Epimerization of glycal derivatives by a cyclopentadienylruthenium catalyst: Application to metalloenzymatic DYKAT

Article abstract of DOI:10.1002/chem.201403720

Epimerization of a non-anomeric stereogenic center in carbohydrates is an important transformation in the synthesis of natural products. In this study an epimerization procedure of the allylic alcohols of glycals by cyclopentadienylruthenium catalyst 1 is presented. The epimerization of 4,6-O-benzylidene-d-glucal 4 in toluene is rapid, and an equlibrium with its d-allal epimer 5 is established within 5 min at room temperature. Exchange rates for allal and glucal formation were determined by 1D ¹H EXSY NMR experiments to be 0.055 s^-1 and 0.075 s ^-1, respectively. For 4-O-benzyl-l-rhamnal 8 the epimerization was less rapid and four days of epimerization was required to achieve equilibration of the epimers at room temperature. The epimerization methodology was subsequently combined with acylating enzymes in a dynamic kinetic asymmetric transformation (DYKAT), giving stereoselective acylation to the desired stereoisomers 12, 13 , and 15. The net effect of this process is an inversion of a stereogenic center on the glycal, and yields ranging from 71% to 83% of the epimer were obtained.

Full text of DOI:10.1002/chem.201403720

DOI: 10.1002/chem.201403720
Full Paper
&
Carbohydrates
Epimerization of Glycal Derivatives by
a Cyclopentadienylruthenium Catalyst: Application to
Metalloenzymatic DYKAT**
Richard Lihammar, Jerk Rçnnols, Gçran Widmalm,* and Jan-E. Bꢀckvall*[a]
Abstract: Epimerization of
a
non-anomeric stereogenic
O-benzyl-l-rhamnal 8 the epimerization was less rapid and
four days of epimerization was required to achieve equilibra-
tion of the epimers at room temperature. The epimerization
methodology was subsequently combined with acylating
enzymes in a dynamic kinetic asymmetric transformation
(DYKAT), giving stereoselective acylation to the desired ste-
reoisomers 12, 13, and 15. The net effect of this process is
an inversion of a stereogenic center on the glycal, and yields
ranging from 71% to 83% of the epimer were obtained.
center in carbohydrates is an important transformation in
the synthesis of natural products. In this study an epimeriza-
tion procedure of the allylic alcohols of glycals by cyclopen-
tadienylruthenium catalyst 1 is presented. The epimerization
of 4,6-O-benzylidene-d-glucal 4 in toluene is rapid, and an
equlibrium with its d-allal epimer 5 is established within
5 min at room temperature. Exchange rates for allal and
1
glucal formation were determined by 1D H EXSY NMR ex-
periments to be 0.055 sꢀ1 and 0.075 sꢀ1, respectively. For 4-
Introduction
Bacterial polysaccharides and other natural products common-
ly contain rare carbohydrates with different stereochemistry
compared to the most common monosaccharides. Such mole-
cules are synthetically accessible through stereocontrolled de
novo synthesis[1,2] or modifications of readily available carbohy-
drates.[3,4] The latter methodology frequently involves stereo-
chemical inversions of secondary alcohols, which regularly
follow two-step procedures: triflation followed by nucleophilic
displacement or oxidation and subsequent stereoselective re-
duction.[5–7] An alternative method would be to use a dynamic
kinetic asymmetric transformation (DYKAT) approach utilizing
a redox epimerization catalyst to epimerize the alcohol and
subsequently trap the inverted alcohol in a transesterification
reaction catalyzed by a stereoselective enzyme (Figure 1).
The enzymatic selectivity in a transesterification of secondary
alcohols is governed by the size of the substituents residing at
the stereogenic center. Assuming that the largest group has
a higher priority than the medium-sized group according to
the CIP rule, lipases show (R)-stereoselectivity, whereas serine
proteases show (S)-stereoselectivity.[8,9]
Figure 1. Schematic picture of the epimerization of a carbohydrate molecule
and the subsequent trapping of one epimer by an enzyme-catalyzed stereo-
selective acylation.
Chemoenzymatic dynamic kinetic resolution (DKR) and
DYKAT have been applied in asymmetric synthesis of a wide
range of secondary alcohols and primary amines with various
transition metal complexes as racemization catalysts (examples
are given in Figure 2).[10] Of these catalysts, cyclopentadienylru-
thenium complex 1, activated by tBuOK, is one of the most ef-
ficient catalysts for racemization of sec-alcohols,[11] and it has
been used in combination with both (R)-selective lipases[11] and
(S)-selective proteases.[12] Catalyst 1 has been thoroughly inves-
tigated with both IR and NMR techniques giving valuable in-
sight into its activation and catalytic mechanisms.[13,14]
[a] R. Lihammar,+ Dr. J. Rçnnols,+ Prof. Dr. G. Widmalm, Prof. Dr. J.-E. Bꢀckvall
Department of Organic Chemistry, Arrhenius Laboratory
Stockholm University, 106 91 Stockholm (Sweden)
E-mail: gw@organ.su.se
jeb@organ.su.se
[+] These authors contributed equally to this work.
[**] DYKAT=Dynamic kinetic asymmetric transformation.
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201403720.
Figure 2. Three examples of ruthenium-based racemization catalysts used in
combination with enzymes in DKR protocols.
Chem. Eur. J. 2014, 20, 14756 – 14762
14756
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Transition metal-catalyzed epimerization of secondary alco-
hols in carbohydrates, utilizing catalyst 1, has been studied
previously.[15] By subjecting partially protected glucose, man-
nose, and allose derivatives to activated 1 in p-xylene at
1208C, it was possible to determine the thermodynamic equi-
librium between the non-anomeric epimers of the carbohy-
drates. Typically, catalyst 1 is able to epimerize secondary alco-
hols at ambient temperature at a rapid rate; however in the
case of these protected carbohydrates, where the secondary
alcohol is in close proximity to several electron-withdrawing
groups the oxidation step of the racemization is much harder
to achieve. Moreover, the neighboring groups increase the
bulk around the secondary alcohol, which further decreases
the efficiency of the epimerization. We envisioned that by per-
forming the epimerization on a protected glycal derivative, the
influence of both of these factors would be lowered and there-
by enable a more rapid epimerization. With a faster reaction
rate it would be possible to study the epimerization by NMR
spectroscopic methods. Furthermore, lower temperatures
would allow combination of the transition metal-catalyzed epi-
merization with an enzymatic resolution, thus giving a DYKAT
that would yield an inversion of a stereogenic center.
Figure 3. Equilibrium between compounds 4 and 5 using 10 mol% of 1;
part of the H NMR spectrum at 258C of the equilibrium mixture, showing
the H2 resonances. Peaks arising from the catalyst–substrate complexes
([Ru]=catalyst 1) are marked with asterisks (*). K1 · K3 =0.90, K2 =0.56.
1
Results and Discussion
Epimerization
Initially, a suitable racemization catalyst was needed and there-
fore screening of ruthenium-based catalysts 1–3 was conduct-
ed. All three catalysts have previously been used successfully
in combination with enzymes yielding enatioenriched allylic
acetates.[16] Epimerization experiments of 4,6-O-benzylidene-d-
glucal, 4, employing catalysts 2 and 3, under standard condi-
tions, only showed traces of epimerized products after 18 h at
608C.[17] Compound 1, on the other hand, was able to rapidly
epimerize 4 at room temperature: aliquots retrieved and ana-
lyzed by 1H NMR spectroscopy after 30 and 60 min showed
a 2:1 mixture of the glucal and allal derivatives 4 and 5 (Keq =
0.504 for 4 Ð 5).
Figure 4. Equilibrium between 4 and 5, facilitated by treatment with tBuOK
activated 1 in [D8]toluene solution at 908C and detected by: a) 1D H EXSY
1
NMR with selective excitation of the H2 resonance at d=4.52 ppm in 4, b)
1
1D H EXSY NMR with selective excitation of the H2 resonance at
1
d=4.74 ppm in 5 and c) the corresponding 1D H NMR spectrum.
Table 1. Exchange rates and energies of the epimerizations of 4 at 908C
and of 8 at 258C.
[a]
Reaction
Keq
k+/10ꢀ3
k /10ꢀ3
DGꢀ
[kJmolꢀ1
]
ꢀ
[sꢀ1
]
[sꢀ1
]
The rapid establishment of an equilibrium inspired an inves-
tigation of the reaction kinetics by NMR spectroscopy. Upon
treatment of compound 4 with 10 mol% of activated catalyst
1 in [D8]toluene solution, the equilibrium between 4 and 5
was established within five minutes at 258C. In addition to the
C3-epimer, allal 5, the catalyst–substrate complexes[18] 4* and
4 Ð 5
8 Ð 9
0.77
1.3
55
0.13
75
0.086
95.7
93.9
1
[a] Determined by integration of the epimers peaks in H NMR spectra.
1
5* were observed in the H NMR spectrum (Figure 3), enabling
the determination of the equilibrium constant for the equilibri-
um between 4* and 5* (K2 in Figure 3). Subsequent additions
of activated catalyst did not increase the complex concentra-
tion significantly.
and 5 occurred and enabled establishment of the exchange
rates. Full exchange matrix analysis gave the exchange rates
k+ =0.055 sꢀ1 and k =0.075 sꢀ1 (Table 1). No significant ex-
ꢀ
change peaks were observed at lower temperatures. At elevat-
ed temperatures 1H NMR data showed a buildup of two ke-
tones: 6[20] and its suggested saturated analogue 7 (Figure 5);
both irreversibly formed through oxidation and redox isomeri-
zation processes, respectively.[21]
At a sample temperature of 908C, the ratio between com-
pounds 4 and 5 was 4:3, and the equilibrium constants were
1
determined to be K1 · K3 =0.71 and K2 =1.10 by H NMR spec-
1
troscopy. 1D H EXSY NMR experiments[19] were performed at
908C with selective excitation of the H2 signals of epimers 4
and 5. With a 3 s mixing period exchange peaks were obtained
(Figure 4), which showed that a fast equilibrium between 4
Epimerization of the 6-deoxygenated compound, 4-O-benzyl
l-rhamnal 8, proceeded at a much lower rate. Equilibration of
the epimers 8 and 9 required four days of reaction time at
Chem. Eur. J. 2014, 20, 14756 – 14762
www.chemeurj.org
14757
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 5. Ketones formed at elevated temperature in the epimerization reac-
tion of compound 4.
Figure 7. Equilibria of compounds 4 and 5, and 8 and 9, respectively, under
catalysis of 1. The dihydropyran rings are schematically depicted in their pre-
ferred conformations.
Table 2. Experimental chemical shifts and coupling constants in ppm
and Hz, respectively, and interatomic distances from molecular models, in
ꢂ, of the glycal hydroxyl groups.
4
5
8
9
dHO3 [ppm]
3JH3,HO3 [Hz]
rO4···HO3 [ꢂ]
1.434
4.51
2.6
1.991
1.87
2.1
1.130
6.38
3.1
2.113
3.09
2.4
Figure 6. Plot of the epimerization equilibrium between 8 (&) and 9 (*)
after treating 8 with tBuOK-activated 1 (10 mol%) in [D8]toluene solution at
258C. Oxidation by-products 10 + 11 (D) are formed in small amounts. Solid
lines are drawn through the data points as an aid to the eye. Populations
In compound 9 the OH proton displayed signs of hydrogen
bonding in the 1H NMR spectrum: The downfield chemical
shift (dHO(9)
@ dHO(8)) and coupling constant JHOCH =3.1
1
are obtained by integration of H NMR spectra.
(Table 2) indicates hydrogen bonding to the benzyl ether
oxygen.[24] Analysis of molecular models of compounds 8 and
9 revealed a favorable geometry for hydrogen bonding in the
latter compound, but not in the former. An intramolecular hy-
drogen bond could thus stabilize epimer 9. This hypothesis
was tested by adding a competitive hydrogen bond donor
(tBuOH) to the epimerization reaction, which potentially could
interfere with this internal hydrogen bond. When tBuOH was
added the equilibrium shifted towards epimer 8 (to give Keq =
1.08), which further supports the existence of an intramolecu-
lar hydrogen bond in 9.[25] A similar trend, although of smaller
magnitude, is present in compounds 4 and 5, where dHO(5)>
dHO(4). The coupling constant JHOCH =1.9 Hz in 5 indicates re-
stricted motion around the C3ꢀO3 bond, with a probable
alignment towards O4 since the distance to the OH group is
short in the molecular model.
room temperature (Figure 6). Rate constants for the epimeriza-
tion could be extracted through a least-sum-square fitting of
the experimental concentrations at different times to a kinetic
model. The kinetic model is based on first-order kinetics re-
garding the disappearance and formation of 8 and 9, respec-
tively (Table 1).
Notably, catalyst–substrate complexes were not observed in
1
the H NMR spectrum of the reaction mixture, suggesting that
the Ru–alkoxide complexes of 8 and 9 are less favored than
the free Ru–OtBu complex. Disfavored formation of the Ru–alk-
oxide complexes may be ascribed to steric bulk created by the
adjacent freely rotating protecting group that could interfere
in the interaction between the ruthenium and hydroxyl group.
The equilibrium mixture consisted of 8 and 9 in a 1:1.3 ratio,
accompanied by 10% of ketones 10 and 11 combined.
Dynamic kinetic asymmetric transformation (DYKAT)
Inspired by the efficient redox epimerization of glycals 4 and 8
at moderate temperatures, we were interested in achieving
a DYKAT by coupling this process with a stereoselective biocat-
alytic reaction. By choosing an (R)-selective lipase, favoring the
transesterification of l-rhamnal epimer 9, an inversion of the
C3 alcohol stereocenter would be obtained (Figure 8). Owing
to its known stability and activity over prolonged reaction
times the lipase Candida antartica lipase B, CALB, was consid-
ered suitable as a transesterification catalyst.
Ring conformation and equilibrium
4
The d-glycals 4 and 5 were assigned to have the H5 conforma-
tion,[22] whereas the 6-deoxy-l-glycals 8 and 9 have the 5H4
conformation; in both cases this results in an antiperiplanar re-
lationship between the axially oriented H4 and H5 atoms. The
conformational assignments rely on the large JH4,H5 coupling
constants (9.6–10.2 Hz) and are supported by previous reports
on similar compounds (Figure 7).[23] Compound 4, in which the
OH group is equatorially oriented, dominated the equilibrium
between 4 and 5, whereas compound 9, in which the OH
group is axially oriented, was the most populated epimer in
the equilibrium between 8 and 9.
The biocatalyst was mixed with tBuOK-activated 1, isopro-
penyl acetate, and Na2CO3 in toluene at 408C and the reaction
reached full conversion after six days, transforming 81% (75%
isolated yield) of l-rhamnal into 12.[26] The remaining 19% con-
Chem. Eur. J. 2014, 20, 14756 – 14762
www.chemeurj.org
14758
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
a relatively high water content
in the reaction solution, increas-
ing the risk of decomposition of
1 in the DYKAT reaction.
Several attempts at DYKAT of
4 were performed at room tem-
perature employing 1 for epime-
rization; however, the reaction
was in all cases terminated
before full conversion was
reached. The best result was ob-
tained after storing the enzyme
preparation together with freshly
activated molecular sieves under
vacuum for three days, and
using 2,2,2-trifluoroethyl buty-
rate as an acyl donor; thereby
reaching 55% conversion (44%
isolated) of butyrate ester 13
after 18 h (Figure 10). At this
point only 13 and 4 could be
Figure 8. a) Empirical enantiopreference for a lipase and a protease. L=Large-sized substituent, M=medium-
sized substituent. b) Glycals preferred in transesterifications.
1
observed by H NMR spectrosco-
py, indicating that the enzyme
was highly selective for transes-
terification of 5, but that the epi-
Figure 9. DYKAT of l-rhamnal 8 (isolated yield in parentheses).
sisted of ketones 10 (8%) and
11 (11%) (Figure 9). These com-
pounds arise from the rutheni-
um alkoxide of 8 (or its epimer
9), where ketone 10 is formed
via dehydrogenation and ketone
11 via isomerization.[21] Notably,
enzymatic acylation of (S)-alco-
hol 8 was not observed, showing
that CALB is highly selective for
Figure 10. DYKAT of 4,6-O-benzylidene-d-glucal, 4. The reaction stopped at 85% conversion due to decomposition
transesterification of the corre- of the epimerization catalyst. Isolated yield in parentheses.
sponding (R)-alcohol.
To achieve a similar inversion
at C3 of the d-glucal derivative 4 to acylated d-allal, a lipase is
unsuitable since the stereogenic center at C3 already possesses
the configuration preferred in a lipase-catalyzed transesterifica-
tion.[27] This was verified by subjecting 4 to CALB and isopro-
penyl acetate in dry toluene, which afforded 17% of acetylated
product after only 1 h of stirring. Instead, a protease, Subtilisin
Carlsberg, displaying the opposite stereoselectivity compared
to a lipase, was chosen for selective transesterification of the
epimer of 4 (i.e. 5). The activity of commercially available Subti-
lisin Carlsberg is low in transesterification reactions, but it can
be enhanced by coating the enzyme with the surfactants octyl
b-d-glucopyranoside and Brij 58 (polyoxyethylene (20) cetyl
ether).[12] These surfactants form a reversed micelle around the
enzyme with the concomitant solubilization of small amounts
of water, thereby providing a stable aqueous microenviron-
ment.[28] The drawback of using this coating technique is the
requirement of a more polar solvent such as THF. Water pres-
ent in this solvent and in the reversed micelles adds up to give
merization had ceased. This was further confirmed by adding
racemic 1-phenylethanol after 18 h of reaction time and ob-
serving acylation with 16% conversion within 2 h. To achieve
a higher conversion, the epimerization catalyst was added in
three portions each of 2.5 mol% at 18 h intervals. Utilizing this
protocol the conversion of 4 to 13 was increased to 78% (71%
isolated) with a small fraction of the glucal ester 14, arising
from a background chemical acylation. About 15% of the start-
ing material 4 remained unconverted.
Also the reverse reaction, that is, the transformation of d-
allal 5 to acetylated d-glucal derivative 15, was performed.
CALB was used as the acylating enzyme and full conversion
was achieved within 18 h of stirring yielding 15 in 83% yield
(Figure 11). The possibility of obtaining carbohydrates with
both axial and equatorial hydroxyl groups highlights the viabil-
ity of the protocol and makes it an interesting method for the
preparation of various carbohydrate isomers.
Chem. Eur. J. 2014, 20, 14756 – 14762
www.chemeurj.org
14759
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
a Bruker 600 MHz spectrometer equipped with a BBI Z-gradient
probe for 1H NMR measurements. Temperatures were calibrated
1
from H peak separation of neat ethylene glycol.[34] Chemical shifts
were referenced to intrinsic [D7]toluene (dH =7.09). 1H chemical
n
shifts and JHH coupling constants were determined with the aid of
the PERCH NMR spin simulation software (PERCH Solutions Ltd.,
Kuopio, Finland). Chemical shifts and coupling constants were al-
tered iteratively until the simulated and experimental spectra ap-
peared highly similar according to visual inspection and the total
root-mean-square value was close to or below 0.1%. NMR chemical
shift assignments for alcohols 5[35] and 9[36] as well as for ketones
6[19] and 10[37] were in agreement with literature data.
Figure 11. DYKAT of d-allal 5 (isolated yield in parentheses).
Conclusion
General procedure for NMR spectroscopic studies of epime-
rization mixtures by NMR spectroscopy
The epimerization of glycals 4 and 8 with cyclopentadienylru-
theium catalyst 1, studied by NMR spectroscopic methods, was
found to be significantly more efficient than that of saturated
carbohydrates.[15] The epimerization rates varied with the sub-
stitution pattern of the glycals; the 4,6-di-O-benzylidene-pro-
tected compounds (4 and 5) reacted faster than the 4-O-
benzyl-6-deoxy compounds (8 and 9). Thanks to the efficacy of
the epimerization, which occurred at ambient temperatures,
the combination of the epimerization protocol with an enzy-
matic resolution was possible, resulting in a DYKAT. Due to ex-
cellent selectivity for acylation of the axial hydroxyl groups at
C3, when employing Subtilisin Carlsberg for the d-glycal system
4/5 and CALB for the l-glycal system 8/9, we were able to ach-
ieve a net inversion of a stereogenic center on the carbohy-
drate derivatives. Also by changing from Subtilisin Carlsberg to
CALB the reverse reaction was achieved, transforming axial
substrate 5 to its equatorial epimer ester 15. To the best of our
knowledge this is the first DYKAT on glycals of this type. The
acyl derivatives 12 and 13 could be employed in syntheses to-
wards a-l-digitoxose-containing natural products[29] such as Ja-
domycin B[30] and Kijanimicin[31] (from 12), and allosamidin[32]
and a-C-glycosides[33] (from 13).
Catalyst
1 (6.4 mg, 0.01 mmol) was dissolved in [D8]toluene
(0.1 mL), and activated by treatment with tBuOK (20 mL, 0.5m in
dry [D8]THF), under an argon atmosphere. The activation of the
catalyst complex was verified by 13C NMR spectroscopy[11] at 258C.
The diastereomerically pure glycal (0.1 mmol) was dissolved in
[D8]toluene (0.4 mL), and added to the activated catalyst. The
sample mixture was entered into the NMR spectrometer, and the
system was tuned, matched, locked and shimmed at 258C, where-
1
after a H NMR spectrum was recorded. The establishment of equi-
1
librium was monitored by subsequent recordings of H NMR spec-
tra.
The 1D H,1H-EXSY spectra of compounds 4 and 5 were acquired
at 908C, with a mixing time of 3 s and total relaxation delay of
10.7 s. Exchange matrix analysis of the EXSY spectra was performed
by employing the EXSYCalc software, version 1.0 (Mestrelab Re-
search).[38] Three-dimensional models of compounds 4, 5, 8, and 9
were built using the Vega ZZ software (release 3.0.1.22).[39]
1
1,5-Anhydro-4-O-benzyl-2,6-dideoxy-l-erythro-hex-1-eno-3-ulose
(10): 1H NMR (CDCl3, 400 MHz): d=7.41–7.25 (5H, m, ArH), 7.37
(1H, d, J12 =5.9 Hz, H1), 5.38 (1H, d, J12 =5.9 Hz, H2), 5.03 (1H, d,
J
gem =ꢀ11.5 Hz, CH2-Ph), 4.66 (1H, d, Jgem =ꢀ11.5 Hz, CH2-Ph), 4.48
(1H, dq, J45 =9.8 Hz, J56 =6.5 Hz, H5), 3.72 (1H, d, J45 =9.8 Hz, H4),
1.43 ppm (3H, d, J56 =6.5 Hz, H6); 13C NMR (CDCl3, 100 MHz): d=
193.0, 162.1, 137.4, 128.42 (2C), 128.41 (2C), 128.0, 105.0, 78.66,
78.65, 73.9, 17.1 ppm; HRMS (ESI): calcd for [M+Na] C13H14NaO3:
241.0831, found 241.0835.
Experimental Section
1,5-Anhydro-4-O-benzyl-2,6-dideoxy-l-erythro-hex-3-ulose (11):
1H NMR (CDCl3, 400 MHz): d=7.39–7.27 (5H, m, ArH), 4.95 (1H,
General methods
J
gem =ꢀ11.4 Hz, CH2-Ph), 4.50 (1H, Jgem =ꢀ11.4 Hz, CH2-Ph), 4.20
(1H, ddd, Jgem =ꢀ11.4 Hz, J1proꢀR,2proꢀR =1.4 Hz, J1proꢀR,2proꢀS =7.4 Hz,
H1pro-R); 3.64 (1H, ddd, 1proꢀS,2proꢀR =2.5 Hz,
gem =ꢀ11.4 Hz,
1proꢀS,2proꢀS =12.6 Hz, H1pro-S), 3.64 (1H dd, J2proꢀS,4 =1.3 Hz, J45
9.4 Hz, H4), 3.59 (1H, dq, J56 =6.0 Hz, J45 =9.4 Hz, H5), 2.71 (1H,
dddd, 2proꢀS,4 =1.3 Hz, 1proꢀS,2proꢀS =12.6 Hz,
1proꢀR,2proꢀS =7.4 Hz, H2pro-S), 2.46 ppm (1H, ddd, Jgem =ꢀ13.9 Hz,
Unless otherwise noted, all reagents and reactants were used as re-
ceived from commercial suppliers. Isopropenyl acetate was dried
over CaCl2 and distilled before use. Dry toluene was obtained from
a VAC-solvent purifier system and dry THF through distillation over
sodium/benzophenone. All other solvents were obtained from
commercial suppliers and used without further purification/drying.
For reactions conducted under dry/inert conditions standard
Schlenk techniques were used (unless otherwise noted). All reac-
tions were monitored by thin-layer chromatography using E. Merck
silica gel 60 F254 plates (TLC analysis). Flash chromatography was
carried out with 60 ꢂ (particle size 35–70 mm) normal phase silica
gel. Optical rotation was measured with a polarimeter equipped
with a Na lamp. HR-MS were recorded on an ESI mass spectrome-
ter. Ruthenium complex [(h5C5Ph5)Ru(CO)2Cl] (1) was synthesized
according to a literature procedure.[11] CALB (Novozym 435) and
CLEA201-UF is commercially available and surfactant-treated Subti-
lisin Carlsberg was prepared according to known procedures.[12]
NMR experiments for conformational analysis were carried out on
J
J
J
=
J
gem =ꢀ13.9 Hz,
J
J
J
J
1proꢀR,2proꢀR =1.4 Hz, J1proꢀS,2proꢀR =2.5 Hz, H2pro-R), 1.38 (3H d, J56
=
6.0 Hz, H6); 13C NMR (CDCl3, 100 MHz): d=205.8, 137.5, 128.5 (2C),
128.3 (2C), 128.0, 85.3, 78.3, 73.4, 67.0, 42.8, 19.3 ppm; HRMS (ESI):
calcd for [M+Na] C13H16NaO3: 243.0996, found 243.0092.
Dynamic Kinetic Asymmetric Transformations
3-O-Acetyl-1,5-anhydro-4-O-benzyl-2,6-dideoxy-l-ribo-hex-1-
enitol (12): To a flame-dried 5 mL micro-vial were added [(h5-
C5Ph5)Ru(CO)2Cl]
1
(4.75 mg, 0.0075 mmol), Na2CO3 (16 mg,
0.15 mmol), and CALB (8 mg). The system was purged with three
argon-vaccum cycles. The vial was transferred to an Ar-filled glove
Chem. Eur. J. 2014, 20, 14756 – 14762
www.chemeurj.org
14760
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
box and dry, degassed toluene (0.6 mL) was added. tBuOK (15 mL,
0.5m in dry THF, 0.0075 mmol) was added and a color change from
yellow to red was observed. After 6 min of stirring the cap was re-
moved and 4-O-benzyl-l-rhamnal 8 (33 mg, 0.15 mmol) was added
in one portion as a solid and the system was quickly sealed with
a teflon septum cap. After an additional 4 min isopropenyl acetate
(45 mL, 0.45 mmol) was added and the vial was taken out of the
glove box and stirred at 408C for six days. When full conversion
was reached the reaction mixture was filtered through Celite with
EtOAc, and the solvent was subsequently evaporated. The residue
was purified by column chromatography (SiO2, gradient from pen-
tane to pentane/EtOAc 9:1) yielding 29 mg (75%) of 12 in the
second fraction.
(2C), 101.7, 101.1, 77.2, 69.0, 68.7, 68.5, 36.4, 18.6 13.7 ppm; HRMS
(ESI): calcd for [M+Na] C17H20NaO5: 327.1203, found 327.1191;
[a]25D =ꢀ838 (c=0.2, CDCl3).
3-O-Acetyl-1,5-anhydro-4,6-O-benzylidene-2-deoxy-d-arabino-
hex-1-enitol (15): To a flame-dried micro-vial were added [(h5-
C5Ph5)Ru(CO)2Cl] 1 (9 mg, 0.015 mmol), Na2CO3 (36 mg, 0.15 mmol),
and CALB (33 mg). The system was purged with three argon-
vacuum cycles after which degassed toluene (1 mL) was added.
tBuOK (32 mL, 0.5m in dry THF, 0.0016 mmol) was added and
a color change from yellow to red was observed. After 6 min of
stirring the cap was removed and 4,6-O-benzylidene-d-allal 5
(78 mg, 0.33 mmol) was added in one portion as a solid and the
system was quickly sealed with a teflon septum cap. After an addi-
tional 4 min isopropenyl acetate (110 mL, 1 mmol) was added and
the reaction mixture was stirred at 408C for 18 h. When full conver-
sion was reached the reaction mixture was filtered through celite
with EtOAc and the solvent was subsequently evaporated. The resi-
due was purified by column chromatography (SiO2, gradient from
pentane to pentane/EtOAc 9:1) yielding 76 mg (83%) of 15 in the
second fraction. Experimental data were in accordance with those
previously reported.[32] 1H NMR (CDCl3, 400 MHz): d=7.52–7.47 (2H,
1H NMR (CDCl3, 400 MHz): d=7.38–7.28 (5H, m, ArH), 6.47 (1H, d,
J
12 =5.9 Hz, H1), 5.52 (1H, dd, J23 =6.0 Hz, J34 =3.7 Hz, H3), 4.87
(1H, dd, J12 =5.9, J23 =6.0 Hz, H2), 4.71 (1H, d, Jgem =ꢀ11.4 Hz, CH2-
Ph), 4.45 (1H, d, Jgem =ꢀ11.4 Hz, CH2-Ph), 4.08 (1H, dq, J45
=
10.2 Hz, J56 =6.3 Hz, H5), 3.42 (1H, dd, J34 =3.7 Hz, J45 =10.2 Hz,
H4), 2.08 (3H, s, Me), 1.37 ppm (3H, d, J56 =6.3 Hz, H6); 13C NMR
(CDCl3, 100 MHz): d=171.0, 148.4, 137.7, 128.6 (2C), 128.4 (2C),
128.1, 97.3, 77.3, 71.9, 70.5, 62.0, 21.5, 17.8 ppm; HRMS (ESI): calcd
for [M+Na] C15H18NaO4: 285.1097, found: 285.1089; [a]25D =ꢀ1638
(c=0.2, CDCl3).
m, ArH), 7.41–7.35 (3H, m, ArH), 6.39 (1H, dd, J12 =6.2 Hz, J13
1.5 Hz, H1), 5.60 (1H, s, CHPh), 5.53 (1H, ddd, J34 =7.7 Hz, J23
=
=
2.1 Hz, J13 =1.5 Hz, H3), 4.81 (1H, dd, J12 =6.2 Hz, J23 =2.1 Hz, H2),
4.39 (1H, dd, J56’ =5.1 Hz, Jgem =ꢀ10.7 Hz, H6’), 4.03 (1H, dd, J34
7.7 Hz, J45 =10.5 Hz, H4), 3.99 (1H, ddd, J45 =10.5 Hz, J56 =10.2 Hz,
1,5-Anhydro-4,6-O-benzylidene-3-O-butanoyl-2-deoxy-d-ribo-
hex-1-enitol (13): To a flame-dried Schlenk flask was added [(h5-
=
C5Ph5)Ru(CO)2Cl]
1
(7.9 mg, 0.0125 mmol), Na2CO3 (53 mg,
J
56’ =5.1 Hz, H5), 3.85 (1H, dd, J56 =10.2 Hz, Jgem =ꢀ10.7 Hz, H6),
0.5 mmol), and surfactant-treated Subtilisin Carlsberg (9 mg). The
system was purged with three argon-vacuum cycles. Distilled and
degassed THF (1 mL) was added followed by tBuOK (25 mL, 0.5m in
dry THF, 0.0125 mmol), giving rise to a color change from yellow to
red. After 6 min stirring 4,6-O-benzylidene-d-glucal 4 (117 mg,
0.5 mmol) was added in one portion as a solid. After an additional
4 min 2,2,2-trifluoroethyl butyrate (150 mL, 1 mmol) was added.
Two batches of activated 1 (7.9 mg in 0.2 mL distilled and de-
gassed THF activated by tBuOK as previously) was added after 18 h
and 36 h, thereafter the reaction was stirred for additionally 4 h.
The reaction mixture was filtered through Celite with EtOAc,
evaporated onto SiO2 (0.2 g), and purified by column chromatogra-
phy (4 g SiO2, eluent pentane to pentane/EtOAc 10:1) yielding
107 mg (71%) of 13 as a white solid.
2.09 ppm (3H, s, Me); 13C NMR (CDCl3, 100 MHz): d=170.9, 145.6,
137.1, 129.4, 128.5 (2C), 126.4 (2C), 101.8, 100.9, 77.1, 69.03, 69.00,
68.5, 21.4 ppm; HRMS (ESI): calcd for [M+Na] C15H16NaO5:
299.0890, found 299.0894; [a]25D =ꢀ958 (c=0.8, CDCl3).
Acknowledgements
This work was supported by grants from the Swedish Research
Council and The Knut and Alice Wallenberg Foundation.
Keywords: carbohydrates
·
dynamic kinetic asymmetric
1H NMR (CDCl3, 400 MHz): d=7.49–7.42 (2H, m, ArH), 7.39–7.32
(3H, m, ArH), 6.50 (1H, d, J12 =6.0 Hz, H1), 5.61 (1H, s, CHPh), 5.46
transformation
spectroscopy
·
enzyme catalysis inversion NMR
·
·
(1H, dd, J23 =5.9 Hz, J34 =4.0 Hz, H3), 5.01 (1H, dd, J12 =6.0 Hz, J23
=
5.9 Hz, H2), 4.46 (1H, dd, J56’ =5.3 Hz, Jgem =ꢀ10.7 Hz, H6’), 4.18
[1] X. Yu, G. O’Doherty in Chemical glycobiology, (Eds.: X. Chen, R. Halcomb,
P. G. Wang), ACS Symposium Series; American Chemical Society, Wash-
ington, DC, 2008, Chapter 1, pp. 3–28.
(1H, ddd, J45 =10.5 Hz, J56 =10.3 Hz, J56’ =5.3 Hz, H5), 3.98 (1H, dd,
J
34 =4.0 Hz, J45 =10.5 Hz, H4), 3.84 (1H, at, J56 10.3 Hz, Jgem
ꢀ10.7 Hz, H6), 2.33 (2H, t, J 7.5 Hz, -CH2CH2CH3), 1.66 (2H, tq, J=
7.5 Hz, J=7.5 Hz, -CH2CH2CH3), 0.93 ppm (3H, t, J=7.5 Hz,
-CH2CH2CH3); 13C NMR (CDCl3, 100 MHz): d=173.3, 147.3, 137.2,
129.2, 128.4 (2C), 126.2 (2C), 101.7, 98.7, 76.2, 68.8, 65.1, 61.8, 36.5,
18.6, 13.8 ppm; HRMS (ESI): calcd for [M+Na] C17H20NaO5:
327.1195, found 327.1203; [a]25D = +1838 (c=0.2, CDCl3).
[2] S. Y. Ko, A. W. M. Lee, S. Masamune, L. A. Reed, K. B. Sharpless, F. J.
Walker, Science 1983, 220, 949–951.
[3] M. Hederos, P. J. J. Konradsson, J. Carbohydr. Chem. 2005, 24, 297–320.
[4] A. E. Christina, V. M. Blas Ferrando, F. de Bordes, W. A. Spruit, H. S. Over-
kleeft, J. D. C. Codꢃe, G. A. van der Marel, Carbohydr. Res. 2012, 356,
282–287.
[5] N. K. Kochetkov, B. A. Dmitriev, O. S. Chizhov, E. M. Klimov, N. N. Maly-
sheva, A. Y. Chernyak, N. E. Bayramova, V. I. Torgov, Carbohydr. Res.
1974, 33, C5–C7.
1,5-Anhydro-4,6-O-benzylidene-3-O-butanoyl-2-deoxy-d-arabino-
1
hex-1-enitol (14): H NMR (CDCl3, 400 MHz): d=7.52–7.45 (2H, m,
[6] P. Sçderman, E. A. Larsson, G. Widmalm, Eur. J. Org. Chem. 2002, 1614–
1618.
[7] H. Mobarak, O. Engstrçm, G. Widmalm, RSC Adv. 2013, 3, 23090–23097.
[8] R. J. Kazlauskas, A. N. E. Weissfloch, A. T. Rappaport, L. A. Cuccia, J. Org.
Chem. 1991, 56, 2656–2665.
[9] R. J. Kazlauskas, A. N. E. Weissfloch, J. Mol. Catal. B 1997, 3, 65–72.
[10] a) R. Marcos, B. Martꢄn-Matute, Isr. J. Chem. 2012, 52, 639–652; b) H. Pel-
lissier, Tetrahedron 2011, 67, 3769–3802; c) J. H. Lee, K. Han, M. J. Kim, J.
Park, Eur. J. Org. Chem. 2010, 999–1015; d) Y. Ahn, S. B. Ko, M. J. Kim, J.
Park, Coord. Chem. Rev. 2008, 252, 647–658.
ArH), 7.41–7.33 (3H, m, ArH), 6.39 (1H, dd, J12 =6.1 Hz, J13 =1.5 Hz,
H1), 5.60 (1H, s, CHPh), 5.56 (1H, ddd, J13 =1.5 Hz, J23 =2.1 Hz, J34
=
7.6 Hz, H3), 4.80 (1H, dd, J12 =6.1 Hz, J23 =2.1 Hz, H2), 4.39 (1H, dd,
J
56’ =4.8 Hz, Jgem =ꢀ10.6 Hz, H6’), 4.04 (1H, dd, J34 =7.6 Hz, J45
=
10.5 Hz, H4), 3.99 (1H, ddd, J45 =10.5 Hz, J56 =10.2 Hz, J56’ =4.8 Hz,
H5), 3.85 (1H, dd, J56 =10.2 Hz, Jgem =ꢀ10.6 Hz, H6), 2.32 (2H, t, J=
7.4 Hz, -CH2CH2CH3), 1.66 (2H, tq, J=7.4 Hz, J=7.4 Hz,
-CH2CH2CH3), 0.93 ppm (3H, t, J=7.4 Hz, -CH2CH2CH3); 13C NMR
(CDCl3, 100 MHz): d=173.6, 145.5, 137.1, 129.3, 128.5 (2C), 126.3
Chem. Eur. J. 2014, 20, 14756 – 14762
www.chemeurj.org
14761
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[11] B. Martꢄn-Matute, M. Edin, K. Bogꢅr, F. B. Kaynak, J.-E. Bꢀckvall, J. Am.
Chem. Soc. 2005, 127, 8817–8825.
[12] a) L. Borꢃn, B. Martꢄn-Matute, Y. Xu, A. Cꢆrdova, J.-E Bꢀckvall, Chem. Eur.
J. 2006, 12, 225–232. For the use of a related catalyst in combination
with S-selective proteases see: b) M. J. Kim, Y. Chung, Y. Choi, H. Lee, D.
Kim, J. Park, J. Am. Chem. Soc. 2003, 125, 11494–11495.
[13] M. C. Warner, J.-E. Bꢀckvall, Acc. Chem. Res. 2013, 46, 2545–2555.
[14] a) J. B. ꢂberg, J. Nyhlꢃn, B. Martꢄn-Matute, T. Privalov, J.-E Bꢀckvall, J. Am.
Chem. Soc. 2009, 131, 6035; b) M. C. Warner, O. Verho, J.-E. Bꢀckvall, J.
Am. Chem. Soc. 2011, 133, 2820–2823.
[15] C. Ramstadius, A. M. Trꢀff, P. Krumlinde, J.-E. Bꢀckvall, I. Cumpstey, Eur. J.
Org. Chem. 2011, 4455–4459.
[16] a) D. Lee, E. A. Huh, M.-J. Kim, H. M. Jung, J. H. Koh, J. Park, Org. Lett.
2000, 2, 2377–2379; b) M. C. Warner, A. Nagendiran, K. Bogꢅr, J.-E. Bꢀck-
vall, Org. Lett. 2012, 14, 5094–5097; c) R. Lihammar, R. Millet, J.-E. Bꢀck-
vall, J. Org. Chem. 2013, 78, 12114–12120.
[17] Epimerization reaction with catalyst 2 was carried in dry toluene. Epi-
merization reaction with catalyst 3 was carried out in dry acetonitrile
using both activated and unactivated 3. Activation was performed by
mixing 3 with iPrOH and Et3N in anhydrous DCM.
[24] a) H. Zhao, Q. Pan, W. Zhang, I. Carmichael, A. S. Serianni, J. Org. Chem.
2007, 72, 7071–7082; b) E. Arunan, G. R. Desiraju, R. A. Klein, J. Sadlej,
S. Scheiner, I. Alkorta, D. C. Clary, R. H. Crabtree, J. J. Dannenberg, P.
Hobza, H. G. Kjaergaard, A. C. Legon, B. Mennucci, D. J. Nesbitt, Pure
Appl. Chem. 2011, 83, 1619–1636.
[25] Epimerization reactions where 6 equivalents of dry tBuOH were added
at the start of the reaction resulted in a 48:52 mixture of 8 and 9 at
258C (Keq =1.08).
[26] A slight increase in reaction temperature, from 258C to 408C, gave
a more rapid epimerization, decreasing the reaction time from eight to
six days.
[27] T. Sakakibara, T. Ito, C. Kotaka, Y. Kajihara, Y. Watanabe, A. Fujioka, Car-
bohydr. Res. 2008, 343, 2740–2743.
[28] K. Sawada, M. Ueda, J. Chem. Technol. Biotechnol. 2004, 79, 369–375.
[29] S. C. Timmons, D. L. Jakeman, Carbohydr. Res. 2007, 342, 2695–2704.
[30] M. Shan, E. U. Sharif, G. A. O’Doherty, Angew. Chem. 2010, 122, 9682–
9685; Angew. Chem. Int. Ed. 2010, 49, 9492–9495.
[31] a) A. K. Mallams, M. S. Puar, R. R. Rossman, A. T. McPhail, P. M. Gross,
R. D. Macfarlane, R. L. Stephens, J. Chem. Soc. Perkin Trans. 1 1983,
1497–1534; b) H. Zhang, J. A. White-Phillip, C. E. Melancon, H.-J. Kwon,
W.-L. Yu, H.-W. Liu, J. Am. Chem. Soc. 2007, 129, 14670–14683.
[32] D. A. Griffith, S. J. Danishefsky, J. Am. Chem. Soc. 1996, 118, 9526–9538.
[33] H. Y. Godage, D. J. Chambers, G. R. Evans, A. J. Fairbanks, Org. Biomol.
Chem. 2003, 1, 3772–3786.
[34] C. Amman, P. Meier, A. E. Merbach, J. Magn. Reson. 1982, 46, 319–321.
[35] A. A. Feast, W. G. Overend, N. A. Williams, J. Chem. Soc. 1965, 7378–
7388.
[36] S. Kçpper, I. Lundt, C. Pedersen, J. Thiem, Liebigs Ann. Chem. 1987,
531–535.
[18] J. B. ꢂberg, M. C. Warner, J.-E. Bꢀckvall, J. Am. Chem. Soc. 2009, 131,
13622–13624.
[19] In an EXSY experiment, chemical exchange is monitored. A peak is se-
lectively excited, whereafter the magnetization is transferred to the
other species involved in the equilibrium during a mixing period. This
gives an exchange peak, in addition to the excited peak, in the result-
ing spectrum. The integral of the exchange peak, and the ratio of the
excited and the exchange peak, depends on the length of the mixing
period, the exchange rate and the relaxation rate.
[20] J. C. Fischer, D. Horton, W. Weckerle, Can. J. Chem. 1977, 55, 4078–
4089.
[21] a) N. Ahlsten, A. Bartoszewicz, B. Martꢄn-Matute, Dalton Trans. 2012, 41,
1660–1670; b) B. Martꢄn-Matute, K. Bogꢅr, M. Edin, F. B. Kaynak, J.-E.
Bꢀckvall, Chem. Eur. J. 2005, 11, 5832–5842.
[37] A. Kirschning, J. Org. Chem. 1995, 60, 1228–1232.
[38] EXSYCalc Version 1.0, J. C. Cobas, M. Martꢄn-Pastor, ꢁMestrelab Re-
search.
[39] A. Pedretti, L. Villa, G. Vistoli, J. Mol. Graphics Modell. 2002, 21, 47–49.
[22] In this half-chair conformation, C4 is above the plane and C5 is below
the plane; for this nomenclature see: G. P. Moss, Pure Appl. Chem. 1996,
68, 2193–2222.
Received: May 28, 2014
[23] a) C. Ernst, M. Piacenza, S. Grimme, W. Klaffke, Carbohydr. Res. 2003,
338, 231–236; b) A. Rivera-Sagredo, J. Jimꢃnez-Barbero, J. Carbohydr.
Chem. 1992, 11, 903–919.
Revised: August 15, 2014
Published online on September 18, 2014
Chem. Eur. J. 2014, 20, 14756 – 14762
www.chemeurj.org
14762
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim