Substrate-induced covalent assembly of a chemzyme and crystallographic characterization of a chemzyme-substrate complex

Article abstract of DOI:10.1021/ja1070224

A substrate induced covalent assembly of a highly organized chemzyme known to be effective in both catalytic asymmetric aziridination and aza Diels-Alder reactions is described and the information gained from which led to an efficient one-pot aziridination protocol. The crystal structures of two chemzyme-iminium complexes were elucidated by X-ray diffraction analysis that provides critical insights into the binding of the substrates with the chemzyme.

Full text of DOI:10.1021/ja1070224

Published on Web 09/23/2010
Substrate-Induced Covalent Assembly of a Chemzyme and
Crystallographic Characterization of a Chemzyme-Substrate
Complex
Gang Hu, Anil K. Gupta, Rui H. Huang, Munmun Mukherjee, and William D. Wulff*
Department of Chemistry, Michigan State UniVersity, East Lansing, Michigan 48824
Received August 5, 2010; E-mail: wulff@chemistry.msu.edu
Abstract: A substrate induced covalent assembly of a highly organized chemzyme known to be effective
in both catalytic asymmetric aziridination and aza Diels-Alder reactions is described and the information
gained from which led to an efficient one-pot aziridination protocol. The crystal structures of two
chemzyme-iminium complexes were elucidated by X-ray diffraction analysis that provides critical insights
into the binding of the substrates with the chemzyme.
Scheme 1
1. Introduction
In the long course of biological evolution, organisms have
developed biosynthetic pathways to a diverse array of organic
compounds by harnessing enzymes that can catalyze a great
number of organic reactions with high selectivities and catalytic
efficiencies under mild conditions. In modern organic synthesis,
nonenzymatic chiral catalysts are playing an ever increasing role,
but their development and rational design is a daunting challenge
since the possibilities and strategies for function in nonproteo-
genic chiral systems are just beginning to be defined. In most
cases, both proteinaceous and nonproteinaceous chiral catalysts,
or chemzymes,1,2 are preformed or generated in situ in advance.
We describe here a novel system involving a chemzyme whose
covalent assembly is induced by its substrate. In addition, a
detailed mapping of the interactions between the chemzyme and
its substrate is elucidated by single-crystal X-ray diffraction
analysis.
We have previously developed a catalyst that gives high
asymmetric inductions for both the aza Diels-Alder3 and
aziridination reactions4 (Scheme 1). This catalyst is prepared
from the vaulted biaryl ligand VANOL 6 or VAPOL 7 by
heating with 3 equiv of triphenylborate at 80 °C for 1 h and
then removing all volatiles under high vacuum. The resulting
catalyst derived from VAPOL is a mixture of the mesoborate 8
under the reaction conditions to the boroxinate 10 (B3) and we
have provided NMR evidence that 10 is the actual catalyst in
these reactions.6
1
(B1) and the pyroborate 9 (B2) as suggested by the H NMR
and 11B NMR spectra as well as by mass spectral analysis.5
The mixture is significantly in favor of the pyroborate species
9 with ratios varying from 4:1 to 20:1 depending on the exact
protocol for catalyst formation.5 More recently, we have reported
that the mixture of mesoborate 8 and pyroborate 9 is converted
2. Substrate-Induced Assembly
The fact that the mixture of mesoborate 8 and pyroborate 9
is converted to the boroxinate 10 under the reaction conditions
in both the aza Diels-Alder and aziridination reactions suggests
that perhaps this is caused by the imine since it is a substrate
common to both reactions. Thus, a series of NMR experiments
were designed to address this issue and to determine if the
preformation of the pyroborate 9 was even necessary. The
identification of these species is facilitated by the fact that
(1) Bjerre, J.; Rousseau, C.; Marinescu, L.; Bols, M. Appl. Microbiol.
Biotechnol. 2008, 81, 1.
(2) Corey, E. J.; Chen, C. P.; Reichard, G. A. Tetrahedron Lett. 1989,
30, 5547.
(3) Newman, C. A.; Antilla, J. C.; Chen, P.; Predeus, A.; Fielding, L.;
Wulff, W. D. J. Am. Chem. Soc. 2007, 129, 7216.
(4) For a review, see: Zhang, Y.; Lu, Z.; Wulff, W. D. Synlett, 2009,
2715.
(5) Zhang, Y.; Desai, A.; Lu, Z.; Hu, G.; Ding, Z.; Wulff, W. D.
Chem.sEur. J. 2008, 14, 3785.
(6) Hu, G.; Huang, L.; Huang, R. H.; Wulff, W. D. J. Am. Chem. Soc.
2009, 131, 15615.
9
10.1021/ja1070224  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 14669–14675 14669

A R T I C L E S
Hu et al.
Scheme 2
the bay protons (Hb) of VAPOL and its derivatives are quite
distinct (Scheme 2), and for VAPOL, this doublet is at δ )
9.77 ppm (CDCl3). Specifically, the question to be asked is
whether an imine can cause the formation of the boroxinate 10
directly from VAPOL and commercially available B(OPh)3 that
is always partially hydrolyzed at room temperature (Scheme
2). As a control, (S)-VAPOL and 3 equiv of B(OPh)3 were
3), an observation consistent with the facile turnover observed
for the reaction.5 The boroxinate 10 can be easily distinguished
from the other boron species by 11B NMR. Most three coordinate
aryl borate compounds have very broad absorptions at 16-18
ppm in the 11B NMR, and thus, it is difficult to distinguish one
borate species from another. However, the increased spherical
symmetry of a four coordinate boron leads to sharp lines in the
spectrum and the negative charge of the borate causes an upfield
shift. The boroxinate 10b has a sharp absorption at 5.51 ppm
(Figure 2, entry 2).
The traditional method for catalyst formation involves heating
(S)-VAPOL with 3 equiv of B(OPh)3 at 80 °C to generate the
pyroborate 9 now known to be the precatalyst. When this
procedure is monitored by NMR, the pyroborate 9 is formed in
53% yield along with a 6% yield of the mesoborate 8 which
has a doublet at δ ) 9.55 ppm and 20% of unreacted VAPOL
(Figure 1, entry 4). Treatment of this mixture with 2 equiv of
the imine 1b at room temperature within a few minutes results
in the conversion of the colorless solution of 8 and 9 to a red
solution of boroxinate 10b in 81% yield with most of the free
VAPOL ligand consumed as indicated by its spectra (Figures 1
and 2, entries 5). When this solution is treated with 2.4 equiv
1
dissolved in CDCl3, and after 10 min, the H NMR revealed
very little reaction (Figure 1, entry 1). A small amount (3%) of
the pyroborate 9 is observed as the doublet at δ ) 9.22 ppm
and this does not change after 24 h. However, if a 1:3 mixture
of VAPOL and B(OPh)3 is treated with 2 equiv of imine 1b at
room temperature, a red solution is immediately formed, and
1
in the time that it takes to perform H NMR and 11B NMR
analysis, the boroxinate complex 10b is formed in 76% yield
as indicated by the doublet at δ ) 10.31 ppm (Figure 1, entry
2). The remainder is the free VAPOL ligand and a small amount
of benzaldehyde from imine hydrolysis and the presence of the
mesoborate 8 and pyroborate 9 can not be detected (Figure 1,
entry 2). In contrast, the addition of the aziridine 3b to a mixture
of (S)-VAPOL and B(OPh)3 does not induce the formation of
a boroxinate species to any significant extent (Figure 1, entry
Figure 1. 1H NMR spectra of the reaction mixture (Scheme 2) in CDCl3 with Ph3CH as internal standard. Entry 1: (S)-VAPOL plus 3 equiv B(OPh)3 at
25 °C for 10 min. Entry 2: (S)-VAPOL plus 3 equiv B(OPh)3 and 2 equiv imine 1b at 25 °C for 10 min. Entry 3: (S)-VAPOL plus 3 equiv B(OPh)3 and
2 equiv aziridine 3b at 25 °C for 10 min. Entry 4: (S)-VAPOL plus 3 equiv B(OPh)3 were heated at 80 °C in toluene for 1 h followed by removal of volatiles
under vacuum for 0.5 h. Entry 5: 2 equiv of imine 1b added to the solution in entry 4 for 10 min at 25 °C. Entry 6: 2.4 equiv of ethyl diazoacetate added
to the NMR tube containing the solution in entry 5 and kept for 9 days at 25 °C. Entry 7: (S)-VAPOL plus 1 equiv triphenoxyboroxine 13 at 25 °C for 10
min. Entry 8: 2 equiv imine 1b added to solution in entry 7 for 10 min at 25 °C.
9
14670 J. AM. CHEM. SOC. VOL. 132, NO. 41, 2010

Chemzyme-Substrate Complex
A R T I C L E S
1
Figure 2. 11B NMR spectra corresponding to the H NMR spectra in Figure 1.
Scheme 3
in boroxine 13 with the VAPOL ligand. To test the efficacy of
triphenoxyboroxine 13 as a source for catalyst formation, an
independent sample of triphenoxyboroxine 13 was prepared9
and treated with an equimolar amount of VAPOL. The 1H and
11B NMR spectra do not reveal the presence of a boroxinate
species but rather the presence of largely unreacted VAPOL
(Figures 1 and 2, entries 7). However, addition of 2 equiv of
imine 1b to this mixture results in the immediate formation of
a red color and the appearance of boroxinate 10b in 55% yield
(Figures 1 and 2, entries 8). This clearly demonstrates that
triphenoxyboroxine 13 can serve as an alternative source of the
boroxinate catalyst without the need to ensure that some water
or water surrogate is present.
The results from the experiments outlined in Figures 1 and 2
indicate that the boroxinate catalyst 10 can be assembled by
the imine 1 from VAPOL and a number of borate sources
including B(OPh)3, the mesoborate 8, the pyroborate 9, and the
boroxine 13. The generation of the catalyst and the subsequent
catalytic cycle are proposed in Scheme 3. In the initiation step,
the presence of partially hydrolyzed borate species or of water
is required (except for boroxine 13) to generate the three
B-O-B linkages in the boroxinate species 10. Once the
boroxinate catalyst 10 is assembled, then presumably the next
step involves the reaction with ethyl diazoacetate 2 to give
boroxinate-aziridine complex 11. Turnover then would require
the loss of aziridine and then incorporation of another molecule
of the imine 1 to regenerate the boroxinate-imine complex 10.
While we have not been able to obtain any direct evidence for
the presence of the protonated boroxinate species 12, it is not
clear if the overall conversion of the boroxinate-aziridine
complex 11 to the boroxinate-imine complex 10 involves loss
of aziridine to give the Brønsted acid 12 which then protonates
the imine to give 10, or if there is an associative reaction
between 11 and the imine 1 which directly leads to the aziridine
3 and the boroxinate-imine complex 10. The fact that this
reaction turns over is consistent with the fact that aziridine is
of ethyl diazoacetate 2, a complex mixture of compounds
containing VAPOL is produced and most of the boroxinate has
been consumed at the end of the reaction (Figure 1, entry 6).
The two large doublets (δ ) 9.83 and 9.58 ppm) have been
identified as the monoalkylation product of VAPOL 14 by ethyl
diazoacetate (Figure 1).5
The formation of the three B-O-B linkages in the boroxine
ring in the boroxinate catalyst 10 requires 3 equiv of water and
the source of this most certainly results from partial hydrolysis
of B(OPh)3.8 All commercial samples of B(OPh)3 that we have
encountered are effective in the aziridination reaction and all
contain varying amounts of partially hydrolyzed boron com-
pounds with triphenoxyboroxine 13 (Scheme 3) identified as
the major one based on NMR analysis of one particular sample.
No further hydrolysis is required to give the boroxinate-imine
complex 10 except only an exchange reaction of a phenol unit
(7) (a) Zhang, Y.; Lu, Z.; Desai, A.; Wulff, W. D. Org. Lett. 2008, 10,
5429. (b) Mukherjee, M.; Gupta, A. K.; Lu, Z.; Zhang, Y.; Wulff,
W. D. J. Org. Chem. 2010, 75, 5643.
(8) As a control, purified B(OPh)3 gives after 24 h only a trace of
boroxinate species 10a utilizing 1 equiv of imine 1a (see supporting
information, section G).
(9) Cole, T. E.; Quintanilla, R.; Rodewald, S. Synth. React. Inorg. Met.-
Org. Chem. 1990, 20, 55.
9
J. AM. CHEM. SOC. VOL. 132, NO. 41, 2010 14671

A R T I C L E S
Hu et al.
Figure 3. Illustration of boroxinate catalyst formation (red solution) in an open vial directly from VAPOL, B(OPh)3 and imine and subsequent aziridination
with EDA.
Scheme 4
significantly slower in inducing boroxinate formation than the
imine (Figure 1, entries 2 and 3).
The fact that the substrate-induced covalent assembly of the
boroxinate 10 requires 3 equiv of water suggests that the reaction
may not be sensitive to normal atmospheric conditions. Indeed,
catalyst assembly can be performed at room temperature in a
vial opened to air and the subsequent catalytic asymmetric
aziridination as well (Figure 3). In the first Frame, a 20 mL
vial is loaded with 0.1 mmol of (S)-VAPOL, 0.3 mmol B(OPh)3,
and 1.0 mmol of imine 1a. When this mixture of three white
solids is dissolved in 2 mL of toluene (used as commercially
supplied), the resulting solution immediately takes on a red color
3. Crystallographic Studies
(Frame 2). NMR analysis indicated that boroxinate species was
generated in approximately 86% yield, similar to Figure 1, entry
The chemzyme-substrate complex 10a derived from the
2. After a few minutes, 1.2 mmol of ethyl diazoacetate is added
and vigorous bubbling immediately ensues as indicated in Frame
3, and after 30 min, the red color has dissipated (Frame 4). After
60 min, the reaction is completed and the product is purified
by column chromatography on silica gel to give aziridine 3a in
82% yield and 92% ee which is essentially the same (82% yield,
94% ee) as previously reported5 for this reaction when initiated
with the precatalyst and carried out under an argon atmosphere
and rigorously dried glassware and solvents (Scheme 1).
However, there are limits to the tolerance of water to this
aziridination reaction. Addition of 1-3 equiv of water during
catalyst formation with commercial B(OPh)3 does not hinder
the reaction of imine 1a, but the addition of 5-10 equiv of
water relative to the ligand results in extremely low conversion.
Therefore, for reactions with very low catalyst loading (0.1-0.5
mol %), it is best to perform the reaction under an inert
atmosphere with dried glassware and solvents under which
conditions up to 500 turnovers have been observed for this
reaction.5 Nevertheless, this simple procedure for one-pot
catalyst generation and aziridination at room temperature under
an argon atmosphere is effective with only 5 mol % catalyst as
illustrated for imines ranging from the electron-rich 1e to the
electron-poor 1f and to the aliphatic imine 1g (Scheme 4).7 It
should be pointed out that the asymmetric inductions in Scheme
4 are very close to those reported for catalysts prepared by first
heating VAPOL and B(OPh)3 to generate the precatalyst except
for the para-methoxy imine 1e were the % ee drops from 98%
benzhydryl (Bh) imine 1a could not be induced to form crystals
suitable for an X-ray diffraction study. In a study directed to
mapping the active site of this chemzyme, it was found that
the tetra-methyldianisylmethyl (MEDAM) imines of the type
1b as well as the tetra-t-butyldianisylmethyl (BUDAM) imines
of the type 1c proved to be superior substrates for the
aziridination reaction giving higher asymmetric inductions and
much greater rates of reaction (11-16 times faster than 1a).7
We were delighted to find that both of the MEDAM imines 1b
and 1d gave single crystals (both yellow) of the boroxinate-imine
complexes 10b and 10d (Scheme 5). All of the data for the
solid-state X-ray structures of 10b and 10d can be found in the
Supporting Information. Structural parameters for 10b and 10d
are available free of charge from the Cambridge Crystallographic
Data Centre under reference number CCDC 784078 and 784079.
The X-ray structure of 10b not only confirms the boroxinate
ion pair motif, but for the first time clarifies how the imine
substrate docks with the catalyst. The structural analysis reveals
the presence of an ion pair consisting of a protonated iminium
ion and the boroxinate anion derived from a boroxine ring in
which the tetra-coordinate boron is spiro-fused to the VAPOL
ligand. This is a remarkable structure where the catalyst and
substrate seem to fit together like a hand and a glove. This close
fit makes possible the large number of noncovalent interactions
observed between the boroxinate anion and the protonated
iminium and these are summarized in Figure 4A.
Foremost among these is a hydrogen bond from the proto-
nated iminium to the oxygen in the boroxinate ring that is
attached to the four coordinate boron (O2) (Scheme 5). The
H-O distance of 2.02 Å (Figure 4A, d1) suggests a strong to
moderate hydrogen bond10,11 and the chemical shift of this
ee to 84% ee and the reason for this is not understood.7b
A
variety of solvents are compatible with this protocol and give
comparable results (Table S3 in Supporting Information).
9
14672 J. AM. CHEM. SOC. VOL. 132, NO. 41, 2010

Chemzyme-Substrate Complex
A R T I C L E S
Scheme 5
proton at 13.7 ppm is consistent with related species in which
proton is on nitrogen not oxygen.6
the free imine on 1H NMR (Figure 5). Thus, the CH-π
interactions of the methyl and t-butyl groups of the MEDAM
and BUDAM imines may contribute, for reasons of increased
binding to the catalyst in the transition state, to the increased
rates and asymmetric inductions of these imines relative to the
benzhydryl imines.7 These interactions may also be related to
the ease with which complex 10b could be crystallized but not
10a.
Another important noncovalent interaction between the
catalyst and the substrate is a π-π stacking interaction between
the protonated benzylidiene iminium moiety and the phenan-
threne rings of the VAPOL ligand. This can be seen in the space-
filling model of the catalyst-substrate complex 10b shown in
Figure 4B where the boroxinate anion is in green. The π-π
stacking interaction occurs with an angle between the two planes
of 4.4° from parallel and with a distance of 3.44 Å that seems
to be in the optimal range for such an interaction.15 As a
consequence of this π-π stacking interaction, it is clear from
Figure 4B that attack by a nucleophile from below would be
blocked. Attack from the top face would correspond to Si-face
addition which is what is observed with catalysts derived from
(S)-VAPOL and (S)-VANOL ligands for both the aziridination
and aza Diels-Alder reactions.3
The crystal structure of 10b perhaps provides some clues as
to the increased reactivity and enantioselectivity of the BUDAM
and MEDAM imines relative to their benzhydryl counterparts.7
There are several CH-π interactions12,13 and perhaps the most
interesting is from an ortho-hydrogen on the MEDAM phenyl
ring to one of the phenyl rings on the back end of the VAPOL
ligand (Figure 4, d3 4.14 Å).14 There is also a CH/π interaction
between one of the MEDAM methyl groups and this same
phenyl ring (Figure 4, d2 4.42 Å). These interactions are easiest
to see in the crystal structure of 10b shown in Figure 4A. There
are other CH-π interactions including one from a MEDAM
methyl to the end ring of a phenanthrene unit in the VAPOL
ligand (Figure 4, d4 3.83 Å). Consistent with these CH-π
interactions placing these methyl groups in the shielding region
of various arene rings is the fact that the methyl groups of the
MEDAM rings are desymmetrized in the 1H NMR spectrum of
complex 10b and are observed as two singlets which are
shielded relative to the free imine 1b by 0.42 and 0.54 ppm
(Figure 5). A similar upfield shift also occurs for the methyl
groups in t-butyl substituents of BUDAM imine-catalyst
complex.6 We have evidence that this shielding effect is not
due to the protonation of imine because the same methyl groups
of the imine-HCl salts do not shift significantly from those of
The crystal structure of the catalyst-substrate complex 10d
with the p-dimethylaminophenyl imine 1d has many of the same
Figure 4. (A) Solid state structure of the chemzyme-substrate complex 10b visualized by the Mercury program (C, gray; O, red; N, blue; B, purple; H,
yellow). Calculated hydrogen atoms and solvent molecules were omitted for clarity. Some secondary interactions were proposed and highlighted: d1 ) 2.02
Å (H-bonding), d2 ) 4.42 Å (CH-π), d3 ) 4.14 Å (CH-π), d4 ) 3.83 Å (CH-π), d5 ) 3.44 Å (π-π stack), θ ) 4.4° (π-π stack). The iminium phenyl
ring is displaced from the middle ring and end ring of phenanthrene by 1.44 and 1.33 Å, respectively (data not shown in figure). (B) Space-filling model of
the complex 10b rotated ∼180° vertically relative to panel A to bring the π-π stacking interaction into view. The boroxinate anion is given in green and
the iminium cation is in traditional colors.
9
J. AM. CHEM. SOC. VOL. 132, NO. 41, 2010 14673

A R T I C L E S
Hu et al.
Figure 5. Chemical shifts of methyl and methoxy groups in various imines and iminiums in CDCl3 with TMS as internal standard.
close contacts present in 10b that can be seen in Figure 4A
including a hydrogen bond between the protonated iminium and
O-2 of the boroxinate which is 2.05 Å (Figure 6, d1). Two of
the CH-π interactions in 10d are similar to those in 10b where
the same methyl groups are interacting with the same aromatic
ring (Figure 6, d2 and d4). A third CH-π interaction (Figure 6,
d3) is present in 10d that may also be present in 10b but is a
little outside of the normal range for CH-π interactions in 10b.
The main difference between the structures 10b and 10d is that
for 10d the interaction of the aryl substituent of the iminium
with the phenanthrene ring of the ligand forms an angle between
the planes of these rings of 47°. This disfavoring of the π-π
stacking interaction in 10d is consistent with the increased
electron density that the dimethylamino groups imparts to the
iminium and may be related to the fact that imine 1d is not a
substrate of the reaction.16
inductions in the aziridination reaction.5,7 The π-π stacking
of the iminium ion involves the conjugated phenyl of the
iminium stacked over the central ring of the phenanthrene of
the VAPOL ligand and apparently the end ring of the phenan-
threne is not needed. Removing the end ring from VAPOL 7
would of course produce the VANOL ligand 6 (Scheme 1)
which would be expected to be capable of similar π-π stacking
observed for VAPOL in the complex 10b bearing in mind that
optimal π-π stacking occurs when the rings are slipped.13
Finally, the chemzyme-substrate complex 10b is capable of
functioning as a catalyst for the aziridination reaction. Crystals
of 10b (10 mol %) catalyze the reaction of the MEDAM imine
1b with ethyl diazoacetate and give the aziridine 3b in 94%
yield and 97% ee in methylene chloride at room temperature
(Scheme 6) which is essentially identical to that previously
reported with methylene chloride as reaction solvent7 and in
the present work with toluene as solvent (Scheme 4). Thus,
while this does not prove that the catalyst-substrate complex
10b is part of the catalytic cycle, it does reveal that 10b is not
an artifact of crystallization and can at least return to the catalytic
cycle. While the crystal structure of 10b gives us a picture of
how the imine substrates bind to the VAPOL boroxinate catalyst
in the solid state, and perhaps in solution as well, the next step
presumably involves the interaction of 10b with ethyl diazoac-
etate and it is not clear what the structure of the chemzyme-
substrate complex 10b would tells us about the structure of the
transition state for this process.
The structure of the chemzyme-substrate complex 10b
provides some insight as to why the VANOL and VAPOL
ligands (Scheme 1) both give nearly identical asymmetric
(10) Taylor, M. S.; Jacobsen, E. N. Angew. Chem., Int. Ed. 2006, 45, 1520.
(11) Jeffrey, G. A.; An Introduction to Hydrogen Bonding; Oxford
University Press: New York, 1997.
(12) Nishio, M. Tetrahedron 2005, 61, 6923.
(13) Plevin, M. J.; Bryce, D. L.; Boisbouvier, J. Nat. Chem. 2010, 2, 466.
(14) Distances for CH/π interactions are defined as described in ref 13.
(15) Meyer, E. A.; Castellano, R. K.; Diederich, F. Angew. Chem., Int.
Ed. Engl. 2003, 42, 1210.
(16) While the reaction of p-methoxyphenyl imine 1e is slow (Scheme 4),
the p-dimethylaminophenyl imine 1d will not react with ethyl
diazoacetate in 24 h.
9
14674 J. AM. CHEM. SOC. VOL. 132, NO. 41, 2010

Chemzyme-Substrate Complex
A R T I C L E S
Figure 6. Crystal structure of B3-imine complex 10d visualized by Mercury program. (A) Capped stick (anion) and ellipsoid (cation) display of the crystal
structure. All calculated hydrogens and solvent molecules are omitted for clarity. Selected parameters are shown: d1 ) 2.05 Å (H-bonding), d2 ) 3.87 Å
(CH-π), d3 ) 3.65 Å (CH-π), d4 ) 4.14 Å (CH-π), distances d2, d3 and d4 are from a methyl group to the center of aromatic ring. (B) Space-filling
drawing of the structure with dcen ) 4.52 Å (distance from the center of phenyl ring of iminium cation to the end ring of phenanthrene), θ ) 48.2° (angle
between iminium cation plane and phenanthrence plane). The boroxinate anion is colored in green to enhance contrast.
“mini-enzyme”.20,21 In the present case, the polyborate core of
Scheme 6
the chemzyme is assembled by an imine whose nitrogen
substituent has been evolved to provide the greatest rates and
highest enantioselectivities in an aziridination reaction.7 Crystal
structures of the chemzyme-substrate complexes provide a
particularly appealing vista of the variety and diversity of
noncovalent interactions that are involved in the binding of the
4. Conclusions
substrate to the chemzyme. These observations should hopefully
serve to initiate new applications of polyborate chemistry and
provide new insights into molecular recognition between a
chemzyme and its substrate that would be important in the
design of function in chemzymes.
Polyborate anions have been well documented in the literature
and show richness in the diversity of their intrinsic structures.17-19
However, the mechanisms by which these structures are formed
have not been investigated in detail and the applications of
polyborate anions in asymmetric catalysis have, to the best of
our knowledge, not been previously reported. The present work
provides structural details of a new class of a highly organized
chemzyme scaffold consisting of a polyborate anionic core
assembled around a chiral vaulted bis-phenol ligand. The
chemzyme assembly is induced by the imine substrate of the
chemzyme. It has been suggested that substrate templating has
played a role in the covalent formation of prebiotic catalysts
and an example has been reported for a dipeptide-zinc
Acknowledgment. This work was supported by a grant from
the National Science Foundation (CHE-0750319). Professor Daniel
Jones is acknowledged for assistance in mass spectrometric analysis
and Dr. Daniel Holmes for insightful discussions on NMR
experiments.
Supporting Information Available: Procedures for the prepa-
ration of new compounds and characterization data for all new
compounds including cif files for 10b and 10d. This material
is available free of charge via the Internet at http://pubs.acs.org.
(17) Beckett, M. A.; Coles, S. J.; Light, M. E.; Fischer, L.; Stiefvater-
Thomas, B. M.; Varma, K. S. Polyhedron 2006, 25, 1011.
(18) Schubert, D. M.; Visi, M. Z.; Knobler, C. B. Inorg. Chem. 2008, 47,
2017.
JA1070224
(20) Kochavi, E.; Bar-Nun, A.; Fleminger, G. J. Mol. EVol. 1997, 45, 342.
(21) Fleminger, G.; Yaron, T.; Eisensstein, M.; Bar-Nun, A. Origins Life
EVol. Biosphere 2005, 35, 369.
(19) Beckett, M. A.; Bland, C. C.; Horton, P. N.; Hursthouse, M. B.; Varma,
K. S. Inorg. Chem. 2007, 46, 3801.
9
J. AM. CHEM. SOC. VOL. 132, NO. 41, 2010 14675