Studies of asymmetric styrene cyclopropanation with a rhodium(II) metallopeptide catalyst developed with a high-throughput screen

Article abstract of DOI:10.1002/chir.22144

Dirhodium metallopeptides have been developed as selective catalysts for asymmetric cyclopropanation reactions. A selective ligand sequence has been identified by screening on-bead metallopeptide libraries in a 96-well plate format. Efficient ligand synthesis and screening allows a 200-member library to be created and assayed in less than three weeks. These metallopeptides catalyze efficient cyclopropanation of aryldiazoacetates, providing asymmetric access to cyclopropane products in high diastereoselectivity. 2013 Wiley Periodicals, Inc.

Full text of DOI:10.1002/chir.22144

CHIRALITY 25:493–497 (2013)
Studies of Asymmetric Styrene Cyclopropanation with a Rhodium(II)
Metallopeptide Catalyst Developed with a High-Throughput Screen
*
RAMYA SAMBASIVAN, AND ZACHARY T. BALL
Department of Chemistry MS60, Rice University, 6100 Main Street, Houston, TX 77005
ABSTRACT
Dirhodium metallopeptides have been developed as selective catalysts for asymmet-
ric cyclopropanation reactions. A selective ligand sequence has been identified by screening on-bead
metallopeptide libraries in a 96-well plate format. Efficient ligand synthesis and screening allows a
200-member library to be created and assayed in less than three weeks. These metallopeptides
catalyze efficient cyclopropanation of aryldiazoacetates, providing asymmetric access to cyclopropane
products in high diastereoselectivity. Chirality 25:493–497, 2013. © 2013 Wiley Periodicals, Inc.
KEY WORDS: asymmetric catalysis; metallopeptides; cyclopropanation; rhodium; peptide
library; catalyst screening
INTRODUCTION
structure can be identified via pyrene fluorescence.11 The
solution-phase catalyst screening is cumbersome because it
requires multiple HPLC separations to isolate the iso-
meric complexes. To speed up the process, we developed
an on-bead, parallel screening method in a 96-well plate
format to assess peptide sequences which were tested on bead
as mono-peptide complexes, Rh2(peptide)(OAc)2. This screen
is based on the hypothesis that optimal sequences in on-bead
mono-peptide catalysts will tend to produce better bis-peptide
catalysts, which can only be synthesized in solution. This
screening technique identified a catalyst for the asymmetric
cyclopropanation of methyl a-phenyldiazoacetate ester and
styrene (Fig. 1). The “lead” on-bead catalyst producing the
(1R,2S)-cyclopropane product in 54% ee at rt exhibited signifi-
cant enhancement in selectivity (87% ee at –50 ꢀC) on moving
to the bis-peptide variant, Rh2(L2.47)2-A (where the suffix
“A” arbitrarily identifies orientational isomers).
Peptide ligands offer advantages over traditional chiral
ligands for selective catalysis: they are modular and functional-
group-rich sources of chirality from the chiral pool. Modern
methods for automated, parallel synthesis facilitate expedient
catalyst screening. In theory, a library screening approach
obviates the tedious process of chiral ligand development for
each and every new reaction, much as high-throughput screen-
ing approaches now routinely deliver lead compounds for biolog-
ical targets. Once a few catalyst libraries have been synthesized,
new reaction development can take advantage of prior efforts.1,2
We recently reported a rhodium(II) metallopeptide catalyst for
asymmetric cyclopropanation that was discovered though a
peptide screening approach.3 In this article, we discuss practical
considerations for library synthesis and screening, and we pres-
ent further studies on the generality of cyclopropanation with
rhodium(II) catalysts containing the chiral peptide ligand,
L2.47 (where “L2.47” is peptide 47/96 of the second round
of optimization in 96-well plates).
EXPERIMENTAL
The use of transition metals with peptide ligands as catalysts
for asymmetric reactions is surprisingly limited. It is often
challenging to design well-defined coordination spheres in
such a functional-group-rich environment. There have been
reports of enantioselective catalysts with unnatural amino acids
containing phosphine4–6 or pyridine7 or phenanthroline8 side
chains which serve as handles to bind to transition metals.
In other cases, entire transition-metal complexes have been
attached to large proteins to form a metalloprotein which can
then aid asymmetric catalysis.9 By turning to the very stable
rhodium(II) carboxylate linkage, we have benefitted from the
dual advantages of a stable ligand sphere using cheap, naturally
available amino acids.
The idea of catalyst screening stemmed from our previous
studies demonstrating the use of 9-mer peptides containing
aspartic acid residues in an internal DxxxD motif (D = aspar-
tate) as chiral ligands for dirhodium-catalyzed asymmetric
Si–H insertions.10 In that work, we individually isolated and
purified catalysts of the type Rh2(peptide)2, prepared by
treatment of a peptide containing aspartates with Rh2(tfa)4
(tfa = trifluoroacetate). There are two isomeric bis-peptide
complexes formed with each peptide sequence, due to the
potential for parallel or antiparallel orientation of the peptides,
which are generally separated by RP-HPLC. The orientational
© 2013 Wiley Periodicals, Inc.
Materials and Methods
All rhodium-catalyzed reactions were carried out in 4-ml vials. Reactions
carried out below 0 ꢀC were conducted in a Neslab CB-80 cold bath. Flash
chromatography was performed with 40-63-mm particle size silica gel.
NMR data was acquired with Bruker Avance 400 MHz or Bruker Avance
500 MHz instrument. 1H and 13C NMR spectra were referenced relative to
residual solvent or TMS. Chiral HPLC analyses were performed on a
Shimadzu SCL-10ADVP instrument with Phenomenex Lux 5u Cellulose-1
(250 ꢁ 4.6 mm analytical) or a Chiralpak IA (250 ꢁ 2 mm ID analytical)
column with a flowrate of 1.5–1.9 ml/min and the detection wavelength
was 220 nm. All reagents were purchased commercially and used without
further purification. All solvents were reagent grade. The synthesis of
diazo substrates10 has been previously reported. For purpose of ee determi-
nation, racemic material was generated using Rh2(OAc)4. The absolute
configuration of product 2a is assigned by comparison of the optical
rotation to previous reports.12,13 The absolute configuration of all
Contract grant sponsor: Robert A. Welch Foundation; Contract grant number:
C-1680.
Contract grant sponsor: NFS CAREER award; Contract grant number:
CHE-1055569.
*Correspondence to: Zachary Ball, Department of Chemistry, Rice
University, 6100 Main Street, Houston, TX 77005. E-mail: zb1@rice.edu
Received for publication 16 October 2012; Accepted 16 November 2012
DOI: 10.1002/chir.22144
Published online 8 June 2013 in Wiley Online Library
(wileyonlinelibrary.com).

494
SAMBASIVAN AND BALL
was added to a solution of the substrates in dichloromethane
(190 ml). The reaction did not proceed at temperatures below rt
and solvent insertion product of the diazo compound was
observed when trifluoroethanol was exclusively used as
the solvent.
1H NMR (400 MHz, CDCl3) d 7.29ꢂ7.18 (m, 5H), 3.78
(s, 3H), 3.35 (s, 3H), 3.23 (t, J = 8.6 Hz, 1H), 2.20 (dd, J = 8,
5.2 Hz, 1H), 1.74 (dd, J = 5.2, 9.2 Hz, 1H). 13C NMR (400
MHz, CDCl3) d 170.4, 167.2, 134.7, 128.6, 128.4, 127.6, 53.0
52.4, 37.4, 32.7, 19.3. GC-MS tR 13.0 min (> 95%), m/z: [M]+
calcd for C13H14O4: 234.3; found: 234.1. Enantiomeric excess
determined by HPLC; Phenomenex Lux 5u Cellulose-1 column,
eluting with 99:1 hexanes/i-PrOH, 1.5 ml/min, (S)-enantiomer
tR 6.1 min; (R)-enantiomer tR 5.7 min.
(1R,2S)-methyl 1-(naphthalen-2-yl)-2-phenylcyclopropanecarboxylate
(2e). The general procedure was employed using methyl 2-(2-
naphthyl)-2-diazoacetate on a 0.004 mmol scale at –50 ꢀC to pro-
vide the product in 82% ee.
Fig. 1. On-bead screening of metallopeptide catalysts for cyclopropanation.
Reaction conditions: diazo (6 mmol), styrene (60 mmol), and catalyst (~0.15 mmol),
in CF3CH2OH. All catalysts are attached to the resin at the C-terminus of the
peptides, and the peptide N-terminus is acetylated. Sequence shown is L2.47.
1H NMR (400 MHz, CDCl3) d 7.75–6.65 (m, 2H), 7.61
(s, 1H), 7.53 (d, J = 8.4 Hz, 1H), 7.42–7.40 (m, 2H), 7.05 – 7.00
(m, 4H), 6.82 – 6.79 (m, 2H), 3.65 (s, 3H), 3.19 (dd, J = 7.4,
9.2, 1H), 2.22 (dd, J = 4.8, 9.2 Hz, 1H), 2.02 (dd, J = 4.8, 7.4
Hz, 1H). 13C NMR (400 MHz, CDCl3) d 174.6, 136.4, 133.2,
132.8, 132.7, 130.7, 130.3, 128.3, 128.0, 127.8, 127.3, 126.6,
126.0, 125.9, 52.9, 37.7, 33.5, 20.9. GC-MS tR 17.6 min (> 95%),
m/z: [M]+ calcd for C21H18O2: 302.3; found: 302.1. Enantio-
meric excess determined by HPLC; Phenomenex Lux 5u
Cellulose-1 column, eluting with 99.5:0.5 hexanes/i-PrOH,
1.5 ml/min, (1R,2S)-enantiomer tR 9.7 min; (1S,2R)-enantiomer
tR = 8.8 min.
other cyclopropane products 2b–2i is assumed by analogy. From our
previously published library screening, the ‘hit’ peptide sequence yield-
ing the highest enantioselectivity in cyclopropanation reaction was
found to be the ligand L2.47, and the catalyst Rh2(L2.47)2-A was
prepared according to reported procedures.3 The characterization data
1
comprising H NMR, 13C NMR, and GC-MS data have been reported for
the racemic product. The chiral compounds have been characterized by
TLC, HPLC, and 1H NMR of the crude reaction mixture. They were all in
agreement with the corresponding data obtained for the racemic product.
The synthesis and characterization of products 2a and 2b–7b have been
reported previously.3
(1R,2S)-methyl 1-(3-(trifluoromethyl)phenyl)-2-phenyl-cyclo-
propanecarboxylate (2f ). The general procedure was employed
using methyl 2-phenyl-(3-trifluoromethyl)-2-diazoacetate on a
0.004-mmol scale at –50 ꢀC to provide the product in 73% ee.
1H NMR (400 MHz, CDCl3) d 7.38–7.36 (m, 1H), 7.27–7.16
(m, 3H), 7.07–7.05 (m, 3H), 6.77–6.75 (m, 2H), 3.68 (s, 3H),
3.17 (dd, J = 7.6, 9.2, 1H), 2.19 (dd, J = 4.8, 9.2 Hz, 1H), 1.91
(dd, J = 4.8, 7.6 Hz, 1H). 13C NMR (400 MHz, CDCl3) d 173.8,
136.2, 135.8, 135.7, 130.8–129.8 (q, J = 128 Hz), 130.1–122.9
(q, J = 1080 Hz), 129.9 (q, J = 16), 128.3, 128.2, 128.1, 126.9,
124.1 (q, J = 16 Hz), 53.0, 37.2, 33.5, 20.3. GC-MS tR 17.6 min
(> 95%), m/z: [M]+ calcd for C18H15F3O2: 320.3; found: 320.1.
Enantiomeric excess determined by HPLC; Phenomenex
Lux 5u Cellulose-1 column, eluting with 99.75:0.25 hexanes/
i-PrOH, 1.5 ml/min, (1R,2S)-enantiomer tR 9.9 min; (1S,2R)-
enantiomer tR 12.8 min.
General Procedure for Cyclopropanation Reactions
A solution of the catalyst Rh2(L2.47)2-A (0.082 mg, 0.5 mol%) in
trifluoroethanol (100 ml) was stirred at –50 ꢀC in a 4-ml vial. A mixture of
methyl a-diazophenylacetate (1 mg, 0.0057 mmol) and styrene (5.92 mg,
0.057 mmol) in trifluoroethanol (100 ml) were stirred at the same tempera-
ture in another 4-ml vial. The catalyst solution was then quickly transferred
to the vial containing the substrates and the reaction mixture was allowed to
stir at –50 ꢀC until the diazo compound reacted fully. The mixture was
concentrated under reduced pressure and purified by flash chromatogra-
phy in a glass pipette on silica gel using 1:99 diethylether/hexane to give
the desired cyclopropane product.
Characterization of the Cyclopropane Products
(1S,2S)-methyl 2-phenyl-1-((E)-styryl) cyclopropanecarboxylate
(2c). The general procedure was employed using methyl
2-((E)-styryl)-2-diazoacetate on a 0.005-mmol scale at –35 C
ꢀ
to provide the product as a white crystalline solid in 54% ee.
1H NMR (500 MHz, CDCl3) d 7.24ꢂ7.20 (m, 4H), 7.18ꢂ7.12
(m, 6H), 6.34 (d, J = 16 Hz, 1H), 6.13 (d, J = 16 Hz, 1H), 3.76
(s, 3H), 3.00 (dd, J = 9.0, 7.0 Hz, 1H), 2.02 (dd, J = 9.0, 5.0,
1H), 1.83 (dd, J = 7.0, 5.0 Hz, 1H). 13C NMR (500 MHz, CDCl3)
d 174.4, 137.3, 135.7, 133.3, 129.3, 128.6, 128.2, 127.5, 127.0,
126.4, 124.3, 52.7, 35.2, 33.5, 18.8. GC-MS tR 19.1 min (> 95%),
m/z: [M]+ calcd for C19H18O2: 278.3; found: 278.2. Enantiomeric
excess determined by HPLC; Phenomenex Lux 5u Cellulose-1
column, eluting with 99:1 hexanes/i-PrOH, 1.5 ml/min,
(1S,2S)-enantiomer tR 7.3 min; (1R,2R)-enantiomer tR 5.9 min.
(1R,2S)-methyl 1-(2-methoxyphenyl)-2-phenylcyclopropane-
carboxylate (2g). The general procedure was employed
using methyl 2-methoxyphenyl-2-diazoacetate on a 0.005-mmol
scale at –50 ꢀC to provide the product in 49% ee.
1H NMR (400 MHz, CDCl3) d 7.16–7.11 (m, 2H), 7.01– 6.98
(m, 3H), 6.85–6.76 (m, 3H), 6.53 (d, J = 8 Hz, 1H), 3.64
(s, 3H), 3.35 (s, 3H), 3.22 (dd, J = 7.6, 9.2, 1H), 1.97 (dd, J =
5.2, 9.2 Hz, 1H), 1.85 (dd, J = 5.2, 7.6 Hz, 1H). 13C NMR
(400 MHz, CDCl3) d 174.7, 159.2, 137.0, 131.8, 128.9, 127.9,
127.3, 126.1, 124.1, 120.0, 110.4, 55.2, 52.7, 34.3, 32.6, 20.8.
GC-MS tR 16.4 min (> 95%), m/z: [M]+ calcd for C18H18O3:
282.3; found: 282.1. Enantiomeric excess determined by
HPLC; Phenomenex Lux 5u Cellulose-1 column, eluting with
99.5:0.5 hexanes/i-PrOH, 1.5 ml/min, (1R,2S)-enantiomer tR
11.3 min; (1S,2R)-enantiomer tR 10.4 min.
(S)-dimethyl 2-phenylcyclopropane-1,1-dicarboxylate (2d). The
general procedure was employed using dimethyl diazomalonate
on a 0.006-mmol scale at rt to provide the product in 38% ee. The
catalyst was dissolved in trifluoroethanol (10 ml) and this solution
Chirality DOI 10.1002/chir

DIRHODIUM METALLOPEPTIDES
495
(1R,2S)-methyl 1-(3-bromophenyl)-2-phenylcyclopropanecarboxylate
(2h). The general procedure was employed using methyl
3-bromophenyl-2-diazoacetate on a 0.004-mmol scale at –50
ꢀC to provide the product in 86% ee.
bead. This would eliminate the need for HPLC purification of
the peptides and the metallopeptide catalysts. We decided to
synthesize peptide libraries in 96-well plates and then subject
the peptides to dirhodium complexation. However, we envis-
aged that complexation with Rh2(tfa)4 would be detrimental to
the effort of identifying a selective catalyst since we again would
face the problem of separating and isolating the isomeric
on-bead bis-peptide complexes. To circumvent this issue,
we decided to complex the on-bead peptides with cis-Rh2
(tfa)2(OAc)2 to form the mono-peptide catalysts and hypothe-
sized that a selective mono-peptide catalyst would exhibit
enhanced selectivity when prepared as a bis-peptide catalyst in
solution. Our first goal was to identify the optimal resin
on which we could synthesize the peptide libraries. From
our solution-phase library screening,10 we had identified L21
(peptide sequence: KZNDAAIDAKZ) as one of the most selec-
tive ligands for the insertion of methyl a-phenyldiazoacetate
into the Si–H bond of phenyl dimethyl silane (Table 1).
We then decided to synthesize this same L21 sequence on
four different types of resins with varying resin loading in
order to narrow down an optimal resin type. Also, we first
established that our catalyst system could be synthesized
on a solid support and that it exhibited reactivity. Similar to
our solution-phase protocol, we complexed the bis-aspartate
containing peptides on the resin bead with cis-Rh2(tfa)2
(OAc)2 in the presence of N,N-diisopropylethylamine using
trifluoroethanol as solvent. The most important factor among
several resins tested was resin loading (see Table 1 for
details). The Novasyn TG amino resin had lower loadings
(0.17–0.29 mmol/g) and yielded product enantioselectivity
(81% at –35 ꢀC) in the model reaction that was similar to
solution-phase catalysts (Table 1). This on-bead catalyst could
be used up to three times without affecting selectivity.
1H NMR (400 MHz, CDCl3) d 7.24–7.21 (m, 2H), 7.11–7.08
(m, 3H), 6.98–6.94 (m, 1H), 6.90–6.87 (m, 1H), 6.80–6.77
(m, 2H), 3.67 (s, 3H), 3.12 (dd, J = 7.6, 9.2, 1H), 2.14 (dd,
J = 5.2, 9.2 Hz, 1H), 1.86 (dd, J = 5.2, 7.6 Hz, 1H). 13C NMR
(400 MHz, CDCl3) d 173.9, 137.4, 135.9, 135.0, 131.0, 130.4,
129.3, 128.2, 128.1, 126.8, 121.8, 53.0, 37.1, 33.4, 20.5. GC-MS
tR 17.7 min (> 95%), m/z: [M]+ calcd for C17H15BrO2: 331.2;
found: 331.1. Enantiomeric excess determined by HPLC;
Phenomenex Lux 5u Cellulose-1 column, eluting with 99.75:0.25
hexanes/i-PrOH, 1.5 ml/min, (1R,2S)-enantiomer tR 11.5 min;
(1S,2R)-enantiomer tR 12.6 min.
(1R,2S)-methyl 1-(4-phenylphenyl)-2-phenylcyclopropanecarboxylate
(2i). The general procedure was employed using methylꢀ4-
phenylphenyl-2-diazoacetate on a 0.004-mmol scale at –50 C
to provide the product in >85% ee.
1H NMR (400 MHz, CDCl3) d 7.52–7.48 (m, 2H), 7.37–7.28
(m, 5H), 7.10–7.05 (m, 5H), 6.82–6.79 (m, 2H), 3.69 (s, 3H),
3.13 (dd, J = 7.2, 9.2, 1H), 2.17 (dd, J = 4.8, 9.2 Hz, 1H), 1.91
(dd, J = 4.8, 7.2 Hz, 1H). 13C NMR (400 MHz, CDCl3) d 174.5,
140.9, 136.5, 132.5, 131.0, 128.9, 128.3, 128.0, 127.8, 127.6
127.4, 127.2, 126.6, 52.9, 37.3, 33.5, 20.8. GC-MS tR 14.1 min
(> 95%), m/z: [M-OMe]+ calcd for C22H18O1: 298.3; found:
298.1. Enantiomeric excess determined by HPLC; Phenomenex
Lux 5u Cellulose-1 column, eluting with 97:3 hexanes/i-PrOH,
1.5 ml/min, (1R,2S)-enantiomer tR 3.8 min; (1S,2R)-enantiomer
tR 4.0 min.
RESULTS AND DISCUSSION
Optimization of Resin Type and Loading
Our first-generation solution-phase screening process in-
volved synthesis of the peptides using standard FMOC proto-
cols, cleavage from the resin, peptide purification, complexation
of the purified peptides with Rh2(tfa)4, and RP-HPLC isolation
of the isomeric bis-peptide catalysts.10 This entire screening
process was somewhat tedious in that it took weeks to screen
a total of six sequences (i.e., 12 metallopeptide catalysts). To
screen more efficiently, we required a new approach and
resorted to an on-bead catalyst screening technique.
Library Screening Strategies
After identifying a suitable resin to synthesize our 96-well
plate peptide library, we then explored various techniques
to make the library. For the optimization of the techniques,
rink amide resin was used to allow peptide cleavage and
purity analysis.
Initial efforts at library synthesis.
a. Kitchen microwave: We initially attempted to take advan-
tage of the benefits of microwave heating14 by
conducting the synthesis using irradiation from a
simple kitchen microwave. We loaded 18 random wells
In the second-generation catalyst screening, the rhodium
(II) metallopeptide catalysts were prepared from peptide
libraries in which the peptides remain attached to the resin
TABLE 1. Optimization of resin type and loading
Entry
Resin
Description
Loading (mmol/g)
ee (%)
1
2
3
4
AAPPTec RAZ001
AAPPTec RAZ051
NovaGel 855037
NovaSyn 855073
aminomethyl polystyrene
“surface active” aminomethyl
hydrazinobenzoyl “safety catch”
TG amino resin
1.12
0.76
0.49
0.29
3
3
51
66(81a)
aReaction at –35 ꢀC.
Chirality DOI 10.1002/chir

496
SAMBASIVAN AND BALL
in the 96-well plate with Rink amide resin and attempted
peptide synthesis on a 5.4-mmol scale. Each deprotection
step was carried out for 4 min and each coupling was
carried out for 6 min. We varied the power setting and
monitored the temperature after the heating with a laser
thermometer. Unfortunately, large variations in tempera-
tures (55–70 ꢀC) were observed, and MALDI analysis indi-
cated the presence of point-deletion impurities. Attempts
to improve the situation by stirring within the microwave
were unsuccessful, largely for technical reasons.
b. Shaker: We were also unsuccessful using an orbiting
incubator for agitation. On MALDI analysis, the pep-
tides synthesized in most wells had an asparagine
deletion, and so we decided not to pursue using the
manual shaker.
Fig. 2. Optimized ligand structure. Z = benzyloxycarbonyl.
affording product 2c with a modest 54% ee. As observed in
silane insertion, an ortho-substituted diazo substrate gave
product with significantlydecreased selectivity (49% ee, Table 3;
entry 6). A non-aryl diazo substrate (dimethyl diazomalonate)
undergoes cyclopropanation only at room temperature,
resulting in product with 38% ee (entry 4). The lack of complete
Manual peptide synthesis with magnetic stirring. A small mag-
netic stir bar was added to each well of the 96-well plate; the
resin was distributed and the peptide synthesis was carried
out using standard FMOC protocols. MALDI and RP-HPLC
analysis of the peptides obtained from each well indicated
the formation of >95% pure peptide in each well.
Encouraged by the purity of the peptides obtained in
the above-mentioned strategy, we synthesized the peptide
library for catalyst screening using the Novasyn TG amino
resin and then subjected the peptides to rhodium(II)
complexation. After metalation with cis-Rh(OAc)2(tfa)2, the
96-well plate was screened successfully against a target
cyclopropanation reaction (Fig. 1). After the reaction, the
reaction wells were concentrated in vacuo to remove excess
styrene and then assayed directly by chiral HPLC. Reaction
yields, which are generally quite high for these reactions,
were not measured due to the small scale used for screening
(1 mg diazo).
TABLE 2. Asymmetric cyclopropanation with
a-diazophenylacetatea
Employing this latter library synthesis method, a screen
uncovered the “hit” sequence, L2.47 (Fig. 2), the details of
which have been previously reported.3 The use of the Rh2
(L2.47)2-A catalyst yielded the product cyclopropane in 92%
ee for the tert-butyl phenyl diazoacetate (Table 2). This class
of catalysts produces cyclopropane products with complete
selectivity for the trans diastereomer.
% ee
R0 = Me
% ee
R0 = t-Bu
Entry
1
Alkene
87 (2a)b
92 (2b)
92 (3b)
2
3
8687 (3a)c
Variation of Diazo Substrates in Cyclopropanation Reactions. In
general, tert-butyl diazoesters yield better enantioselectivity
than their methyl ester equivalents (Table 2; entries 1–6). The
improvement was most pronounced in the case of a-methyl
styrene (Table 2; entry 3).
The rhodium(II) metallopeptide catalyst Rh2(L2.47)2-A
afforded good enantioselectivities in the cyclopropanation
reaction of a-phenyldiazoacetate with a variety of olefins.
We then wanted to assess the selectivity of the same catalyst
with other diazo variants and subjected the catalyst to
cyclopropanation reactions of styrene with a number of diazo
substrates. The enantioselectivities are as shown in Table 3.
In most of these reactions, the cyclopropane was obtained as
the major product. A small amount (<15%) of the solvent
insertion product with trifluoroethanol was formed as observed
from NMR of the crude reaction mixture.
6987 (4a)c
9087 (5a)c
95 (4b)
97 (5b)
4
5
6
9087 (6a)c
72 (7a)c
95 (6b)
80(7b)d
Many aryldiazoacetates gave good enantioselectivities
similar to the parent compound, methyl a-phenyldiazoacetate.
The naphthyl-, 3-bromo- and 4-phenyl- substrate gave selec-
tivites of >80% (Table 3; entries 4,7,8). Chemoselective
cyclopropanation of a styryldiazo substrate was also possible,
Chirality DOI 10.1002/chir
aEntries in the right column have been previously reported.3
bThe absolute configuration for entry 1 was established by comparison to pub-
lished data; that of other products is assumed by analogy.
cReaction at –35 ꢀC.
dReaction at –25 ꢀC.

DIRHODIUM METALLOPEPTIDES
497
TABLE 3. Asymmetric cyclopropanation with styrene
substrate generality observed here is a common feature of
chiral catalysts generally, and provides motivation for efficient
ligand discovery approaches to address new problematic sub-
strates as they arise.
CONCLUSIONS
Thus, on-bead catalyst screening has a number of advan-
tages: facile ligand screening, parallel synthesis of hundreds
of catalysts, and easy catalyst removal by filtration. Employing
this screening technique, over 100 catalysts can be screened
and a “hit” catalyst identified in a matter of days. We have
established concepts for chiral ligand discovery using peptides
as a general and efficient path to address the challenging
problem of catalyst development for asymmetric synthesis,
and we believe these ideas will be applicable to important reac-
tions of other transition metals.
Entry
1
Starting material
Product
eea (%)
87
2a
ACKNOWLEDGMENTS
We acknowledge financial support from the Robert A. Welch
Foundation research grant C-1680 and from an NSF CAREER
award (CHE-1055569).
2
3
2c
2d
54
LITERATURE CITED
38b
1. Wee AGH. Rhodium(II)-catalyzed reaction of diazocompounds in the service
of organic synthesis of natural and non-natural products. Curr Org Synth
2006;3:499–555.
2. Evans PA. Modern rhodium-catalyzed organic reactions, Weinheim:
Wiley-VCH; 2005.
4
5
6
7
8
2e
2f
82
73
49
86
85c
3. Sambasivan R, Ball ZT. Screening rhodium metallopeptide libraries “on
bead”: asymmetric cyclopropanation and a solution to the enantiomer
problem. Angew Chem Int Ed 2012;51:8568–8572.
4. Gilbertson SR, Wang XF, Hoge GS, Klug CA, Schaefer J. Synthesis of
phosphine-rhodium complexes attached to a standard peptide synthesis
resin. Organometallics 1996;15:4678–4680.
5. Gilbertson SR, Wang XF. The parallel synthesis of peptide based phosphine
ligands. Tetrahedron 1999;55:11609–11618.
6. Gilbertson SR, Wang XF. The combinatorial synthesis of chiral phosphine
ligands. Tetrahedron Lett 1996;37:6475–6478.
7. Coquière D, Bos J, Beld J, Roelfes G. Enantioselective artificial
metalloenzymes based on
a bovine pancreatic polypeptide scaffold.
2g
2h
2i
Angew Chem Int Ed 2009;48:5159–5162.
8. Bos J, Fusetti F, Driessen AJM, Roelfes G. Enantioselective artificial
metalloenzymes by creation of a novel active site at the protein dimer
interface. Angew Chem Int Ed 2012;51: 7472–7475.
9. Thomas CM, Ward TR. Artificial metalloenzymes: proteins as
hosts for enantioselective catalysis. Chem Soc Rev 2005;34:
337–346.
10. Sambasivan R, Ball ZT. Metallopeptides for asymmetric dirhodium cataly-
sis. J Am Chem Soc 2010;132:9289–9291.
11. Sambasivan R, Ball ZT. Determination of orientational isomerism in rho-
dium(ii) metallopeptides by pyrene fluorescence. Org Biomol Chem
2012;10:8203–8206.
12. Davies HML, Lee GH. Dirhodium(II) tetra(N-(dodecylbenzenesulfonyl)
prolinate) catalyzed enantioselective cyclopropenation of alkynes. Org
Lett 2004;6:1233–1236.
aThe absolute configuration of 2a was established by comparison to published
data. Other products are assumed by analogy.
13. Bruzinski PR. Novel chiral rhodium(II) carboxylates as catalysts in
asymmetric cyclopropanations. Ph.D. Thesis (9801276). SUNY at Buffalo,
Buffalo, NY, 1997.
bReaction at rt in trifluoroethanol/CH2Cl2 (1:19).
cEnantiomeric peaks slightly overlapping in the chiral HPLC trace. This
measure may underestimate the selectivity of this substrate.
14. Murray JK, Gellman SH. Parallel synthesis of peptide libraries using
microwave irradiation. Nat Protoc 2007;2:624–631.
Chirality DOI 10.1002/chir