving hydrogen atoms during dynamics, and an integration time step of 2 fs
was used. The simulations were carried out with a GPU version of the PMEMD pro-
gram implemented in the Amber16 package, during which snapshots were saved
every 10 ps. Analysis was performed with the cpptraj module available in
AmberTools. For clustering, all-atom root-mean-square deviations of catalyst and
hydrophobic residues (styrene residues were not included) were used in conjunction
with the k-means algorithm with the maximum number of clusters set to 20. All
images were produced in VMD and xmgrace. The initial files for each simulation sys-
tem and the resulting simulation trajectories have been uploaded to Figshare
(10.6084/m9.figshare.4757161).
SUPPLEMENTAL INFORMATION
Supplemental Information includes Supplemental Experimental Procedures, 16 fig-
ures, and 11 tables can be found with this article online at http://dx.doi.org/10.1016/
j.chempr.2017.04.004.
AUTHOR CONTRIBUTIONS
E.K., E.P., C.J.H., M.L.O., M.L.C., and L.A.C. conceived the work and designed the
experiments. L.A.C., E.P., Z.X., A.D, M.R.G., and M.D.N. synthesized the materials
and performed the kinetic analysis. A.G., K.C.-J., M.L.O., and M.L.C. designed, per-
formed, and interpreted the computational work. All authors contributed to writing
the paper.
ACKNOWLEDGMENTS
Computational assistance from Benjamin Noble is acknowledged and appreciated.
Funding from the US Army International Technology Center Pacific (ITC-PAC
FA5209-14-C-0017 to L.A.C.), the Defense Science Institute (L.A.C.), a Veski Innova-
tion Fellowship (L.A.C.), and the MRSEC Program of the National Science
Foundation (award DMR-1121053 to C.J.H.) is gratefully acknowledged. M.D.N. ac-
knowledges a John Stocker Postgraduate Fellowship, an Australian Nanotechnology
Network Travel Fellowship, and an Australian Postgraduate Award. M.L.C. and
M.L.O. acknowledge generous allocations of supercomputing time on the
National Facility of the Australian National Computational Infrastructure and an
Australian Research Council Future Fellowship. A.G. acknowledges an Endeavour
Research Fellowship. Unilever and UCSB have patents covering the catalysts de-
signed herein.
Received: November 9, 2016
Revised: March 21, 2017
Accepted: April 11, 2017
Published: May 11, 2017
REFERENCES AND NOTES
1. Hasan, F., Shah, A.A., and Hameed, A. (2006).
Industrial applications of microbial lipases.
Enzyme Microb. Technol. 39, 235–251.
3. Polga´r, L. (2005). The catalytic triad of serine
peptidases. Cell. Mol. Life Sci. 62, 2162–2172.
5. Breslow, R. (1995). Biomimetic chemistry and
artificial enzymes - catalysis by design. Acc.
Chem. Res. 28, 146–153.
4. Breslow, R., and Overman, L.E. (1970). ‘Artificial
enzyme’ combining a metal catalytic group
and a hydrophobic binding cavity. J. Am.
Chem. Soc. 92, 1075–1077.
2. Hedstrom, L. (2002). Serine protease
mechanism and specificity. Chem. Rev. 102,
4501–4524.
6. Overberger, C.G., Salamione, J.C., and
Yaroslavsky, S. (1967). Cooperative effects in
the esterolytic action of synthetic
744 Chem 2, 732–745, May 11, 2017

macromolecules containing imidazole and
hydroxyl functions. J. Am. Chem. Soc. 89,
6231–6236.
19. Rodionov, V., Gao, H., Scroggins, S., Unruh,
D.A., Avestro, A.-J., and Frechet, J.M.J. (2010).
Easy access to a family of polymer catalysts
from modular star polymers. J. Am. Chem. Soc.
132, 2570–2572.
33. Fersht, A.R., and Sperling, J. (1973). The
charge relay system in chymotrypsin and
chymotrypsinogen. J. Mol. Biol. 74,
137–149.
7. Bjerre, J., Rousseau, C., and Marinescu, L.
(2008). Artificial enzymes, ‘Chemzymes’:
current state and perspectives. Appl.
Microbiol. Biotechnol. 81, 1–11.
34. Warshel, A. (1978). Energetics of enzyme
catalysis. Proc. Natl. Acad. Sci. USA 75, 5250–
5254.
20. Burton, A.J., Thomson, A.R., Dawson, W.M.,
Brady, R.L., and Woolfson, D.N. (2016).
Installing hydrolytic activity into a completely
de novo protein framework. Nat. Chem. 8,
837–844.
8. Agnus, Y., Louis, R., and Weiss, R. (1979).
Bimetallic copper(I) and copper(II) macrocyclic
complexes as mimics for type-3 copper pairs in
copper enzymes. J. Am. Chem. Soc. 101, 3381–
3384.
35. Craik, C.S., Roczniak, S., Largman, C., and
Rutter, W.J. (1987). The catalytic role of the
active site aspartic acid in serine proteases.
Science 237, 909–913.
21. Fracaroli, A.M., Siman, P., Nagib, D.A., Suzuki,
M., Furukawa, H., Toste, F.D., and Yaghi, O.M.
(2016). Seven post-synthetic covalent reactions
in tandem leading to enzyme-like complexity
within metal–organic framework crystals.
J. Am. Chem. Soc. 138, 8352–8355.
9. Jang, B.-B., Lee, K.-P., Min, D.-H., and Suh, J.
(1998). Immobile artificial metalloproteinase
containing both catalytic and binding groups.
J. Am. Chem. Soc. 120, 12008–12016.
36. Szeltner, Z., Rea, D., Juha´sz, T., Renner, V.,
Mucsi, Z., Orosz, G., Fulop, V., and Polga´r, L.
¨ ¨
(2002). Substrate-dependent competency of
the catalytic triad of prolyl oligopeptidase.
J. Biol. Chem. 277, 44597–44605.
22. Kofoed, J., and Reymond, J.L. (2005).
Dendrimers as artificial enzymes. Curr. Opin.
Chem. Biol. 9, 656–664.
10. Xue, G., De Hont, R., Munck, E., and Que, L.
¨
(2010). Million-fold activation of the [Fe2(m-O)2]
diamond core for C–H bond cleavage. Nat.
Chem. 2, 400–405.
37. Frisch, M.J., Trucks, G.W., Schlegel, H.B.,
Scuseria, G.E., Robb, M.A., Cheeseman, J.R.,
Scalmani, G., Barone, V., Petersson, G.A.,
Nakatsuji, H., et al. (2010). Gaussian 09,
Revision B.01 (Gaussian Inc.).
23. Caminade, A.M., Ouali, A., Keller, M., and
Majoral, J.-P. (2012). Organocatalysis with
dendrimers. Chem. Soc. Rev. 41, 4113–4125.
11. Li, Y., Li, Y., Wang, B., Luo, Y., Yang, D., Tong,
P., Zhao, J., Luo, L., Zhou, Y., Chen, S., et al.
(2013). Ammonia formation by a thiolate-
bridged diiron amide complex as a
24. Ema, T., Tanida, D., Matsukawa, T., and Sakai,
T. (2008). Biomimetic trifunctional
38. Izgorodina, E.I., Lin, C.Y., and Coote, M.L.
(2007). Energy-directed tree search: an
efficient systematic algorithm for finding
the lowest energy conformation of
molecules. Phys. Chem. Chem. Phys. 9, 2507–
2516.
nitrogenase mimic. Nat. Chem. 5, 320–326.
organocatalyst showing a great acceleration
for the transesterification between vinyl ester
and alcohol. Chem. Commun. 8, 957–959.
12. Ke, Z., Abe, S., Ueno, T., and Morokuma, K.
(2012). Catalytic mechanism in artificial
metalloenzyme: QM/MM study of
25. Coulembier, O., Josse, T., Guillerm, B.,
Gerbaux, P., and Dubois, P. (2012). An
imidazole-based organocatalyst designed for
bulk polymerization of lactide isomers:
inspiration from nature. Chem. Commun.
(Camb.) 48, 11695–11697.
phenylacetylene polymerization by rhodium
complex encapsulated in apo-ferritin. J. Am.
Chem. Soc. 134, 15418–15429.
39. Case, D.A., Cerutti, D.S., Cheateham, T.E., III,
Darden, T.A., Duke, R.E., Giese, T.J., Gohlke,
H., Goetz, A.W., Greene, D., Homeyer, N., et al.
(2016). AMBER16 Package (University of
California, San Francisco).
13. Kitano, H., Sun, Z.-H., and Ise, N. (1983).
Functionalized polymer lattices. 2. Catalytic
effects of imidazole-containing lattices on
hydrolyses of phenyl esters. Macromolecules
16, 1306–1310.
26. Merrifield, R.B. (1963). Solid phase peptide
synthesis, I. The synthesis of a tetrapeptide.
J. Am. Chem. Soc. 85, 2149–2154.
40. Wang, J., Wang, W., Kollman, P.A., and Case,
D.A. (2006). Automatic atom type and bond
type perception in molecular mechanical
calculations. J. Mol. Graph. Model. 25,
247–260.
14. Wulff, G., Gross, T., and Schonfeld, R. (1997).
¨
27. Merrifield, R.B. (1969). Solid-phase peptide
synthesis. Adv. Enzymol. Relat. Areas Mol. Biol.
32, 221–296.
Enzyme models based on molecularly
imprinted polymers with strong esterase
activity. Angew. Chem. Int. Ed. 36, 1962–1964.
41. Wang, J., Wolf, R.M., Caldwell, J.W., Kollman,
P.A., and Case, D.A. (2004). Development and
testing of a general amber force field.
28. Eisenthal, R., Danson, M.J., and Hough, D.W.
(2007). Catalytic efficiency and kcat/KM: a useful
comparator? Trends Biotechnol. 25, 247–249.
15. Huerta, E., Stals, P.J.M., Meijer, E.W., and
Palmans, A.R.A. (2013). Consequences of
folding a water-soluble polymer around an
organocatalyst. Angew. Chem. Int. Ed. 52,
2906–2910.
J. Comput. Chem. 25, 1157–1174.
29. Paetzel, M., and Dalbey, R.E. (1997). Catalytic
hydroxyl/amine dyads within serine proteases.
Trends Biochem. Sci. 22, 28–31.
42. Cieplak, P., Cornell, W.D., Bayly, C., and
Kollman, P.A. (1995). Application of the
multimolecule and multiconformational RESP
methodology to biopolymers: charge
derivation for DNA, RNA, and proteins.
J. Comput. Chem. 16, 1357–1377.
16. Wulff, G., and Liu, J. (2012). Design of
biomimetic catalysts by molecular imprinting in
synthetic polymers: the role of transition state
stabilization. Acc. Chem. Res. 45, 239–247.
30. Topf, M., Va´rnai, P., and Richards, W.G. (2002).
Ab initio QM/MM dynamics simulation of the
tetrahedral intermediate of serine proteases:
insights into the active site hydrogen-bonding
network. J. Am. Chem. Soc. 124, 14780–14788.
17. Terashima, T., Mes, T., De Greef, T.F.A.,
Gillissen, M.A.J., Besenius, P., Palmans, A.R.A.,
and Meijer, E.W. (2011). Single-chain folding of
polymers for catalytic systems in water. J. Am.
Chem. Soc. 133, 4742–4745.
43. Loncharich, R.J., Brooks, B.R., and Pastor, R.W.
(1992). Langevin dynamics of peptides: the
frictional dependence of isomerization rates of
N-acetylalanyl-N0-methylamide. Biopolymers
32, 523–535.
31. Feldman, A.R., Lee, J., Delmas, B., and Paetzel,
M. (2006). Crystal structure of a novel viral
protease with a serine/lysine catalytic dyad
mechanism. J. Mol. Biol. 358, 1378–1389.
18. Chi, Y., Scroggins, S.T., and Frechet, J.M.J.
(2008). One-pot multi-component asymmetric
cascade reactions catalyzed by soluble star
polymers with highly branched non-
44. Ryckaert, J.-P., Ciccotti, G., and Berendsen,
H.J.C. (1977). Numerical integration of the
cartesian equations of motion of a system with
constraints: molecular dynamics of n-alkanes.
J. Comput. Phys. 23, 327–341.
32. Ekici, O.D., Paetzel, M., and Dalbey, R.E. (2008).
Unconventional serine/threonine proteases:
variations on the catalytic Ser/His/Asp
interpenetrating catalytic cores. J. Am. Chem.
Soc. 130, 6322–6323.
triad configuration. Protein Sci. 17, 2023–2037.
Chem 2, 732–745, May 11, 2017 745