Spiroligozymes for transesterifications: Design and relationship of structure to activity

Article abstract of DOI:10.1021/ja3069648

Transesterification catalysts based on stereochemically defined, modular, functionalized ladder-molecules (named spiroligozymes) were designed, using the inside-out design strategy, and mutated synthetically to improve catalysis. A series of stereochemically and regiochemically diverse bifunctional spiroligozymes were first synthesized to identify the best arrangement of a pyridine as a general base catalyst and an alcohol nucleophile to accelerate attack on vinyl trifluoroacetate as an electrophile. The best bifunctional spiroligozyme reacted with vinyl trifluoroacetate to form an acyl-spiroligozyme conjugate 2.7 × 10³-fold faster than the background reaction with a benzyl alcohol. Two trifunctional spiroligozymes were then synthesized that combined a urea with the pyridine and alcohol to act as an oxyanion hole and activate the bound acyl-spiroligozyme intermediate to enable acyl-transfer to methanol. The best trifunctional spiroligozyme carries out multiple turnovers and acts as a transesterification catalyst with k₁/k_uncat of 2.2 × 10³ and k₂/k_uncat of 1.3 × 10². Quantum mechanical calculations identified the four transition states of the catalytic cycle and provided a detailed view of every stage of the transesterification reaction.

Full text of DOI:10.1021/ja3069648

Article
pubs.acs.org/JACS
Spiroligozymes for Transesterifications: Design and Relationship of
Structure to Activity
Mahboubeh Kheirabadi,† Nihan C
̧
elebi-Olcu
̧
m,‡ Matthew F. L. Parker,† Qingquan Zhao,† Gert Kiss,‡
̈
̈
K. N. Houk,*,‡ and Christian E. Schafmeister*,†
†Department of Chemistry, Temple University, 1901 North 13th Street, Philadelphia, Pennsylvania 19122, United States
‡Department of Chemistry and Biochemistry, University of California at Los Angeles, Los Angeles, California 90095-1569, United
States
S
* Supporting Information
ABSTRACT: Transesterification catalysts based on stereo-
chemically defined, modular, functionalized ladder-molecules
(named spiroligozymes) were designed, using the “inside-out”
design strategy, and mutated synthetically to improve catalysis.
A series of stereochemically and regiochemically diverse
bifunctional spiroligozymes were first synthesized to identify
the best arrangement of a pyridine as a general base catalyst
and an alcohol nucleophile to accelerate attack on vinyl
trifluoroacetate as an electrophile. The best bifunctional
spiroligozyme reacted with vinyl trifluoroacetate to form an
acyl-spiroligozyme conjugate 2.7 × 103-fold faster than the background reaction with a benzyl alcohol. Two trifunctional
spiroligozymes were then synthesized that combined a urea with the pyridine and alcohol to act as an oxyanion hole and activate
the bound acyl-spiroligozyme intermediate to enable acyl-transfer to methanol. The best trifunctional spiroligozyme carries out
multiple turnovers and acts as a transesterification catalyst with k1/kuncat of 2.2 × 103 and k2/kuncat of 1.3 × 102. Quantum
mechanical calculations identified the four transition states of the catalytic cycle and provided a detailed view of every stage of the
transesterification reaction.
to synthesize multifunctional spiroligomers that could carry out
a transesterification reaction. We call this class of spiroligomer
catalysts the “spiroligozymes”, for short.
INTRODUCTION
■
Natural enzymes are proficient catalysts.1,2 In the past few
years, chemists have tried to mimic enzymes and to synthesize
specially designed organocatalysts to optimize efficiency of
many different transformations.3−10 Despite many attempts to
make such catalysts, the rational design of enzyme mimics is
largely an unrealized dream.
Esterases and proteases are enzymes that catalyze the
hydrolysis of carboxyl esters and amides. Their high proficiency
in hydrolysis is attributable to the high-level structural and
electrostatic preorganization in their active sites.17−19 They
efficiently catalyze multiple chemical steps with minimal
structural reorganization of the catalytic functionalities while
maintaining optimal transition-state stabilization for each step
during catalysis.17,19 Recently, Houk and co-workers have used
explicit solvent, periodic boundary molecular dynamics
simulations to show that a naturally evolved cysteine protease
cathepsin K has a well-defined active site with tightly
maintained hydrogen bonds for the catalytic Cys-His-Asn
triad in the dynamic environment.20 The catalytic triad of Ser-
His-Asp, originally described in chymotrypsin, is a classic
structural motif in enzymology.21 It was originally postulated
that the catalytic triad enhances nucleophilicity of the hydroxyl
group in the seryl residue to attack the acylating agents.
Enhanced electron density on the serine hydroxyl necessitates
deprotonation of the hydroxyl by His as a general base and
neutralization of the generated charge being formed on the
Recently, we have developed the synthesis of highly
functionalized and shape programmable macromolecules
formed by coupling cyclic, stereochemically pure building
blocks through pairs of amide bonds.11,12 These molecules form
spiro-fused oligomers (named spiroligomers or bis-peptides)
upon which functional groups can be displayed in controlled
three-dimensional arrangements (Figure 1). We recently
demonstrated a simple example of this strategy where the
spiro linkage was used to place a phenyl group to stabilize the
transition state of an aldol reaction by the hydrophobic effect.13
Inspired by the success of the “inside-out” approach for the
design of artificial enzymes14−16 we have sought to apply this
approach to the construction of catalytic spiroligomers. This
involves designing spiroligomer scaffolds with fused-ring
structures and stereochemical configurations that are capable
of presenting catalytic functional groups in three-dimensional
(3D) constellations suggested by theozyme calculations (Figure
1). To test our ability to design and synthesize more complex
catalytic spiroligomers using the “inside-out” design, we set out
Received: July 19, 2012
Published: September 19, 2012
© 2012 American Chemical Society
18345
dx.doi.org/10.1021/ja3069648 | J. Am. Chem. Soc. 2012, 134, 18345−18353

Journal of the American Chemical Society
Article
identified peptides that carry out enantioselective acyl transfer
reactions.9,25 Inspired by these results we sought to design
transesterification catalysts by systematically incorporating and
organizing increasing numbers of functional groups arranged on
a modular spiroligomer backbone in a way that would allow us
to perturb the structures of the catalysts and observe the
relationship between their structures and their catalytic rates.
EXPERIMENTAL SECTION
■
Theozyme Calculations. Calculations were performed with
Gaussian 09.26 Transition states were located for the nucleophilic
activation of vinyl trifluoroacetate in the presence of catalytic side
chain mimics of enzymes and spiroligozymes. Methanol, 4-
methylimidazole, and acetate were used to mimic the nucleophilic
triad of serine hydrolases and proteases, serine, histidine, and
aspartate/glumate, respectively. The nucleophilic dyad of sprilogo-
zymes was represented by methanol and 3- or 4-methylpyridine,
depending on the attachment position of the pyridine to the
spiroligozyme backbone. Geometry optimizations were carried out at
the B3LYP/6-31G(d) level of theory.27−30 Solvent and medium effects
were taken into account in the optimizations using the integral-
equation-formalism polarizable continuum model (IEF-PCM).31−33
A
dielectric constant of 4 was used to represent the electrostatic
environment of an enzyme active site for TS1-triad, and a dielectric
continuum representing chloroform (ε = 4.7) was employed to solvate
the spiroligozyme model TS1-dyad. Energies were refined by single-
point calculations with the M06-2X/6-31++G(2d,p)34 using IEF-PCM
with radii and nonelectrostatic terms for the SMD solvation model35 as
implemented in Gaussian 09.26 Single-point energies in solution
including all computed corrections were augmented with the
correction to Gibbs free energy from unscaled B3LYP frequencies
computed for 1 mol/L solute at 298 K. The free energies are reported
in the text with respect to the separated substrate and preorganized
catalyst complex.
Figure 1. “Inside-out” approach to catalytic spiroligomers. The
program CANDO23 is used to predict and design a spiroligomer
formed from the four building blocks to complement the nucleophilic
transition state “theozyme” shown on the right.
Molecular Dynamics Simulations. Molecular dynamics (MD)
simulations were performed with AMBER 1136 to evaluate whether the
designed catalytic contacts were maintained in a dynamic environment
in the presence of explicit solvent molecules. A recently described MD
protocol for the evaluation of designed enzymes was employed.20 This
protocol predicted successfully the active and inactive designs for a set
of Kemp elimination designs on the basis of the stability of the
designed active sites. The spiroligozyme parameters were generated
with the antechamber module of AMBER 1136 using the general
AMBER force field (GAFF),37 with partial charges set to fit the
electrostatic potential generated at HF/6-31G(d) by RESP.38,39 The
charges were calculated according to the Merz−Singh−Kollman
scheme40,41 using Gaussian 03.26 The spiroligozymes were immersed
in a chloroform box42 ensuring a solvent layer of at least 10 Å around
the spiroligozymes. This resulted in the addition of up to 244 solvent
molecules. A two-stage minimization approach was used. Initially, the
positions of solvent molecules were minimized followed by an
unrestrained minimization of all atoms. The systems were then heated
in six 50 K, 50 ps steps from 0 to 300 K at constant volume periodic
boundary conditions. Harmonic restraints of 10 kcal/mol were applied
to the spiroligozyme, and the Langevin equilibration scheme was used
to control and equalize the temperature. A 1 fs time step was used
during the heating stages to allow potential inhomogeneities to self-
adjust. Each system was then equilibrated initially for 2 ns with a 2 fs
time step at constant volume, followed by a 2 ns equilibration with a 2
fs time step at a constant pressure of 1 atm. A 20 ns production
molecular dynamics simulation was carried out for each system using
PMEMD43 using the isothermal−isobaric ensemble (NPT). Long-
range electrostatic effects were modeled using the particle-mesh-Ewald
method.44 Trajectories were saved every 100 steps (0.2 ps). This
resulted in a total of 100,000 frames from the production MD run.
Post-MD data analysis was performed using the ptraj module of
AMBER 11.36 In order to investigate the effect solvent on the
preorganization of the catalytic functional groups, simulations were
also carried out in a box of dichloromethane molecules.45
imidazole moiety by another acid catalyzed interaction of His-
Asp or His-Glu in the so-called “charge relay system”.21
Subsequent discovery of similar triads in the subtilisin protease
revealed that a specific oxyanion hole in the active site stabilizes
the negative charge being formed on the substrate ester
carbonyl oxygen in the tetrahedral intermediate.22
Recently, Baker and Houk have generated four artificial
enzymes for ester hydrolysis by building nucleophilic dyads and
oxyanion holes into protein scaffolds using the “inside-out”
enzyme design strategy.24 The most active of the variants
catalyzed the cleavage of a p-nitrophenyl ester with a catalytic
efficiency (kcat/Km) of 400 M−1 s−1 and a rate acceleration of
approximately 3.7 × 103 −fold over the background. Kinetic
experiments showed that the protein is rapidly acylated as
programmed by the design, but the subsequent deacylation step
is slow and limits the overall catalytic efficiency.
A great deal of work has been carried out by chemists that
demonstrates that converting intermolecular general acid/base
reactions to intramolecular reactions on small organic scaffolds
can enhance reaction rates by several orders of magnitude.17
These model systems convert intermolecular general acid-/
base-catalyzed reactions to small (5−8 membered) ring
hydrogen-bonded transition states that greatly enhance reaction
rates by creating enormous effective molarities of the general
acid/base in the reaction. Sakai and co-workers demonstrated a
trifunctional small-molecule catalyst displaying a pyridine, an
alcohol, and a urea that accelerates a transesterification reaction
between vinyl trifluoroacetate and methanol by an impressive
3.2 × 104-fold for k1/kuncat and 3.8 × 103-fold for k2/kuncat,.10
Miller and co-workers screened small peptide libraries and
18346
dx.doi.org/10.1021/ja3069648 | J. Am. Chem. Soc. 2012, 134, 18345−18353

Journal of the American Chemical Society
Article
Conformational Analysis. An extensive conformational search
was carried out for each designed spiroligozyme using Macromodel46 to
explore their structural preferences. The conformational search was
carried out with the OPLS-2005 force field47 using a mixed torsional
and low mode sampling procedure as implemented in Macromodel.
Polak−Ribiere Conjugate Gradient minimization method48 was used
with convergence criterion for energy set to <0.001 kJ/mol. The
solvent effects were taken into account using the GB/SA solution
model,49 which treats the solvents as a fully equilibrated analytical
continuum starting near the van der Waals surface of the solute, with
parameters for chloroform. The conformational search produced 1557
conformers in the range of 5 kcal/mol. The resulting structures were
clustered, and a total of 25 lowest-energy conformers were selected for
further optimization using density functional theory. In order to
provide an extended sampling and a more complete representation of
the conformational space, a set of 20 representative snapshots was also
selected from the 20 ns MD production run for additional analysis
with density functional theory. OPLS-2005 and GAFF are specifically
parametrized for the accurate treatment of small organic mole-
cules.37,47 The selected 45 structures were subjected to full geometry
optimization with M06-2X/6-31G(d)34 in a dielectric continuum
treated with IEF-PCM solvation model31−33 representing chloroform
as the solvent. Single-point energies including radii and non-
electrostatic terms for the SMD solvation model35 were calculated at
the same level of theory. All 45 conformers for each spiroligozyme
were sorted with respect to their energies in solution including all
computed corrections. All density functional theory calculations were
performed using Gaussian 09.26
Scheme 1. Synthesis of Bifunctional Spiroligomers Together
with Their Relative Rate of Acylation by Vinyl
Trifluoroacetate
Calculation of the Energy Profile. All stationary points for the
spiroligozyme-catalyzed acyl transfer reaction from vinyl trifluoroace-
tate to methanol were located using Gaussian 09.26 ONIOM MO:MO
type calculations50−52 were used for the geometry optimizations, as
high-accuracy model chemistries scale unfavorably with the size of the
molecule. Two layers were defined within the structure, and they were
treated at different levels of theory. The high level was composed of 4-
methylpyridine, o-tolylmethanol, 1-(3,5-bis(trifluoromethyl)phenyl)-3-
phenylurea groups of the spiroligozyme and the substrates (vinyl
trifluoroacetate and methanol) and was treated with density functional
B3LYP/6-31G(d).27−30 The low level included the atoms of the
spiroligozyme backbone and was treated at the semiempirical PM6
level of theory.53 The transition states were characterized using the
harmonic vibrational frequencies. Due to a convergence failure in the
ONIOM optimization of TS3, this transition state was located using a
relaxed potential energy surface scan for the formation of the C−O
bond with 0.01 Å increments between d(C−O) = 1.93 Å and d(C−O)
=
a
BAA-2, BAA-4, and BAA-7 refer to each of the bis-amino acid
monomers 2a−b, 4a−e and 7a−c that were used to construct each
compound 1a−g.
1.71 Å (see the SI). The energies of all located stationary points were
refined with single-point calculations on the complete system at the
M06-2X/6-31G(d) level of theory34 in a dielectric continuum
representing chloroform using the IEF-PCM31−33 with radii and
nonelectrostatic terms for SMD solvation model.35 Single-point
energies in solution including all computed corrections are reported
in the text.
tert-butyl groups were removed using 95% trifluoroacetic acid (TFA)
and 5% tri-isopropylsilane (TIS) as scavengers. Alloc protection of the
resulting amino acids using allyl chloroformate and triethylamine
produced compounds 7a−c with an overall yield of 65%. The
bifunctional spiroligomers 1a−g were synthesized by converting the
alloc protected amino acids 7a−c to the 1-hydroxy-7-azabenzotriazole
(HOAt) esters and combining them with the amino acids 4a−e
together with diisopropylethylamine to form an amide bond.
Simultaneous deprotection of the alloc protecting groups and
formation of the diketopiperazine by addition of catalytic amount of
tetrakis(triphenylphosphine)-palladium(0) with an excess of borane
dimethylamine complex as scavenger resulted in compounds 1a−g.
Synthesis of 15a and 15b. Synthesis of 15a and 15b was carried
out starting from bis-amino acid 5a (Scheme 2). Simultaneous alloc
protection of the primary amine and carboxylic acid of 5a was
accomplished using allyl chloroformate and DIPEA and DMAP and
resulted in the bis-N-alloc protected bis-amino acid. Removal of the
Boc and tert-butyl ester groups was achieved using 95% TFA and 5%
TIS. N-alloc protection of the resulting amino acid using allyl
chloroformate produced compound 9 with an overall yield 95%.
Spiroligomer 10 was synthesized by converting alloc protected amino
Synthesis of 1a−1g. The spiroligomers 1a−g, each displaying a
benzyl alcohol and a pyridyl or benzyl group, were synthesized as
shown in Scheme 1. The synthesis of the diastereomeric bis-amino
acids 2a and 2b have been previously described.11 Each primary amine
of 2a and 2b was reductively alkylated with the corresponding benzyl
and pyridyl aldehydes 3a−c (Scheme 1 table) with sodium
cyanoborohydride at room temperature. The resulting functionalized
bis-amino acids were combined with phenyl isocyanate and triethyl-
amine to form phenyl hydantoins. These hydantoins were treated with
33% HBr in acetic acid to remove carboxybenzyl and tert-butyl
protecting groups to produce amino acids 4a−e after crystallization
from methanol. Allyl-carbamate (alloc) protected amino acids 7a−c
were synthesized starting from N-Boc, tert-butyl protected bis-amino
acids 5a−c. Each free amine of 5a−c was reductively alkylated with 2-
formylbenzyl-allyl-carbonate 6 and sodium cyanoborohydride at room
temperature. The resulting functionalized bis-amino acids were
exposed to 1-naphthyl isocyanate and triethylamine at room
temperature to afford 1-naphthyl hydantoins. Subsequently, Boc and
18347
dx.doi.org/10.1021/ja3069648 | J. Am. Chem. Soc. 2012, 134, 18345−18353

Journal of the American Chemical Society
Article
Scheme 2. Synthesis of trifunctional spiroligomers
acid 9 to the HOAt ester and combining it with the amino acid 4a
together with DIPEA. Quantitative formation of the amide bond was
observed over 12 h. Subsequent addition of tetrakis-
(triphenylphosphine)palladium(0) and phenylsilane resulted in
simultaneous alloc deprotection and diketopiperazine formation within
12 h. Acidic workup with 2 M HCl in ethylacetate gave product 10
with an overall yield of 77% starting from intermediate 9. Reductive
alkylation of amino acid 10 with aldehyde 6 afforded bifunctional
amino acid 11. Trimer 13 was synthesized by converting amino acid
11 to the HOAt ester and combing it with functionalized amino acid
12 together with DIPEA using the acyl transfer reaction that we
recently described.11,54 To promote diketopiperazine formation, an
excess of HOAt/DIC was added to the reaction mixture. The crude
product was subjected to 95% TFA and 5% TIS to remove the Boc
and tert-butyl groups. Purification of the mixture produced amino acid
13 in 29% yield from compound 11. Coupling of the urea benzoic
acids bearing bis-amino acids 14a and 14b was achieved by
preactivation of each of the amino acids with HATU and DIPEA for
10 min and then combining them with the amino acid 11. The
reaction was carried out for 4 h, after which a solution of Pd(PPh3)4
and BH3.NMe2 in dichloromethane was added to the reaction mixture
to remove the alloc protecting group and produce trifunctional
spiroligomers 15a and 15b with a 25% yield from 13.
RESULTS AND DISCUSSION
■
We first began by exploring computationally the impact of
replacing the carboxylate and imidazole pair of a serine esterase
catalytic triad with a simpler pyridine group to investigate how
pyridine compares to naturally occurring His-Glu dyads in
activating an alcohol as a nucleophile. We generated model
transition states (theozymes) to determine the optimal
arrangement of the catalytic functional groups for effective
catalysis. Two model transition states (Figure 2, TS1-triad and
TS1-dyad) were optimized for the nucleophilic addition of an
activated alcohol to vinyl trifluoroacetate using density
functional theory.
Figure 2. Activation of an alcohol as a nucleophile. Transition states
[M06-2X/6-31++G(2d,p)(chloroform)//B3LYP/6-31G(d) (chloro-
form), activation free energies (ΔG⧧) in kcal/mol. Spr: spiroligozyme]
for the nucleophilic addition of model Ser-His-Asp/Glu triad and
pyridine−alcohol dyad to vinyl trifluoroacetate.
TS1-triad mimics the nucleophilic Ser-His-Glu triad of a
typical serine hydrolase active site. The calculations predict a
concerted activation/addition mechanism, in which the partially
activated oxygen (dO−H = 1.21 Å) attacks the ester (dC−O = 2.05
Å) with an activation free energy of 8.6 kcal/mol with respect
to the preorganized catalyst.
kcal/mol, ∼6 kcal/mol higher than the barrier for TS1-triad.
The mechanism is still concerted, but with a late transition
state: the forming C−O bond is significantly shorter (dC−O
=
1.78 Å), and the proton is closer to pyridine (dN−H = 1.20 Å).
In order to investigate the electronic effects of the attachment
position of the pyridine to the spiroligozyme backbone, we
located TS1-dyad-3pyridyl, which differs from TS1-dyad by
the location of the methyl substituent. TS1-dyad-3pyridyl and
TS1-dyad features a pyridine-activated alcohol. The
nucleophilic addition occurs with a free energy barrier of 14.7
18348
dx.doi.org/10.1021/ja3069648 | J. Am. Chem. Soc. 2012, 134, 18345−18353

Journal of the American Chemical Society
Article
TS1-dyad are very similar in geometry and in energy (see the
SI).
the base and the benzyl alcohol on opposite faces of the 5−6−
5-fused ring system, and the base cannot form an internal
hydrogen bond with the benzyl alcohol. Spiroligomer 1b
becomes acylated slightly faster than 1a, but its time-dependent
acylation data fit better to a kinetic model that was second
order in spiroligomer, suggesting that the acylation results from
intermolecular activation of the benzyl alcohol by the pyridyl
group of a second spiroligomer. Spiroligomer 1g bears the same
stereochemistry as that of 1c but has a 4-pyridyl group rather
than the 3-pyridyl group of 1c. Compound 1g is acylated more
than 10 times faster than 1c and exhibits a krel value of 2700
over the background reaction. Spiroligomer 1f was designed to
increase the separation between the 4-pyridyl base and the
alcohol as in 1e and was 5-fold slower compared to 1g.
The relative rate constants of 1a−g suggest the importance
of the correct alignment of the catalytic dyad for effective
catalysis. We used molecular dynamics (MD) simulations to
study the preorganization of the pyridine and alcohol
functionalities in a dynamic environment solvated with
chloroform. The results show that the molecules are highly
flexible and span a number of alternative conformations (see
the SI). The H-bonding interaction of the dyad is disrupted
mainly by the favorable interactions of benzyl alcohol with the
backbone amide oxygens of the spiroligozyme. MD simulations
are indicative of other stable spiroligozyme conformations with
substantial deviations from the catalytic arrangement of the
dyad. In order to shed light onto the effect of structural
preferences on reactivity patterns, we performed an extensive
conformational analysis of the designed spiroligozymes. We
focused our attention on spiroligozymes 1c and 1g; their
structures differ only by the attachment position of pyridine,
but 1g is acylated ∼10-fold faster than 1c. A high-level quantum
mechanical analysis of the conformational space of 1c and 1g
revealed that the most stable conformers of these spiroligo-
zymes display arrangements that have more buried oxygen
nucleophiles and would be expected to be less catalytically
active than those with more exposed oxygen nucleophiles. In
1c, pyridine and benzyl alcohol can form a H-bond in three
different orientations (see the SI). The most stable conformer
1c-1 displays an excellent agreement in the arrangement of the
catalytic dyad with the QM theozyme (Figure 4a, RSMD = 0.32
Å). However, this orientation unfavorably requires the
approach of the ester from the hindered face of the dyad.
The conformation that has the most exposed oxygen
nucleophile and a pyridine−alchohol hydrogen bond: 1c-2,
lies 2.9 kcal/mol higher in energy than the lowest-energy
conformer 1c-1. The catalytic dyad shows a relatively more
distorted arrangement, compared to the ideal geometry of the
theozyme (Figure 4a, rmsd = 0.48 Å), but allows the approach
of the ester from the less hindered face. As observed in MD
simulations, the most stable conformer of 1g, 1g-1, shows an
alternative favorable interaction, in which benzyl alcohol forms
a H-bond with the amide carbonyl of the spiroligozyme rather
than with pyridine (Figure 4b, rmsd = 2.10 Å). An analogous
conformer of 1c is also found to be more stable than the
reactive conformer 1c-2 (see the SI). Nevertheless, unlike 1c,
the relative positioning of the pyridine and benzyl alcohol in 1g
permits only a single H-bonding conformation for the catalytic
dyad: 1g-2. 1g-2 is 2.3 kcal/mol higher in energy than 1g-1,
and overlays well with the QM theozyme (Figure 4b, rmsd =
0.43 Å).
The computed activation free energies with the pyridine−
alcohol dyad encouragingly fall in a reasonable range for a
catalytic mechanism. However, the relatively high barrier of
TS1-dyad compared to that of the triad model suggests that the
success of such transformations may be limited to the use of
more reactive acyl donors with enhanced electrophilicities.
Indeed, vinyl acetate was not consumed in the presence of
spiroligozyme 1g (data not shown), whereas 1g showed high
reactivity relative to benzyl alcohol with vinyl trifluoroacetate.
In order to explore how the nucleophilicity of an alcohol
could be enhanced by holding it close to a pyridyl group as a
general base, we modeled spiroligomers 1a−g that would
present an alcohol and a pyridyl (1b−g) or benzyl (1a) pair of
functional groups in different relative geometries. Some of the
spiroligomers (1c−g) could form hydrogen-bonded 18−19
membered rings in order to approximate geometries of the
theozyme (TS1-dyad), while others (1a and 1b) would act as
control molecules. The modeling suggested that the coupling of
the two pro4 bis-amino acids 4a−e and 7a−c through a
diketopiperazine formed by coupling their proline amino acids
could approximate the required geometry.
Kinetic experiments were carried out by combining the
spiroligomers 1a−g with an excess of vinyl trifluoroacetate in
CDCl3 at 22 °C using 19F NMR (see the SI). Figure 3 shows
Figure 3. Time-dependent acylation of spiroligomers 1a−g .
the conversion of spiroligomers to the trifluoroacetate esters
over time. These data demonstrate that the stereochemistry of
the spiroligomers and the regiochemistry of the pyridyl group
tune the relative 3D approach of a pyridyl base to the alcohol,
which leads to large differences in activity of the alcohol as a
nucleophile. Spiroligomer 1a is a control molecule that displays
a phenyl group instead of a pyridyl group, and the lack of this
base results in a very slow rate of acylation; thus, 1a is defined
to have a krel of 1 (Scheme 1 table). Spiroligomer 1c bears a 3-
pyridyl group that modeling demonstrates can approach in
close proximity to the benzyl alcohol and form a hydrogen
bond with it and increase its nucleophilicity; it is acylated 250
times faster than 1a. Spiroligomers 1d and 1e were designed to
slightly increase the distance between the 3-pyridyl group and
the benzyl alcohol relative to 1c by altering the relative
stereochemistry of stereocenters 1 and 3 to that of stereo-
centers 2 and 4. Both of these spiroligomers show a slightly
slower rate of acylation relative to 1c. Spiroligomer 1b holds
The results of the MD simulations and conformational
analyses show that the designed spiroligozymes present
18349
dx.doi.org/10.1021/ja3069648 | J. Am. Chem. Soc. 2012, 134, 18345−18353

Journal of the American Chemical Society
Article
Figure 4. Low-energy conformers [M06-2X/6-31G(d) (chloroform), relative electronic energies in kcal/mol with respect to the lowest-energy
conformer] of (a) 1c and (b) 1g (spiroligozymes, shown in green sticks) overlaid with the corresponding model transition state (theozyme, shown
in gray lines).
functional groups in quantum mechanically computed catalytic
arrangements, but these functional groups sample multiple
conformational states in which a subset is catalytically active.
The calculations suggest that eliminating the unproductive
conformations of spiroligozymes could significantly enhance
the acylation rates. The higher activity of 1g relative to that of
1c can be explained, in part, by (1) the lower energetic penalty
to adopt a reactive conformation and (2) the closer agreement
of the catalytic geometry with the optimal arrangement of the
functional groups in the transition state.
Inspired by the rapid formation of acyl−enzyme intermediate
in enzyme-catalyzed transesterification reactions, we sought to
evaluate spiroligomer 1g as a catalyst for the transesterification
of vinyl trifluoroacetate with methanol. Figure 5 shows the
formation of methyl trifluoroacetate over time with 10 mol % of
1g. Reactions were carried out under pseudo-first-order
conditions with excess methanol. Spiroligomer 1g shows
slightly faster methanolysis than the background reaction of
methanol with vinyl trifluoroacetate. The slow formation of
product is attributed to the lack of an “oxyanion hole” to
stabilize the transition states that follow the formation of the
trifluoroacyl−catalyst (1 g) complex.
In order to improve the rate of acyl transfer from the
trifluoroacyl−1g intermediate in the second step of trans-
esterification, the trifunctionalized spiroligomers 15a and 15b
were synthesized. Each bears a urea group as a potential donor
of two hydrogen bonds to activate the acyl−spiroligomer
intermediate for methanolysis. The modular synthesis of
functionalized spiroligomers makes it straightforward to
synthesize 15a and 15b by extending the synthesis of 1a−g.
Analysis of the methanolysis reaction of vinyl trifluoroacetate
in the presence of 10 mol % trifunctional spiroligomers 15a−
15b demonstrates catalysis of the transesterification reaction.
Spiroligomer 15a bears a 4-substituted urea benzoyl group and
shows an acceleration of methanolysis relative to the bifunc-
tional spiroligomer 1g and the background reaction (Figure 5).
Spiroligomer 15b displays a 3-substituted urea benzoyl and
demonstrates an even greater acceleration of transesterification
over the background reaction. These results prompted us to
calculate second-order rate constants for the 15b-catalyzed
transesterification. The reaction was carried with 6 mol %
catalyst under pseudo first-order conditions using 1 M
methanol, 0.6 mM 15b, and different concentration of vinyl
trifluoroacetate ranging from 10 mM to 53 mM. Using the
steady-state approximation, the rate constants for acylation of
compound 15b (k1) and the subsequent deacylation of the
intermediate 15b-Ac (k2) can be calculated.10 kuncat is measured
as the rate of transesterification when methanol is combined
with vinyl trifluoroacetate in the absence of base or catalyst.
Calculation of second-order rate constants revealed a 2200-fold
rate enhancement for k1 and 130-fold acceleration of k2. The
resulting ratio of k1 and kuncat is consistent with the krel (2700)
observed for the direct acylation of 1g.
Figure 5. A plot of the time-dependent formation of trifluoromethy-
lacetate from trifluorovinylacetate. (Conditions: All reactions are run
with 9 mM vinyl trifluoroacetate and 1 mM catalyst. Methanol
concentrations: 1.2 M MeOH with 15b, 1 M MeOH with 15a, 1.1 M
MeOH with 1g; 1 M MeOH for the blank reaction.)
18350
dx.doi.org/10.1021/ja3069648 | J. Am. Chem. Soc. 2012, 134, 18345−18353

Journal of the American Chemical Society
Article
Figure 6. Energy profile [M06-2X/6-31G(d)(chloroform)//ONIOM(B3LYP/6-31G(d):PM6); PCM-corrected energies in kcal/mol] for the
transesterification of vinyl trifluoroacetate catalyzed by 15b.
In order to assess whether the newly developed trifunctional
spiroligozymes function as designed, we calculated the
stationary points along the transesterification pathway of vinyl
trifluoromethylacetate catalyzed by 15b (Figure 6, see the SI for
3D representations of all stationary points). The trans-
esterification mechanism involves the pyridine−alcohol dyad
that provides the essential nucleophile. Vinyl trifluoroacetate
forms a complex with the spiroligozyme through H-bonding
interactions with the urea. At TS1, the benzyl alcohol is
partially deprotonated by the general base pyridine, and attacks
vinyl trifluoromethylacetate with an activation energy of 8.6
kcal/mol (Figure 6). TS1 overlays well with the theozyme
model involving urea as an oxyanion hole mimic (rmsd = 0.69
Å, see the SI). TS1 leads to the first tetrahedral intermediate,
which is 7.9 kcal/mol higher in energy than the starting
complex. The resulting oxyanion is stabilized by the urea, and
the protonated pyridine forms an H-bond with the oxygen of
the leaving group. This conformational reorganization of
pyridine allows general acid-assisted loss of vinyl alcohol
(TS2) in the next step (Figure 7). TS2 has an activation barrier
of 10.8 kcal/mol and generates the acylated spiroligozyme vinyl
alcohol complex, which is slightly exothermic by 0.4 kcal/mol.
The exchange of vinyl alcohol and methanol leads to an
energetically stable acylspirologozyme intermediate, which is
5.1 kcal/mol lower in energy than the starting complex. The
reaction then proceeds with the addition of methanol to the
acylated spiroligozyme via TS3. TS3, shown in Figure 6, is very
similar in geometry to TS2 (Figure 7). Methanol is partially
deprotonated by pyridine; the forming C−O bond is short
indicative of a late transition stateand the developing
negative charge on oxygen is stabilized by the urea. The attack
of methanol activated by the general base pyridine is 16.0 kcal/
mol uphill from the acyl−spiroligozyme intermediate. TS3
forms the tetrahedral intermediate of the deacylation step,
which is 3.0 kcal/mol higher in energy relative to that of the
starting complex. Subsequent conformational rearrangement of
pyridine leads to TS4 (Figure 7). Protonated pyridine now acts
a general acid and activates the benzyl alcohol as the leaving
group. This promotes the facile cleavage of the covalent bond
between the spiroligozyme and the substrate (ΔE = 6.0 kcal/
mol), releasing the transesterification product (ΔE = −2.2 kcal/
mol).
The experimental activation free energy of acylation, ΔG⧧ ,
1
(15b k1 = 0.17 M−1 s−1) is 18.3 kcal/mol, and ΔG⧧ which
corresponds to the deacylation step (15b k2 = 0.01 M−21 s−1) is
20.0 kcal/mol.55 The highest calculated barriers for the
acylation and deacylation steps are 10.8 and 10.9 kcal/mol
respectively (Figure 6, TS2 and TS3). However, these
calculated barriers cannot be directly compared to ΔG⧧ and
1
ΔG⧧ because the very large size of the system precludes the
2
calculation of important contributions such as entropies of
complex reactant binding, product release, and substrate
exchange processes, energetic penalties related to conforma-
tional preferences of the species formed in the course of the
reaction, and explicit solvent interactions. We estimate the
entropy of the complex reactant binding range to be between 4
and 6 kcal/mol,56 and the conformational energetic penalties
are on the order of 3 kcal/mol (see Figure 4). These additional
18351
dx.doi.org/10.1021/ja3069648 | J. Am. Chem. Soc. 2012, 134, 18345−18353

Journal of the American Chemical Society
Article
gozymes have been shown to be effective catalysts of
methanolysis of the ester, vinyl trifluoroacetate.
ASSOCIATED CONTENT
* Supporting Information
■
S
Computational details, Cartesian coordinates, absolute energies
and 3D representations of all optimized structures, snapshots
from molecular dynamics simulations, full citation for reference
26, and distance−angle scatter plots for pyridine−alcohol H-
bond. This material is available free of charge via the Internet at
http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
meister@temple.edu; houk@chem.ucla.edu
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by the Defense Threat Reduction
Agency (DOD-DTRA) (HDTRA1-09-1-0009) and the Na-
tional Institute of General Medical Sciences, National Institutes
of Health (GM36700 and GM067866). We thank the UCLA
Institute for Digital Research and Education (IDRE), the
Shared Research Computing Services (ShaRCS) Pilot Project
for the University of California Systems, and the Extreme
Science and Engineering Discovery Environment (XSEDE) for
computer time. This research was supported by an allocation of
advanced computing resources provided by the National
Science Foundation (TG-CHE100059). The computations
were performed on Kraken at the National Institute for
Computational Sciences (http://www.nics.tennessee.edu/).
C.E.S. and M.K. thank Dr. Charles DeBrosse for advice and
assistance with NMR. We also thank Dr. David Dalton for his
insight and helpful discussions.
Figure 7. Transition structures [Optimized using ONIOM(B3LYP/6-
31G(d):PM6). All distances are given in Å. Values in parentheses are
the corresponding partial bond distances for TS4 and TS3 (R′ =
CH3).] for the transesterification of vinyl trifluoroacetate catalyzed by
15b.
terms (except for solvent interactions) would positively add to
the calculated barriers to bring them in line with the
experimental activation free energies, ΔG⧧ and ΔG⧧ .
1
2
REFERENCES
■
The optimized intermediates and transition structures clearly
demonstrate that the newly developed biomimetic spiroligo-
zymes are able to provide correctly positioned catalytic arrays
for effective catalysis. Conformational reorganization of
catalytic functional groups maintaining optimized TS stabiliza-
tion along the reaction pathway contributes to catalysis, as
described for their natural counterparts.19 Active sites of natural
esterases are also highly preorganized with well-defined and
well-maintained interactions in the dynamic environment.
However, for di- and trifunctional spiroligozymes, avoiding
nonproductive conformations through better preorganization
of the reactive groups should lead to further lowering of the
transition-state energies. As for 1c and 1g, the MD simulations
for 15b depict alternative interactions of the catalytic
functionalities with the spiroligozyme backbone. The most
stable conformer of 15b is 2.4 kcal/mol lower in energy than
the reactive conformer (see the SI).
(1) Radzicka, A.; Wolfenden, R. Science 1995, 267, 90.
(2) Zhang, X.; Houk, K. N. Acc. Chem. Res. 2005, 38, 379.
(3) Cheong, P. H. Y.; Legault, C. Y.; Um, J. M.; Celebi-Olcum, N.;
Houk, K. N. Chem. Rev. 2011, 111, 5042.
(4) MacMillan, D. W. C. Nature 2008, 455, 304.
(5) Knowles, R. R.; Jacobsen, E. N. Proc. Natl. Acad. Sci. U.S.A. 2010,
107, 20678.
(6) Tanaka, F.; Barbas, C. F. J. Am. Chem. Soc. 2002, 124, 3510.
(7) Houk, K. N.; Cheong, P. H. Y. Nature 2008, 455, 309.
(8) Alberg, D. G.; Poulsen, T. B.; Bertelsen, S.; Christensen, K. L.;
Birkler, R. D.; Johannsen, M.; Jorgensen, K. A. Bioorg. Med. Chem. Lett.
2009, 19, 3888.
(9) Jarvo, E. R.; Copeland, G. T.; Papaioannou, N.; Bonitatebus, P. J.;
Miller, S. J. J. Am. Chem. Soc. 1999, 121, 11638.
(10) Ema, T.; Tanida, D.; Matsukawa, T.; Sakai, T. Chem. Commun.
2008, 957.
(11) Brown, Z. Z.; Schafmeister, C. E. Org. Lett. 2010, 12, 1436.
(12) Schafmeister, C. E.; Brown, Z. Z.; Gupta, S. Acc. Chem. Res.
2008, 41, 1387.
(13) Zhao, Q.; Lam, Y. H.; Kheirabadi, M.; Xu, C.; Houk, K. N.;
Schafmeister, C. E. J. Org. Chem. 2012, 77, 4784.
(14) Rothlisberger, D.; Khersonsky, O.; Wollacott, A. M.; Jiang, L.;
Dechancie, J.; Betker, J.; Gallaher, J. L.; Althoff, E. A.; Zanghellini, A.;
Dym, O.; Albeck, S.; Houk, K. N.; Tawfik, D. S.; Baker, D. Nature
2008, 453, 190.
CONCLUSION
■
Quantum mechanics and molecular dynamics with empirical
force fields have been used to design and analyze a novel type
of spiroligozyme, a spirocyclic peptidomimetic catalyst with a
relative rigid structure that can be designed and synthesized to
present functional groups in an arrangement that exhibits
multifunctional catalysis. Here, several generations of spiroli-
(15) Jiang, L.; Althoff, E. A.; Clemente, F. R.; Doyle, L.;
Rothlisberger, D.; Zanghellini, A.; Gallaher, J. L.; Betker, J. L.;
18352
dx.doi.org/10.1021/ja3069648 | J. Am. Chem. Soc. 2012, 134, 18345−18353

Journal of the American Chemical Society
Article
Tanaka, F.; Barbas, C. F.; Hilvert, D.; Houk, K. N.; Stoddard, B. L.;
Baker, D. Science 2008, 319, 1387.
(16) Siegel, J. B.; Zanghellini, A.; Lovick, H. M.; Kiss, G.; Lambert, A.
R.; Clair, J. L. S.; Gallaher, J. L.; Hilvert, D.; Gelb, M. H.; Stoddard, B.
L.; Houk, K. N.; Michael, F. E.; Baker, D. Science 2010, 329, 309.
(17) Kirby, A. Acc. Chem. Res. 1997, 30, 290.
(47) Banks, J. L.; Beard, H. S.; Cao, Y.; Cho, A. E.; Damm, W.; Farid,
R.; Felts, A. K.; Halgren, T. A.; Mainz, D. T.; Maple, J. R.; Murphy, R.;
Philipp, D. M.; Repasky, M. P.; Zhang, L. Y.; Berne, B. J.; Friesner, R.
A.; Gallicchio, E. J. Comput. Chem. 2005, 26, 1752.
(48) Polak, E.; Ribiere, G. Ser. Rouge 1969, 16, 35.
(49) Still, W. C.; Tempczyk, A.; Hawley, R. C.; Hendrickson, T. J.
Am. Chem. Soc. 1990, 112, 6127.
(18) Warshel, A.; Sharma, P. K.; Kato, M.; Xiang, Y.; Liu, H.; Olsson,
(50) Svensson, M.; Humbel, S.; Morokuma, K. J. Chem. Phys. 1996,
105, 3654.
(51) Vreven, T.; Morokuma, K. J. Comput. Chem. 2000, 21, 1419.
(52) Dapprich, S.; Komaromi, I.; Byun, K. S.; Morokuma, K.; Frisch,
M. J. Mol. Struct. (THEOCHEM) 1999, 462, 1.
M. H. M. Chem. Rev. 2006, 106, 3210.
(19) Smith, A. J. T.; Muller, R.; Toscano, M. D.; Kast, P.; Hellinga, H.
W.; Hilvert, D.; Houk, K. N. J. Am. Chem. Soc. 2008, 130, 15361.
(20) Kiss, G.; Rothlisberger, D.; Baker, D.; Houk, K. N. Protein Sci.
̈
2010, 19, 1760.
(53) Stewart, J. J. P. J. Mol. Model. 2007, 13, 1173.
(54) Brown, Z. Z.; Schafmeister, C. E. J. Am. Chem. Soc. 2008, 130,
14382.
(55) The experimental activation free energies of acylation and
deacylation steps were calculated for 15b at room temperature using
the Eyring−Polanyi equation (see Supporting Information).
(56) Lu, B.; Wong, C. F. Biopolymers 2005, 79, 277.
(21) Hedstrom, L. Chem. Rev. 2002, 102, 4501.
(22) Wright, C. S.; Alden, R. A.; Kraut, J. Nature 1969, 221, 235.
(23) Schafmeister, C. E. CANDO: Computer Aided Nanostructure
Design and Optimization software package - under development;
2012.
(24) Richter, F.; Blomberg, R.; Khare, S. D.; Kiss, G.; Kuzin, A. P.;
Smith, A. J. T.; Gallaher, J. L.; Pianowski, Z.; Helgenson, R. C.;
Grjasnow, A.; Xiao, R.; Seetharam, J.; Su, M.; Vorobev, S.; Lew, S.;
Forouhar, F.; Kornhaber, G. J.; Hunt, J. F.; Montelione, G. T.; Tong,
L.; Houk, K. N.; Hivert, D.; Baker, D. J. Am. Chem. Soc. 2012,
DOI: 10.1021/ja3037367.
(25) Miller, S. J. Acc. Chem. Res. 2004, 37, 601.
(26) Frisch, M. J.; et al. Gaussian 09, revision C.01; Gaussian, Inc.:
Wallingford, CT, 2010;. Frisch, M. J.; et al. Gaussian 03, revision C.02;
Gaussian, Inc.: Wallingford, CT, 2004. See the Supporting
Information for the complete reference.
(27) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
(28) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B: Condens. Mater. Phys.
1998, 37, 785.
(29) Vosko, S. H.; Wilk, L.; Nusair, M. Can. J. Phys. 1980, 58, 1200.
(30) Stephens, P. J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J. J.
Phys. Chem. 1994, 98, 11623.
(31) Cossi, M.; Barone, V.; Cammi, R.; Tomasi, J. Chem. Phys. Lett.
1996, 255, 327.
(32) Cances
3032.
́
, E.; Mennucci, B.; Tomasi, J. J. Chem. Phys. 1997, 107,
(33) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. Rev. 2005, 105,
2999.
(34) Zhao, Y.; Truhlar, D. G. Theor. Chem. Acc. 2008, 120, 215.
(35) Marenich, A. V.; Cramer, C. J.; Truhlar, D. G. J. Phys. Chem. B
2009, 113, 6378.
(36) Case, D. A.; Darden, T. A.; Cheatham, I. T. E.; Simmerling, C.
L.; Wang, J.; Duke, R. E.; Luo, R.; Walker, R. C.; Zhang, W.; Merz, K.
M.; Roberts, B.; Wang, B.; Hayik, S.; Roitberg, A.; Seabra, G.;
́
Kolossvary, I.; Wong, K. F.; Paesani, F.; Vanicek, J.; Liu, J.; Wu, X.;
Brozell, S. R.; Steinbrecher, T.; Gohlke, H.; Cai, Q.; Ye, X.; Wang, J.;
Hsieh, M. J.; Cui, G.; Roe, D. R.; Mathews, D. H.; Seetin, M. G.; Sagui,
C.; Babin, V.; Luchko, T.; Gusarov, S.; Kocalenko, A.; Kollman, P. A.
AMBER 11; University of California: San Francisco, 2010.
(37) Wang, J.; Wolf, R. M.; Caldwell, J. W.; Kollman, P. A.; Case, D.
A. J. Comput. Chem. 2004, 25, 1157.
(38) Wang, J. M.; Cieplak, P.; Kollman, P. A. J. Comput. Chem. 2000,
21, 1049.
(39) Bayly, C. I.; Cieplak, P.; Cornell, W. D.; Kollman, P. A. J. Phys.
Chem. 1993, 97, 10269.
(40) Besler, B. H.; Merz, K. M.; Kollman, P. A. J. Comput. Chem.
1990, 11, 431.
(41) Singh, U. C.; Kollman, P. A. J. Comput. Chem. 1984, 5, 129.
(42) Cieplak, P.; Caldwell, J. W.; Kollman, P. A. J. Comput. Chem.
2001, 22, 1048.
(43) Duke, R. E.; Pedersen, L. G. PMEMD; University of North
Carolina: Chapel Hill,: 2003.
(44) Darden, T. A.; York, D.; Pedersen, L. J. Chem. Phys. 1993, 98,
10089.
(45) Blas, J. R.; Mar
Am. Chem. Soc. 2002, 124, 12796.
(46) Macromodel version 9.9; Shrodinger, LLC: New York, NY, 2011.
́
quez, M.; Sessler, J. L.; Luque, F. J.; Orozco, M. J.
̈
18353
dx.doi.org/10.1021/ja3069648 | J. Am. Chem. Soc. 2012, 134, 18345−18353