Catalytic enantioselective acyl transfer: the case for 4-PPY with a C-3 carboxamide peptide auxiliary based on synthesis and modelling studies

Article abstract of DOI:10.1039/C6OB01991A

A series of 4-pyrrolidinopyridine (4-PPY) C-3 carboxamides containing peptide-based side chains have been synthesised and evaluated in the kinetic resolution of a small library of chiral benzylic secondary alcohols. A key design element was the incorporation of a tryptophan residue in the peptide side chain for promoting π-stacking between peptide side chain and the pyridinium ring of the N-acyl intermediate, in which modelling was used as a structure-based guiding tool. Together, a catalyst containing a LeuTrp-N-Boc side chain (catalyst 8) was identified that achieved s-values up to and in slight excess of 10. A transition-state model based on the modelling is proposed to explain the origin of enantioselectivity. This study establishes the usefulness of modelling as a structure-based guiding tool for enantioselectivity optimization as well as the potential for developing scalable peptide-based DMAP-type catalysts for large-scale resolution work.

Full text of DOI:10.1039/C6OB01991A

Organic &
Biomolecular Chemistry
View Article Online
View Journal
PAPER
Catalytic enantioselective acyl transfer: the case
for 4-PPY with a C-3 carboxamide peptide
auxiliary based on synthesis and modelling
studies†
Cite this: DOI: 10.1039/c6ob01991a
Rudy E. Cozett,a Gerhard A. Venter,a,b Maheswara Rao Gokadaa and Roger Hunter*a
A series of 4-pyrrolidinopyridine (4-PPY) C-3 carboxamides containing peptide-based side chains have
been synthesised and evaluated in the kinetic resolution of a small library of chiral benzylic secondary
alcohols. A key design element was the incorporation of a tryptophan residue in the peptide side chain for
promoting π-stacking between peptide side chain and the pyridinium ring of the N-acyl intermediate, in
which modelling was used as a structure-based guiding tool. Together, a catalyst containing a LeuTrp-N-
Boc side chain (catalyst 8) was identified that achieved s-values up to and in slight excess of 10. A transi-
tion-state model based on the modelling is proposed to explain the origin of enantioselectivity. This study
establishes the usefulness of modelling as a structure-based guiding tool for enantioselectivity optimiz-
ation as well as the potential for developing scalable peptide-based DMAP-type catalysts for large-scale
resolution work.
Received 9th September 2016,
Accepted 28th October 2016
DOI: 10.1039/c6ob01991a
www.rsc.org/obc
catalyst for alcohol resolution that have been developed
include amidines and bicyclic imidazoles,9 phosphorus-
Introduction
The use of chiral 4-dimethylaminopyridine (DMAP) derivatives based,10 N-heterocyclic carbenes11 and vicinal diamines,12
in the kinetic resolution (KR) of racemic secondary alcohols which have all been extensively covered in some excellent
via Lewis-base catalysis has gained considerable momentum recent reviews.3,13 Recently, 4-PPY KR has been demonstrated
since the first example by Vedejs and Chen reported twenty to have application to a medicinal chemistry scale-up of an
years ago using an asymmetric centre attached to the α-carbon enriched tetrazole hemiaminal ester prodrug for clinical evalu-
at C-2 of the pyridine ring.1 Although they achieved high ation, in which interestingly Connon’s4g simple monamide at
s-values (>10), catalyst reactivity was low, with reactions requir- the 3-position turned out to be the best transfer catalyst.14
ing both a Lewis acid and a stoichiometric amount of catalyst
Peptide catalysts have also been impressively demonstrated,
for full conversion that was attributed to steric congestion at first by Miller,15 and later by Schreiner16 and others,17 using
the “ortho” position in accordance with earlier findings by an N-methyl histidine residue as the Lewis base to afford high
Russian and German investigators.2 Since then,3 chiral DMAP- s-values in the KR of secondary alcohols, although these types
based methodology has blossomed to using the full gamut of of peptide catalysts are not easily accessed in large amounts.
chirality types from central,4 axial5 to planar.6
Moreover, and in the context of the present study, to date only
In more recent work, Seidel7 has elegantly demonstrated two studies18 have appeared regarding developing a peptide
that a suitable chiral environment can be generated using auxiliary (as opposed to mono-amides4g,19) in conjunction
achiral DMAP in conjunction with a chiral thiourea anion with a DMAP (or 4-PPY) template in the KR of secondary alco-
binder to bring about the kinetic resolution of amines via a hols, one of them involving a tryptophan residue.18b The lack
chiral ion-pair,8 resulting in excellent s-values. Other types of of examples is probably due to the challenge involved of pre-
dictively engineering an appropriate interaction between
peptide auxiliary and template so as to create a suitable chiral
aDepartment of Chemistry, University of Cape Town, Rondebosch, 7701, South Africa.
E-mail: Roger.Hunter@uct.ac.za; Fax: +27 21 6505195; Tel: +27 21 650 2544
bScientific Computing Research Unit, University of Cape Town, Rondebosch, 7701,
South Africa
environment around the N-acylation centre. However, such a
catalyst system would have the obvious advantage over other
reported catalysts of being able to produce large quantities of
either catalyst enantiomer (as well as diastereomers) for
scale-up studies. Furthermore, in spite of the advances
made in the field, generally there have been few mechanistic
†Electronic supplementary information (ESI) available:
A full Experimental
section, spectroscopic data of catalysts, HPLC data for the kinetic resolutions
and molecular modelling results. See DOI: 10.1039/c6ob01991a
This journal is © The Royal Society of Chemistry 2016
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
studies,4a,g,5f,20 regarding the origin of enantioselectivity with of commercially available 4-chloronicotinic acid with pyrroli-
DMAP-based catalysts. In this communication we report on dine and isolation of the product as its potassium salt 2 in a
the use of molecular modelling as a directing tool in the devel- 95% isolated yield. The latter was then coupled in a pyridine/
opment of a scalable 4-PPY 3-carboxamide catalyst. The suc- water mix using EDC and HOBt with either Trp-OMe or Leu-
cessful catalyst was used in the kinetic resolution of a small Trp-OMe in yields of around 45% in each case. The Leu-Trp-
library of secondary alcohols, in which strong evidence for OMe was prepared by conventional means using an FMoc pro-
π-stacking20a of a tryptophan residue with the N-acylpyridinium tecting group for the intermediate and also via an EDC coup-
intermediate was obtained.
ling, Scheme 2. Yields for the coupling weren’t optimized
further, but this is likely to be possible on any scale-up. EDC
coupling in the 4-PPY derivative synthesis proved to give
superior yields compared to those using acid chloride method-
ology (∼25% yield).
Results and discussion
The derivatives 3 and 4 were fully characterised by NMR
spectroscopies, IR spectroscopy and HRMS. Chiral HPLC of 4
revealed a single peak, which, coupled with a single set of res-
onances in its 13C NMR spectrum, was taken as sufficient evi-
dence that epimerisation had not taken place. Resolution of
1-(2-naphthyl)ethanol under the standard resolution con-
ditions of 1 equivalent of racemic alcohol, isobutyric anhy-
dride (0.7 eq.) and catalyst (5 mol%), with triethylamine
(0.9 eq.) in DCM at −78 °C, led to a decent conversion with
each of the catalysts, depending on the substrate, of between
30–60% after three hours, revealing a high acyl-transfer activity
of the DMAP derivative. Following chromatographic separ-
ation, the usual procedure according to the method of
Kagan,21 of measuring the ees of the product ester and
residual alcohol using chiral HPLC resulted in s-values of 2.3
(C = 42%) for catalyst 3 and an encouraging 5.3 (C = 34%) for
catalyst 4. No reaction was observed using triethylamine in the
absence of catalyst, in accordance with other studies.
Encouraged by the result for 4, it was decided to use it in a
broader substrate study, results of which are shown in Table 1.
The s-values were generally low (1.3–3.3), apart from the
naphthyl alcohol in entry 1 (5.3), so it was decided to proceed
to computer modelling of the N-acyl intermediate (4N-Ac) in
Our design strategy was inspired by the work of Fuji and
Kawabata,4a,b and Yamada.4c,i Both authors demonstrated
interesting stacking interactions in the N-acylpyridinium ion
intermediate in order to achieve appropriate pyridine ring
facial discrimination. In Fuji’s case this was achieved via
naphthyl-pyridinium π–π stacking, while Yamada’s case was
postulated to involve a pyridinium-thiocarbonyl group cation-π
interaction. In our case it was hoped to invoke a π-stacking
interaction involving the π-excessive indole ring of a trypto-
phan residue in the peptide chain interacting with the elec-
tron-deficient N-acylpyridinium ion intermediate 1, Scheme 1.
The first two derivatives targeted were the 4-PPY 3-carbox-
amides 3 and 4 with Trp and LeuTrp peptide side-chains,
respectively (Scheme 2). The idea of introducing Leu as the
first amino acid at the carboxyl attachment point was twofold
as: (i) to improve dichloromethane (DCM) solubility in the
kinetic resolution, and (ii) to possibly provide some screw for
discrimination of one of the pyridine faces. Modelling later on
vindicated this notion. A 4-pyrrolidino substituent rather than
its dimethylamino counterpart facilitated an improved solubi-
lity of the catalyst in DCM, the solvent of choice for the kinetic
resolutions. Synthesis of 3 and 4 was relatively straightforward
based on methodology used by Yamada,4i first involving SNAr
Scheme 1 π-Stacking model for a DMAP-peptide acylation transfer catalyst.
Scheme 2 Synthesis of peptide-PPY catalysts 3 and 4.
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2016

View Article Online
Organic & Biomolecular Chemistry
Paper
Table 1 The kinetic resolution of various racemic sec-alcohols catalysed by 4
Entry
1
Substrate
Ca
eeb alcohol
31.1
eec ester
59.4
sd
Configuratione
34
5.3
(S)
2
3
49
4
38.6
21.9
40.1
30.3
3.3
2.3
(S)
(S)
4
5
6
63
60
27
12.4
48.4
17.3
7.1
32.2
46.3
1.3
3.0
3.2
(S)
(S)
(S)
a Conversion C = 100 × (ee of recovered alcohol)/(ee of recovered alcohol + ee of ester). b ee of recovered alcohol as measured by chiral HPLC on
an AD, OD, or IC column. c ee of the hydrolysed ester (2 M NaOH in MeOH/H2O as measured on AD, OD, or IC column). d Selectivity factor
lnð1 ꢀ CÞð1 ꢀ eeÞ
s ¼
where ee refers to the recovered alcohol. e The absolute configuration of the faster-reacting enantiomer (ester) and deter-
lnð1 ꢀ CÞð1 þ eeÞ
mined by a comparison of [α]2D0 values with those reported in the literature.
order to identify potential structural elements for improve- expected to play in stabilising the active conformer of the cata-
ment (Fig. 1).
lyst. The HF-3c method was developed by Grimme, which is
For this purpose, an initial set of 1000 diverse conformers intended for pre-screening applications of large molecules.24
of 4N-Ac, based on a root-mean-square deviation (RMSD) cri- Subsequently, from these 100 structures, the ten with the
terion to assure effective spanning of conformational space, lowest energy were further refined using density functional
was first generated using the Open Babel toolkit,22 which theory (DFT), in which the B3LYP25–27 hybrid functional and
employs a genetic algorithm for this task. The structure of def2–SVP28 basis set was used in the calculations. In the DFT
each conformer was optimized using molecular mechanics calculations, dispersion was taken into account by using the
and the MMFF94 force field.23 A subset of the 100 lowest well-established D3 dispersion correction of Grimme, with
energy conformers was then extracted and further optimized Becke–Johnson damping to assure the correct asymptotic be-
using the efficient HF-3c method.24 This is a quantum mech- haviour.29,30 To speed up these calculations, the resolution of
anical calculation with three empirical corrections that, aside identity (RI) approximation was used to approximate Coulomb
from corrections for basis set inefficiencies, also includes a integrals31 as well as a chain of states method for numerical
dispersion correction. The latter was considered to be impor- integration of the exchange contribution to the hybrid func-
tant in context, given the role that dispersion interactions were tional.32 Finally, since the dipeptides are charged, a further
correction was made for solvent stabilization effects during the
geometry optimizations, using the conductor-like screening
polarizable continuum model (COSMO) with DCM.33 To
further limit computational effort, the effect of an explicitly
included anion on the structures, was not taken into account.
This level of QM theory is given the shorthand B3LYP–D3(BJ)/
def2–SVP. All DFT geometry optimizations were performed
with tight convergence criteria and the lowest energy struc-
tures were confirmed as minima by the absence of imaginary
frequencies in the calculated vibrational spectrum. The
vibrational spectra were calculated numerically using two-
Fig. 1 N-Acyl-pyridinium cation 4N-Ac of catalyst 4.
This journal is © The Royal Society of Chemistry 2016
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
sided (central) differences with a step size of 0.001 Bohr and it between the centroids of the pyridinium and indole benzene
was assumed that imaginary frequencies up to 50 cm−1 were ring is 4.435 Å, which is beyond the 3.3–3.8 Å range suggested
due to numerical errors and thus not indicative of a saddle by Janiack for qualifying as π-stacking.35 Likewise, the angle
point. The ORCA 3.0.2 software package was used for the between the benzene ring normal and the vector between the
quantum mechanical calculations.34
centroids, describing the parallel displacement, is approxi-
Fig. 2 shows the resultant lowest energy conformer, while mately 48°. This latter value is again much greater than the 25°
the full set of ten conformers are given in the ESI.† Based on a found in an ideal gas phase optimized π-stacked benzene-pyri-
Boltzmann distribution of the energies at 298.15 K, the lowest dine complex that has a parallel displaced geometry.36
energy conformer had a weight of 90%.
Several points of the conformational analysis deserve lizing interaction, it was felt that things were moving in the
mentioning: right direction in that, as shown by the right-hand frame of
Although this therefore does not suggest a substantial stabi-
• The N-acylpyridinium moiety adopts an s-trans confor- Fig. 2, H-2 of the pyridinium ring (shown in cyan) lies over the
mation [τ(C-2pyri, N-1pyri, CN–acyl, ON–acyl) = −172.2°] in which pyrrole ring (shown in magenta) of the indole ring.
the carbonyl oxygen points away from the more substituted
Further evidence for the interaction suggested from the
pyridine C-3 side, even though this results in the bulky isopro- modelling was sought from 1H NMR studies. Here, the aim
pyl group being brought closer to the C-3 chain. While it may was to study NMR shift differences between the H-2, H-5 and
seem counterintuitive that the bulky iso-propyl group is on the H-6 resonances for 4 and its pyridinium-ion 4N-Me in the hope
same side as the peptide substituent at C-3, this is in agree- of finding evidence for such an interaction. N-Methylation4g
ment with reports by Spivey5b and Connon.4g Spivey has rather than acylation4a turned out to be more appropriate in
suggested that while the preferred orientation of the N-acyl terms of sample handling. As a reference for 4/4N-Me it was
group is not sterically driven, the orientation of the iso-propyl decided to replace the Trp amino acid of the peptide chain
group may be due to a stereoelectronic stabilisation due to with a Leu, so Leu-Leu-OMe was prepared via an EDC coupling
partial conjugation with the C-3 substituent.5b We tested this as before, which in turn was coupled with the 4-PPY nicotinate
by also performing a geometry optimization of the analogous potassium salt 2, also using EDC and in pyridine/water as
s-cis conformer. The resultant structure had τ(C-2pyri, N-1pyri
,
before, affording the dipeptide Leu-Leu catalyst 5 in 27% yield
CN–acyl, ON–acyl) = −8.6°, but was 41.4 kJ mol−1 higher in energy after chromatography ultimately. In turn, both dipeptides were
than the minimum (and we return to this point again when methylated to completion (TLC) with methyl iodide at room
discussing compound 8N-Ac).
temperature in DCM to produce methylated derivatives that
• The peptide chain curves underneath the pyridinium ring were isolated by filtration in the case of 5N-Me and evaporated
– defined as one looks at it with the C-3 substituent on the directly in the case of 4N-Me. Spectroscopic characterisation
right-hand side. This would imply exclusive attack of the (see ESI†) revealed
incoming secondary alcohol from “the top” or re face of the δ 3.79 ppm in the 1H NMR spectra of 4N-Me and one at
acyl carbonyl group in this case. 4.06 ppm for the 5N-Me sample confirming that methylation
a new N-methyl singlet for 3H at
• The screw of the chain results in the indole ring being had taken place in each case. The chemical shifts for the three
brought close to the acyl site, effectively blocking the si (under- pyridine ring hydrogens in the 1H NMR spectra of each of 4
neath) face of the N-acyl carbonyl group. Here, a hydrogen bond and 5 were each identified, together with those of their methyl-
can be seen between the indole NH and the first carbonyl group ated derivatives, based on chemical shift and coupling con-
oxygen in the C-3 chain (due to Leu) at a distance of 2.335 Å stant considerations as well as a COSY spectrum for 4N-Me. The
and an angle of 118.1° – see the left-hand drawing in Fig. 2 – chemical shift difference between each corresponding hydro-
supporting the possibility of an interaction between the indole gen of the pyridine ring (H-2, H-5, and H-6) in unmethylated
and the pyridinium ring as originally postulated. The distance catalysts 4 (Leu-Trp) and 5 (Leu-Leu) was measured as
Fig. 2 Two different views of the lowest energy conformer of 4N-Ac
.
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2016

View Article Online
Organic & Biomolecular Chemistry
Paper
Δδ (δ5 – δ4), and similarly ΔδM (δ5
− δ4 ) for the methylated together this suggests a stacking-type interaction in which H-2
N-Me
N-Me
derivatives 4N-Me and 5N-Me. The results are shown in Table 2.
comes closest (of the three pyridine ring protons) to the indole
The results were indeed very informative. Comparing all ring shielding zone. Furthermore, a COSY spectrum of 4N-Me
three individual pyridine resonances for 4 and 5 revealed a revealed the NHLeu hydrogen (of the first peptide of the chain)
0.014 ppm downfield shift (Δδ) for H-2 going from 4 to 5, and to be significantly deshielded compared to that in 4 (6.73 to
virtually no changes for H-5 and H-6 (Δδ ∼ 0). Turning to the 8.78 ppm). The deshielding indicates an increase in the
data for the salts, however, gave a much different picture. δ+ character of the H of NHLeu, which in turn implies a greater
While both H-2 resonances for 4N-Me and 5N-Me were delocalisation of the N lone pair into the carbonyl, presumably
deshielded relative to those for their neutral counterparts 4 to compensate for the electron-withdrawing effect of the adja-
and 5 (0.15 ppm for the 4 series and 0.61 ppm for the 5), the cent pyridinium ring. Such an increase in δ− character at the
H-6 resonances for the salts were actually shielded compared first carbonyl oxygen would be consistent with the important
to those for 4 and 5 (0.47 ppm for the 4 series and 0.37 ppm hydrogen bond between the indole NH and the first carbonyl
for the 5). Understandably, in view of proximity as well as res- oxygen of the chain as identified by the modelling in Fig. 2.
onance and inductive considerations, H-5 was far less respon- While the NMR and modelling views have subtle differences
sive and was only slightly deshielded going from neutral com- regarding indole and pyridinium ring alignment, taken together
pound to salt (0.08 ppm in the 4 series and 0.19 ppm in the 5). these results provide compelling support for a significant inter-
While the shielding of H-6 going from 5 to 5N-Me (8.14 ppm to action between the π-deficient pyridinium and π-excessive
7.77 ppm) was surprising, the most striking relationship emer- indole ring as indicated in Fig. 2, and in line with modern
ging from the data in the context of the modelling results was thinking on the structural requirements required to promote
the relatively large difference between the chemical shifts of enantioselectivity in chiral DMAP Lewis base catalysis.
H-2 in the salts (ΔδM = 0.60). Comparing with the fairly negli-
Based on these conclusions, in the next phase of the
gible ΔδM values for H-5 and H-6 revealed a significant shield- project it was decided to extend the peptide chain from dipep-
ing of H-2 in the LeuTrp salt 4N-Me, in agreement with the tide to tripeptide retaining the indole in the middle position.
modelling image in Fig. 2 showing a π-stacking-type inter- This was intended to explore whether extra length in the chain
action between H-2 (cyan) and the pyrrole ring of the trypto- might help with buttressing on the C-5/C-6 side (“left-hand”
phan indole ring. Further support was obtained from a NOESY side). The terminal residue was chosen as either Leu or Trp,
experiment, which showed a correlation between H-2 and the syntheses of which are shown in Schemes 3 and 4. For the
indole NH and an adjacent hydrogen (α to NH as H-5 or on the Leu-Trp-Trp catalyst 6 a Boc protecting group was used
A-ring H-10 – see the ESI† for numbering) as well as to NHTrp
.
throughout the peptide coupling sequence, and coupling the
Weaker correlations were identified between the pyridine H-5 tripeptide Leu-Trp-Trp-OMe to the 4-PPY nicotinate salt 2 in a
and H-6 hydrogens and the indole A-ring hydrogens. Taken pyr/H2O mix and EDC as before, furnished a 53% yield of 6.
Conversely, for the Leu-Trp-Leu catalyst a combination of Boc
and FMoc protecting groups was utilised in which the final
coupling to afford the actual catalyst 7 was 44%. Spectroscopic
Table 2 1H NMR chemical shifts (ppm) of the pyridine hydrogens for 4,
data, notably the single set of resonances in the 13C NMR
5, and their methylated analogues
spectra, suggested a single diastereomer for each catalyst,
which was supported by chiral HPLC data for 7, which revealed
a >98% purity as a single enantiomer and diastereomer.
Kinetic resolution experiments were then carried out as
before and the results are shown in Table 3 for tripeptide cata-
lysts 6 and 7:
Unfortunately, the s-values for tripeptide catalysts 6 and 7
showed lower selectivity than those of the dipeptide catalyst 4.
Additionally, the Leu-Trp-Leu catalyst 7 showed much slower
conversion so was only tested against the naphthyl substrate
(entry 5 in Table 3), which gave C = 5.8% at −78 °C for 3 h; s =
3.3, and for which the LeuTrp dipeptide 4 had given the best
result of s = 5.3 at around 50% conversion. A longer time of
12 h with 7 (entry 6) did raise the conversion to an acceptable
Ha,b
δ4
δ5
Δδc
δ4N-Me
δ5N-Me
ΔδM
d
level (46%), but the s-value dropped to 2.4. Similarly, the
addition of a Trp residue to the dipeptide terminus of 4 dimin-
ished resolution efficiency (catalyst 6, entries 1–4), possibly by
presenting competing conformational options with conflicting
resolution outcomes. The possibility of the intermediacy of
peptide aggregates also can’t be discounted in this case in
view of the significant lowering in reaction rate. In any event,
H-2
H-5
H-6
8.12
6.43
8.14
8.26
6.47
8.14
0.14
0.04
0.0
8.27
6.51
7.67
8.87
6.66
7.77
0.60
0.15
0.10
a CDCl3 as solvent. b All pyridine resonances were unambiguously
assigned by NMR spectroscopy (1H, and COSY). c The Δδ value for each
resonance corresponds to δ5 − δ4. d The ΔδM values correspond to δ5
N-Me
− δ4
for the methylated series.
N-Me
This journal is © The Royal Society of Chemistry 2016
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
Scheme 3 Synthesis of Leu-Trp-Trp tripeptide catalyst 6.
Scheme 4 Synthesis of Leu-Trp-Leu tripeptide catalyst 7.
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2016

View Article Online
Organic & Biomolecular Chemistry
Paper
Table 3 The kinetic resolution of various racemic sec-alcohols catalysed by 6 or 7
Entrya
1
Catalyst
Substrate
C
eeb alcohol
10.7
eec ester
28.2
sd
Configuratione
6
28
2.0
(S)
2
3
6
6
35
53
12.1
7.6
22.3
6.7
1.8
1.2
(S)
(S)
4
5
6
6
7
7
51
21.6
3.2
21.1
52.5
30.4
1.9
3.3
2.4
(S)
(S)
(S)
5.8
46f
26.4
a Conditions = alcohol (1.0 eq.), cat (5 mol%), (i-PrCO)2O (0.7 eq.), NEt3 (0.9 eq.), −78 °C, DCM, 3 h. b ee of recovered alcohol as measured by
chiral HPLC on a AD, OD, or IC column. c ee of the hydrolysed ester (2 M NaOH in MeOH/H2O as measured on a AD, OD, or IC column).
lnð1 ꢀ CÞð1 ꢀ eeÞ
d Selectivity factor s ¼
where ee refers to the recovered alcohol. e The absolute configuration of the faster-reacting enantiomer
(ester) and determined by a comparison of [α]D20 values with those reported in the literature. f T = 12 h.
lnð1 ꢀ CÞð1 þ eeÞ
“bigger isn’t always better” seemed to be the message so we mer, τ(C-2pyri, N-1pyri, CN–acyl, ON–acyl) = 179.0°, 14.0 kJ mol−1
returned to modelling for scientific direction. Modelling of the higher in energy. Apart from the configurational reversal, it is
N-acylpyridinium ion of Leu-Trp dipeptide 4 (Fig. 2) suggested interesting to note that the difference in energy between the
that C-6 and C-7 of the indole A-ring could be fruitful avenues two conformations is much lower in 8N-Ac (14.0 kJ mol−1) than
of functionalization to pursue in order of the possibility of pro- that in 4N-Ac (41.4 kJ mol−1). Comparison of the two structures
viding extra buttressing, but this would require a new and shows, in the case of 8N-Ac, that the conformation switch can
more complex synthesis. Hence, as a second choice of pursuit be related to a change in coordinates that is restricted mainly
it was decided to remove the hydrogen bond shown in Fig. 2, to the N-acyl group, whereas for 4N-Ac there are several
and N-protect in which a Boc protection was chosen in view of additional changes in the relative coordinates of other atoms
its likely ease of introduction as well as the steric bulk of the (i.e. in the peptide chain), not directly involved in the torsional
t-butyl group (catalyst 8).
change. This leads to the loss of a number of stabilizing intra-
To gain further insight into the possible consequence of molecular interactions when the conformation of 4N-Ac
this choice, the calculation procedure outlined before was changes from s-trans to s-cis. Specifically, the greater driving
again repeated to produce ten low-energy conformers of 8N-Ac
.
force for structural change in the 4N-Ac case could be primarily
Fig. 3 shows the resultant lowest energy conformer, again with due to the close approach of the 2nd carbonyl group oxygen
the full set of ten conformers given in the ESI.† Based on a (Leu) of the peptide chain to the oxygen atom of the twisted
Boltzmann distribution of the energies at 298.15 K, the lowest N-acyl group in 4N-Ac s-cis, the ON–acyl⋯Opeptide distance being
energy conformer of 8N-Ac had a weight of 97%.
4.24 Å. Other changes include an adjustment in the relative
Gratifyingly, these calculations presented a more promising orientation of the indole ring, the loss of intramolecular inter-
conformational picture for improving the s-factor with salient action between this carbonyl group and the C-2 hydrogen
features as follows:
atom of the pyridinium ring as well as the absence of the
• As with the LeuTrp-NH catalyst 4 (4N-Ac), the dipeptide hydrogen bond between the indole NH and the first carbonyl
chain of 8 (8N-Ac) coils underneath the pyridinium ring, with group oxygen, already mentioned earlier (see the ESI† for
the C-3 substituent on the “right”. However, the N-acyl group further details of how these interactions were identified and
is now in an s-cis conformation with τ(C-2pyri, N-1pyri, CN–acyl
,
can be illustrated). Conversely, in 8N-Ac s-cis, the corresponding
ON–acyl) = −14.8°. Further calculations showed that this is ON–acyl⋯Opeptide distance is 7.28 Å, due to the different orien-
indeed the most favourable orientation, with the s-trans confor- tation of the 2nd carbonyl group that points away from the
This journal is © The Royal Society of Chemistry 2016
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
Fig. 3 Two different views of the lowest energy conformer of 8N-Ac
.
N-acyl group, which obviates the need for any further adjust- in THF at 20 °C to minimise the risk of any epimerisation.
ments as the conformation changes. Following a basic work-up, the crude product was purified by
• The first carbonyl group of the side chain of 8N-Ac had a column chromatography using 10% MeOH/DCM, and catalyst
dihedral angle τ(C-2pyri, C-3pyri, C, O) = −120.5°. While not per- 8 was isolated in a 89% yield as a pale-yellow solid, Scheme 5.
pendicular to the plane of the pyridine ring exactly, the CvO
The structure of catalyst 8 was unequivocally assigned on
group effectively points “up” compared to “down” for that of the basis of its spectroscopic data using a combination of 1D,
4N-Ac, the latter due to hydrogen bonding. This result was 2D 1H and 13C NMR as well as IR and high resolution mass
deemed to have great significance later.
spectrometry. Importantly, a single set of resonances was
• Boc introduction into the dipeptide results in a shift of observed in both spectra with no sign of any diastereomers. A
the indole ring to lie more associated with (“underneath”) the new upfield singlet in the 1H NMR spectrum present at δH 1.64
pyridinium ring of 8N-Ac (see Fig. 3, right-hand frame), conse- integrating for nine protons confirmed the introduction of the
quently also minimising steric strain between the pyrrolidino Boc group. Two samples were prepared, which by chiral HPLC
group and the Boc t-butyl group. The distance between the gave peak areas of 98 and 99% respectively (see the ESI†). This
centroids of the pyridinium and indole benzene ring is was deemed to be good enough for the resolution experiments,
3.734 Å, and the angle between the benzene ring normal and in which the 98% catalyst was used for the bulk results, since
the vector between the centroids, approximately 30°. In this this was the one produced in quantity.
regard, one can now consider this to show an improved degree
As before (Table 1) KR experiments with catalyst 8 were run,
of π-stacking.34 However, the Boc group points “north” away this time on an extended library and the results are shown in
from C-4 rather than emerging on the C-5/C-6 pyridine ring Table 4, in descending order of s-values within any series
side where buttressing with the substrate might occur.
(single vs. double ring substrates):
Encouraged by these results, N-Boc dipeptide catalyst 8 was
Introduction of the Boc group significantly enhanced the
synthesised by reacting 4 with di-tert-butyl dicarbonate (3 eq.) s-values, with three entries (entries 1, 2, and 9) exceeding the
Scheme 5 Synthesis of catalyst 8.
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2016

View Article Online
Organic & Biomolecular Chemistry
Paper
Table 4 The kinetic resolution of racemic sec-alcohols catalysed by 8
Entrya
1
Substrate
CHPLC
ee (alcohol)c
57.1
ee (ester)
72.0
s-Valued
Configuratione
b
44
10.8
(S)
2
3
4
5
50
44
42
30
68.5
53.8
50.2
29.4
68.8
69.9
69.1
68.0
10.9
9.6
8.9
7.0
(S)
(S)
(S)
(S)
6
7
20
45
16.2
35.7
65.3
43.1
5.6
3.5
(S)
(S)
8
9
36
45
13.9
59.3
24.6
71.1
1.9
(S)
(S)
10.7
10
11
50
56
63.2
50.6
62.4
8.1
3.7
(S)
(S)
39.8f
12
11
3.6
30.6
1.9
(S)
a Conditions = alcohol (1.0 eq.), cat (5 mol%), (i-PrCO)2O (0.7 eq.), NEt3 (0.9 eq.), −78 °C, DCM, 3 h. b ee of recovered alcohol as measured by
chiral HPLC on an AD, OD, or IC column. c ee of the hydrolysed ester (2 M NaOH in MeOH/H2O as measured on an AD, OD, or IC column).
lnð1 ꢀ CÞð1 ꢀ eeÞ
d Selectivity factor s ¼
where ee refers to the recovered alcohol. e The absolute configuration of the faster-reacting enantiomer
(ester) and determined by a comparison of [α]2D0 values with those reported in the literature. f Ester not hydrolysed but measured directly.
lnð1 ꢀ CÞð1 þ eeÞ
synthetically useful benchmark value for s of 10. Entries 1–8 with no compromising of the conversion, which stayed close
show the influence of substitution on the aromatic ring to 50%. By comparison, substitution of the substrate phenyl
(entries 1, 3, 5, 7 and 8) of 1-phenylethanol or in the chain ring with electron-donating or releasing groups favoured the
(entries 2 and 6). Notably, the standard substrate 1-phenyletha- former rather than the latter (entry 5 with a p-OMe, s =
nol (entry 4) used throughout studies in the literature (entry 4) 7.0 versus entry 7 with a p-Br, s = 3.5 and entry 8 with p-NO2,
returned a value of 8.9, almost a threefold improvement com- s = 1.9), suggesting the importance of a π-stacking interaction
pared to that of catalyst 3 without the Boc group (s = 3.3, entry in view of the s-value increasing with increasing electron
2, Table 1) and one similar or better for chiral DMAP catalysts density in the substrate phenyl ring promoting a tighter inter-
with DMAP C-3 amide-based auxiliaries. The introduction of a action with the electron-deficient pyridinium ring. Moving the
steric effect in the form of methyl groups either on the ring or stereogenic centre one down the chain in the form of 1-phenyl-
chain resulted in very good s-values around 10 (entries 1–3) 2-propanol returned an encouraging value of 5.6 for an
This journal is © The Royal Society of Chemistry 2016
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
aliphatic alcohol (entry 6). Similarly, in the naphthalene nyl carbon in TS-I allows for the substrate’s aromatic ring to
series, with the chiral hydroxyethyl side-chain at the naphtha- slide away from the leucine carbonyl group pointing “up”,
lene 2-position, the s-value rose from 5.3 (entry 1, Table 1) with placing it over to the left-hand side (pyridine C-5/C-6). This
catalyst 4 to 8.1 (entry 10, Table 4) with 8. Furthermore, shift- results in the minimisation of steric strain between the phenyl
ing the side-chain to the 1-position resulted in the s-value ring and the leucine carbonyl group pointing upwards as shown
increasing to a respectable value of 10.7. Conversely, extending in TS-I. By comparison, the same considerations applied to
the same chiral side-chain to the 9-position of anthracene TS-II for the (R)-enantiomer result in an increase in phenyl/CO
reduced the s-value to a low 4.0 probably due to its greater steric strain. These findings are in agreement with an argument
symmetrical nature with a choice of rings for π-stacking, but put forward by Yamada in his thiocarbonyl-based catalysis stu-
still with good conversion (54%). This reduction in value con- dies.4i They also offer an explanation of why the s-value
tinued with propargylic alcohol in entry 12 to a very low 1.9, increases significantly going from catalyst 4 to 8. In the former
presumably now to inefficient π-stacking with the triple bond, the carbonyl group is pointing “down” due to H-bonding while
as suggested by the low conversion (11%) under the same con- the NH is pointing “up” but is further out (Fig. 2).
ditions and time. These latter results indicate the relative
This model doesn’t take into account the likely influence
importance of the electronic effect over the steric one. In all exerted by the counter anion associated with the pyridinium
cases, the fast-reacting enantiomer was the S-enantiomer as ion in terms of both hydrogen bonding and steric strain.5f,19b
determined by measuring the sign of the optical rotations and Though the s-values in this study with our best catalyst 8 are
comparing them to literature values.
only just touching on the synthetically useful benchmark of
Fig. 4 outlines a possible transition-state model for the two 10, the modelling data strongly suggests that further progress
enantiomers of 1-phenylethanol reacting with 8N-Ac based on can be made using this structure-based approach. The ques-
the modelling data. Importantly, π-stacking is postulated to tion of the intermediacy of peptide aggregates is an obvious
occur between the substrate aromatic ring and the top face of transition-state complication that cannot be totally discounted
the catalyst pyridine ring, with the N-Boc indole moiety block- and one that wasn’t considered by the modelling. However,
ing the underneath face as suggested by the modelling data the fact that reactions involved clear solutions (DCM) at
(Fig. 4 – only the CvO group of the chain is shown). The −78 °C, and were fast compared to other kinetic resolutions
minimum energy conformation of the α-hydroxyalkyl substitu- with similar catalysts, very likely indicates that reactions with
ent containing the stereogenic centre is taken as having the catalyst 8 proceed via a monomeric species under the conditions
methyl group perpendicular and pointing away from the plane described. All things considered, obvious possibilities for
of the pyridine ring in order to minimise allylic strain between improvement would include: (i) introducing H-bonding effects
the methyl group of the alcohol and the ortho-hydrogens of the in the peptide side chain so as to promote a tighter binding
substrate aryl ring in accordance with Birman and Houk’s between catalyst, substrate and counter-ion to the
theoretical findings on Birman’s amidine catalyst.37 These N-acylpyridinium ion as noted by Connon4g and Sunoj,38 as well
considerations result in transition states I and II for the (S)- as (ii) increasing the tightness of the π-stacking by placing appro-
and (R)-enantiomers respectively. Importantly, stereoelectronic priate substituents (non-stereogenic) on the indole ring and/or at
alignment between the hydroxyl oxygen and the N-acyl carbo- C-55e,39 of the pyridinium ring. Finally, as shown by Spivey,5e
changing the N-pyrrolidino group to a long-chain dialkylamino
substituent might also have a favourable effect on the s-values.
Conclusions
In conclusion, the results of this study have demonstrated the
usefulness of applying rational design principles based on
modelling data to developing peptides in a modular way as
desirable green, and synthetically stable scalable auxiliaries in
chiral DMAP-type Lewis base enantiocatalysis. Such an
approach avoids having to use wasteful resolution protocols
for catalyst enantiomer synthesis.5d,6i
Acknowledgements
We thank the South African National Research Foundation
(NRF) for financial support towards this project. Computations
were performed using facilities provided by the University of
Fig. 4 Proposed transition-state models for the two enantiomers of
1-phenylethanol interacting with N-acylpyridinium ion 8N-Ac
.
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2016

View Article Online
Organic & Biomolecular Chemistry
Paper
Cape Town’s ICTS High Performance Computing team: http://
hpc.uct.ac.za.
I. Sotofte, P. Norrby and M. Johannsen, J. Org. Chem., 2005,
70, 10890; (i) R. P. Wurz, E. C. Lee, J. C. Ruble and G. C. Fu,
Adv. Synth. Catal., 2007, 349, 2345–2352; ( j) B. Hu,
M. Mend, Z. Wang, W. Du, J. S. Fossey, X. Hu and
W.-P. Deng, J. Am. Chem. Soc., 2010, 132, 17041–17044;
References
(k) For
a
helicenoidal chiral DMAP catalyst, see:
1 E. Vedejs and X. Chen, J. Am. Chem. Soc., 1996, 118, 1809–
1810.
M. R. Crittall, H. S. Rzepa and D. R. Carbery, Org. Lett.,
2011, 13, 1250–1253 and M. R. Crittall, N. W. G. Fairhurst
and D. R. Carbery, Chem. Commun., 2012, 48, 11181–11183;
(l) S. Y. Lee, J. M. Murphy, A. Ukai and G. C. Fu, J. Am.
Chem. Soc., 2012, 134, 15149–15153.
7 D. Seidel, Synlett, 2014, 783–794.
8 J. Briere, S. Oudeyer, V. Dalla and V. Levacher, Chem. Soc.
Rev., 2012, 41, 1696–1707.
9 (a) X. Li, H. Jiang, E. W. Uffman, L. Guo, Y. Zhang, X. Yang
and V. B. Birman, J. Org. Chem., 2012, 77, 1722–1737;
(b) Z. Zhang, M. Wang, F. Xie, H. Sun and W. Zhang, Adv.
Synth. Catal., 2014, 356, 3164–3170.
2 (a) G. Höfle, W. Steglich and H. Vorbrüggen, Angew. Chem.,
Int. Ed. Engl., 1978, 17, 569–583; (b) L. I. Bondarenko,
A. I. Kirichenko, L. M. Litvinenko, I. N. Dimitrenko and
V. D. Kobets, Zh. Org. Khim., 1981, 17, 2588–2594.
3 R. P. Wurtz, Chem. Rev., 2007, 107, 5570–5595.
4 (a) T. Kawabata, M. Nagato, K. Takasu and K. Fuji, J. Am.
Chem. Soc., 1997, 119, 3169–3170; (b) T. Kawabata,
K. Yamamoto, Y. Momose, H. Yoshida, Y. Nagaoka and
K. Fuji, Chem. Commun., 2001, 2700–2701; (c) S. Yamada,
T. Misono and S. Tsuzuki, J. Am. Chem. Soc., 2004, 126,
9862–9872; (d) D. Díez, M. J. Gil, R. F. Moro, N. M. Garrido, 10 T. A. Duffey, J. A. MacKay and E. Vedejs, J. Org. Chem.,
I. S. Marcos, P. Basabe, F. Sanz, H. B. Broughton and 2010, 75, 4674–4685.
J. G. Urones, Tetrahedron: Asymmetry, 2005, 16, 2980–2985; 11 (a) Y. Suzuki, K. Yamaguchi, K. Muramatsu and M. Sato,
(e) T. Poisson, M. Penhoat, C. Papamicael, G. Dupas,
V. Dalla, F. Marsais and V. Levacher, Synlett, 2005, 2285–
Chem. Commun., 2004, 2770–2771; (b) T. Kano, K. Sasaki
and K. Maruoka, Org. Lett., 2005, 7, 1347–1349.
2288; (f) T. Poisson, M. Penhoat, C. Papamicael, G. Dupas, 12 (a) Y. Kawamata and T. Oriyama, Chem. Lett., 2010, 39,
V. Dalla, F. Marsais and V. Levacher, Synlett, 2005, 2285–
2288; (g) C. O. Dálaigh, S. J. Hynes, J. E. O’Brien,
T. McCabe, D. J. Maher, G. W. Watson and S. J. Connon,
382–384; (b) E. P. Kündig, A. Enríquez-Garcia, T. Lomberget
and G. Bernadinelli, Angew. Chem., Int. Ed., 2006, 118, 104–
107.
Org. Biomol. Chem., 2006, 4, 2785–2793; (h) H. V. Nguyen, 13 (a) E. Vedejs and M. Jure, Angew. Chem., Int. Ed., 2005, 44,
D. C. D. Butler and C. J. Richards, Org. Lett., 2006, 8, 769–
772; (i) S. Yamada, T. Misono, Y. Iwai, A. Masumizu and
Y. Akiyama, J. Org. Chem., 2006, 71, 6872–6880.
3974–4001; (b) A. C. Spivey and S. Arseniyadis, Top. Curr.
Chem., 2010, 291, 233–280; (c) H. Pellissier, Adv. Synth.
Catal., 2011, 353, 1613–1666; (d) C. E. Muller and
P. R. Schreiner, Angew. Chem., Int. Ed., 2011, 50, 6012–6042.
5 (a) A. C. Spivey, P. Charbonneau, T. Fekner,
D. H. Hochmuth, A. Maddaford, C. Malardier-Jugroot, 14 D. W. Piotrowski, A. S. Kamlet, A.-M. R. Dechert-Schmitt,
A. J. Redgrave and M. A. Whitehead, J. Org. Chem., 2001,
66, 7394–7401; (b) A. C. Spivey, A. Maddaford, D. P. Leese
J. Yan, T. A. Brandt, J. Xiao, L. Wei and M. T. Barrila, J. Am.
Chem. Soc., 2016, 138, 4818–4823.
and A. J. Redgrave, J. Chem. Soc., Perkin Trans. 1, 2001, 15 (a) S. J. Miller, G. T. Copeland, N. Papaioannou,
1785–1794; (c) K. Jeong, S. Kim, H. Park, K. Chang and
K. S. Kim, Chem. Lett., 2002, 1114–1115; (d) A. C. Spivey,
F. Zhu, M. B. Mitchell, S. G. Davey and R. L. Jarvest, J. Org.
Chem., 2003, 68, 7379–7385; (e) A. C. Spivey, D. P. Leese,
F. Zhu, S. G. Davey and R. L. Jarvest, Tetrahedron, 2004, 60,
4513–4525; (f) E. Larionov, M. Mahesh, A. C. Spivey,
A. C. Y. Wei and H. J. Zipse, J. Am. Chem. Soc., 2012, 134,
9390–9399; (g) G. Ma, J. Deng and M. P. Sibi, Angew. Chem.,
Int. Ed., 2014, 44, 12012–12015.
6 (a) J. C. Ruble and G. C. Fu, J. Org. Chem., 1996, 61, 7230–
7231; (b) J. C. Ruble, H. A. Latham and G. C. Fu, J. Am.
Chem. Soc., 1997, 119, 1492–1493; (c) J. C. Ruble,
J. Tweddell and G. C. Fu, J. Org. Chem., 1998, 63, 2794–
T. E. Horstmann and E. M. Ruel, J. Am. Chem. Soc., 1998,
120, 1629–1630; (b) G. T. Copeland, E. R. Jarvo and
S. J. Miller, J. Org. Chem., 1998, 63, 6784–6785;
(c) E. R. Jarvo, G. T. Copeland, N. Papaioannou,
P. J. Bonitatebus and S. J. Miller, J. Am. Chem. Soc., 1999,
121, 11638–11643; (d) M. M. Vasbinder, E. R. Jarvo and
S. J. Miller, Angew. Chem., Int. Ed., 2001, 40, 2824–2827;
(e) G. T. Copeland and S. J. Miller, J. Am. Chem. Soc., 2001,
123, 6496–6502; (f) M. B. Fierman, D. J. O’Leary,
W. E. Steinmetz and S. J. Miller, J. Am. Chem. Soc., 2004,
126, 6967; (g) S. Miller, Acc. Chem. Res., 2004, 37, 601–610;
(h) E. A. Colby Davie, S. M. Mennen, Y. Xu and S. J. Miller,
Chem. Rev., 2007, 107, 5759–5812.
2795; (d) C. E. Garrett and G. C. Fu, J. Am. Chem. Soc., 1998, 16 (a) L. Wanka, C. Cabrele, M. Vanejews and P. R. Schreiner,
120, 7479–7483; (e) C. E. Garrett, M. M. Lo and G. C. Fu,
J. Am. Chem. Soc., 1998, 120, 10276–10276; (f) B. Tao,
J. C. Ruble, D. A. Hoic and G. C. Fu, J. Am. Chem. Soc.,
1999, 121, 5091–5092; (g) S. Bellemin-Laponnaz,
J. Tweddell, J. C. Ruble, F. M. Breitling and G. C. Fu, Chem.
Commun., 2000, 1009–1010; (h) J. G. Seitzberg, C. Dissing,
Eur. J. Org. Chem., 2007, 1474–1490; (b) C. E. Müller,
L. Wanka, K. Jewell and P. R. Schreiner, Angew. Chem., Int.
Ed., 2008, 47, 6180–6183; (c) C. E. Müller, D. Zell and
P. R. Schreiner, Chem. – Eur. J., 2009, 15, 9647–9650;
(d) R. Hrdina, C. E. Müller and P. R. Schreiner, Chem.
Commun., 2010, 46, 2689–2690.
This journal is © The Royal Society of Chemistry 2016
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
17 (a) K. Ishihara, Y. Kosugi and M. Akakura, J. Am. Chem. 24 R. Sure and S. Grimme, J. Comput. Chem., 2013, 34, 1672–
Soc., 2004, 126, 12212–12213; (b) F. Formaggio, A. Barazza, 1685.
A. Bertocco, C. Toniolo, Q. B. Broxterman, B. Kaptein, 25 A. D. Becke, J. Chem. Phys., 1993, 98, 5648.
E. Brasola, P. Pengo, L. Pasquato and P. Scrimin, J. Org. 26 C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens.
Chem., 2004, 69, 3849–3856; (c) Y. Kosugi, M. Akakura and
Matter, 1988, 37, 785–789.
K. Ishihara, Tetrahedron, 2007, 63, 6191–6202; 27 S. H. Vosko, L. Wilk and M. Nusair, Can. J. Phys., 1980, 58,
(d) X.-L. Geng, J. Wang, G.-X. Li, P. Chen, S.-F. Tian and
J. Qu, J. Org. Chem., 2008, 73, 8558–8562.
18 (a) S. Tabanella, I. Valancogne and R. F. W. Jackson, Org.
1200–1211.
28 F. Weigend and R. Ahlrichs, Phys. Chem. Chem. Phys., 2005,
7, 3297–3305.
Biomol. Chem., 2003, 1, 4254–4261; (b) H. Schedel, K. Kan, 29 S. Grimme, J. Antony, S. Ehrlich and H. Krieg, J. Chem.
Y. Ueda, K. Mishiro, K. Yoshida, T. Furuta and T. Kawabata,
Beilstein J. Org. Chem., 2012, 8, 1778–1787.
19 (a) T. Kawabata, R. Stragies, T. Fukaya, Y. Nagaoka,
Phys., 2010, 132, 154104.
30 S. Grimme, S. Ehrlich and L. Goerigk, J. Comput. Chem.,
2011, 32, 1456–1465.
H. Schedel and K. Fuji, Tetrahedron Lett., 2003, 44, 1545– 31 F. Neese, J. Comput. Chem., 2003, 24, 1740–1747.
1548; (b) G. Priem, B. Pelotier, S. J. F. Macdonald, 32 F. Neese, F. Wennmohs, A. Hansen and U. Becker, Chem.
M. S. Anson and I. B. Campbell, J. Org. Chem., 2003, 68,
3844–3848; (c) B. Pelotier, G. Priem, S. J. F. Campbell, 33 S. Sinnecker, A. Rajendran, A. Klamt, M. Diedenhofen and
M. S. Anson, R. J. Upton and I. B. Campbell, Tetrahedron F. Neese, J. Phys. Chem. A, 2006, 110, 2235–2245.
Phys., 2009, 356, 98–109.
Lett., 2005, 46, 9005–9007; (d) Y. Ueda, W. Muramatsu, 34 F. Neese, WIREs Comput. Mol. Sci., 2012, 2, 73–78.
K. Mishiro, T. Furuta and T. Kawabata, J. Org. Chem., 2009, 35 (a) C. Janiak, J. Chem. Soc., Dalton Trans., 2000, 3885–3896;
74, 8802–8805; (e) H. Mandai, S. Irie, M. Akehi, K. Yuri,
M. Yoden, K. Mitsudo and S. Suga, Heterocycles, 2013, 87,
329–340.
(b) C. R. Martinez and B. L. Iverson, Chem. Sci., 2012, 3,
2191–2201.
36 J. Řezáč, K. E. Riley and P. Hobza, J. Chem. Theor. Comput.,
2011, 7, 2427–2438.
20 (a) Y. Wei, L. Held and H. Zipse, Org. Biomol. Chem., 2006,
4, 4223–4230; (b) L. Mesas-Sanchez and P. Diner, Chem. – 37 X. Li, P. Liu, K. N. Houk and V. Birman, J. Am. Chem. Soc.,
Eur. J., 2015, 21, 5623–5631. 2008, 130, 13836–13837.
21 H. B. Kagan and J. C. Fiaud, Top. Stereochem., 1988, 18, 38 C. B. Shinisha and R. B. Sunoj, Org. Lett., 2009, 11, 3242–
249–330. 3245.
22 N. M. O’Boyle, M. Banck, C. A. James, C. Morley, 39 (a) N. De Rycke, G. Berionni, F. Couty, H. Mayr,
T. Vandermeersch and G. R. Hutchison, J. Cheminform.,
2011, 3, 1–14.
23 T. A. Halgren, J. Comput. Chem., 1996, 17, 490–519.
R. Goumont and O. R. P. David, Org. Lett., 2011, 13, 530–
533; (b) T. Poisson, S. Oudeyer and V. Levacher, Tetrahedron
Lett., 2012, 53, 3284–3287.
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2016