Computationally-Led Ligand Modification using Interplay between Theory and Experiments: Highly Active Chiral Rhodium Catalyst Controlled by Electronic Effects and CH–π Interactions

Article abstract of DOI:10.1002/adsc.201701191

A chiral ligand for the rhodium-catalyzed asymmetric 1,4-addition of an arylboronic acid to a coumarin substrate that could markedly reduce catalyst loading was developed using interplay between theoretical and experimental approaches. Evaluation of the transition states for insertion and for hydrolysis of intermediate complexes (which were emphasized in response to the experimental results) using DFT calculations at the B97D/6-31G(d) level with the LANL2DZ basis set for rhodium revealed that: (i) the electron-poor nature of the ligands and (ii) CH–π interactions between the ligand and coumarin substrates played significant roles in both acceleration of insertion and inhibition of ArB(OH)₂ decomposition (protodeboronation). The computationally-designed ligand, incorporating the above information, enabled a decrease in the catalyst loading to 0.025 mol% (S/C=4,000), which is less than one one-hundredth relative to past catalyst loadings of typically 3 mol%, with almost complete enantioselectivity. Furthermore, the gram-scale synthesis of the urological drug, (R)-tolterodine (l)-tartrate, was demonstrated without the need of intermediate purification. (Figure presented.).

Full text of DOI:10.1002/adsc.201701191

Title: Computationally-led Ligand Modification using Interplay between
Theory and Experiments: Highly Active Chiral Rhodium Catalyst
Controlled by Electronic Effects and CH–π Interactions
Authors: Toshinobu Korenaga, Ryo Sasaki, Toshihide Takemoto,
Toshihisa Yasuda, and Masahito Watanabe
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Adv. Synth. Catal. 10.1002/adsc.201701191
Link to VoR: http://dx.doi.org/10.1002/adsc.201701191

10.1002/adsc.201701191
Advanced Synthesis & Catalysis
FULL PAPER
DOI: 10.1002/adsc.201((will be filled in by the editorial staff))
Computationally-led Ligand Modification using Interplay
between Theory and Experiments: Highly Active Chiral
Rhodium Catalyst Controlled by Electronic Effects and CH–π
Interactions
Toshinobu Korenaga,a,* Ryo Sasaki,a Toshihide Takemoto,b Toshihisa Yasuda,b
Masahito Watanabeb
a
Department of Chemistry and Biological Sciences, Faculty of Science and Engineering, Iwate University, 4-3-5 Ueda,
Morioka, Iwate 020-8551, Japan
E-mail: korenaga@iwate-u.ac.jp
Central Research Laboratory, Technology and Development Division, Kanto Chemical Co., Inc., Soka, Saitama 340-
b
0003, Japan
Received: ((will be filled in by the editorial staff))
Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/adsc.201######.((Please
delete if not appropriate))
Abstract. A chiral ligand for the rhodium-catalyzed The computationally-designed ligand, incorporating above
asymmetric 1,4-addition of arylboronic acid to a coumarin information, decreased the catalyst loading up to 0.025
substrate that could markedly reduce catalyst loading was mol% (S/C = 4,000), which is less than one one-hundredth
developed using interplay between theoretical and relative to past catalyst loadings of typically 3 mol%, with
experimental approaches. Evaluation of the transition states almost complete enantioselectivity. Furthermore, the gram-
for insertion and for hydrolysis of intermediate complexes scale synthesis of the urological drug, (R)-Tolterodine (L)-
(which were emphasized in response to the experimental tartrate, was demonstrated without the need of intermediary
results) using DFT calculations at the B97D/6-31G(d) with purification.
the LANL2DZ basis set for rhodium revealed that: a) the
electron-poor nature of the ligands; and b) CH–π interactions
between the ligand and coumarin substrates played
significant roles in both acceleration of insertion and
inhibition of ArB(OH)2 decomposition (protodeboronation).
Keywords: Asymmetric 1,4-addition; DFT calculations;
Transition states; Chiral phosphine ligand; Pharmaceutical
molecule
reactions because it can be performed with
computationally-deep considerations of the catalytic
cycle, including transition states based on
experimental results. Therefore, the method would
enable computationally-led rational design of the
catalyst for improving catalytic activity or selectivity.
In particular, some successful examples of
improvement of stereoselectivity by computationally-
led design of catalysts were reported.[1,7] In contrast,
only a few results have been reported for the
Introduction
The development of an efficient catalytic reaction
requires many screenings of existing catalysts or
many tests of modified catalysts for activity. When
the modification or a de novo design of a catalyst is
attempted, success will depend on a chemist’s
intuition as well as knowledge, experience, and the
experimental results. Computational chemistry could
aid in these approaches,[1] but a computationally-led
design for the development of effective catalysts is
not commonplace because: a) specialized skills are
required to carry out the necessary calculations
correctly and b) sufficiently accurate calculations are
not always affordable.[2] This is particularly the case
in organometallic systems which are very
complicated due to the complex and multistep
catalytic cycles.[3] Furthermore, because compromises
of a calculation method are necessary at present, it is
important to validate the calculation results by
comparison with reference data.[2b,4] Recently, the
interplay between theoretical and experimental
approaches has been reported to contain profound
insight into catalytic reactions, including transition
metal catalysts.[5,6] This interplay method has the
potential for reactivity prediction of catalytic
improvement
of
catalytic
activity
using
computational ligand design with this interplay
method. Schoenebeck designed and synthesized
bisphophines-bearing CF3 groups to enhance the
reductive elimination of Ph-CF3 from palladium,
although that in itself was not the catalytic reaction.[8]
Yoshizawa and Nishibayashi demonstrated activity
improvement of a molybdenum catalyst for ammonia
synthesis using the interplay method.[9] Introduction
of electron-donating methoxy groups to a ligand
designed
by
calculations
accelerated
the
molybdenum-catalyzed nitrogen fixation to give ca.
1.5 times the amount of ammonia. Anderson and
Straker et al. reported excellent computational ligand
design.[10] They developed a highly reactive rhodium
catalyst
for
the
enantioselective
and
1
This article is protected by copyright. All rights reserved.

10.1002/adsc.201701191
Advanced Synthesis & Catalysis
diastereoselective cycloisomerization of ynamide-
vinylcyclopropanes to [5.3.0]-azabicycles. The
computationally-designed ligand, bearing fluorine
groups, improved the reaction time (e.g., 1 h → 5
min) and enantioselectivity of the catalytic reaction.
The above examples show that the interplay method
is very useful for rational catalyst design, but the
examples, thus far, have not yet used the interplay
method to reduce catalyst loadings. Therefore, we
will use the interplay method to develop the ligand
for practically usable catalyst at scales viable for the
pharmaceutical industry.
This work demonstrates development of a chiral
rhodium catalyst to drastically reduce the catalyst
loading to 0.025 mol% (S/C = 4,000) in the
asymmetric 1,4-adddition of 3a to coumarins 2 by
computationally-led chiral ligand design using the
interplay between theoretical and experimental
approaches. Furthermore, a synthetic protocol for
(R)-tolterodine synthesis from (R)-1aa is also
demonstrated to utilize the industrial synthesis.[23]
Results and Discussion
We were interested in rhodium-catalyzed asymmetric
1,4-addition of arylboronic acids to enone substrates
(the so-called Hayashi–Miyaura coupling)[11,12]
because the C-C bond formation reaction produces
chiral 1,4-adducts (1), which include many useful
chiral synthetic intermediates.[13] Even though many
effective chiral rhodium catalysts for the reaction
have been developed, it is hard to utilize the reaction
for industrial scale syntheses.[14] Although some
examples of gram or kilogram scale reactions have
been reported, the catalyst loading was over 1 mol%
in all cases.[14,15] Because a rhodium complex is very
expensive, the large amount of rhodium complex
typically required does not justify the cost in
Firstly, chiral ligands were screened for asymmetric
1,4-addition of 3a to 2a in the presence of 0.1 mol%
rhodium catalyst (S/C = 1000). The reactions of 2a
with 1 equiv. of 3a were performed in 1,4-dioxane,
which was a typical solvent for the reaction, with
aqueous solution of NaHCO3 at 25 °C for 6 h using
known chiral ligands for rhodium (rhodium/ligand =
1.0) (Table 1). Conventional chiral ligands (BINAP,
SEGPHOS, and MeO-BIPHEP) were useless for the
1,4-addition reaction (entries 2-4). The electron-poor
difluorphos[24] and chiral diene ligands were also not
effective (entries 5 and 1). When highly electron-poor
MeO-F12-BIPHEP was used, the reaction proceeded
to give (R)-1aa in 23% yield with >99% ee (entry 6).
However, the use of even more electron-poor MeO-
3,5-(CF3)2-BIPHEP decreased the yield of (R)-1aa to
8% and with an inadequate enantioselectivity of 92%
ee (entry 9). The results indicated that the electron-
poor nature of the chiral diphosphine is important for
enhancing the catalytic activity of rhodium,[25] but
that is not enough as a requirement for the 1,4-
addition reaction to progress. Next, we focused on
both the yields and the residual quantity of 3a (Table
1). A certain amount of 3a was always consumed
regardless if the ligands were ineffective at catalyzing
the 1,4 addition (entries 1–5).[12c] The consumption of
3a was further accelerated using highly electron-poor
phosphines. Although the reaction using MeO-F12-
BIPHEP or MeO-3,5-(CF3)2-BIPHEP gave only low
yields of the 1,4-addition product, (R)-1aa, no
residual 3a was observed in the reaction solution
(entries 6 and 9). These results revealed that the
reason for incomplete 1,4-addition was not only the
low catalytic activity of the rhodium/diphosphine
system, but also the disappearance of reagent 3a.
However, increasing amount of 3a to 3.5 equiv.
improved the yield to only 38% (entry 7).
Furthermore, increasing reaction temperature to 50°C
accelerated decomposition of 3a and gave no product
of (R)-1aa (entry 8). Therefore, further improvement
proved difficult using the known ligands.
industrial
synthesis.
The
rhodium-catalyzed
asymmetric 1,4-addition to 6-methylcoumarin (2a)
with phenylboronic acid (3a) is a typical example of
an unfavorable reaction for industrial scale use. The
reaction gives (R)-6-methyl-4-phenylchroman-2-one
(1aa), which can be converted readily to Detrusitol®
((R)-tolterodine (L)-tartrate), an important urological
drug (Scheme 1).[16] Since the reaction using
rhodium/(R)-SEGPHOS catalyst was demonstrated
by Hayashi,[17] the same reaction was reported by us
using the rhodium/(R)-MeO-F12-BIPHEP catalyst,[18]
by Carnell using a rhodium/chiral diene ligand,[19]
and by Mino using the rhodium/(R)-BICMAP (2,2ʼ-
diphenylphosphino-1,1ʼ-bi-5,6-dihydrobenzofuran)
catalyst.[20] Although the reactions gave (R)-1aa in
high yield with high enantioselectivity, over 99% ee,
the catalyst loadings were 3 to 6 mol% (S/C = 33 to
17) in all above examples. Similar examples using
coumarin (2b) instead of 2a also required high
catalyst loadings.[17-20,21,22] Furthermore, the typically
required amount of 3a is 10 equivalents because
consumption of 3a by protodeboronation[12c] arises
from the low reactivity of 2a.
Table 1. Ligand Screening for Rhodium-catalyzed
Asymmetric 1,4-Addition of 3a to 2a (S/C = 1,000) [a]
Scheme 1. Synthesis of Detrusitol® via rhodium-catalyzed
asymmetric 1,4-addition of 3a to 2a.
2
This article is protected by copyright. All rights reserved.

10.1002/adsc.201701191
Advanced Synthesis & Catalysis
of complex 6 gives (R)-1aa and regenerates the active
species 4 in dimer form (4•••4). The low-reactivity of
2a slows the reaction rate of insertion (kins), resulting
in a change of the rate-determining step to
insertion,[28] while hydrolysis of complex 5 is
actualized (kins < khyd) to enhance the consumption of
3a, unlike the reaction of cyclohexenone instead of
2a. If the insertion reaction can be accelerated by an
appropriate ligand, the reaction rate of insertion (kins)
would be enhanced and the hydrolysis of complex 5
(khyd) will be disturbed. However, finding the
appropriate chiral ligand using random or intuitive
screenings has limitations, as shown in Table 1.
Therefore, the rational design of a ligand is essential
for accelerating this catalytic reaction.
Entry (R)-Ligand
Yield [%] ee [%] 3a[b]
1
diene*
1
0
0
0
0
-
ca. 80%
ca. 80%
ca. 50%
ca. 50%
ca. 90%
none
2
BINAP
-
3
SEGPHOS
MeO-BIPHEP
difluorphos
-
4
-
5
-
6
MeO-F12-BIPHEP 23
MeO-F12-BIPHEP 38
MeO-F12-BIPHEP 0
>99
>99
-
7[c]
8[d]
9
none
Figure 1. Catalytic cycle of rhodium-catalyzed asymmetric
1,4-addition of 3a to 2a.
none
MeO-3,5-(CF3)2-
BIPHEP
8
92
none
To design the effective ligand, the insertion step and
hydrolysis of 5 were estimated by density functional
theory (DFT) calculations. Although some theoretical
consideration of enantioselectivity in rhodium-
catalyzed asymmetric 1,4-addtion were reported,[29]
there are few reports evaluating the catalytic reaction
rate by theoretical calculations.[30] Furthermore no
examples were reported for hydrolysis of
intermediate Rh-Ph complex 5.
[a]Reaction conditions: 2a (4.32 mmol), [RhCl(C2H4)2]2
(2.16 μmol, 0.1 mol% of Rh), (R)-ligand (4.32 μmol, 0.1
mol%), PhB(OH)2 (4.32 mmol, 1.0 equiv.) and sat.
NaHCO3 (5.8 mL) in 1,4-dioxane (4.3 mL) under argon
atmosphere at 25 °C for 6 h, unless otherwise specified.
1
[b]Remaining quantity of 3a was calculated by H NMR.
[c]PhB(OH)2 (15.1 mmol, 3.5 equiv.). [d]At 50 °C.
All structures, including transition states, were fully
optimized and characterized using frequency
calculations at the B97D functional[31] with the 6-
31G(d) basis set[32] for the organic molecules and the
LanL2DZ[32] basis set (with effective core potentials)
for rhodium using Gaussian 09, revision E.01.[33] The
B97D functional was chosen by the results of
comparison with other functions (B3LYP and
M062X: See Supporting Information). Gibbs free
energies (298.15 K, 1 atm) were initially computed
for the gas phase. Relative Gibbs free energies in
solvent were obtained using single-point energy
calculations of the optimized structures at the same
level with the SCRF method based on CPCM (1,4-
dioxane)[34] followed by the addition of thermal
corrections, which were calculated using the
Although abnormal consumption of 3a in Hayashi–
Miyaura coupling is known as protodeboronation of
the arylboronic acid,[12c] it was evident that this was
not simply hydrolysis of 3a with H2O because the
abnormal consumption of 3a was not observed in the
reaction with a highly reactive cyclohexenone
substrate.[26] Therefore, we had an insight into the
catalytic cycle of the asymmetric 1,4-addition of 3a
to 2a according to the catalytic cycle proposed by
Hayashi et al.[27] (Figure 1). The active species of
[RhOH(ligand)]2 4, which is generated from
[RhCl(ligand)]2 with base and H2O, undergoes
transmetalation with 3a to give complex 5. The
insertion reaction of 5 coordinated with 2a (5•••2a)
forms the C-C bond asymmetrically. The hydrolysis
3
This article is protected by copyright. All rights reserved.

10.1002/adsc.201701191
Advanced Synthesis & Catalysis
geometrical optimization mentioned above. As the
initial calculation structures of Rh/MeO-F12-BIPHEP
complexes, a partial X-ray structure of [RhCl(meo-
f12-biphep)]2 was used.[26a]
The energy diagrams for both the insertion of 2a to
complex 5, giving the (R)-product as a favorable
enantiomer,[35] and the hydrolysis of complex 5, i.e.,
decomposition process of 3a, are summarized in
Figure 2. The models used existing chiral ligands:
(R)-MeO-F12-BIPHEP (black) or (R)-MeO-BIPHEP
(blue).
Figure 2. Energy diagrams of insertion or hydrolysis of 5. Relative Gibss free energies (kcal/mol) obtained by single
point energy calculations with the SCRF method based on CPCM (1,4-dioxane) are shown. The energies of TS≠ins
described in plain text were for transition states giving R-products. The energies for transition states giving S-products
were described in italic with dot line and #.
The insertion, or the hydrolysis, of 5 proceeds via
TS≠ins, or TS≠hyd, after coordination with the enone
substrate (5•••sub), or H2O (5•••2H2O) (See
Supporting Information, Figure S1). Although the
activation energy of insertion had been calculated
from the difference between complex 5 and TS≠ins [29]
,
it is necessary to carefully consider the starting point
of insertion step to evaluate the reaction rate. After
formation of 5 by transmetalation of PhB(OH)2 (3a)
to rhodium in the presence of NaHCO3, 5 was
stabilized by coordination with NaOB(OH)2 to give
complex 5•••NaOB(OH)2,[30] which was the most
stable complex in the 5•••donating molecules (2a,
H2O, 1,4-dioxane, NaHCO3, and NaOB(OH)2)
(Scheme
2).
The
most
stable
complex
5•••NaOB(OH)2 would be the mutual short-lived
resting-state in both insertion and hydrolysis
processes, in other words, the complex is starting
point of both insertion and hydrolysis.
Scheme 2. Complex 5 with donating molecules.[a],[b]
4
This article is protected by copyright. All rights reserved.

10.1002/adsc.201701191
Advanced Synthesis & Catalysis
Next, the insertion step was evaluated. Although the
activation free energy for insertion (ΔG≠ins) was
evaluated by (energy of TS≠ins) – (energy of 5•••sub),
the hydrolysis of 5, in agreement with the
experimental result (Table 1). Therefore, suppression
of the hydrolysis of 5 in a real system would need a
negative ΔΔG≠ value obtained by decreasing the
ΔG≠C-C or increasing the ΔG≠hyd values.
the kins would be evaluated using ΔG≠C-C (not ΔG≠ins
)
which was calculated by (energy of TS≠ins) – (energy
of 5•••NaOB(OH)2). In the case of MeO-F12-BIPHEP
(black in Figure 2), complex 5 was stabilized by
NaOB(OH)2 (5•••NaOB(OH)2) at -18.66 kcal/mol.
The ligand exchange of NaOB(OH)2 to 2a via
unsaturated complex 5 to give 5•••sub increased the
energy level to -8.52 kcal/mol. The insertion
proceeded via TS≠ins to give intermediate complex 6.
The ΔG≠ins was 11.42 kcal/mol, and it was larger than
the case of highly-reactive cyclohexenone by 2.03
kcal/mol (gray in Figure 2), indicating that substrate
2a was a less reactive substrate than cyclohexenone.
On the contrary, the ΔG≠ins[MeO-F12-BIPHEP, 2a]
was smaller than that of MeO-BIPHEP by 2.71
kcal/mol, indicating that the electron-poor nature of
As is evident from the results (Table 1), the
introduction of too strong electron-withdrawing
groups to the ligand does not lead to an improvement
in the reaction yield, and thus, the structures of the
TS≠ins that give the favorable enantiomer were
dissected to design the ligand that would decrease the
ΔΔG≠ value. The structures of TS≠inss, bearing (a)
(R)-MeO-BIPHEP, (b) (R)-MeO-F12-BIPHEP, and (c)
(R)-MeO-3,5-(CF3)2-BIPHEP, are depicted in Figure
3. Steric (or electronic) repulsion was observed
between π-electron of 6-methylcoumarin and lone
pair of the meta-fluorine on the pendant aryl groups
of MeO-F12-BIPHEP, resulting in the aryl group
being tilted away from 6-methylcoumarin (D(Rh-P-
Cipso-Cortho) = 51.2° for MeO-F12-BIPHEP vs. D(Rh-
P-Cipso-Cortho) = 41.4° for MeO-BIPHEP). The larger
trifluoromethyl group in MeO-3,5-(CF3)2-BIPHEP
further tilted the aryl group (D(Rh-P-Cipso-Cortho) =
66.8°). In contrast, the attractive CH-π interaction
between ortho- and meta-protons in the pendant
phenyl group and the aromatic plane of 6-
methylcoumarin was observed in the MeO-BIPHEP
system: this was because the C∙∙∙H distances were
2.69 or 2.73 Å, respectively, which are shorter than
sum of their van der Waals radii of C and H (2.9 Å).
Although the activation energy for insertion (ΔG≠ins
or ΔG≠C-C) of the MeO-BIPHEP case was larger than
that of electron-poor MeO-F12-BIPHEP, the CH-π
interaction ought to contribute to the stabilization of
its transition state.[7e,38,39] In other words, if such an
attractive interaction is introduced into an electron-
poor ligand system, it is expected that the activation
energy will further decrease.
MeO-F12-BIPHEP
accelerated
the
insertion
reaction.[36] Furthermore, in the case of MeO-
BIPHEP, more stabilized 5•••NaOB(OH)2 increased
ΔG≠C-C, as compared to that of MeO-F12-BIPHEP
(27.61 vs. 21.56 kcal/mol). As a result, MeO-
BIPHEP was shown to be ineffective for insertion.
The energy trends were in agreement with the
experimental results in Table 1.
Next, in view of the ΔG≠hyd, the favorable route for
hydrolysis of
5
was the pathway involving
bimolecular H2O (not unimolecular or termolecular
H2O), revealed by the estimation of transition
states.[37] In a similar way to ΔG≠C-C, ΔG≠hyd was
calculated by (energy of TS≠hyd) – (energy of
5•••NaOB(OH)2) to evaluate khyd. On calculating
ΔΔG≠ (=ΔG≠C-C – ΔG≠hyd) to compare ΔG≠C-C and
ΔG≠hyd values, the ΔΔG≠ value in the case of MeO-
F12-BIPHEP was 0.92 kcal/mol. The result indicates
that the insertion reaction is more unfavorable than
Figure 3. Transition states of insertion of rhodium/(R)-ligand with 6-methylcoumarin giving (R)-1aa as a favorable
enantiomer.
We designed chiral ligands L1–L3 for the rhodium-
catalyzed 1,4-addition to coumarin substrates (Figure
4). The pendant aryl groups of L1 (3-fluoro-4-
(trifluoromethyl)-3-pyridinyl group), which were
inspired by the consideration of the structure of TS≠ins
as mentioned above, have a meta-proton in the
trifluoromethylphenyl
group)
and
L2
(6-
expectation of
a
CH–π interaction with the
5
This article is protected by copyright. All rights reserved.

10.1002/adsc.201701191
Advanced Synthesis & Catalysis
coumarin’s plane. The both pendant aryl groups for
L1 (σ* = 1.21) and L2 (σ* = 1.20) show similar
electron-withdrawing natures as compared with the
3,4,5-trifluorophenyl group (σ* = 1.11) in MeO-F12-
BIPHEP.[40] It should be noted that there are no
examples of tertiary phosphines bearing the above
asymmetrical fluoroaromatic groups. Among them,
no phosphorous compounds bearing the group for L1
(3-fluoro-4-trifluoromethylphenyl group) had been
reported. On the contrary, the pendant aryl of L3, 3,5-
difluoro-4-trifluoromethyl phenyl group, was chosen
by conventional intuition that the electronic poor
ligand would accelerate the insertion. The group has
no meta-proton, but possesses a strong electron-
withdrawing ability (σ* = 1.46).[40] The effect of
fluorine substituents in pendant aryls on
enantioselectivity is expected to cause no problems
because all energies of TS≠ins giving the (S)-product
in both the MeO-F12-BIPHEP and MeO-BIPHEP
cases (dotted line with # in Figure 2) were higher by
over 4 kcal/mol than the TS≠ins giving the (R)-
product.[41] Before the syntheses of these ligands, the
ligand effects on insertion or hydrolysis of 5 were
estimated by DFT calculations in a similar manner as
mentioned above. The structures of TS≠ins, giving (R)-
1aa, bearing (a) (R)-L1, (b) (R)-L2, (c) (R)-L3 are
depicted in Figure 5. As expected, the attractive CH–
π interaction between the ortho- and meta-protons in
the pendant aryl group and the aromatic plane of the
6-methylcoumarin was observed in both (R)-L1 and
(R)-L2 systems, and not observed in the (R)-L3
system.
Figure 4. Design of chiral diphosphine ligands bearing
fluorofunctional groups.
Figure 5. Transition states of insertion of rhodium/designed-ligand with 6-methylcoumarin giving (R)-1aa.
The ΔG≠ins, ΔG≠C-C, and ΔG≠hyd values were calculated to 12.54 kcal/mol (Figure 6)[43]. The stabilization by
in a similar manner to Figure 2 (Table 2). As expected, CH–π interaction of the TS≠ins giving the (R)-product
the ΔG≠ins of (R)-L1 or (R)-L2 (10.61 or 11.27 resulted in the expansion of the energy difference
kcal/mol) were lower than that of MeO-F12-BIPHEP from TS≠ins giving the (S)-product by up to 9.07
system (11.42 kcal/mol). In particular, ΔG≠ins of (R)- kcal/mol (Figure 6).
L1 was decreased by 0.81 kcal/mol from that of
Table 2. Calculation Results for (R)-L1, (R)-L2, and (R)-
L3.[a]
MeO-F12-BIPHEP,
whose
energy
difference
corresponded to a 3.9-fold acceleration of insertion at
25 °C.[42] It is obvious that the decrease of ΔG≠ins is
due to the CH–π interaction contribution because
rotation of the pendant aryl group of (R)-L1 to the
opposite side, (i.e., no CH–π interaction as in the
MeO-F12-BIPHEP system) increases the ΔG≠ins value
≠ins[b]
5•••NaOB(OH)2
Ligand
ΔG
ΔG≠C-C[c] ΔG≠hyd[d] ΔΔG≠[e]
(R)-L1 -18.21
10.61 20.59 21.76 -1.17
6
This article is protected by copyright. All rights reserved.

10.1002/adsc.201701191
Advanced Synthesis & Catalysis
(R)-L2 -17.44
(R)-L3 -17.35
11.27 21.79
11.46 22.24
23.28 -1.49
22.40 -0.16
Scheme 3. Syntheses of chiral ligands L1 – L3.
The electronic properties of ligands L1, L2, and L3
were estimated using the νCO values of cis-
[RhCl(CO)(ligand)] complexes, which were easily
synthesized according to the typical method (Table
3).[44] The νCO values revealed that the electronic
properties of the ligands L1 – L3 were approximately
as expected. The electronic ability of L1 (2036 cm-1)
were identical with MeO-F12-BIPHEP, and that of L2
were slightly more electron-poor (2038 cm-1). More
electron-poor nature of L3 (2044 cm-1) was equal to
MeO-3,5-(CF3)2-BIPHEP.
[a]Relative free energies (kcal/mol) from complex 5
obtained by single point energy calculations with the SCRF
method based on CPCM (1,4-dioxane) are shown. [b]ΔG≠ins
(kcal/mol) = (energy of TS≠ins) – (energy of 5•••2a).
[c]ΔG≠C-C (kcal/mol) = (energy of TS≠ins) – (energy of
5•••NaOB(OH)2). [d]ΔG≠hyd (kcal/mol) = (energy of TS≠hyd
– (energy of 5•••NaOB(OH)2). [e]ΔΔG≠ =ΔG≠C-C – ΔG≠hyd.).
)
Table 3. Electronic Properties of Chiral Ligands[a]
CO (cm-1)
Ligand

MeO-BIPHEP
2014[b]
2036[b]
2036
MeO-F12-BIPHEP
Figure 6. Comparison of ΔG≠ins in TS≠ins[(R)-L1].
L1
2038
L2
The ΔG≠C-C of (R)-L1 (20.59 kcal/mol) also decreased
by 0.97 kcal/mol compared to that of MeO-F12-
BIPHEP (21.56 kcal/mol), whose energy difference
corresponded to a 5.1-fold acceleration of C–C
formation at 25 °C.[42] ΔG≠hyd of (R)-L1 was increased
to 21.76 kcal/mol; therefore, ΔΔG≠ was decreased to -
1.17 kcal/mol and the yield of the asymmetric 1,4-
addition was expected to improve. On the contrary,
the ΔG≠C-C of (R)-L2 (21.79 kcal/mol) did not
improve from the case of MeO-F12-BIPHEP.
However, the ΔG≠hyd of (R)-L2 was increased to
23.28 kcal/mol; therefore, the ΔΔG≠ was decreased to
-1.49 kcal/mol beyond the case of (R)-L1. Although
the improvement of the catalytic activity of
rhodium/(R)-L2 was not expected, inhibition of
hydrolysis of 5 was expected. On the other hand,
ligand (R)-L3 was not expected for the asymmetric
1,4-addition because the relatively high ΔG≠ins (11.46
kcal/mol) and ΔG≠C-C (22.24 kcal/mol) values would
decelerate the insertion step. To conclude the
predictions from the computations, good results can
be expected by the use of (R)-L1 or (R)-L2,
particularly (R)-L1, for the asymmetric 1,4-addition
of 3a to 2a. To test the theoretical predictions, the
ligands (R)-L1, (R)-L2, and (R)-L3 were synthesized.
The ligands were synthesized from known (R)-
tetrachlorophosphine intermediate 8, derived from
(R)-7 (Scheme 3).[26a] The reaction of (R)-8 with the
corresponding Grignard reagents gave the ligands (R)-
L1, (R)-L2, and (R)-L3 in moderated yields.
2044
L3
MeO-3,5-(CF3)2-BIPHEP
2044[b]
[a]The νCO values of [RhCl(phosphine)(CO)] in CHCl3. [b]ref.
44.
Similar to the experiments carried out for Table 1 data,
asymmetric 1,4-addition reactions of 3a to 2a using
the ligands L1 – L3 were performed (Table 4). As
predicted by the calculations described above, ligands
(R)-L1 and (R)-L2 improved the yield of (R)-1aa, and
ligand (R)-L3 deteriorated the yield as compared to
MeO-F12-BIHEP (entries 1, 9, 10 in Table 4 vs. entry
6 in Table 1). Especially, the yield of the reaction
using ligand (R)-L1 was increased up to 53% yield of
(R)-1aa without loss of enantioselectivity (entry 1).
The reaction conditions were further investigated by
using the most effective (R)-L1 ligand. The yields of
(R)-1aa were influenced by the organic solvent.
Although the reactions in Et2O or CH2Cl2 did not
progress much (entries 2 or 3), xylene and toluene
solvents improved the yields slightly, up to 65%, with
>99% ee (entry 4 and 5). Fortunately, toluene is a
more suitable solvent for an industrial synthesis than
1,4-dioxane which is a suspected carcinogen.[45]
When the amount of 3a was increased to 3.0 equiv.,
the yield of (R)-1aa was over 90% (despite S/C =
2,000) with the longer reaction time of 36 h (entry 6),
in contrast with the case using MeO-F12-BIPHEP
(Table 1, entry 7). The reaction using 0.033 mol%
(S/C = 3,000) or 0.025 mol% (S/C = 4,000)
proceeded using 3.5 equiv. 3a to give 91% or 76%
yield of (R)-1aa with >99% ee, respectively (entries 7
or 8). In view of cost in industrial application, 3.5
equiv. of 3a would be the maximum value, and S/C =
7
This article is protected by copyright. All rights reserved.

10.1002/adsc.201701191
Advanced Synthesis & Catalysis
2,000 or more is the amount of rhodium catalyst similar to that of 2a (entries 5–7). Although the
needed to satisfy the requirements for industrial reactions of 7-methoxycoumarin (2c) or 6-
synthesis.[46] Therefore, we judged that the ethoxycoumarin (2d) with 3a using conventional
appropriate catalyst and reaction conditions could be ligand had been known to be very less reactive,[17,20]
discovered.
the corresponding products were produced in 73% or
60% yield, respectively (entries 8, 9). However, the
reaction of 6-hydroxycoumarin (2e) with 3a did not
proceed at all because phenolic hydroxyl groups
deactivate the rhodium catalyst.[47]
Table 4. Rhodium-catalyzed Asymmetric 1,4-Addition of
3a to 2a using Ligands L1 – L3 [a]
Table 5. Rhodium-catalyzed Asymmetric 1,4-Addition of
Some Coumarin Derivatives using (R)-1[a]
Entry Ligand S/C Solvent Yield [%] ee [%] 3a[b]
1
2
3
4
5
(R)-L1 1,000 dioxane 53
>99
none
(R)-L1 1,000 Et2O
< 1
6
ca. 90%
ca. 90%
none
Entry
Ar of 3
Yield [%] ee [%]
71 (R)-1ab >99
82 (R)-1ac >99
2a
2a
2a
2a
(R)-L1 1,000 CH2Cl2
>99
>99
>99
>99
>99
>99
>99
99
1
2
3
4
4-Me-C6H4 3b
3-Me-C6H4 3c
(R)-L1 1,000 xylene 60
(R)-L1 1,000 toluene 65
none
3,5-Me2-C6H3 3d 87 (R)-1ad >99
6[c] (R)-L1 2,000 toluene 92
7[d] (R)-L1 3,000 toluene 91
8[e] (R)-L1 4,000 toluene 76
none
4-Cl-C6H4 3e
25 (R)-1ae >99
none
O
O
2b
none
5
6
95 (R)-1ba 99
2b
2b
3a
3b
9
(R)-L2 1,000 dioxane 33
(R)-L3 1,000 dioxane 2
none
76
(R)- >99
1bb
10
ca. 50%
7
8
82 (R)-1bc 99
73 (R)-1ca >99
3c
[a]Reaction conditions: 2a (4.32 mmol), [RhCl(C2H4)2]2
(2.16 μmol, 0.1 mol% of Rh), (R)-ligand (4.32 μmol, 0.1
mol%), 3a (4.32 mmol, 1.0 equiv.) and sat. NaHCO3 (5.8
mL) in organic solvent (4.3 mL) under argon atmosphere at
25 °C for 6 h, unless otherwise specified. [b]Remaining
MeO
O
O
2c
EtO
3a
O
O
O
1
quantity of 3a was calculated by H NMR. [c]2a (4.32
2d
mmol), [RhCl(C2H4)2]2 (1.08 μmol, 0.05 mol% of Rh), (R)-
L1 (2.16 μmol, 0.05 mol%), 3a (13.0 mmol, 3.0 equiv.)
and sat. NaHCO3 (5.8 mL) in toluene (4.3 mL) for 36 h.
[d]2a (6.48 mmol), [RhCl(C2H4)2]2 (1.08 μmol, 0.033 mol%
of Rh), (R)-L1 (2.16 μmol, 0.033 mol%), 3a (22.7 mmol,
3.5 equiv.) and sat. NaHCO3 (8.7 mL) in toluene (6.5 mL)
for 48 h. [e]2a (8.64 mmol), [RhCl(C2H4)2]2 (1.08 μmol,
0.025 mol% of Rh), (R)-L1 (2.16 μmol, 0.025 mol%), 3a
(30.2 mmol, 3.5 equiv.) and sat. NaHCO3 (11.6 mL) in
toluene (8.6 mL) for 60 h.
HO
9
60 (R)-1da >99
3a
3a
O
2e
2a
10
0 (R)-1ea -
[a]Reaction conditions: 2a (4.32 mmol), [RhCl(C2H4)2]2
(1.08 μmol, 0.1 mol% of Rh), (R)-L1 (2.16 μmol, 0.1
mol%), 3a (13.0 mmol, 3.0 equiv.), and sat. NaHCO3 (5.8
mL) in organic solvent (4.3 mL) under argon atmosphere at
25 °C for 36 h.
We carried out the asymmetric 1,4-addition of various
coumarin substrates 2 with 3.0 equiv. of arylboronic
acids 3 in the presence of 0.05 mol% of rhodium/(R)-
L1 catalyst (S/C = 2000) (Table 5). The arylboronic
acids bearing electron-donating groups (3b–3d) gave
high to moderate yields of the corresponding (R)-1
with >99% ee (entries 1-3). In contrast, the yield of
the reaction with arylboronic acid bearing an electron-
withdrawing group (3e) was low because
decomposition of 3e was too fast (entry 4). The
reaction activity of non-substituted coumarin (2b) was
Although the amount of both the rhodium/(R)-L1
catalyst and 3a for the catalytic synthesis of (R)-1aa
could be largely reduced, it is important to follow the
transformation reaction to a drug molecule for
industrial synthesis.[23] Hayashi et al. reported the
synthesis of (R)-tolterodine as follows: reduction of
(R)-1aa, which was prepared from 2a with 10 equiv
of 3a in the presence of 3.0 mol% of rhodium/(R)-
SEGPHOS
catalyst
in
1,4-dioxane,
using
8
This article is protected by copyright. All rights reserved.

10.1002/adsc.201701191
Advanced Synthesis & Catalysis
diisobutylaluminum hydride; followed by 10 mol% of In summary, we succeeded in the development of a
Pd-catalyzed reductive amination with chiral ligand for the rhodium-catalyzed asymmetric
diisopropylamine in MeOH under H2 atmosphere (50 1,4-addition to a coumarin substrate with low catalyst
psi) to produce (R)-tolterodine.[17] This method is loading. The ligand was designed through the
problematic for industrial application because of the interplay between theoretical and experimental
large amounts of rhodium catalyst and 3a used, but approaches. The intimate theoretical calculations,
also because of the large amount of palladium catalyst including estimation of transition states, were based
used for reductive amination. The large amount of on experimental results and enabled rational ligand
palladium under a hydrogen atmosphere may increase design. The resulting ligand (R)-L1 decreased the
the danger for a serious accident on the industrial catalyst loading of rhodium catalyst to less than one
scale. Furthermore, it is desirable for chromatographic one-hundredth of past catalyst loadings with almost
purifications to be avoided in the industrial process, complete enantioselectivity. To our knowledge, this
particularly silica-gel column chromatography.[17,48] study is the first successful example of drastically
We tried the transformation of (R)-1aa to (R)- decreasing catalyst loading using the interplay
tolterodine to 1) avoid using hydrogen gas in a between theoretical and experimental approaches. The
catalytic reductive amination reaction and 2) avoid L1 could be found out by using the interplay method
purification of the synthetic intermediates (Scheme 4). because the unusual 3-fluoro-4-trifluoromethylphenyl
The reduction of 15.07 g (>99% ee) of (R)-1aa with group, which had never been used for the phosphine
DIBAL was carried out in a similar manner to ref. 17. derivatives, was hard to notice by intuition without
After extraction and removal of the organic solvent, computationally assist. We are now trying the
crude product 9 was obtained in 16.74 g. The computationally-led design method to develop
reductive amination of 9 with diisopropylamine another effective catalyst in a different catalytic
without H2 gas succeeded when using Ir catalyst and system.
HCO2H. The reaction could be performed using crude
product 9 in the presence of only ca. 0.1 mol% of
Cp*IrCl[8-quinolinolate] catalyst[49] to give 20.76 g of
Experimental Section
crude product 10. According to 1H NMR
General experimental methods.
spectroscopy, after extraction and removal of the
organic solvent, the conversion yield of 10 from (R)-
1aa was 70%. Finally, a portion of crude product 10
(1.07 g) without further purification was reacted with
(L)-tartrate. The generated solid was filtered and
washed with cooled EtOH, followed by drying under
vacuum to give 0.94 g of pure Detrusitol® ((R)-
Tolterodine (L)-tartrate) with >99% ee. Although the
total yield of 60% from (R)-1aa to Detrusitol® was
not very high, the synthetic method -- without using
hydrogen gas and without the need of intermediary
purification -- will enable large to industrial scale
syntheses.
All reactions were carried out under an argon atmosphere
with dry solvents under anhydrous conditions, unless
otherwise noted. All solvents were purchased from Kanto
Chemical Co. and then were stored in Schlenk tubes under
an argon atmosphere. H2O was purified by distillation prior
to use. Reagents were purchased at the highest commercial
quality and used without further purification, unless
otherwise noted. Preparative column chromatography was
carried out by using silica gel (Kanto Chemical Co. 60N,
1
63-210 μm). H NMR spectra were measured at 400 MHz
using tetramethylsilane (TMS) as an internal standard (δ 0
ppm). 13C NMR spectra were measured at 101 MHz, and
chemical shifts are given relative to chloroform-d (δ 77.16
ppm). 19F NMR spectra were measured at 376 MHz, and
chemical shifts are given relative to CCl3F using C6F6 as
secondary reference (-162.9 ppm). 31P NMR spectra were
measured at 162 MHz, and chemical shifts are given
relative to 85% H3PO4 externally. IR spectra were
measured at resolution 4.0 cm-1 or 0.1 cm-1 using JASCO
FTIR-4200.
Synthesis
of
(R)-(6,6’-dimethoxybiphenyl-2,2’-
diyl)bis[bis(3-fluoro-4-
trifluoromethylphenyl)phosphine] ((R)-L1).
A flame-dried 10 mL screw-cap test tube was flushed with
argon and charged with triphosgene (594 mg, 2.0 mmol)
was added a solution of Aliquat®336 (47 mg, 0.12 mmol)
in toluene (3.3 mL). And then, the solution was stirred at
30 °C for 2 days. To the Schlenk flask containing (R)-7
(278 mg, 1.0 mmol) under argon was added 7.5 mL of dry,
deoxygenated CH2Cl2 and cooled to -78 °C. After cooling,
the solution in screw-cap tube was added to Schlenk flask
containing (R)-7 via syringe, and the solution was allowed
to warm to r.t. The bright yellow solution was stirred for 2
days. Concentration under vacuum gave the crude (R)-8. A
flame-dried 100 ml three-necked, round-bottomed flask
was flushed with argon and charged with magnesium
turnings (365 mg, 15.0 mmol), LiCl (317 mg, 7.5 mmol)
and 18.3 ml of Et2O. A solution of DIBAL in hexane (1.0
M, 100 μL, 0.10 mmol) was added and stirred for 5 min.
Then 4-bromo-2-fluorobenzotrifluoride (850μL, 6.0 mmol)
Scheme 4. Gram-scale synthesis of Detrusitol® ((R)-
tolterodine (L)-tartrate) from (R)-1aa.
Conclusion
9
This article is protected by copyright. All rights reserved.

10.1002/adsc.201701191
Advanced Synthesis & Catalysis
was added and the reaction mixture was stirred for 1 h.
Then, a solution of (R)-8 (1.0 mmol) in 1.8 ml of THF was
added dropwise over 5 min. The solution was stirred at
40 °C for 4 h, and then saturated aqueous NH4Cl solution
was added. After extracted with EtOAc (three times), the
organic layer was dried over Na2SO4, filtrated, and
concentrated under reduced pressure. The resulting solid
was purified by silica gel column chromatography
(Hexane/EtOAc = 8/1) to give (R)-L1 as a white solid (413
mg, 0.45 mmol, 45% yield) (See Supporting Information
for characterization details).
Biomol. Chem. 2014, 12, 2745-2753; d) B. J. Rooks, M.
R. Haas, D. Sepulveda, T. Lu, S. E. Wheeler, ACS
Catal. 2015, 5, 272-280; e) S. E. Wheeler, T. J. Seguin,
Y. Guan, A. C. Doney, Acc. Chem. Res. 2016, 49,
1061-1069; f) A. C. Doney, B. J. Rooks, T. Lu, S. E.
Wheeler, ACS Catal. 2016, 6, 7948-7955; g) J. de
Oliveira Silva, R. A. Angnes, V. H. Menezes da Silva,
B. M. Servilha, M. Adeel, A. A. C. Braga, A. Aponick,
C. R. D. Correia, J. Org. Chem. 2016, 81, 2010-2018;
h) A. J. Neel, A. Milo, M. S. Sigman, F. D. Toste, J. Am.
Chem. Soc. 2016, 138, 3863-3875.
Syntheses of the L2 and L3 were described in Supporting
Information
[8] M. C. Nielsen, K. J. Bonney, F. Schoenebeck, Angew.
Chem., Int. Ed. 2014, 53, 5903-5906.
General procedure for rhodium-catalyzed asymmetric
1,4-addition of phenylboronic acid to 6-methylcoumarin.
[9] H. Tanaka, Y. Nishibayashi, K. Yoshizawa, Acc. Chem.
Res. 2016, 49, 987-995.
A 20 mL Schlenk flask was flushed with argon and charged
with [RhCl(C2H4)2]2 (0.42 mg, 1.08 μmol, S/C=2,000), (R)-
L1 (2.00 mg, 2.16 μmol), and toluene (0.8 mL) was stirred
at room temperature for 10 min. This mixture was
transferred to a 50 mL Schlenk flask flushed with argon
and charged with 6-methylcoumarin (692 mg, 4.32 mmol),
PhB(OH)2 (1.58 g, 13.0 mmol), toluene (3.5 mL) and sat.
NaHCO3 aq. (5.8 mL) via cannula. The resulting mixture
was stirred at 25 °C for 36 h. The reaction mixture was
extracted with EtOAc. The organic layer were dried over
Na2SO4 and concentrated under reduced pressure. The
crude material was purified by silica gel column
chromatography with hexane/EtOAc = 4 : 1 to give (R)-6-
methyl-4-phenylchroman-2-one ((R)-1aa) as a white solid
(951 mg, 3.99 mmol, 92% yield, >99% ee) (See Supporting
Information for characterization details).
[10] R. N. Straker, Q. Peng, A. Mekareeya, R. S. Paton, E.
A. Anderson, Nature Commun. 2016, 7, 10109.
[11] Y. Takaya, M. Ogasawara, T. Hayashi, M. Sakai, N.
Miyaura, J. Am. Chem. Soc. 1998, 120, 5579-5580.
[12] Recent reviews: a) M. M. Heravi, M. Dehghani, V.
Zadsirjan, Tetrahedron: Asymmetry 2016, 27, 513-588;
b) M. Mauduit, O. Basle, H. Clavier, C. Crevisy, A.
Denicourt-Nowicki in Comprehensive Organic
Synthesis II, 2nd ed. (Eds.: P. Knochel, G. A.
Molander) Elsevier, Amsterdam, 2014; Vol. 4, pp 189-
341; c) P. Tian, H.-Q. Dong, G.-Q. Lin, ACS Catal.
2012, 2, 95-119; d) G. Berthon-Gelloz, T. Hayashi in
Boronic Acids, 2nd ed. (Ed.: D. G. Hall) Wiley-VCH,
Weinheim, Germany, 2011, Vol. 1, pp 263-313; e) G.
Berthon, T. Hayashi in Catalytic Asymmetric Conjugate
Reactions, (Ed.: A. Cordova) Wiley-VCH, Weinheim,
Germany, 2010, pp 1-70.
Acknowledgements
This work was supported in part by JSPS KAKENHI (Grant-in
Aid for Scientific Research (C), no. 16K05764 (T.K.)) and Grants
at Iwate University (T.K.).
[13] a) M. Jean, B. Casanova, S. Gnoatto, P. van de Weghe,
Org. Biomol. Chem. 2015, 13, 9168-9175; b) H. J.
Edwards, J. D. Hargrave, S. D. Penrose, C. G. Frost,
Chem. Soc. Rev. 2010, 39, 2093-2105.
References
[1] K. N. Houk, P. H.-Y. Cheong, Nature 2008, 455, 309-
313.
[14] J. Magano, J. R. Dunetz, Chem. Rev. 2011, 111, 2177-
[2] a) D. J. Tantillo, Acc. Chem. Res. 2016, 49, 1079; b) C.
Poree, F. Schoenebeck, Acc. Chem. Res. 2017, 50, 605-
608.
2250.
[15] a) J. Magano, J. R. Dunetz in New Trends in Cross-
coupling:Theory and Applications (Ed.: J. J. Spivey)
Royal Society of Chemistry, Cambridge, UK, 2015, Vol.
21, pp 697-778; b) A. Parker in Transition Metal-
Catalyzed Couplings in Process Chemistry (Eds.: J.
Magano, J. R. Dunetz) Wiley-VCH, Weinheim,
Germany, 2013, pp 121-134; c) G. P. Howell, Org.
Process Res. Dev. 2012, 16, 1258-1272; d) K. Lukin, Q.
Zhang, M. R. Leanna, J. Org. Chem. 2009, 74, 929-
931; e) S. Brock, D. R. J. Hose, J. D. Moseley, A. J.
Parker, I. Patel, A. J. Williams, Org. Process Res. Dev.
2008, 12, 496-502.
[3] a) T. Sperger, I. A. Sanhueza, I. Kalvet, F. Schoenebeck,
Chem. Rev. 2015, 115, 9532-9586; b) X. Zhang, L. W.
Chung, Y.-D. Wu, Acc. Chem. Res. 2016, 49, 1302-
1310.
[4] K. Rohmann, M. oelscher, W. Leitner, J. Am. Chem.
Soc. 2016, 138, 433-443.
[5] See Acc. Chem. Res. special issue, 2016
“Computational Catalysis for Organic Synthesis”.
[6] a) A. S. K. Tsang, I. A. Sanhueza, F. Schoenebeck,
Chem. - Eur. J. 2014, 20, 16432-16441; b) G.-J. Cheng,
X. Zhang, L. W. Chung, L. Xu, Y.-D. Wu, J. Am. Chem.
Soc. 2015, 137, 1706-1725.
[16] a) Y. Yoshinaga, J. Pharm. Soc. Japan 2009, 129,
241-245; b) P. Abrams, Expert opinion on pharm. 2001,
2, 1685-1701.
[7] a) J. C. Ianni, V. Annamalai, P.-W. Phuan, M. Panda,
M. C. Kozlowski, Angew. Chem., Int. Ed. 2006, 45,
5502-5505; b) K. C. Harper, M. S. Sigman, Science
2011, 333, 1875-1878; c) G. Jindal, R. B. Sunoj, Org.
[17] G. Chen, N. Tokunaga, T. Hayashi, Org. Lett. 2005, 7,
2285-2288.
10
This article is protected by copyright. All rights reserved.

10.1002/adsc.201701191
Advanced Synthesis & Catalysis
[18] T. Korenaga, R. Maenishi, K. Osaki, T. Sakai,
H.-L. Qin, E. A. B. Kantchev, Dalton Trans. 2015, 44,
Heterocycles 2010, 80, 157-162.
2737-2746.
[19] Y. Luo, A. J. Carnell, Angew. Chem., Int. Ed. 2010, 49, [31] S. Grimme, J. Comput. Chem. 2006, 27, 1787-1799.
2750-2754.
[32] For Gaussian basis sets: a) M. J. Frisch, J. A. Pople, J.
[20] T. Mino, K. Miura, H. Taguchi, K. Watanabe, M.
Sakamoto, Tetrahedron: Asymmetry 2015, 26, 1065-
1068.
S. Binkley, J. Chem. Phys. 1984, 80, 3265-3269; b) W.
J. Hehre, L. Radom, P. v. R. Schleyer, J. A. Pople in Ab
initio Molecular Orbital Theory; John Wiley: New
York, USA, 1986 , and references cited therein.
[21] C. Defieber, J.-F. Paquin, S. Serna, E. M. Carreira,
Org. Lett. 2004, 6, 3873-3876.
[33] Gaussian 09, Revision E.01; Gaussian, Inc.,
Wallingford, CT, 2009.
[22] No reaction product was afforded: K. Vandyck, B.
Matthys, M. Willen, K. Robeyns, L. Van Meervelt, Org. [34] V. Barone, M. Cossi, J. Phys. Chem. A 1998, 102,
Lett. 2006, 8, 363-366. 1995-2001.
[23] Problems occurring in a catalytic reaction for [35] Transition states in insertion step of Rh-diphosphine
industrial scale: C. A. Busacca, D. R. Fandrick, J. J.
Song, C. H. Senanayake in Applications of Transition
Metal Catalysis in Drug Discovery and Development
(Eds.: M. L. Crawley, B. M. Troust) John Wiley &
Sons, Hoboken, USA, 2012, pp 1-24.
complexes: see refs. 29d, 29h, 29i, 29j.
[36] Q. Chen, B.-L. Lin, Y. Fu, L. Liu, Q.-X. Guo, Res.
Chem. Intermed. 2005, 31, 759-767.
[37] ΔG≠hyd(2H2O) = 20.64 kcal/mol vs. ΔG≠hyd(H2O) =
23.24 kcal/mol or ΔG≠hyd(3H2O) = 21.08 kcal/mol in the
cases of MeO-F12-BIPHEP. See Figure S1 in
Supporting Information.
[24] J.-P. Genet, T. Ayad, V. Ratovelomanana-Vidal, Chem.
Rev. 2014, 114, 2824-2880.
[25] Review for ligand electronic effects in homogeneous
catalysis: M. L. Clarke, J. J. R. Frew, Organomet. Chem.
2009, 35, 19-46.
[38] a) A. J. Neel, M. J. Hilton, M. S. Sigman, F. D. Toste,
Nature 2017, 543, 637-646; b) E. H. Krenske, K. N.
Houk, Acc. Chem. Res. 2013, 46, 979-989.
[26] a) T. Korenaga, K. Osaki, R. Maenishi, T. Sakai, Org
Lett 2009, 11, 2325-2328; b) T. Korenaga, R. Maenishi,
K. Hayashi, T. Sakai, Adv. Synth. Catal. 2010, 352,
3247-3254.
[39] The B97D functionals afforded good results for
consideration of CH-π interaction: S. Tsuzuki, Annu.
Rep. Prog. Chem., Sect. C: Phys. Chem. 2012, 108, 69-
95.
[27] T. Hayashi, M. Takahashi, Y. Takaya, M. Ogasawara,
[40] Taft’s σ* values were evaluated according to the
following reference: Y. Nagai, H. Matsumoto, T.
Nakano, H. Watanabe, Bull. Chem. Soc. Jap. 1972, 45,
2560-2565.
J. Am. Chem. Soc. 2002, 124, 5052-5058.
[28] The rate-determining step of the reaction using
reactive substrate is transmetalation: A. Kina, H.
Iwamura, T. Hayashi, J. Am. Chem. Soc. 2006, 128,
3904-3905.
[41] The energy difference by 3.1 kcal/mol is a value that
corresponds to the experimentally measured 99% ee at
25 °C.
[29] a) G.-F. Zha, H.-L. Qin, E. A. B. Kantchev, Org.
Biomol. Chem. 2017, 15, 2226-2233; b) J. D. Sieber, D.
Rivalti, M. A. Herbage, J. T. Masters, K. R. Fandrick,
D. R. Fandrick, N. Haddad, H. Lee, N. K. Yee, B. F.
Gupton, C. H. Senanayake, Org. Chem. Front. 2016, 3,
1149-1153; c) Y.-G. Li, L. Li, M.-Y. Yang, H.-L. Qin,
E. A. B. Kantchev, RSC Adv. 2015, 5, 5250-5255; d)
H.-L. Qin, X.-Q. Chen, Y.-Z. Huang, E. A. B. Kantchev,
Chem. - Eur. J. 2014, 20, 12982-12987; e) E. A. B.
Kantchev, Chem. Sci. 2013, 4, 1864-1875; f) S.
Gosiewska, J. A. Raskatov, R. Shintani, T. Hayashi, J.
M. Brown, Chem. - Eur. J. 2012, 18, 80-84; g) E. A. B.
Kantchev, Chem. Commun. 2011, 47, 10969-10971; h)
A. Poater, F. Ragone, R. Mariz, R. Dorta, L. Cavallo,
Chem. - Eur. J. 2010, 16, 14348-14353; i) P. Mauleon, I.
Alonso, M. Rodriguez Rivero, J. C. Carretero, J. Org.
Chem. 2007, 72, 9924-9935; j) T. Itoh, T. Mase, T.
Nishikata, T. Iyama, H. Tachikawa, Y. Kobayashi, Y.
Yamamoto, N. Miyaura, Tetrahedron 2006, 62, 9610-
9621.
[42] The ratio of reaction rate was calculated by equation
of e(ΔG≠[MeO-F12-BIPHEP]-ΔG≠[L1])/RT
.
[43] In the calculations of L2, similar results were obtained.
See Figure S2 in supporting information.
[44] T. Korenaga, A. Ko, K. Uotani, Y. Tanaka, T. Sakai,
Angew. Chem. Int Ed Engl 2011, 50, 10703-10707.
[45] U.S. Environmental Protection Agency. Integrated
Risk Information System (IRIS). 1,4-Dioxane (CASRN
123-91-1), 2013.
[46] S. Huebner, J. G. de Vries, V. Farina, Adv. Synth.
Catal. 2016, 358, 3-25.
[47] F.-X. Chen, A. Kina, T. Hayashi, Org. Lett. 2006, 8,
341-344.
[48] Separation processes account for 40−70% of both
capital and operating costs in the chemical and
pharmaceutical industries: P. Marchetti, M. F. Jimenez
Solomon, G. Szekely, A. G. Livingston, Chem. Rev.
2014, 114, 10735-10806, and references therein.
[30] Evaluation of transmetalation as a rate determining
step by DFT calculations: a) ref. 29a. b) Y.-G. Li, G. He,
11
This article is protected by copyright. All rights reserved.

10.1002/adsc.201701191
Advanced Synthesis & Catalysis
[49] a) M. Watanabe, J. Hori, K. Murata, Patent JP5719115,
2015; b) M. Watanabe, J. Hori, K. Murata, Patent
EP2228377, 2014; c) M. Watanabe, J. Hori, K. Murata,
Patent US9211533, 2015.
12
This article is protected by copyright. All rights reserved.

10.1002/adsc.201701191
Advanced Synthesis & Catalysis
FULL PAPER
Computationally-led Ligand Modification using
Interplay between Theory and Experiments: Highly
Active Chiral Rhodium Catalyst Controlled by
Electronic Effects and CH–π Interactions
Adv. Synth. Catal. Year, Volume, Page – Page
Toshinobu Korenaga,* Ryo Sasaki, Toshihide
Takemoto, Toshihisa Yasuda, Masahito Watanabe
13
This article is protected by copyright. All rights reserved.