Modulating the performance of an asymmetric organocatalyst by tuning its spatial environment in a metal-organic framework

Article abstract of DOI:10.1021/jacs.7b07921

Systematically tuning the spatial environment around the active sites of synthetic catalysts is a difficult challenge. Here, we show how this can be accomplished in the pores of multicomponent metal-organic frameworks. This relies on embedding a catalytic unit in a pore of the MUF-77 framework and then tuning its environment by introducing different functional groups to the surrounding linkers. This approach benefits from the structural regularity of MUF-77, which places each component in a precise location to circumvent disorder. Prolinyl groups, which are catalytically competent toward asymmetric aldol reactions, were selected as the catalytic unit. Since every prolinyl group is positioned in an identical environment, correlations between the pore architecture and the activity of these single-site catalysts can be elucidated. Systematic engineering of the pore structure, which is achieved by installing modulator groups on the framework linkers, impacts on the reaction rate and the enantiomeric excess of the aldol products. Furthermore, the spatial environment around the proline catalyst can override its innate stereochemical preference to dictate the preferred enantiomer of the reaction product. These results offer a new way to design three-dimensional active site environments for synthetic catalysts.

Full text of DOI:10.1021/jacs.7b07921

Article
pubs.acs.org/JACS
Modulating the Performance of an Asymmetric Organocatalyst by
Tuning Its Spatial Environment in a Metal−Organic Framework
Lujia Liu,† Tian-You Zhou,† and Shane G. Telfer*
MacDiarmid Institute for Advanced Materials and Nanotechnology, Institute of Fundamental Sciences, Massey University,
Palmerston North 4442, New Zealand
S
* Supporting Information
ABSTRACT: Systematically tuning the spatial environment around the active
sites of synthetic catalysts is a difficult challenge. Here, we show how this can be
accomplished in the pores of multicomponent metal−organic frameworks. This
relies on embedding a catalytic unit in a pore of the MUF-77 framework and then
tuning its environment by introducing different functional groups to the
surrounding linkers. This approach benefits from the structural regularity of
MUF-77, which places each component in a precise location to circumvent
disorder. Prolinyl groups, which are catalytically competent toward asymmetric
aldol reactions, were selected as the catalytic unit. Since every prolinyl group is
positioned in an identical environment, correlations between the pore architecture
and the activity of these single-site catalysts can be elucidated. Systematic
engineering of the pore structure, which is achieved by installing modulator groups
on the framework linkers, impacts on the reaction rate and the enantiomeric excess of the aldol products. Furthermore, the
spatial environment around the proline catalyst can override its innate stereochemical preference to dictate the preferred
enantiomer of the reaction product. These results offer a new way to design three-dimensional active site environments for
synthetic catalysts.
maintained upon functionalization. The pores in these
frameworks feature a precise and predictable three-dimensional
array of chemical groups. Given that these MOFs tolerate a
diverse range of substituents on the linkers, they offer a unique
platform to create pockets of chemical space with predefined
characteristics that can be tuned with precision at the molecular
scale.
Catalysts that are active and selective toward organic
transformations are highly sought after. One method of
enhancing the performance of abiological catalysts is to engage
the reaction participants with specific noncovalent contacts
with, for example, functional groups on the catalyst or the
solvent.25−32 This relies on the sensitivity of the reaction
outcome to perturbations to the energies of the competing
transition states, which in turn is a function of the catalyst
structure and its environment. An unmet challenge in this
domain is the implementation of a well-defined array of
noncovalent interactions in all three dimensions around a
catalytic site. We propose that this can be achieved by
exploiting the ability to engineer the chemical microenviron-
ment of the multicomponent MUF-77 framework.
INTRODUCTION
■
The size, shape and chemical characteristics of the pores in
metal−organic frameworks (MOFs) are central to their
functional properties.1 One established method of tuning the
pore characteristics of MOFs is to introduce functional groups
to the organic linkers. This is beautifully demonstrated by
several isoreticular families of frameworks.2−4 The arrangement
of two or more different functional groups in a MOF pore
engenders more sophisticated chemical characteristics. One
method to engineer such pores is to coassemble a set of linkers
that share the same backbone into a multivariate MOF.5
Although such materials can exhibit interesting functional
properties,6,7 the linkers are distributed in the lattice with a
degree of randomness, which leads to pore heterogeneity.8−11
This will frequently be a disadvantage for applications such as
catalysis and sensing, which benefit from architectural
regularity. An alternative strategy of using two12−19 or
three4,20−23 topologically distinct linkers to build up multi-
component MOFs24 can preserve pore homogeneity. Since the
linkers can be discriminated during framework assembly, they
become positioned in unique lattice sites. We have shown that
by installing functional groups on the organic linkers of the
multicomponent MOFs MUF-7 and MUF-77, [Zn4O-
(bdc)1/2(bpdc)1/2(trig)4/3] (bdc = 1,4-benzenedicarboxylate,
bpdc = 4,4′-biphenyldicarboxylate, trig = 1,3,5-benzenetriben-
zoate or a truxene analogue) their pore architectures can be
programmed.4,20 An important consequence of the lack of
disorder in MUF-7 and MUF-77 is that structural regularity is
To implement this strategy, we first selected a catalytic unit
to embed in MUF-77. Earlier work has established that effective
heterogeneous catalysts can be generated by appending known
organocatalytic units to the components of MOFs and related
Received: August 3, 2017
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.7b07921
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
porous materials.33−38 We thus envisaged that a catalytic group
could be installed on one of the three linkers of MUF-77
(Figure 1). If the linker that bears the catalytic unit has the
Figure 1. A schematic depiction of a pore in MUF-77 equipped with a
site for catalysis and modulator groups that are positioned to influence
the course of the reaction.
same backbone as one of the components of the MUF-77
framework it will specifically occupy the position of its parent
linker in the lattice. A diverse range of organocatalytic moieties
is amenable to this role, for example proline39,40 or urea,41 or
groups that coordinate metal ions.42−44 Here, we opted for the
prolinyl group since it has the advantages of being readily
available in an enantiopure form, nontoxic, resistant to oxygen
and water, and high activity under mild conditions (e.g., the
addition of strong acids or bases is not required). The second
aspect of our design strategy is the installation of modulator
groups on the two linkers that do not bear the catalytic unit
(Figure 1). We anticipated that, via noncovalent interactions
with the reaction participants, these groups may be able to
influence the reaction rate or alter the regioselectivity or
stereoselectivity of product formation.
Figure 2. (a) The synthetic route to the Boc-protected MUF-77
derivatives and the thermolytic deprotection reaction that converts
them into active catalysts. (b) The chemical structures of the tritopic
truxene linkers used in this study. (c) The structures of R2bdc-Pro and
R2bpdc-Pro. (d) The structures of the ditopic modulator ligands used
in this study.
To deprotect the prolinyl units in [Zn4O(bdc-Pro-
Boc)1/2(bpdc)1/2(hmtt)4/3] and [Zn4O(bdc)1/2(bpdc-Pro-
Boc)1/2(hmtt)4/3] we simply heated the materials to 200 °C
under vacuum to release the Boc groups as isobutylene and
CO2 using our previously established protocol (Figure 2a).45,46
The resulting MOFs are referred to as [Zn4O(bdc-
Pro)1/2(bpdc)1/2(hmtt)4/3] and [Zn4O(bdc)1/2(bpdc-
Pro)1/2(hmtt)4/3]. A combination of SCXRD (Table S1),
RESULTS AND DISCUSSION
■
Catalyst Synthesis and Structure. Our initial step was to
introduce catalytically active prolinyl groups to MUF-77. Boc-
protected (S)-prolinyl groups were appended to bdc and bpdc
linkers by amide coupling reactions to give H2bdc-ProBoc and
H2bpdc-ProBoc,39 respectively (Figure 2a). The incorporation
of bdc-ProBoc and bpdc-ProBoc into MUF-77 analogues was
achieved using standard reaction conditions, viz the solvother-
mal reaction of three linkers (H2bdc, H2bpdc and
5,5′,10,10′,15,15′-hexamethyltruxene-2,7,12-tricarboxylic acid
(H3hmtt)) with Zn(NO3)2 in DEF. The structures of the
r e s u l t i n g f r a m e w o r k s , [ Z n 4 O ( b d c - P r o -
Boc)1/2(bpdc)1/2(hmtt)4/3] and [Zn4O(bdc)1/2(bpdc-Pro-
Boc)1/2(hmtt)4/3], were determined by X-ray crystallography
(SCXRD, Table S1). The lattice topology (ith-d) of both
structures match the parent MUF-77 framework. Both the
reflection and refinement statistics suggest that the framework
model is best refined in the centrosymmetric Pm-3 space group
despite the enantiopure (S)-prolinyl groups. This is due to the
indistinct contributions made to the electron density maps by
the chiral portion of the ligands. The existence of the intact
ProBoc groupsalong with the phase purity of the MOFs
was confirmed by 1H NMR spectroscopy on dissolved samples.
The integrals of the peaks of each of the three linkers perfectly
matches those anticipated from the MOF stoichiometry
(Figures S7, S15).
1
powder X-ray diffraction (PXRD, Figures S25−S26), H NMR
spectroscopy (Figures S8, S14), and microscopy (Figures S58−
S60) shows that the Boc groups are completely cleaved in this
step and that the integrity of the lattice is maintained. We note
t h a t a t t e m p t s t o s y n t h e s i z e [ Z n 4 O ( b d c -
Pro)1/2(bpdc)1/2(hmtt)4/3] and [Zn4O(bdc)1/2(bpdc-
Pro)1/2(hmtt)4/3] directly from H2bdc-Pro and H2bpdc-Pro,
respectively, led to undesired reactions of the prolinyl groups
and limited catalytic activity (Figures S9, S15).
The structures of [Zn4O(bdc-Pro)1/2(bpdc)1/2(hmtt)4/3] and
[Zn4O(bdc)1/2(bpdc-Pro)1/2(hmtt)4/3] were determined by
single-crystal X-ray diffraction and are shown in Figure 3a. As
previously observed for other members of the MUF-77 series,
the three linkers are discriminated during framework assembly
and each linker occupies a specific lattice site. Since the proline
groups can occupy any one of four symmetry-related sites on
the bdc or bpdc linkers it is remarkable that the first four atoms
that extend away from the aryl ring could be found in the
electron density difference map. The observation of these
B
DOI: 10.1021/jacs.7b07921
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
oriented toward the interior of a tetrahedral cavity defined by
the bdc, bpdc and four hmtt linkers. All of the pores in these
frameworks are crystallographically identical. An importantly
corollary of this is that every prolinyl group is poised to operate
in an equivalent environment in [Zn4 O(bdc-
Pro)1/2(bpdc)1/2(hmtt)4/3] and [Zn4O(bdc)1/2(bpdc-
Pro)1/2(hmtt)4/3].
We subsequently designed modulator ligands to incorporate
in the proline-containing frameworks. Our first consideration
was that modulator linkers without suitable symmetry would
not preserve the desired single-site nature of these catalysts.
Second, the size of the functional groups appended to the
modulator backbone is important. These groups must be large
enough to exert an influence on the reaction occurring at the
catalytic site but not so large that access by the reaction
substrates is blocked. We used molecular modeling to assess
this steric requirement. With these considerations in mind the
ditopic modulators shown in Figure 2d were designed and
synthesized. These linkers preserve a C2 symmetry axis
orthogonal to their long axes to maintain compatibility with
the symmetry of the pore environment. To produce C3-
symmetric tritopic modulators the methyl groups of hmtt were
replaced by ethyl, butyl groups and octyl groups (hett, hbtt and
hott, Figure 2b).
A model of a hypothetical catalytic site, which was produced
by simple molecular modeling, is presented in Figure 3b. This
schematic depicts an intermediate in the asymmetric aldol
reaction of acetone and 4-nitrobenzaldehyde that is covalently
bound to the prolinyl group in one of the pores of [Zn4O(bdc-
Pro)1/2(pcdc)1/2(hmtt)1/2]. This model emphasizes the poten-
tial for contacts between the van der Waals surfaces of the
linkers and the reaction participants when reactions involving
typical substrates take place at the proline site in this pore. The
proximity of the phenyl ring of pcdc and the methyl groups of
hmtt to the catalytic ensemble suggests that theseand
similargroups ought to be capable of imparting control
over the catalyzed reaction via noncovalent interactions (that
are either attractive or repulsive in nature). Although not
shown, solvent molecules present in the pore are also likely to
play a role. Here, we also make a brief comment about dynamic
motion of the framework. On the basis of measurements on
related MOFs,47,48 the bdc-Pro and bpdc-Pro linkers in MUF-
77 are not expected to be static. Rather, they are likely to flip
around their long axes on very short (∼picosecond) time scales.
A 180° flip of the catalytic linker will project the prolinyl group
into a crystallographically identical pore. During this rotational
motion the prolinyl group will sample the environment of the
large dodecahedral pore of the MUF-77 framework. However,
since all prolinyl groups in the framework will experience the
same dynamic motion and time-averaged environment and
their homogeneity will not be compromised.
The synthesis of MUF-77 type frameworks from sets of bdc/
bpdc/truxene components (Figures 2b and 2d) was accom-
plished using conditions similar to those developed for the
parent MOFs. In each case, a prolinyl group was installed on
either the bdc linker (bdc-Pro) or the bpdc linker (bpdc-Pro).
The ancillary ditopic ligand and the truxene linker served as
modulator linkers. The combinations of linkers that were
incorporated in the catalytic frameworks are summarized in
Table 1. Powder X-ray diffraction (Figures S25−S26) together
with 1H NMR spectroscopy on dissolved samples (Figures S7−
S23) demonstrate that the global structure of MUF-77 is
preserved upon the introduction of these functionalized
Figure 3. (a) The structures of [Zn4O(bdc-Pro)1/2(bpdc)1/2(hmtt)4/3
]
and [Zn4O(bdc)1/2(bpdc-Pro)1/2(hmtt)4/3] as deduced by X-ray
crystallography. The prolinyl groups are shown in representative
positions. Hydrogen atoms are omitted. (b) A schematic illustration of
[Zn4O(bdc-Pro)1/2(pcdc)1/2(hmtt)4/3] bound to an aldol reaction
intermediate to emphasize the potential contacts between the
participants of the catalysis reaction and the modulator groups. The
aldol intermediate is shown in orange, the methyl groups of hmtt in
red, and the phenyl ring of the pcdc linker in violet.
atoms is made possible by the high crystallographic order of the
lattice and the stability of the framework toward removal of
occluded solvent molecules (that would otherwise produce
background scatter). The atoms that could not be found
crystallographically were place in calculated positions. In Figure
3a the prolinyl groups are shown in representative positions
C
DOI: 10.1021/jacs.7b07921
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Preliminary screening experiments revealed that the reaction
rate increases with added water, as observed for homogeneous
catalysts,54 and is independent of whether the reaction mixture
is stood or tumbled. We settled on the following standard
catalysis conditions: standing a heterogeneous mixture of the
catalyst, acetone (≈1 mL, large excess), H2O (20% v/v) and
the aldehyde (0.04 M) at 20 °C. By running the reaction with
increasing quantities of catalyst we established that the
saturation loading of the catalyst, beyond which additional
catalyst does not increase the reaction rate, is reached around a
loading of ≈30 mol % of proline units (relative to the amount
of aldehyde). At 10 mol % the reaction typically proceeded to
completion over a period of ≈48 h. This became the standard
catalyst quantity, which is in the typical range for organo-
catalysts. The reaction mixture was automatically sampled by
HPLC every 120 min for 12 h and then again after 24 h (Figure
S35). An observed rate constant, kobs, was calculated based on
the consumption of the aldehyde over the initial 12 h of the
reaction (Figure 4a). The linearity of the pseudo-first-order
kinetics plot over the entire course of the reaction was
established by monitoring catalysis by [Zn4O(bdc-
Pro)1/2(bpdc)1/2(hmtt)4/3] until complete consumption of the
aldehyde (Figure S53). The ee was determined from HPLC
chromatograms recorded after 24 h. A control reaction
Table 1. Aldol Reaction of Acetone and p-
Nitrobenzaldehyde Catalyzed by MUF-77 Analogues with
a
Installed bdc-Pro or bpdc-Pro Linkers
b
c
#
catalyst linker set
ee (%)
kobs
1
bdc-Pro/bpdc/hmtt
bdc/bpdc-Pro/hmtt
bdc-Pro/bpdc/hett
bdc/bpdc-Pro/hett
bdc-Pro/bpdc/hbtt
bdc/bpdc-Pro/hbtt
bdc-Pro/bpdc/hott
bdc/bpdc-Pro/hott
ndc/bpdc-Pro/hmtt
ndc/bpdc-Pro/hbtt
ndc/bpdc-Pro/hott
bdc-Pro/pcdc/hmtt
bdc-Pro/dppdc/hmtt
bdc-ProBoc/bpdc/hmtt
bdc/bpdc-ProBoc/hmtt
−3.9 0.6
19.2 1.1
−3.9 0.6
17.0 1.0
4.2 0.6
18.3 2.6
1.8 0.3
15.7 2.1
24.2 1.4
18.1 1.1
19.1 1.1
−10.3 1.4
−26.5 1.6
n.r.
0.74 0.03
0.52 0.03
0.69 0.04
0.32 0.02
0.62 0.02
0.15 0.01
0.42 0.02
0.11 0.01
0.50 0.02
0.16 0.01
0.13 0.01
0.052 0.003
0.27 0.01
n.r.
2
3
4
5
6
7
8
9
10
11
12
13
14
15
n.r.
n.r.
homogeneous catalysts
16
17
Me2bdc-Pro
8.7 1.2
3.94 0.10
0.49 0.01
Me2bpdc-Pro
30.3 1.8
a
Standard reaction conditions: 10 mol % catalyst (prolinyl units
relative to total aldehyde), 0.04 M 4-nitrobenzaldehyde in acetone
b
with 20 vol % H2O at 20 °C. n.r. = no reaction. Enantiomeric excess
of product 1. A positive sign indicates the R enantiomer of the product
c
is dominant and vice versa. Observed rate constant based on the
consumption of 4-nitrobenzaldehyde with units of L mol−1 day−1.
Detailed data and information on standard deviation estimates are
available in the SI.
components and that all samples are delivered in a phase-pure
form. Any framework defects occur at exceedingly low levels, as
1
can be inferred from (i) H NMR spectroscopy on dissolved
samples which showed the linkers to be present in the exact
ratio defined by the framework stoichiometry, (ii) the absence
of any visible impurities in these NMR spectra, (iii) the match
between the simulated and experimental N2 adsorption
isotherms (Figures S32−S34) since it is well established that
defective MOFs will lead to anomalous gas adsorption
behavior,49−53 and (iv) the SCXRD refinement statistics and
thermal displacement parameters, which are consistent with the
full occupancy of all framework atoms.
These observations establish MUF-77 as a powerful and
versatile platform for the development of a broad family of
isoreticular heterogeneous catalysts in which the pore environ-
ment around the site of catalysis can be systematically tuned.
Catalysis Reactions. We found that the parent materials,
[Zn4O(bdc-Pro)1/2(bpdc)1/2(hmtt)4/3] and [Zn4O-
(bdc)1/2(bpdc-Pro)1/2(hmtt)4/3], catalyze the asymmetric
aldol transformation of 4-nitrobenzaldehyde and acetone to 1
(Table 1 Entries 1 and 2). This reaction was chosen as a model
since it has been investigated in a range of other heterogeneous
systems. We emphasize that our aim was not to produce an
outstanding catalyst per se, but to use this reaction to develop
the central concept of this report.
Figure 4. (a) A plot showing that the consumption of 4-
nitrobenzaldehyde during its aldol reaction with acetone to produce
1 catalyzed by [Zn4O(bdc)1/2(bpdc-Pro)1/2(hmtt)4/3] exhibits pseudo-
first-order kinetics. The orange squares are the measured data and the
blue line is the least-squares minimized line of best fit to these points.
C is the concentration of 4-nitrobenzaldehyde at time t, while C0 is the
initial concentration. (b) Optical micrographs of crystals of [Zn4O-
(bdc)1/2(bpdc-Pro)1/2(hmtt)4/3] before and after catalyzing the
production of 1 demonstrate that their morphology and transparency
is retained during catalysis (scale bar = 200 μm).
D
DOI: 10.1021/jacs.7b07921
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
demonstrated that, as anticipated, frameworks with Boc-
protected proline groups are catalytically inactive (Table 1,
Entries 14 and 15). Further, to demonstrate the heterogeneous
nature of the reaction, after several hours the catalysis mixture
was filtered through a 0.45 μm filter to remove the MOF
crystals. This halted the reaction, indicating that no catalytically
active species are present in the supernatant.
Once this standard procedure had been established we
developed a careful set of protocols to enable valid comparisons
to be made between the various catalysts. We first verified that
catalysts obtained from (i) independent synthetic batches, and
(ii) different samples from the same batch led to reproducible
kobs and ee values (Tables S4−S5). This also allowed standard
deviations of kobs and ee to be estimated. We then examined the
activity of the catalysts with modulator groups, and their kobs
and ee values are listed in Table 1. All of these catalysts
produced linear pseudo-first-order kinetics plots (Figures S38−
S48). As a further quality control measure, kobs and the ee for
the parent catalysts (Table 1, Entries 1 and 2) were
redetermined several times during the period when data on
the full set of catalysts was being acquired and were found to
fall within one standard deviation of the listed values.
We found that [Zn4O(bdc-Pro)1/2(bpdc)1/2(hmtt)4/3] and
[Zn4O(bdc)1/2(bpdc-Pro)1/2(hmtt)4/3] are recyclable catalysts.
The crystals are unchanged after catalyzing the production of 1,
as indicated by PXRD (Figure S54), NMR spectroscopy
(Figure S55), and optical microscopy (Figures 4b, S60).
Following one reaction cycle, the crystals can be isolated by
removal of the reaction supernatant, washed with fresh solvent
and reused to catalyze a subsequent reaction. A series of five
runs was conducted using the same catalyst (Figures S51−S52).
The reaction rate remained constant over successive runs for
[Zn4O(bdc)1/2(bpdc-Pro)1/2(hmtt)4/3], while a slight decrease
was observed for [Zn4O(bdc-Pro)1/2(bpdc)1/2(hmtt)4/3]. A
similar drop in performance has been noted previously for
some other heterogeneous catalysts with pyrrolidinyl
groups.35,36,39 In addition to demonstrating the recyclability
of these catalysts, these results also illustrate their tolerance
toward water, which comprises 20% of the reaction medium.
This high degree of water stability was established for the
parent MUF-77 framework in an earlier report.20 It is
somewhat uncommon for zinc(II)-carboxylate MOFs, and
thus is an attractive characteristic of the MUF-77 catalyst
family.
(2) The topography and chemistry of the catalytic pores in
MUF-77 can be programmed by installing modulator
linkers. Similar manipulations of catalytic pores have only
been reported for a limited number of synthetic
materials. For example, Long et al. observed that alkyl
or fluoro groups installed adjacent to the catalytic metal
sites of a series of Fe-MOF-74 analogues exerted a
significant influence over the catalytic process,62 and
Deria et al. reported the dependence of the product
distribution of an acyl transfer reaction on lattice
topology in a series of Zr-porphyrin MOFs.63 MUF-77
analogues enable pore environments to be controlled
with greater precision and diversity to build on these
earlier breakthroughs.
(3) The rate of the reaction catalyzed by bpdc-Pro is
unaffected by its incorporation into the MUF-77
framework (Entry 2 vs 17). This demonstrates that
heterogenization of a catalytic unit in a MOF does not
necessarily introduce kinetic barriers such as hindered
mass transport. The reaction rate of homogeneous
Me2bdc-Pro is higher than the MOFs containing the
bdc-Pro linker (Entry 1 vs 16). We tentatively attribute
this to an acceleration of the homogeneous reaction by
beneficial noncovalent interactions between the carbox-
ymethyl group and the reaction participants.
(4) Modulator groups placed on linkers remote from the site
of catalysis have a decisive impact on the asymmetric
aldol reaction (Table 1). These results embody the
successful realization of the central concept introduced
by this study. When catalysis takes place at the bdc-Pro
site and the tritopic linker is hmtt, replacing bpdc by the
pcdc modulator results in a significant improvement in
the ee of the aldol product 1 (Entries 1 vs 12). A further
pronounced enhancement of the ee of 1 is induced by
the dppdc modulator (Entry 13). Similar effects are
apparent for catalysis taking place on the bpdc linker: the
incorporation of ndc as a modulator boosts the ee of the
̀
reaction vis-a-vis bdc (Entry 9 vs Entry 2). Alkyl group
modulators on the truxene ligand also have a discernible
impact on the aldol reaction. The reaction occurring at
bdc-Pro produces the S enantiomer of the product in
excess when hmtt and hett are installed (Entries 1 and
3); however, the preference switches to the R enantiomer
with hbtt and hott (Entries 5 and 7). When using bpdc-
Pro as the catalytic linker, the ee is sensitive to the
identity of the alkyl substituents on the truxene while the
bdc linker is in place (Entries 2, 4, 6 and 8). This
sensitivity is amplified when the bdc is replaced by ndc
(Entries 9−11).
A series of telling conclusions may be drawn from the
structural and catalysis data acquired in this study:
(1) The installation of prolinyl groups on the linkers to
produce catalytically active analogues of MUF-77 does
not perturb the structural characteristics of the parent
framework and they remain well ordered. Therefore,
these materials are genuine single-site55−60 catalysts and
all the prolinyl groups in a given MOF exist in an
identical microenvironment. This allows reliable struc-
ture−activity relationships to be established among this
series of catalysts. Previously, important advances have
been made by showing that the outcome of catalytic
reactions in MOF pockets can be altered by modifying
the pore structure.11,56,61 However, if each pore is not
identical, for example due to incomplete postsynthetic
modification reactions or framework disorder and
defects, then the observed catalytic activity will stem
from the ensemble of different pore environments.
(5) These results encouraged us to extend this study to the
aldol reaction between cyclopentanone and p-nitro-
benzaldehyde (Tables 2, S6). We selected four catalysts:
the two parent catalysts together with those featuring
modulator groups that had been found to exert
considerable influence over the aldol reaction with
acetone (dppdc and ndc). Syn and anti diastereoisomers
are possible for 2 (Figure S56), and both of them exist as
a pair of enantiomers. We found that [Zn4O(bdc-
Pro)1/2(bpdc)1/2(hmtt)4/3] catalyzes the formation of 2
with a syn:anti ratio of 1.74:1, which changes to a ratio of
2.14:1 when bpdc is replaced by the dppdc modulator
(Entries 1 and 3). Moreover, the ee of the syn
diastereomer changes from −2.8% to 3.4%, and the ee
E
DOI: 10.1021/jacs.7b07921
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(Entries 5 and 7). Further, if the prolinyl group is positioned on
the bpdc linker the R enantiomer is preferentially generated by
all catalysts. Similar observations were made for the aldol
reaction that produces 2: the preferred enantiomer of both
diastereomers can be reversed by switching the location of the
catalytic group in the MOF.
Table 2. Aldol Reaction of Cyclopentanone and p-
Nitrobenzaldehyde Catalyzed by MUF-77 Analogues with
a
Installed bdc-Pro or bpdc-Pro Linkers
Achieving stereochemical reversal in enantioselective trans-
formations is a demanding challenge in asymmetric catal-
ysis.65,66 It is valuable since both enantiomers of a reaction
product can be obtained when only one hand of the catalyst is
available, which is often the case for catalysts derived from
natural products. To our knowledge, there are only two
reported cases where the innate enantioselectivity of a catalyst
is reversed by its inclusion in a MOF.67,68 There are surprisingly
few examples of a reversal in enantioselectivity upon the
heterogenization of a homogeneous catalyst in other materi-
als.66,69 Our results go beyond these pioneering examples by
establishing that the position of a catalytic site in a MOF can
induce enantioselectivity reversal. Furthermore, the innate
stereochemical preference of the catalyst can be restored by
altering the catalysis microenvironment with wholly achiral
modulators. These observations underscore the utility of being
able to tune the catalysis microenvironment in multicomponent
MOFs.
b
c
#
catalyst linker set
dr syn:anti
ee (%) syn | anti
1
2
3
4
bdc-Pro/bpdc/hmtt
bdc/bpdc-Pro/hmtt
bdc-Pro/dppdc/hmtt
ndc/bpdc-Pro/hmtt
1.74:1
0.55:1
2.14:1
0.64:1
−2.8 | −35.8
37.2 | 16.8
3.4 | −10.2
40.3 | 24.3
a
Standard reaction conditions: 10 mol % catalyst (prolinyl units
relative to total aldehyde), 0.04 M 4-nitrobenzaldehyde in water-
saturated cyclopentanone at 20 °C. The product distribution was
b
c
determined by HPLC (Figure S57). Diastereomeric ratio. Enantio-
meric excess of each of the two diastereomers of 2. Detailed
information is available in the SI.
of the anti diastereomer is shifted from −35.8% to
−10.2%. Similarly, when the catalyst is appended to the
bpdc linker, the replacement of bdc by the ndc
modulator has a noticeable impact on the product
distribution (Entries 2 and 4).
SUMMARY
The foregoing results demonstrate that a transformation
catalyzed in a MOF pore can be influenced by a precise and
tunable spatial array of modulator groups positioned remote
from the site of catalysis. The modulator groups appear to act
in synergy, rather than in a linear, additive fashion, and the
properties of these catalysts is reminiscent of the emergent gas
adsorption properties displayed by the MUF-7 series.4 We
anticipate that the reaction mechanism parallels that generally
accepted for related small-molecule catalysts.64 While emphasis
of the current study was not to optimize the catalyst
performance per se, but to provide proof-of-concept, ongoing
investigations are aimed at developing detailed insights into the
key noncovalent interactions between the framework and the
reaction participants. Elucidating these interactions with
precision will extremely challenging since (i) the prolinyl
group undergoes dynamic motion, (ii) the magnitude of the
variation in ee values indicates that the difference in
diastereoisomeric transition states along the reaction pathway
is very small, (iii) these energy differences will be influenced by
a plethora of weak interactions of the catalytic intermediates
with the modulator groups, and (iv) solvent effects may be
important. Relevant investigations are underway in our
laboratory. This knowledge will guide the rational design of
modulator groups in next-generation catalysts that are able to
exert an even more pronounced influence over the catalytic
process to generate even higher reaction rates and enantiose-
lectivities.
■
This work establishes multicomponent frameworks as a
powerful new class of heterogeneous catalysts in which
modulator groups can be positioned in the pore to influence
the catalytic transformation. Owing to the single-site nature of
the catalytic pore, unambiguous correlations between frame-
work modifications and the catalytic activity can be elucidated.
The ability to design a three-dimensional microenvironment
around the site of catalysis distinguishes multicomponent
MOFs from conventional catalysts. Certain parallels exist
between these MOFs and enzymes in that catalysis takes
place at a specific site in a well-defined pocket and the reaction
rate and stereochemical outcome are determined by
interactions of the reaction components with a precisely
positioned array of functional groups.70,71 Taking inspiration
from protein engineering, where tuning the active site topology
and chemistry is a well-established method for enhancing
enzyme performance,72 systematic pore programming in
multicomponent metal−organic frameworks offers opportuni-
ties to develop synthetic catalysts beyond the proof of principle
examples presented here. For a given target reaction, we
anticipate that a suitably engineered MOF catalyst can be
optimized by, inter alia, modifying the catalytic unit, the nature
and disposition of the modulator groups, and the lattice
structure. These parameters define a vast combinatorial matrix
that is capable of identifying high-performance catalysts.
One unanticipated series of observations is that for the
p r o d u c t i o n o f a l d o l p r o d u c t
1 [ Z n 4 O ( b d c -
ASSOCIATED CONTENT
■
Pro)1/2(bpdc)1/2(trux)4/3] (trux = hmtt, hett) favor the S
enantiomer while homogeneous Me2bdc-Pro favors the R
enantiomer (Entries 1, 3, 16). Strikingly, this indicates that the
spatial environment around the prolinyl group in these
heterogeneous catalysts overrides its inherent enantioselectivity
to control the stereochemical pathway of the reaction. The
innate stereochemical preference40 of the bdc-Pro catalyst can
be restored by the installation of hbtt or hott modulator groups
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.7b07921.
Synthesis, characterization, X-ray crystallography details,
kinetic analyses and additional tables and figures (PDF)
Crystal data (CIF)
F
DOI: 10.1021/jacs.7b07921
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(21) Dutta, A.; Wong-Foy, A. G.; Matzger, A. J. Chem. Sci. 2014, 5,
3729−3734.
AUTHOR INFORMATION
■
Corresponding Author
*s.telfer@massey.ac.nz
(22) Lee, S. J.; Doussot, C.; Baux, A.; Liu, L.; Jameson, G. B.;
Richardson, C.; Pak, J.; Trousselet, F.; Coudert, F.-X.; Telfer, S. G.
Chem. Mater. 2016, 28, 368−375.
ORCID
(23) Qin, J.-S.; Du, D.-Y.; Li, M.; Lian, X.-Z.; Dong, L.-Z.; Bosch, M.;
Su, Z.-M.; Zhang, Q.; Li, S.-L.; Lan, Y.-Q.; Yuan, S.; Zhou, H.-C. J. Am.
Chem. Soc. 2016, 138, 5299−5307.
Lujia Liu: 0000-0001-9452-4371
Shane G. Telfer: 0000-0003-1596-6652
(24) Qin, J.-S.; Yuan, S.; Wang, Q.; Alsalme, A.; Zhou, H.-C. J. Mater.
Chem. A 2017, 5, 4280−4291.
Author Contributions
†L.L. and T.-Y.Z. contributed equally.
(25) Neel, A. J.; Hilton, M. J.; Sigman, M. S.; Toste, F. D. Nature
2017, 543, 637−646.
Notes
The authors declare no competing financial interest.
(26) Toste, F. D.; Sigman, M. S.; Miller, S. J. Acc. Chem. Res. 2017,
50, 609−615.
(27) Davis, H. J.; Phipps, R. J. Chem. Sci. 2017, 8, 864−877.
(28) Giuliano, M. W.; Miller, S. J. Top. Curr. Chem. 2016, 157−201.
(29) Raynal, M.; Ballester, P.; Vidal-Ferran, A.; van Leeuwen, P. W.
N. M. Chem. Soc. Rev. 2014, 43, 1734−1787.
ACKNOWLEDGMENTS
■
We are grateful to the RSNZ Marsden Fund for supporting this
research (contract 14-MAU-024). We acknowledge the
contribution of NeSI high-performance computing facilities
and the MMIC to this research. We thank David Lun for his
valuable technical assistance and Prof. Geoffrey Jameson, A/
Prof. Guy Jameson and A/Prof. Gareth Rowlands for insightful
discussions. L.L. is grateful for support from the China
Scholarship Council for an outstanding Ph.D. student award
and the RSNZ for the Hatherton Award.
(30) Raynal, M.; Ballester, P.; Vidal-Ferran, A.; van Leeuwen, P. W.
N. M. Chem. Soc. Rev. 2014, 43, 1660−1733.
(31) Meeuwissen, J.; Reek, J. N. H. Nat. Chem. 2010, 2, 615−621.
(32) Miller, S. J. Acc. Chem. Res. 2004, 37, 601−610.
(33) Yoon, M.; Srirambalaji, R.; Kim, K. Chem. Rev. 2012, 112,
1196−1231.
(34) Chughtai, A. H.; Ahmad, N.; Younus, H. A.; Laypkov, A.;
Verpoort, F. Chem. Soc. Rev. 2015, 44, 6804−6849.
(35) Xu, H.; Gao, J.; Jiang, D. Nat. Chem. 2015, 7, 905−912.
(36) Wang, C. A.; Zhang, Z. K.; Yue, T.; Sun, Y. L.; Wang, L.; Wang,
W. D.; Zhang, Y.; Liu, C.; Wang, W. Chem. - Eur. J. 2012, 18, 6718−
6723.
REFERENCES
■
(1) Furukawa, H.; Cordova, K. E.; O’Keeffe, M.; Yaghi, O. M. Science
2013, 341, 974.
(2) Eddaoudi, M.; Kim, J.; Rosi, N.; Vodak, D.; Wachter, J.; O’Keeffe,
M.; Yaghi, O. M. Science 2002, 295, 469−472.
(37) Rostamnia, S.; Doustkhah, E. RSC Adv. 2014, 4, 28238−28248.
(38) Zhang, J.; Han, X.; Wu, X.; Liu, Y.; Cui, Y. J. Am. Chem. Soc.
2017, 139, 8277−8285.
(3) Sim, J.; Yim, H.; Ko, N.; Choi, S. B.; Oh, Y.; Park, H. J.; Park, S.;
Kim, J. Dalton Trans. 2014, 43, 18017−18024.
(39) Lun, D. J.; Waterhouse, G. I. N.; Telfer, S. G. J. Am. Chem. Soc.
2011, 133, 5806−5809.
(4) Liu, L.; Konstas, K.; Hill, M. R.; Telfer, S. G. J. Am. Chem. Soc.
2013, 135, 17731−17734.
(40) Bonnefoy, J.; Legrand, A.; Quadrelli, E. A.; Canivet, J.;
Farrusseng, D. J. Am. Chem. Soc. 2015, 137, 9409−9416.
(41) Roberts, J. M.; Fini, B. M.; Sarjeant, A. A.; Farha, O. K.; Hupp, J.
T.; Scheidt, K. A. J. Am. Chem. Soc. 2012, 134, 3334−3337.
(42) Jeong, K. S.; Go, Y. B.; Shin, S. M.; Lee, S. J.; Kim, J.; Yaghi, O.
M.; Jeong, N. Chem. Sci. 2011, 2, 877.
(43) Cho, S.-H.; Ma, B.; Nguyen, S. T.; Hupp, J. T.; Albrecht-
Schmitt, T. E. Chem. Commun. 2006, 2563−2565.
(44) Thacker, N. C.; Lin, Z.; Zhang, T.; Gilhula, J. C.; Abney, C. W.;
Lin, W. J. Am. Chem. Soc. 2016, 138, 3501−3509.
(45) Deshpande, R. K.; Minnaar, J. L.; Telfer, S. G. Angew. Chem., Int.
Ed. 2010, 47, 4598−4602.
(46) Gupta, A. S.; Deshpande, R. K.; Liu, L.; Waterhouse, G. I. N.;
Telfer, S. G. CrystEngComm 2012, 14, 5701−5704.
(47) Gould, S. L.; Tranchemontagne, D.; Yaghi, O. M.; Garcia-
Garibay, M. A. J. Am. Chem. Soc. 2008, 130, 3246−3247.
(48) Jiang, X.; Duan, H.-B.; Khan, S. I.; Garcia-Garibay, M. A. ACS
Cent. Sci. 2016, 2, 608−613.
(49) Shearer, G. C.; Chavan, S.; Bordiga, S.; Svelle, S.; Olsbye, U.;
Lillerud, K. P. Chem. Mater. 2016, 28, 3749−3761.
(50) Ghosh, P.; Colon, Y. J.; Snurr, R. Q. Chem. Commun. 2014, 50,
11329−11331.
(5) Osborn Popp, T. M.; Yaghi, O. M. Acc. Chem. Res. 2017, 50,
532−534.
(6) Xia, Q.; Li, Z.; Tan, C.; Liu, Y.; Gong, W.; Cui, Y. J. Am. Chem.
Soc. 2017, 139, 8259−8266.
(7) Deng, H.; Doonan, C. J.; Furukawa, H.; Ferreira, R. B.; Towne, J.;
Knobler, C. B.; Wang, B.; Yaghi, O. M. Science 2010, 327, 846−850.
(8) Park, T.-H.; Koh, K.; Wong-Foy, A. G.; Matzger, A. J. Cryst.
Growth Des. 2011, 11, 2059−2063.
(9) Burrows, A. D. CrystEngComm 2011, 13, 3623−3642.
(10) Kong, X.; Deng, H.; Yan, F.; Kim, J.; Swisher, J. A.; Smit, B.;
Yaghi, O. M.; Reimer, J. A. Science 2013, 341, 882−885.
(11) Fracaroli, A. M.; Siman, P.; Nagib, D. A.; Suzuki, M.; Furukawa,
H.; Toste, F. D.; Yaghi, O. M. J. Am. Chem. Soc. 2016, 138, 8352−
8355.
(12) Lin, Q.; Bu, X.; Kong, A.; Mao, C.; Zhao, X.; Bu, F.; Feng, P. J.
Am. Chem. Soc. 2015, 137, 2235−2238.
(13) Furukawa, H.; Ko, N.; Go, Y. B.; Aratani, N.; Choi, S. B.; Choi,
E.; Yazaydin, A. O.; Snurr, R. Q.; O’Keeffe, M.; Kim, J.; Yaghi, O. M.
Science 2010, 329, 424−428.
(14) Helten, S.; Sahoo, B.; Bon, V.; Senkovska, I.; Kaskel, S.; Glorius,
F. CrystEngComm 2015, 17, 307−312.
(15) Grunker, R.; Bon, V.; Muller, P.; Stoeck, U.; Krause, S.; Mueller,
U.; Senkovska, I.; Kaskel, S. Chem. Commun. 2014, 50, 3450−3452.
(16) Klein, N.; Senkovska, I.; Gedrich, K.; Stoeck, U.; Henschel, A.;
Mueller, U.; Kaskel, S. Angew. Chem., Int. Ed. 2009, 48, 9954−9957.
(17) Yuan, S.; Lu, W.; Chen, Y. P.; Zhang, Q.; Liu, T. F.; Feng, D.;
Wang, X.; Qin, J.; Zhou, H. C. J. Am. Chem. Soc. 2015, 137, 3177−
3180.
(51) Liang, W.; Coghlan, C. J.; Ragon, F.; Rubio-Martinez, M.;
D’Alessandro, D. M.; Babarao, R. Dalton Trans. 2016, 45, 4496−4500.
(52) Thornton, A. W.; Babarao, R.; Jain, A.; Trousselet, F.; Coudert,
F. X. Dalton Trans. 2016, 45, 4352−4359.
(53) Bueken, B.; Reinsch, H.; Reimer, N.; Stassen, I.; Vermoortele,
F.; Ameloot, R.; Stock, N.; Kirschhock, C. E. A.; De Vos, D. Chem.
Commun. 2014, 50, 10055−10058.
(18) Yuan, S.; Qin, J.-S.; Zou, L.; Chen, Y.-P.; Wang, X.; Zhang, Q.;
Zhou, H.-C. J. Am. Chem. Soc. 2016, 138, 6636−6642.
(19) Lee, S. J.; Doussot, C.; Telfer, S. G. Cryst. Growth Des. 2017, 17,
3185−3191.
(54) Zotova, N.; Franzke, A.; Armstrong, A.; Blackmond, D. G. J. Am.
Chem. Soc. 2007, 129, 15100−15101.
(55) Pelletier, J. D. A.; Basset, J.-M. Acc. Chem. Res. 2016, 49, 664−
677.
(20) Liu, L.; Telfer, S. G. J. Am. Chem. Soc. 2015, 137, 3901−3909.
G
DOI: 10.1021/jacs.7b07921
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(56) Cohen, S. M.; Zhang, Z.; Boissonnault, J. A. Inorg. Chem. 2016,
55, 7281−7290.
̆
(57) Metzger, E. D.; Brozek, C. K.; Comito, R. J.; Dinca, M. ACS
Cent. Sci. 2016, 2, 148−153.
(58) Li, Z.; Schweitzer, N. M.; League, A. B.; Bernales, V.; Peters, A.
W.; Getsoian, A. B.; Wang, T. C.; Miller, J. T.; Vjunov, A.; Fulton, J. L.;
Lercher, J. A.; Cramer, C. J.; Gagliardi, L.; Hupp, J. T.; Farha, O. K. J.
Am. Chem. Soc. 2016, 138, 1977−1982.
(59) Manna, K.; Ji, P.; Lin, Z.; Greene, F. X.; Urban, A.; Thacker, N.
C.; Lin, W. Nat. Commun. 2016, 7, 12610.
(60) Rogge, S. M. J.; Bavykina, A.; Hajek, J.; Garcia, H.; Olivos-
Suarez, A. I.; Sepulveda-Escribano, A.; Vimont, A.; Clet, G.; Bazin, P.;
Kapteijn, F.; Daturi, M.; Ramos-Fernandez, E. V.; Llabres i Xamena, F.
X.; Van Speybroeck, V.; Gascon, J. Chem. Soc. Rev. 2017, 46, 3134−
3184.
(61) Yuan, S.; Chen, Y.-P.; Qin, J.-S.; Lu, W.; Zou, L.; Zhang, Q.;
Wang, X.; Sun, X.; Zhou, H.-C. J. Am. Chem. Soc. 2016, 138, 8912−
8919.
(62) Xiao, D. J.; Oktawiec, J.; Milner, P. J.; Long, J. R. J. Am. Chem.
Soc. 2016, 138, 14371−14379.
́ ́
(63) Deria, P.; Gomez-Gualdron, D. A.; Hod, I.; Snurr, R. Q.; Hupp,
J. T.; Farha, O. K. J. Am. Chem. Soc. 2016, 138, 14449−14457.
̈
̧elebi-Olcu̧ m, N.;
̈
(64) Cheong, P. H.-Y.; Legault, C. Y.; Um, J. M.; C
Houk, K. N. Chem. Rev. 2011, 111, 5042−5137.
(65) Bartok
́
, M. Chem. Rev. 2010, 110, 1663−1705.
(66) Escorihuela, J.; Burguete, M. I.; Luis, S. V. Chem. Soc. Rev. 2013,
42, 5595−5617.
(67) Zheng, M.; Liu, Y.; Wang, C.; Liu, S.; Lin, W. Chem. Sci. 2012, 3,
2623.
(68) Nießing, S.; Czekelius, C.; Janiak, C. Catal. Commun. 2017, 95,
12−15.
(69) García, R. A.; Van Grieken, R.; Iglesias, J.; Sherrington, D. C.;
Gibson, C. L. Chirality 2010, 22, 675−683.
(70) Lillerud, K. P.; Olsbye, U.; Tilset, M. Top. Catal. 2010, 53, 859−
868.
(71) Nath, I.; Chakraborty, J.; Verpoort, F. Chem. Soc. Rev. 2016, 45,
4127−4170.
(72) Toscano, M. D.; Woycechowsky, K. J.; Hilvert, D. Angew. Chem.,
Int. Ed. 2007, 46, 3212−3236.
H
DOI: 10.1021/jacs.7b07921
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX