Kinetic resolution of esters via metal catalyzed methanolysis reactions

Article abstract of DOI:10.1039/b805236k

Some chiral lanthanide complexes of the Schiff base adducts of: a) bis(2-pyridylcarboxaldehyde) and (1R),(2R)-trans-1,2-diaminocyclohexane (Pyr-R,R′-chxn: 3); b) 6-methyl-2-pyridylcarboxaldehyde and (1R),(2R)-trans 1,2-diaminocyclohexane (MePyr-chxn, 4); and c) 2,6-pyridyldicarboxaldehyde and (1R),(2R)-trans-1,2-diaminocyclohexane ((Pyr-R,R′-chxn)₂, 5) have been screened for their utility to promote kinetic resolution via metal catalyzed alcoholyses of the p-nitrophenyl esters of chiral d- and l-Boc-protected glutamine and phenylalanine. Solvents were varied to optimize the kinetic selectivity values, defined as k ₂^L/k₂^D or k₂ ^D/k₂^L, for the methanolysis and in some cases, ethanolysis of these substrates. At ambient temperature the greatest selectivity was found for the ethanolysis of Boc-Gln-OPNP, catalyzed by 3:Yb ³⁺:(^-OEt) (k₂^L/k₂ ^D = 7.2). The greatest selectivity for Boc-Phe-OPNP is k ₂^D/k₂^L = 3.9 for its methanolysis promoted by 5:La³⁺:(^-OMe). A kinetic method is introduced for the determination of both d and l rate constants for catalyzed alcoholysis from a single kinetic experiment. The activation parameters ΔH ^? and ΔS^? were determined for the metal catalyzed methanolysis and ethanolysis of the Boc-Gln-OPNP substrates, and selectivity factors were found to increase at lower temperatures. A low temperature time course for the ethanolysis of racemic Boc-Gln-OPNP catalyzed by 3:Yb³⁺:(^-OEt) at -15 °C indicated that after 3 hours 60% residual d-enantiomer was observed having an enantiomeric excess of >95% ee. The activation parameters for the ethanolysis of the same substrate catalyzed by (Pyr-R,R′-chxn)₂:La³⁺:(^-OEt) predict a k₂^D/k₂^L = 40.4 at -40 °C with a large ee of >99% with ～80% of l isomer remaining at that temperature which has been experimentally confirmed. The Royal Society of Chemistry.

Full text of DOI:10.1039/b805236k

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Kinetic resolution of esters via metal catalyzed methanolysis reactions
Christopher I. Maxwell, Kalpa Shah, Pavel V. Samuleev, Alexei A. Neverov and R. Stan Brown*
Received 28th March 2008, Accepted 10th April 2008
First published as an Advance Article on the web 27th May 2008
DOI: 10.1039/b805236k
Some chiral lanthanide complexes of the Schiff base adducts of: a) bis(2-pyridylcarboxaldehyde) and
(1R),(2R)-trans-1,2-diaminocyclohexane (Pyr-R,Rꢀ-chxn: 3); b) 6-methyl-2-pyridylcarboxaldehyde and
(1R),(2R)-trans 1,2-diaminocyclohexane (MePyr-chxn, 4); and c) 2,6-pyridyldicarboxaldehyde and
(1R),(2R)-trans-1,2-diaminocyclohexane ((Pyr-R,Rꢀ-chxn)2, 5) have been screened for their utility to
promote kinetic resolution via metal catalyzed alcoholyses of the p-nitrophenyl esters of chiral D- and
L-Boc-protected glutamine and phenylalanine. Solvents were varied to optimize the kinetic selectivity
L
D
L
values, defined as k2 /k2 or k2D/k2 , for the methanolysis and in some cases, ethanolysis of these
substrates. At ambient temperature the greatest selectivity was found for the ethanolysis of
Boc-Gln-OPNP, catalyzed by 3:Yb3+:(−OEt) (k2 /k2 = 7.2). The greatest selectivity for
L
D
Boc-Phe-OPNP is k2D/k2 = 3.9 for its methanolysis promoted by 5:La3+:(−OMe). A kinetic method is
L
introduced for the determination of both D and L rate constants for catalyzed alcoholysis from a single
kinetic experiment. The activation parameters DH‡ and DS‡ were determined for the metal catalyzed
methanolysis and ethanolysis of the Boc-Gln-OPNP substrates, and selectivity factors were found to
increase at lower temperatures. A low temperature time course for the ethanolysis of racemic
Boc-Gln-OPNP catalyzed by 3:Yb3+:(−OEt) at −15 ◦C indicated that after 3 hours 60% residual
D-enantiomer was observed having an enantiomeric excess of >95% ee. The activation parameters for
the ethanolysis of the same substrate catalyzed by (Pyr-R,Rꢀ-chxn)2:La3+:(−OEt) predict a k2D/k2
=
L
40.4 at −40 ◦C with a large ee of >99% with ∼80% of L isomer remaining at that temperature which has
been experimentally confirmed.
(Boc-Gln-OPNP, 1) and phenylalanine (Boc-Phe-OPNP, 2) pro-
Introduction
moted by various metal complexes of the bis(2-pyridyl carboxalde-
hyde) Schiff base of (1R),(2R)-trans-1,2-diaminocyclohexane12
(Pyr-R,Rꢀ-chxn:M, 3:M).
Kinetic resolution of racemates is an emerging tool for the
separation of enantiomers1 although it has an inherent problem
in that the yield of a given enantiomeric product necessarily
decreases with time due to the buildup in concentration of the less
reactive enantiomer of the starting material. This is particularly
so when the chiral kinetic discrimination between the L and D
substrates (defined as the kinetic selectivity, kL/kD or kD/kL) is
not large. Several successful kinetic resolutions are known using
man-made catalysts2 and the use of enzymes for the purpose
of selective hydrolysis has been investigated for many years.3
Although transesterification reactions are well-known4 and recent
work has provided the mechanistic intricacies of metal ion
catalyzed transesterifications,5 the use of this technique for kinetic
resolution of esters is still very under-developed.6–11
Our recent extensive mechanistic investigations of the rapid
methanolysis reactions of both activated and non-activated es-
ters in the presence of [La3+(−OCH3)]2, Eu3+(−OCH3) and the
Zn2+(−OCH3)-complex of 1,5,9-triazacyclododecane5 suggested
that these metal ion/alkoxides, when complexed to chiral ligands,
might provide useful catalysts for the kinetic resolution of
esters and related compounds. Herein we describe our proof-of-
principle studies of the kinetic resolution of p-nitrophenyl esters
of the D and L N-tert-butoxycarbonyl derivatives of glutamine
In addition, we have screened some metal ion complexes
of two other Schiff base variants, namely those of 6-methyl-
2-pyridylcarboxaldehyde and (1R),(2R)-trans-1,2-diaminocyclo-
hexane13 (MePyr-chxn:M, 4:M) and the macrocyclic tetra-Schiff
base formed from 2,6-pyridyldicarboxaldehyde and (1R),(2R)-
trans-1,2-diaminocyclohexane ((Pyr-R,Rꢀ-chxn)2:M, 5:M). Herein
we report the results of these studies along with the rate constants
kL and kD for the catalyzed reactions of the enantiomers at room
temperature. We also present a useful method involving a single
kinetic experiment to determine the kL and kD rate constants for a
given ester from which one can readily calculate the ee vs. percent
conversion curves. Finally, in two cases we have determined the
activation parameters for the catalyzed transesterification and
from these predicted, and subsequently experimentally verified,
enantiomeric excesses (ee’s) of greater than 99% at selected reduced
temperatures.
Experimental
Materials
Methanol (99.8% anhydrous), sodium methoxide (0.5
M
solution in methanol), Boc-L-glutamine 4-nitrophenyl ester
(98%+), Boc-D-glutamine 4-nitrophenyl ester (98%+), Boc-L-
phenylalanine 4-nitrophenyl ester (98%+), Boc-D-phenylalanine
Department of Chemistry, Queen’s University, Kingston, Ontario, Canada
K7L 3N6. E-mail: rsbrown@chem.queensu.ca; Fax: +1 613 533 6667;
Tel: +1 613 533 2400
2796 | Org. Biomol. Chem., 2008, 6, 2796–2803
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
4-nitrophenyl ester (98%+), 6-methyl-2-pyridinecarboxaldehyde,
2-pyridinecarboxaldehyde and (+)-and (−)-trans-1,2-cyclo-
hexanediamine (99%) were obtained from Aldrich and used as
received. Mx+(−OTf)x salts, where M = Zn, Yb, Tm, Nd, La,
Eu, Ho were all obtained from Aldrich. Anhydrous ethanol was
obtained from Commercial Alcohols, Brampton, Ontario. Pyr-
chxn and MePyr-chxn were prepared as reported,12,13 as was ligand
5, termed here (Pyr-R,Rꢀ-chxn)2.14
and kL were obtained. The rate constants for each enantiomer were
then averaged to give the reported rate constant. This methodology
controls for variations between prepared samples and effects due
to the order of addition of the enantiomers but has an estimated
2% error attributed to dilution of the catalysts in the subsequent
experiments which is incorporated into the standard deviations
reported.
The single kinetic run method to obtain both the kD and kL was
conducted as per the following example. Two UV/vis cuvettes,
each containing 0.1 mmol dm−3 Pyr-R,Rꢀ-chxn:Yb3+(−OCH3) cat-
alyst, were prepared as described above. To one cuvette was added
25 lL of the enantiomerically pure D-Boc-Gln-OPNP substrate
(4.0 mmol dm−3 in acetonitrile), and the cuvette was placed into
the sample chamber of a dual beam UV/vis spectrophotometer.
To the second cuvette was added 25 lL of a stock solution of
racemic substrate (4.0 mmol dm−3 in acetonitrile) after which it
was placed in the reference cell position. The resulting biphasic
kinetic traces were analyzed to determine kL and kD by NLLSQ
fitting the absorbance vs. time data to eqn (1);
General methodology for determination of enantiomeric excess by
kinetic studies
Stock solutions (50 mmol dm−3) of the catalyst Mx+(−OTf)x
where M = Zn, Yb, Tm, Nd, La, Eu, Ho (50 mmol dm−3),
ligand (50 mmol dm−3), and NaOMe (25 mmol dm−3) were
prepared in anhydrous methanol. Stock solutions (4 mmol dm−3)
of each enantiomer of the substrate (Boc-Gln-OPNP and Boc-
Phe-OPNP) were prepared in anhydrous acetonitrile.
For each kinetic run the catalyst was formulated in situ by
the addition of 25 lL of each M(OTf)x, ligand and NaOMe
stock solution and subsequently 50 lL of the stock solution of
substrate to methanol such that the final volume was 2.5 mL. The
5:La3+-complex formation was relatively slow as judged by the
increase in catalytic rate constant as a function of time, so
the freshly made complex solutions were allowed to stand at
room temperature overnight prior to use. The reaction rates were
followed at 324 nm (for the formation of p-nitrophenol) using
a Cary Bio-100 spectrophotometer with the cell compartment
thermostated at 25.0 0.1 ◦C. First order rate constants (kobs) were
evaluated from fits of the absorbance vs. time profiles to a standard
exponential model. Specifically, two identical sample solutions in
1 cm cuvettes were prepared. To one sample the L enantiomer
was added and the reaction was allowed to go to completion to
determine the first order rate constant for its disappearance (kL).
Subsequently an aliquot of the stock solution of the D enantiomer
was added to the same cuvette and a first order rate constant
for its disappearance (kD) was similarly obtained. For the other
sample solution, the D enantiomer was added followed by the L
enantiomer and two first order rate constants corresponding to kD
Abst = Abs0 + 0.5(DAbs)e−kLt − 0.5(DAbs)e−kDt
where DAbs = Abs∞ − Abs0.
(1)
Determination of activation parameters
Following the general methodology outlined above, the kobs
constants for transesterification of each enantiomer of Boc-Gln-
OPNP were obtained in MeOH and EtOH solvent at four to
◦
five temperatures between 25 ◦C and 5 C with the Pyr-R,Rꢀ-
chxn:Yb3+(−OR) catalyst, and four to seven temperatures between
40 ◦C to 10 ◦C for the (Pyr-R,Rꢀ-chxn)2:La(−OR) catalyst. Three
reactions of each substrate were followed at each temperature and
the corresponding kobs values were averaged to give the values
reported in Tables 5 and 6. The k2 rate constants were then fit to a
standard Boltzman equation k = ATe(−DH/RT + DS/R) based on a 1/k
weighting: the activation parameters, DH‡ and DS‡ are given in
Table 7.
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 2796–2803 | 2797
©

View Article Online
General methodology for determination of enantiomeric excess by
HPLC
A Hewlett Packard Series 1050 HPLC using a Chiralcel OD-H
column (Daicel Chemical Industries) was used for the determina-
tion of the enantiomeric excess produced by 5:La3+(−OEt), Pyr-
R,Rꢀ-chxn:Yb3+(−OEt) and Pyr-S,Sꢀ-chxn:Yb3+(−OEt) catalyzed
reactions of a racemic mixture of L- and D-Boc-Gln-OPNP (formu-
lated by adding equal amounts of the authentic enantiomers). All
chromatograms were run with a mobile phase of 10% isopropanol–
90% hexanes at 2.00 ml min−1 and monitored at a wavelength
of 270 nm. The retention times for the starting materials and
catalysts were determined from authentic samples. A 25 lL aliquot
of a 4 mmol dm−3 solution of the commercial D-Boc-Gln-OPNP
substrate was found to contain about 6–7% of the corresponding
L-enantiomer as an impurity.
Fig. 1 Plots of kobs vs. [Pyr-R,Rꢀ-chxn:Yb3+:0.5(NaOMe)] for methanol-
ysis of D-Boc-Gln-ONp (●) and L-Boc-Gln-ONp (᭺) (0.05 mmol dm−3
)
In a typical experiment to monitor the enantiomeric excess of
the reaction as a function of time, a reaction vial equipped with a
rubber septum was charged with Yb(OTf)3 (5 lL, 50 mmol dm−3),
ligand Pyr-R,Rꢀ-chxn (5 lL, 50 mmol dm−3), NaOEt (5 lL,
25 mmol dm−3) and an internal standard of toluene (50 lL,
1.63 mol dm−3) in 0.5 mL of ethanol. An identical reference
vial was prepared with substrate but no catalyst. The vials were
placed in the freezer in CaCl2–acetone bath held at a temperature
of −15 ◦C. After 15 minutes of cooling, the reaction was
initiated by adding the racemic Boc-Gln-OPNP substrate (50 lL,
10 mmol dm−3) to each vial. After addition of substrate the final
volume in each vial was 0.635 mL, with final concentrations of
3.9 × 10−1 mmol dm−3 Yb(OTf)3, 3.9 × 10−1 mmol dm−3 Pyr-
R,Rꢀ-chxn, 2.0 × 10−1 mmol dm−3 NaOEt, 8 × 10−1 mmol dm−3 of
the racemic Boc-Gln-OPNP substrate and 0.13 mol dm−3 toluene.
Immediately after the addition of substrate, a 10 lL aliquot of
the reference mixture without catalyst was injected to provide
a time zero point. Subsequent injections of 10–25 lL aliquots
from the reaction vial containing catalyst were made at 25 minute
in methanol at 25 ◦C.
Fig. 2 Plots of kobs vs. [Pyr-R,Rꢀ-chxn:Yb3+:0.5(NaOMe)] for methanol-
ysis of D-Boc-Phe-ONp (●)and L-Boc-Phe-ONp (᭺) (0.05 mmol dm−3) in
methanol at 25 ◦C.
◦
intervals. For experiments run at −40 C, 5:La3+(−OEt) catalyst
mixtures at the concentrations described above were prepared in
six vials which were placed in a freezer at −40 ◦C. At three times,
duplicatereactions were stopped withthe addition of HClO4 (7.8 ×
10−1 mmol dm−3) and LiCl (3.9 mmol dm−3) for future HPLC
analysis. The relative concentrations of each residual starting
enantiomer were determined based on the peak areas of the L
and D enantiomers corrected by dividing the peak area of interest
by that of the internal standard. The ee was calculated using eqn
(2), where AL and AD are the corrected peak areas of the L and D
enantiomers.
sspH15 corresponding to the pKa of the catalyst system when the
[Pyr-chxn:M3+:(HOCH3)]/[Pyr-chxn:M3+:(−OCH3)] ratio is unity.
The fact that the plots in Fig. 1 and 2 are linear with intercepts of
zero rules out the involvement of free methoxide in the production
of the p-nitrophenol product and also rules out involvement of
higher order species such as dimers.
Given in Table 1 are the second order rate constants for
methanolysis of the L and D substrates promoted by La3+(−OCH3),
Eu3+(−OCH3), and Yb3+(−OCH3) complexes of Pyr-R,Rꢀ-chxn as
well as the La3+-complex of 5 in pure methanol at 25 ◦C. The
gradients of the lines in Fig. 1 and 2 are taken as the second
order (k2) constants, although we know that only half of the
complex is present as the active CH3O−-containing form. At
ambient temperature, none of the three complexes is particularly
enantioactive with the Boc-Phe-OPNP substrate and all the
(2)
Results and discussion
L
Shown in Fig. 1 and 2 are plots of the kobs values vs. [Pyr-R,R-
chxn:Yb3+:0.5(−OCH3)] for the catalyzed methanolysis reaction of
L- and D-Boc-Gln-OPNP and L- and D-Boc-Phe-OPNP. Similar
plots (not shown) are obtained for the reaction of L- and D-Boc-
Gln-OPNP promoted by La3+:(Pyr-R,Rꢀ-chxn)2 in the presence of
0.5 eq. of NaOCH3. Previous work5 established that the active
forms of these and related complexes contain one methoxide per
metal ion, but the use of the 0.5 eq. buffers the medium at the
selectivity values, defined as k2 /k2D, are less than unity. On the
L
D
other hand, the k2 /k2 ratios are all greater than unity with the
Pyr-R,Rꢀ-chxn:Ln3+ complexes for the Boc-Gln-OPNP substrate,
L
D
while the 5:La3+ catalyst has the opposite preference where k2 /k2
is <1 for reasons that are not clear.
We introduce here a faster way to screen the catalytic activity
by which one can determine both the kD and kL constants and
selectivity factors for a given catalyst in a single reaction. This
2798 | Org. Biomol. Chem., 2008, 6, 2796–2803
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
Table 1 Second order rate constants (k2) for the methanolysis of L- and D-Boc-Phe-OPNP and Boc-Gln-OPNP catalyzed by Pyr-R,Rꢀ-
chxn:Ln3+:0.5(−OCH3) and by 5:(Pyr-R,Rꢀ-chxn)2:0.5(−OCH3) complexes at 25 ◦Ca
Substrate
Boc-Phe-OPNP
Boc-Gln-OPNP
k2L/dm3
k2D/dm3
k2L/dm3
k2D/dm3
mol−1 s−1
mol−1 s−1
k2L/k2
mol−1 s−1
mol−1 s−1
k2L/k2
Db
Db
Catalyst
Pyr-R,Rꢀ-chxn:La3+:(−OCH3), sspH = 8.3
Pyr-R,Rꢀ-chxn:Eu3+:(−OCH3),sspH = 6.5
Pyr-R,Rꢀ-chxn:Yb3+:(−OCH3),sspH = 6.5
5:La3+:(−OCH3),sspH = 8.8
7.5
4.9
10.8
6.4
0.70
(18% ee, 43% L)
0.77
(13% ee, 42% L)
0.50
(25% ee, 47% L)
0.35
31.5
63.8
340
27.0
25.0
89
1.17
(8% ee, 37% D)
2.6
(44% ee, 55% D)
3.8
(58% ee, 62% D)
0.36
14.5
2.5
24.6
7.0
2.8
7.8
(47% ee, 56% L)
(47% ee, 56% L)
a k2 defined from the gradient of the plot of kobs vs. [ligand:M3+:0.5(−OCH3)]. b ee values and residual yield of predominant stereoisomer calculated at
optimal catalytic conversion (OCC) using eqn (3)–(5)
is similar to the ‘quasi-racemate’ method for determining small
kinetic isotope effects,17 which we have adapted to dual-beam
UV/vis spectrophotometry. In the typical example, illustrated
here with the methanolysis of Boc-Gln-OPNP promoted by the
Yb3+:(−OCH3) complex of Pyr-R,Rꢀ-chxn, the reference and sam-
ple cells of the spectrometer are respectively charged with equal
amounts of the racemic and enantiomerically pure substrates,
along with identical amounts of catalyst and the absorbance
difference vs. time curve (Fig. 3) is monitored until the completion
of the reaction. Characteristically these have an ‘up/down’ or
‘down/up’ behaviour depending on which enantiomer reacts more
rapidly, and if both react at the same rate, then there is no rise/fall
behaviour. The kL and kD rate constants are easily obtained from
NLLSQ fits of the Abs vs. time curve to the expression in eqn 1.
For the example chosen here, kL = 0.028 s−1 and kD = 0.0068 s−1
and the kL/kD = 4.1, which is experimentally the same as what was
obtained from the single run experiments in Table 1, row 3. The
determination of tOCC and the [residual D] and [residual L] isomers
at that time are given as:
tOCC = ln(kL/kD)/(kL − kD)
(3)
(4)
(5)
[L]/[L0] = exp(−kLtOCC
)
[D]/[D0] = exp(−kDtOCC
)
The ee at tOCC is defined as: ([L] − [D])/([L] + [D])100. These
values are given in Tables 1–6 along with the percentage of the
dominant enantiomer remaining at the OCC.
The effect of changing the metal ion and ligand to metal ion
ratio was investigated in methanol with 0.5 mmol dm−3 of metal
ion (added as the corresponding triflate), one or two equivalents
of Pyr-R,Rꢀ-chxn and 0.25 mmol dm−3 of added NaOCH3. The
same series of reactions was investigated in a solvent comprising
90% CH3CN–10% methanol and the results are also presented in
Table 2. Of the various metal ions Yb3+ and Tm3+ have the greatest
selectivity factors, and although increasing the ligand/Mx+ ratio
from one to two may increase, decrease or have no effect on the
reaction rate in selected cases, this does not seem to have any
profound effect on the selectivity factors. Only in the case of Nd3+
does the change in the solvent to 9 : 1 acetonitrile increase the
rate and the selectivity factor appreciably: in all other cases the
kL/kD ratio drops or remains constant. The same thing is seen
for La3+:(Pyr-R,Rꢀ-chxn)2 where the 9 : 1 acetonitrile : methanol
solvent system increases the selectivity factor to 3.5 from 2.8 in
methanol.
DAbs vs. time curve in Fig. 3 also visually provides the time (tOCC
)
at which the rates of catalytic conversion of the D and L isomers
are the same (vertical dashed line) which we will define here as the
‘optimal catalytic conversion (OCC)’. The exact equations for the
Some additional screening was done to investigate the effect
of 9 : 1 solvent : methanol compositions on the selectivity
factor for the Yb3+:(−OCH3) complex of Pyr-R,Rꢀ-chxn with L-
and D-Boc-Gln-OPNP using THF, ether, dichloromethane, 1,2-
dichloroethane and toluene. Only THF and toluene with 2 : 1
ligand : Yb3+ ratio gave selectivity factors approaching those in
methanol, the respective values being 3.9 and 3.4.
Fig. 3 A DAbs vs. time plot for the methanolysis of D-Boc-Gln-OPNP
and racemic Boc-Gln-OPNP (0.04 mmol dm−3 in sample and reference
cell respectively) promoted by 0.1 mmol dm−3 Pyr-R,Rꢀ-chxn:Yb3+ in the
presence of 0.05 mmol dm−3 NaOCH3. Dashed vertical line defines the
time (∼65 s) where the catalyzed rates of conversion of the D- and L-
isomers are equal (tOCC).
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 2796–2803 | 2799
©

View Article Online
Table 2 Observed first order rate constants for the catalysis methanolysis of 4 × 10−5 mol dm−3 L- and D-Boc-Gln-OPNP in methanol and 90%
acetonitrile–methanol promoted by 0.5 mmol dm−3 of metal ion, one or two equivalents of Pyr-R,Rꢀ-chxn and 0.25 mmol dm−3 of added NaOCH3
Mx+
Lig : Mx+
103kL/s−1
103kD/s−1
kL/kD
ee, with residual yieldb
Yb3+
CH3OH
9 : 1(A : M)a
Tm3+
CH3OH
9 : 1(A : M)a
9 : 1(A : M)a
Nd3+
CH3OH
9 : 1(A : M)a
Zn2+
1 : 1
2 : 1
2 : 1
1 : 1
2 : 1
1 : 1
2 : 1
1 : 1
2 : 1
1 : 1
1 : 1
2 : 1
1 : 1
2 : 1
155.1
192.0
163
102.0
170.5
152.0
201.8
24.7
12.0
28.3
21.2
21.0
40.8
40.5
41.2
27.2
45.2
53.2
71.3
12.3
7.0
12.7
11.0
9.5
33.8
34.3
3.8
4.7
4.0
3.8
3.8
2.9
2.8
2.0
1.7
2.3
1.9
2.2
1.6
1.5
58%, 62% D
65%, 66% D
60%, 63% D
58%, 62% D
58%, 62% D
48%, 57% D
47%, 57% D
33%, 50% D
26%, 47% D
38%, 52% D
31%, 49% D
37%, 52% D
22%, 45% D
21%, 45% D
CH3OH
9 : 1(A : M)a
9 : 1(A : M)a
52.5
52.5
a 90% acetonitrile–methanol solution. b ee values and residual yield of predominant stereoisomer calculated at optimal catalytic conversion (OCC) using
eqn (3)–(5).
Complex 4
cell was added 0.05 mmol dm−3 of L-Boc-Gln-OPNP, and its
transesterification reaction was followed to completion. To the
same cell was then added 0.05 mmol dm−3 of the D isomer of Boc-
Gln-OPNP, and this reaction was again followed to completion.
The entire process was repeated with a fresh solution of the catalyst
formulated in the same way, and the reaction of the D and then
L isomer of the substrate was followed to completion. Given in
Table 4 are the averages of the various first order rate constants
determined in this way, and it can be seen that the selectivity
factors are highest in ethanol. For (Pyr-R,Rꢀ-chxn)2:La3+ with
0.5 eq. of added methoxide, the second order rate constants in
ethanol for transesterification of L- and D-Boc-Gln-OPNP are
As one expects a more sterically demanding complex might lead
to greater chiral discrimination, we undertook single run kinetic
studies of the rates of methanolysis of the Yb3+(−OCH3) and
La3+(−OCH3) complexes of MePyr-chxn. The results in Table 3
indicate, that only in the cases of the Yb3+ complexes in the
presence of 2 ligands is there an appreciable selectivity and under
these conditions the reactions are markedly slower in the presence
of a single equivalent of ligand. This probably means that the
complexes are not fully formed when the ligand/M3+ ratio is 1.
Even so, the dramatic reduction in reaction rate, coupled with a
poorer selectivity than was realized with Pyr-R,Rꢀ-chxn:Yb3+ and
La3+:(Pyr-R,Rꢀ-chxn)2, made us discontinue study with ligand 4.
k2 = 2.0 dm3 mol−1 s−1 and k2 = 11.1 dm3 mol−1 s−1 for a
L
D
stereoselectivity factor of 5.5.
Comparison of selectivity factors in methanol, ethanol and
n-propanol
Table 4 First order rate constants for the alcoholysis reaction of
L- and D-Boc-Gln-OPNP promoted by 0.5 mmol dm−3 Pyr-R,Rꢀ-
chxn:Yb3+:0.5(−OR) in various alcohols at T = 25 ◦C
As our earlier results indicated that the metal ion catalysts
are generally active in other light alcohols such as ethanol
and propanol, we compared the selectivity factors obtained in
these three solvents. For this series of reactions the kL and kD
constants were obtained in a single cell containing 0.5 mmol dm−3
each of Yb3+ (introduced as its triflate) and of Pyr-R,Rꢀ-chxn,
and 0.25 mmol dm−3 of added NaOCH3 which immediately
equilibrates to form the NaOR of the ROH solvent. Into this
Solvent
103kL/s−1
103kD/s−1
kL/kD
ee, with residual yielda
MeOH
EtOH
PrOH
155 10
51.5 1.5
6.4 1.9
40.8 1.0
7.2 0.4
2.0 0.7
3.8
6.9
3.2
58%, 62% D
75%, 73% D
52%, 59% D
a ee values and residual yield of predominant stereoisomer calculated at
optimal catalytic conversion (OCC) using eqn (3)–(5).
Table 3 First order rate constants for methanolysis of substrates promoted by Yb3+(−OCH3) and La3+(−OCH3) complexes of ligand 4 in MeOH,
T = 25 ◦C, [M3+] = 0.5 mmol dm−3, [−OCH3] = 0.25 mmol dm−3
Subst.
M
Lig : M
103kL/s−1
103kD/s−1
kL/kD
ee with residual yielda
1
Yb
La
Yb
La
1 : 1
2 : 1
1 : 1
2 : 1
1 : 1
2 : 1
1 : 1
2 : 1
31.0
19.3
26.2
35.8
0.30
0.20
7.8
27.0
8.5
26.3
34.0
0.28
0.13
7.5
1.1
2.3
1
7%, 39% D
38%,52% D
0.2%,37% D
2.6%,38% D
3.4%,38% D
21%,45% D
1.9%,38% D
4.0%,35% L
1
2
1.1
1.5
1
8.3
9.0
0.9
a ee values and residual yield of predominant stereoisomer calculated at optimal catalytic conversion (OCC) using eqn (3)–(5).
2800 | Org. Biomol. Chem., 2008, 6, 2796–2803
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
Table 7 Activation parameters for the ethanolysis and methanolysis
of L- and D-Boc-Gln-OPNP promoted by Pyr-R,Rꢀ-chxn:Yb3+:(−OR),
3:Yb3+:(−OR) and (Pyr-R,Rꢀ-chxn)2:La3+:(−OR), 5:La3+:(−OR)a
Activation parameters for the methanolysis and ethanolysis
reactions of L- and D-Boc-Gln-OPNP promoted by
Pyr-R,Rꢀ-chxn:Yb3+:(−OR) and (Pyr-R,Rꢀ-chxn)2:La3+:(−OCH3)
System
DH/kJ mol dm−3
DS/J mol dm−3 K−1
Alt◦hough none of the complexes provided excellent selectivity at
25 C, any difference in the activation enthalpies (DH‡) for the
catalytic transesterification of the D and L isomers must be mani-
fested in a changing kL/kD ratio as a function of temperature, with
3:Yb3+:(−OR)
MeOH
L
D
L
D
33.9 1.0 39.3 1.4 −73.7 3.6
−66.0 4.8
EtOH
54.3 2.4 67.6 3.8 −16.4 8.4
−4.4 12.4
5:La3+:(−OR)
MeOH
L
D
L
D −39
8
the magnitude of the change being a reflection of the DDH‡ (L vs. D)
.
69.0 2.4 47.7 2.3
67.3 3.9 45.4 3.3
23.7 7.9
Knowledge of the activation parameters for the catalyzed reactions
is essential for calculating the temperature where the best compro-
mise between ee and maximal residual enantiomer is realized. The
activation parameters were determined in methanol and ethanol
following the procedures described in the previous sections. Given
in Tables 5 and 6 are the observed first order rate constants of
EtOH
−9.1 12.0
−73 11
a Error limits are given as maximum deviation from the mean of two or
three independent runs for each constant.
the catalyzed reactions along with the selectivity factors at various
temperatures. The errors limits are computed as the maximum
deviation from the mean of two or three determinations of the
constant. When the first order rate constants are divided by the
[active catalyst]18 ([Pyr-R,Rꢀ- chxn:Yb3+:(−OR)] = 0.25 mmol dm−3
or[Pyr-R,Rꢀ-chxn)2:La3+:(−OR)] = 0.1 mmol dm−3 in ethanol and
0.05 mmol dm−3 in methanol), the so-produced second order rate
constants (k2) can be used to compute the activation parameters
for the catalyzed transesterifications of L and D Boc-Gln-OPNP
given in Table 7.
The activation parameters refer to the overall catalyzed process
which we have shown previously5 with La3+2(−OCH3)2 and a
1,5,9-triazacyclononane:Zn2+(−OCH3) complex, involves the steps
shown in eqn (6) and (7) with a pre-equilibrium catalyst and
substrate binding followed by the formation of an anionic tetra-
hedral intermediate (TO−) that is stabilized by complexation. With
good leaving groups such as p-nitrophenoxy, the breakdown of the
tetrahedral intermediate is fast,5 so the k1 step for formation of the
metal-stabilized intermediate is rate limiting. Thus, for the metal
catalyzed alcoholyses operative here, the activation parameters
Table 5 First order rate constants for the alcoholysis of L- and D-
Boc-Gln-OPNP (4 × 10−5 mol dm−3) promoted by active catalyst
(0.25 mmol dm−3 Pyr-R,Rꢀ-chxn:Yb3+:(−OR), (3:Yb3+:(−OR))) in ethanol
and methanol at various temperaturesa
T/◦C
103kL/s−1
103kD/s−1
k/kD
ee with residual yieldb
EtOH
25
20
10
5
51.5 1.5
36.9 2.4
13.9 1.0
11.0 0.8
7.2 0.4
4.6 0.1
1.9 0.1
1.0 0.2
7.2
8.5
7.3
75%, 73% D
78%, 74% D
76%, 73% D
83%, 79% D
11.0
MeOH
25
20
10
5
155 10
40.8 1.0
31.0 0.7
17.3 0.3
11.1 1.0
3.8
3.7
4.1
4.7
58%, 62% D
58%, 62% D
60%, 63% D
65%, 66% D
115
1
70.7 2.0
52.6 1.8
a Error limits are given as maximum deviation from the mean of 2 or
3 independent runs for each constant. b ee values and residual yield
of predominant stereoisomer calculated at optimal catalytic conversion
(OCC) using eqn (3)–(5).
obs
refer to the overall second order rate constant given as k2 = Kbk1.
At 25 ◦C the transesterification reactions in ethanol for the D and
L isomers with both catalysts are about three to five times slower
than the corresponding reactions in methanol but the selectivities
are larger in ethanol. In both alcohols, the selectivity factors with
the catalysts get larger as the temperatures drop, indicating that
one can reduce the temperature to a point that this would be a
viable kinetic resolution technique for selected substrates.
Table 6 First order rate constants for the alcoholysis of L- and D-Boc-
Gln-OPNP (0.04 mmol dm−3) promoted by active catalyst (0.1 mmol dm−3
in ethanol and by 0.05 mmol dm−3 (Pyr-R,Rꢀ-chxn)2:La3+:(−OR),
(5:La3+:(−OR)), in methanol) at various temperaturesa
T/◦C
103kD/s−1
103kL/s−1
kD/kL
ee with residual yielda
(6)
(7)
EtOH
40
35
30
25
3.8 0.2
2.7 0.2
2.5 0.1
1.9 0.3
1.95 0.1
1.1 0.1
0.95 0.07
0.69 0.02
2.0
2.4
2.6
3.0
32%, 50% L
42%, 54% L
55%, 44% L
56%, 47% L
In order to demonstrate experimentally that the enantiomeric
excesses of the reaction do improve markedly at low temperature,
we determined, using HPLC, the time course for disappearance
of each enantiomer of a racemic Boc-Gln-OPNP mixture (8 ×
10−1 mmol dm−3) in the presence of each of the chiral Pyr-R,Rꢀ-
chxn:Yb3+(−OR) and Pyr-S,Sꢀ-chxn:Yb3+(−OR) complexes (3.9 ×
10−1 mmol dm−3) in EtOH at −15 ◦C over the course of three hours.
Given in Fig. 4 is the time course promoted by the Pyr-R,Rꢀ
complex showing the residual amounts of the D and L isomers
and the enantiomeric excesses in the total product mixture as a
function of time. After three hours, essentially all the starting L-
isomer has reacted while about 60% of the enantiomerically pure
D-isomer is left (experimental ee after 180 min is 97%) A similar
plot (not shown) obtained with the Pyr-S,Sꢀ-chxn:Yb3+(−OCH3)
MeOH
40
35
30
25
20
15
10
17.0 0.2
12.0 0.9
9.7 1.1
7.8 0.7
4.9 0.6
3.59 0.02
2.5 0.1
9.3 0.3
5.7 0.7
4.4 0.9
2.8 0.3
1.4 0.3
0.82 0.02
0.54 0.01
1.8
2.1
2.2
2.8
3.5
4.3
4.6
29%, 48% L
36%, 51% L
38%, 52% L
47%, 56% L
56%, 61% L
62%, 65% L
64%, 66% L
a ee values and residual yield of predominant stereoisomer calculated at
optimal catalytic conversion (OCC) using eqn (3)–(5).
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 2796–2803 | 2801
©

View Article Online
L isomer remaining. We confirmed this prediction by running the
reaction at −40 ◦C in six vials, quenching the reaction in duplicate
vials at 720, 1440 and 2160 minutes by the addition of an excess
of perchloric acid and LiCl, and analyzing the reaction mixtures
by HPLC. The observed experimental points for residual D and L
isomers are placed on Fig. 5 as ꢀ and ● symbols. The experiment
confirms the prediction that a so-constituted catalytic mixture run
for 36 hours at −40 ◦C yields ∼79% of the L isomer with an
enantiomeric excess of >99%.
Conclusion
We have described a study of the use of some chiral lanthanide
metal ion complexes to effect kinetic resolutions via transesterifi-
cation of the nitrophenyl esters of N-Boc protected amino acids of
glutamine and phenylalanine under mild conditions of essentially
neutral ‘pH’ (in alcohol). The choice of the p-nitrophenyl esters is
dictated by the ease with which the kinetics can be determined by
UV/visible spectrophotometry. Only the La3+ complex of ligand
5 ((Pyr-R,Rꢀ-chxn)2) is moderately effective at room temperature
Fig. 4 Time course for the ethanolysis of racemic Boc-Gln-OPNP
promoted by Pyr-R,Rꢀ-chxn:Yb3+(−OCH3) at −15 ◦C : solid bar, residual
D-isomer; dotted bar, residual L-isomer; solid line, enantiomeric excess.
complex reveals a correspondingly fast reaction of the D-isomer,
with about 54% of the residual enantiomerically pure L-isomer
remaining after about 195 minutes, ee ∼100%).
for selective cleavage of the phenyl alanine derivative (kL/kD
=
0.26). The best selectivity for the glutamine derivative comes
from the Yb3+:Pyr-R,Rꢀ-chxn complex in ethanol where the kL/kD
ratio is 11 at 5 ◦C, but activation parameter analysis indicates
that a far better selectivity can be achieved with both catalysts
at lower temperatures. The activation parameters allow one to
predict an optimum reaction condition for a given concentration
of catalyst where maximum values of the ee and residual unreacted
enantiomer can be achieved in a conveniently accessible time. In a
selected example we showed that a catalyst having an unimpressive
As a final demonstration of the accuracy of prediction of ee for
the ethanolysis of a racemic Boc-Gln-OPNP mixture at reduced
temperature, the activation parameters for the reaction of the D
and L enantiomers with (Pyr-R,Rꢀ-chxn)2:La3+:(−OEt) given in
Table 7 were used to find an optimum temperature and [catalyst]
at which the compromise of high ee and maximum amount
of residual substrate are achieved in a conveniently accessible
reaction time. An appropriate compromise was calculated to be
at −40 ◦C with 1 mmol dm−3 catalyst and the computed plots of
residual D and L isomer concentrations, ee and DAbs. vs. time are
given in Fig. 5.
k2D/k2 = 3 at 25 ◦C is predicted to have a k2D/k2 = 40.4 at
L
L
◦
−40 C and we experimentally confirmed a predicted large ee of
>99% with ∼80% of the L isomer remaining.19
In order to more conveniently screen the catalytic systems at var-
ious temperatures, we have presented a simplified kinetic method
where the rate constants for both enantiomers can be determined
in a single experiment. Application of this methodology allows
rapid determination of the activation parameters of the reactions
for a given catalyst with each of the two enantiomers. This is
essential for calculation of a temperature where one can obtain
high enantiomeric excesses in reasonable times with appreciable
amounts of the less reactive isomer remaining.
Acknowledgements
Fig. 5 Time course for various parameters arising from the ethanolysis
of D-Boc-Gln-OPNP and (1.0 mmol dm−3) and rac-Boc-Gln-OPNP, each
promoted by 1.0 mmol dm−3 (Pyr-R,Rꢀ-chxn)2:La3+:(−OEt) in ethanol at
−40 ◦C. Plots are constructed using first order rate constants calculated
from activation parameter values (kD = 4.53 × 10−5 s−1, kL = 1.12 ×
10−6 s−1). Shown are the ee (dotted line, right axis), and the residual
fractions of D and L enantiomers vs. time (··–, ---, left axis). Solid line
is the computed DAbs. vs. time plot where the right y-axis represents
the maximum absorbance expected for cleavage of one isomer. ꢀ and ●
symbols are duplicates of experimentally observed amounts of residual of
D and L isomers at times 720, 1440 and 2160 min.
The authors gratefully acknowledge the financial support of the
Natural Sciences and Engineering Research Council of Canada
as well as Queen’s University and the Summer Work Experience
program administered through Queen’s University. In addition,
they thank Dr Igor Kozin of this department for his invaluable
assistance with the GC and HPLC analytical methods.
Notes and references
1 H. B. Kagan and J. C. Fiaud, Top. Stereochem., 1988, 18, 249.
2 E. Vedejs and M. Jure, Angew. Chem., Int. Ed., 2005, 44, 3974; A.
Ghanem and H. Y. Aboul-Eneim, Chirality, 2005, 17, 1; D. E. J. E.
Robinson and S. D. Bull, Tetrahedron: Asymmetry, 2003, 14, 1407; M.
Keith, J. F. Larrow and E. N. Jacobsen, Adv. Synth. Catal., 2001, 343,
5; A. H. Hoyveda and M. T. Didiuk, Curr. Org. Chem., 1998, 2, 489.
The computed data indicate that at −40 ◦C the k2D/k2 = 40.4
L
and at 2200 minutes the predicted ee is >99%, with 84% of residual
2802 | Org. Biomol. Chem., 2008, 6, 2796–2803
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
3 B. G. Davis and V. Borer, Nat. Prod. Rep., 2001, 18, 618; H. Groger,
Adv. Synth. Catal., 2001, 343, 547; S. M. Roberts, J. Chem. Soc., Perkin
Trans. 1, 2001, 1475; M. T. Reetz, Angew. Chem., Int. Ed., 2001, 113,
292.
4 J. Otera, Chem. Rev., 1993, 93, 1449; M. C. Cleij, W. Drenth and R. J. M.
Nolte, J. Org. Chem., 1991, 56, 3883.
5 A. A. Neverov, T. McDonald, G. Gibson and R. S. Brown,
Can. J. Chem., 2001, 79, 1704; A. A. Neverov, G. Gibson and R. S.
Brown, Inorg. Chem., 2003, 42, 228; A. A. Neverov, N. E. Sunderland
and R. S. Brown, Org. Biomol. Chem., 2005, 3, 65.
6 M. Tokunaga, J. Kiyosu, Y. Obora and Y. Tsuji, J. Am. Chem. Soc.,
2006, 128, 44817.
7 M. Tokunaga, H. Aoyama, J. Kiyosu, Y. Shirogane, T. Iwasawa, Y.
Obora and Y. Tsuji, J. Organomet. Chem., 2007, 692, 472.
8 C. Dro, S. Bellemin-Laponnaz, R. Welter and L. H. Gade, Angew.
Chem., Int. Ed., 2004, 43, 4479.
9 J. S. You, X.-Q. Yi, X.-Y. Su, T. Wang, Q.-X. Xiang, M. Yang and R.-
G. Xie, J. Mol. Catal. A: Chem., 2003, 202, 17.
15 For the designation of pH in non-aqueous solvents we use the forms
described by Bosch and co-workers16 based on the recommendations of
the IUPAC, Compendium of Analytical Nomenclature. Definitive Rules
1997, 3rd edn, Blackwell, Oxford, UK, 1998. If one calibrates the
measuring electrode with aqueous buffers and then measures the pH
of an aqueous buffer solution, the term wwpH is used; if the electrode is
calibrated in water and the ‘pH’ of the neat buffered methanol solution
then measured, the term swpH is used; and if a correction factor of −2.24
(in the case of methanol) is subtracted from the latter reading, then the
term sspH is used.
16 Given that the autoprotolysis constant of methanol is 10−16.77 (mol
s
dm−3)2, neutral pH in methanol is 8.4: E. Bosch, F. Rived, M. Rose´s
s
and J. Sales, J. Chem. Soc., Perkin Trans. 2, 1999, 1953.
17 G. Bergson, O. Matsason and S. Sjoberg, Chem. Scr., 1977, 11, 25; A. J.
Bennet and M. L. Sinnott, J. Am. Chem. Soc., 1986, 108, 7287.
18 In order to compute the accurate activation parameters, the first order
rate constants are divided by the concentration of the active catalyst
1
which is of the total amount of ligand:M3+ put in solution.
2
10 D. Ryabov, G. M. Kazankov, S. A. Kurzeev, P. V. Samuleev and V. A.
Polyakov, Inorg. Chim. Acta, 1998, 280, 57.
11 M. C. Cleij, F. Mancin, P. Scrimin, R. Tecilla and U. Tonellato,
Tetrahedron, 1997, 53, 357; C. Cleij, P. Scrimin, P. Tecilla and U.
Tonellato, Langmuir, 1996, 12, 2956; P. Scrimin, P. Tecilla and U.
Tonellato, J. Org. Chem., 1994, 59, 4194.
12 J. F. Larrow and E. N. Jacobsen, J. Org. Chem., 1994, 59, 1939.
13 D.-H. Jo, Y.-M. Chiou and L. Que, Inorg. Chem., 2001, 40, 3181.
14 S. W. Bligh, N. Choi, W. J. Cummins, E. G. Evagorou, J. D.
Kelley and M. McPartlin, J. Chem. Soc., Dalton Trans., 1994, 23,
3369.
19 The fact that an enantioactive catalyst shows little selectivity at a given
experimental temperature does not necessarily indicate that it is a poor
catalyst at all temperatures. This can only be verified by activation
parameter analysis using Eyring or Arhenius plots. If the gradients for
these plots, which define the DH‡ for the catalyzed reaction of each
isomer are very close to each other, then the selectivity will be the
same at all accessible temperatures. On the other hand, if the gradients
are substantially different, the two plots will generally cross at the
isokinetic temperature. The selectivity value, and in fact the D(R) vs.
L(S) selectivity itself, then depends on how close and on which side of
the isokinetic temperature the experiment is run.
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 2796–2803 | 2803
©