Enzyme-purification and catalytic transformations in a microstructured PASSflow reactor using a new tyrosine-based Ni-NTA linker system attached to a polyvinylpyrrolidinone-based matrix

Article abstract of DOI:10.1039/b712804e

The synthesis of a Ni-nitrilotriacetic acid (Ni-NTA) attached via a new tyrosine-based linker matrix on monolithic crosslinked poly(vinyl benzyl chloride)/poly(vinylpyrrolidinone) is described. This matrix is incorporated inside a microstructured PASSflow reactor which was used for automatic purification and immobilisation of His₆-tagged proteins. These could be used as stable and highly active biocatalysts for the synthesis of (R)-benzoin (6), (R)-2-hydroxy-1-phenylpropan-1-one (7) and 6-O-acetyl-d-glucal (17) in a flow-through mode. The Royal Society of Chemistry.

Full text of DOI:10.1039/b712804e

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Enzyme-purification and catalytic transformations in a microstructured
PASSflow reactor using a new tyrosine-based Ni-NTA linker system attached
to a polyvinylpyrrolidinone-based matrix†
Gerald Dra¨ger,*a Csilla Kiss,a Ulrich Kunzb and Andreas Kirschning*a
Received 20th August 2007, Accepted 18th September 2007
First published as an Advance Article on the web 5th October 2007
DOI: 10.1039/b712804e
The synthesis of a Ni-nitrilotriacetic acid (Ni-NTA) attached via a new tyrosine-based linker matrix on
monolithic crosslinked poly(vinyl benzyl chloride)/poly(vinylpyrrolidinone) is described. This matrix is
incorporated inside a microstructured PASSflow reactor which was used for automatic purification and
immobilisation of His6-tagged proteins. These could be used as stable and highly active biocatalysts for
the synthesis of (R)-benzoin (6), (R)-2-hydroxy-1-phenylpropan-1-one (7) and 6-O-acetyl-D-glucal (17)
in a flow-through mode.
solution along the polymeric matrix. The reactor itself is integrated
Introduction
into an HPLC-based solvent-system, allowing automated reactor
Continuous flow reactors using immobilised enzymes have been
loading, reaction control and reactor regeneration (Fig. 1).
widely used in industrial applications.1 Commonly, enzymes are
For the present purposes we had to find the best polymeric
immobilised onto a solid support by either physical adsorption,
phase and linker system for functionalising the interior of the
membrane entrapment, polymeric gel entrapment or covalent
reactor with Ni-NTA groups. The immobilised nitrilotriacetic acid
bonding. Immobilisation has been achieved on different carrier
(NTA) is a phase suitable for metal ion affinity chromatography
materials such as controlled pore glass, microporous silica,
(IMAC). IMAC matrices can be used for purification of His6-
nanoporous aluminium oxide, polyaminostyrene, nylon, poly(2-
tagged proteins from crude cell extracts and have served as carriers
for tagged enzymes in biocatalysis.7 Besides the known lysine
linker, we planned to develop a new linker system based on
tyrosine.
hydroxyethylmethacrylate), polyurethane, and sepharose, just to
name the most important ones.2 Additionally, crosslinked enzyme
membranes have been described.3 A major advantage of immo-
bilised enzymes is their insolubility, allowing easy product sepa-
ration and reuse. Often they show increased stability compared to
their homogeneous counterparts.
Technically, we first tested the new, functionalised polymers in
form of a powder (obtained by precipitation polymerization)6 and
then transferred the conditions to a monolithic glass/polymer
composite material shaped in the form of Raschig rings (Fig. 2).
Finally, we incorporated the new IMAC-phases inside a PASSflow
microreactor which contains the same composite material.
So far little work has been devoted to the development of an
immobilisation strategy within a continuous flow reactor system
that allows purification of the target enzymes by means of Ni-
NTA-based immobilisation as well as synthetic use of the enzyme
probe under continuous flow conditions. In order to achieve this
goal, the solid phase, the linker system, as well as the reactor, have
to be optimised to form an ideal enabling technology platform.4
In this report we disclose a microreactor concept for im-
mobilised enzymes, which is based on our recently devel-
oped PASSflow (polymer-assisted solution phase synthesis under
flow conditions) system.5,6 The reactor concept is based on a
pressure-resistant reactor housing filled with a porous, monolithic
glass/polymer composite material which is appropriately func-
tionalised for immobilising different chemical species. The com-
posite is set up of a continuous polymeric phase (interconnected
beads 1–10 lm) and a porous glass body (irregular microchannels;
10–100 lm diameter), which leads to a forced convective flow of the
Results and discussion
Preparation and immobilisation of lysine- and tyrosine-based
Ni-NTAs on optimised polymer
First, we prepared two NTA-linker systems. While one linker is
based on an established literature procedure8 using lysine, the other
NTA-linker utilises tyrosine as central linker element. Besides
optimising the linker for the chosen polymer phase, the latter
approach allows us to circumvent the current patent situation.
The preparation of both linker systems is described in Scheme 1
and Scheme 2. Thus, Cbz-protected L-lysine (1) was alkylated with
bromoacetic acid to yield the NTA functionality. Protection of the
three carboxylic acids as methyl esters was necessary to enhance
the solubility in N-methylpyrrolidinone (NMP) required for the
solid phase attachment. Finally, Cbz cleavage was accomplished
by hydrogenation using Pd/C.
aInstitut fu¨r Organische Chemie and Zentrum fu¨r Biomolekulare Wirk-
stoffchemie (BMWZ), Leibniz Universita¨t Hannover, Schneiderberg 1b,
30167, Hannover, Germany. E-mail: draeger@oci.uni-hannover.de; Fax: +49
(0)511-7623011
bInstitut fu¨r Chemische Verfahrenstechnik der Technischen Universita¨t
Clausthal, Leibnizstrasse 17, D-38678, Clausthal-Zellerfeld, Germany
Immobilisation was first addressed using poly(vinyl benzyl
chloride) (VBC) crosslinked with 5% divinyl benzene (DVB).
Amino ester 2 was reacted with Merrifield resin and sodium
† Electronic supplementary information (ESI) available: IR spectra of
functionalised polymers and SDS-Page. See DOI: 10.1039/b712804e
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 3657–3664 | 3657
©

View Article Online
Fig. 1 Typical PASSflow reactor used for this study, and schematic view of the PASSflow-NTA reactor setup.
Fig. 2 Types of solid phases used.
Scheme 2 Synthesis of tyrosine-based NTA linker on crosslinked
poly(vinylpyrrolidinone)/poly(vinyl benzyl chloride) polymer. Reagents
and conditions: a) K2CO3, TBAI, BrCH2CO2Me, 3-pentanone, 108 ◦C
(55%); b) Pd/C, H2, MeOH, rt (99.5%); c) polymer with ArCH2Cl groups
in Passflow reactor, CsI (0.5 eq.), DMF, 60 ◦C, 72 h (∼30% of theoretical
loading, 1.7 mmol g−1); d) LiOH, MeOH–H2O (1 : 1), 50 ◦C, 60 h; e) NiCl2,
H2O, DMSO, rt, 10 min.
Scheme 1 Synthesis of lysine-based NTA linker on several samples of
crosslinked polymers. Reagents and conditions: a) BrCH2CO2H, NaOH,
H2O, 40 ◦C, 12 h (87%); b) H2SO4, MeOH, reflux, 12 h; c) Pd/C, H2,
MeOH, rt, 12 h (98% for two steps); d) polymer with ArCH2Cl groups,
NaOtBu, NMP, 60 ◦C, 72 h [64% for Merrifield resin (5% DVB) determined
by gravimetry]; e) LiOH, MeOH–H2O (1 : 1), 50 ◦C, 60 h; f) NiCl2, H2O,
DMSO. Cbz = benzyloxycarbonyl; NMP = N-methylpyrrolidinone.
DVB, VBC and N-vinylpyrrolidinone (10: 70: 20 vol %).10 This
resin was prepared by precipitation polymerisation in the pore
volume of highly porous glass rods or Raschig rings to yield a
polymeric matrix inside the glass pores, as well as an excess of
the corresponding polymeric powder. The rods were embedded in
a solvent- and pressure-resistant PASSflow housing.6 The glass-
supported polymer was then treated with amino ester 2 under
the conditions described above, followed by ester cleavage (LiOH,
MeOH–THF). This was followed by treatment with a solution
of Cl2SiMe2 in CH2Cl2 for silyl capping of silanol groups on the
glass. This step prevents non-specific binding of proteins to the
glass inside the Raschig rings and PASSflow reactor, respectively.
After Ni2+ complexation we accomplished the formation of a Ni-
NTA phase 5 attached to a novel monolithic, sufficiently polar
polymeric backbone.
tert-butylate in N-methylpyrrolidinone, affording resin 3. Com-
pletion of loading was judged by IR-spectroscopy. Mild ester
hydrolysis was accomplished (LiOH, MeOH–H2O 1 : 1), yielding
resin 4.9 Finally, treatment with an aqueous solution of NiCl2
yielded Ni-NTA on the polystyrene-derived matrix 5, which could
be visualized by the colour change of the polymer from light yellow
to light blue. Several experiments failed to purify or immobilise
His6-tagged proteins on this polymer, which led us to conclude
that the polarity of the polymeric phase prevented exposure of the
Ni-NTA to the aqueous phase.
In the same manner we prepared samples of tyrosine-based
NTA-linker 7 (prepared from O-benzyltyrosine 6 by a standard
three-step sequence). Triester 7 was attached to the crosslinked
copolymer composed of DVB, VBC and vinyl pyrrolidinone as
In order to increase the polarity of the solid phase, we
prepared several crosslinked polymeric samples composed of
3658 | Org. Biomol. Chem., 2007, 5, 3657–3664
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
described for the established linker 2 (Scheme 2). The best results
were achieved for a monomeric composition of DVB/VCB/N-
vinyl pyrrolidinone (10 : 70 : 20 vol %), leading to maximum
loading of 30% for 8 (analysed by IR and gravimetrically).10
Loading of benzaldehyde lyase (BAL) and optimisation of benzoin
reaction
Studies on the protein loading and optimisation of the reaction
conditions were first conducted with the polymer powder obtained
from the precipitation polymerisation protocol. Thus, the Ni-
NTA-functionalised polymers were employed for the stepwise
purification and immobilisation of His6-tagged proteins, and in
the following could directly be used in enzymatic transformations.
As a model enzyme we settled on recombinant benzaldehyde
lyase (BAL, EC 4.1.2.38)11 a thiamine pyrophosphate-dependent
enzyme which utilises acyl anion equivalents as intermediates.12
Both optimised Ni-NTA phases proved to be suitable for immo-
bilising His6-tag-modified BAL from a cell lysate (from E. coli;
recombinant expressed BAL with His6-tag). The tyrosine-based
polymer showed improved capacities compared to the lysine-based
phase (Fig. 3). Quantification of the immobilised protein was
done with the Bradford assay13 after excessive washing (phosphate
buffer, pH 7 and imidazole buffer, 10 mM) followed by protein
elution with an imidazole buffer solution (250 mM). The purity of
the protein fractions were determined to be >98% by SDS-Page.14
Scheme 3 Optimised conditions for benzoin reaction of benzaldehyde
(10) on polymeric powder.
Fig. 4 Benzoin reaction (according to Scheme 3) with lysine-based linker
employing different amounts of benzaldehyde 10 (ꢀ = 15 lL; ᭜ = 30 lL;
ꢁ = 100 lL; ᭹ = 200 lL); all reactions were monitored by GC.
samples (15, 30, 100 and 200 lL) were quantitatively consumed
within a reasonable time (Fig. 5). When 500 lL benzaldehyde (10)
was added the reaction did not lead to completion (only 74%)
within 300 min.
Fig. 3 Amount of immobilised BAL for different volumes of cell lysate
(solid line = lysine-linker; dashed line = tyrosine linker).
For the benzoin reaction with benzaldehyde 10, it turned out
that addition of 10% of DMSO to the reaction buffer (phosphate
buffer, pH 7)15 gave best results (Scheme 3). Although BAL usually
operates best at pH 9.5, we had to lower the pH to 7 because we
encountered formation of nickel hydroxide at pH > 8.0. Under
these optimised reaction conditions the leaching of protein was
not detectable with the Bradford assay.16
Fig. 5 Benzoin reaction (according to Scheme 3) with tyrosine-based
linker employing different amounts of benzaldehyde 10 (ꢀ = 15 lL; ᭜ =
30 lL; ꢁ = 100 lL; ᭹ = 200 lL; = 500 lL); all reactions were monitored
by GC.
The lysine-based polymer loaded with His6-tagged BAL rapidly
transformed 15 and 30 lL (0.15 and 0.3 mmol) of benzaldehyde
(Fig. 4), while larger amounts (100 and 200 lL) required prolonged
reaction times. At these higher concentrations partial enzyme
inhibition exerted by the product was observed. In fact, we noted
that it was difficult to completely remove the product from the
polymeric phase.
The immobilized BAL could be reused without loss of activity
after washing the polymer either with MTB-ether or ethyl acetate
after each run and reactivation with phosphate buffer. In this way
four consecutive batch transformations were conducted (Fig. 6).
Benzoin reaction under continuous flow conditions
In contrast, the tyrosine-based polymer showed better proper-
ties under the standardised reaction conditions. All benzaldehyde
Immobilising the His6-tagged BAL inside a PASSflow reactor
loaded with functionalized polymer in principle allows one to
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 3657–3664 | 3659
©

View Article Online
enzyme is commonly used to cleave carboxyl protecting groups
such as tert-butyl, methyl, benzyl, allyl and chloroethyl esters.20 Af-
ter immobilization of His6-tagged BsubpNBE inside the PASSflow
reactor, enzyme leaching was studied using p-nitrophenyl acetate
(14), which was readily cleaved to yield 15 (Scheme 5). After having
passed through the reactor, a sample of the reaction mixture was
analyzed by HPLC and stored at 37 ◦C for 1 h to eventually
allow further enzymatic reaction catalyzed by cleaved protein.
After this time, the sample was again analyzed and compared
in order to evaluate enzyme leaching. It turned out that the co-
solvent used for dissolving the reactant is crucial for binding;
DMSO leads to detectable leaching but methyl tert-butyl ether
was fully compatible, and no transfer of enzymatic activity to the
reaction medium was observed.
Fig. 6 Repeated benzoin reactions (according to Scheme 3) with ty-
rosine-based linker; all reactions were monitored by GC.
Next, we used tri-O-acetyl-D-glucal (16) as substrate. In so-
lution, 16 is completely deacetylated within minutes by Bsub-
pNBE. Under flow-through conditions with immobilized enzyme
this reaction is much slower. After complete consumption of 16
(60 h), 6-O-acetyl-D-glucal (17) could be isolated in good yields
along with traces of 18 (Scheme 5).
conduct the benzoin reaction in a circular mode (using the reactor
in a batch mode) or in one run in a continuous flow manner.
Attachment of the lysine- and tyrosine-based NTA-linker to the
polymeric phase (DVB/VBC/N-vinylpyrrolidinone = 10 : 70 : 20)
inside the PASSflow reactor (which contained approx. 200 mg of
polymer), silylation of the glass matrix17 and the trapping of Ni2+
were carried out according to the optimised protocols described
above for the powder. The reaction mixtures were commonly
circulated at flow rates of 1 mL min−1.
Next, we studied the cross-benzoin reaction between benzalde-
hyde 10 and acetaldehyde 12 which yielded (R)-benzoin (11)
and (R)-2-hydroxy-1-phenylpropan-1-one (13)18 in very good yield
under flow-through conditions (Scheme 4).
Scheme 5 Flow-through acetate cleavage with His6-tag BsubpNBE
attached to polymeric phase via tyrosine-linked Ni-NTA.
Scheme 4 Flow-through synthesis of (R)-benzoin (11) and (R)-2-
hydroxy-1-phenylpropan-1-one (13) with His6-tag BAL attached to poly-
meric phase via tyrosine-linked Ni-NTA.
Conclusions
In summary, we have presented a mild and rapid method for
simultaneous purification and immobilization of His6-tagged
proteins allowing facile biocatalysis in a flow-through mode. For
this purpose we optimised the solid phase as well as introduced an
improved linker system for IMAC. This microstructured reactor
system should be broadly applicable for the purification of His6-
tagged proteins as well as their synthetic use.
In a similar manner the reactor was loaded with His6-tagged
BAL and the eluted protein was analysed with the Bradford assay
and SDS Page (see ESI†). When a solution of benzaldehyde (30 lL)
in phosphate buffer (5 mL) was pumped through the PASSflow
◦
reactor (flow rate = 1 mL min−1) at 37 C, full consumption of
benzaldehyde was observed when analysing the solution that left
the reactor (Scheme 4). After extraction with ethyl acetate (R)-
benzoin (11) was isolated in 99% yield. This reaction was repeated
with the same reactor another three times, giving the same results.
Experimental
General remarks
Regioselective ester hydrolysis under continuous flow conditions
IR spectra were recorded with a Bruker Vektor 22 FT-IR
spectrophotometer (GoldenGate ATR unit). Optical rotations
were measured with a Perkin Elmer 341 polarimeter. NMR
Esterase-catalyzed deacetylation was studied using p-nitrobenzyl
esterase from Bacillus subtilis (BsubpNBE19). This highly active
3660 | Org. Biomol. Chem., 2007, 5, 3657–3664
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
spectra were recorded on a Bruker ARX-400 (1H, 400 MHz;
13C, 100 MHz) spectrometer. All spectra were measured using
standard Bruker pulse sequences. Multiplicities are described
using the following abbreviations: s = singlet, d = doublet, t =
triplet, q = quartet, m = multiplet, br = broad. ESI mass
spectra were recorded on a LCT mass spectrometer (Micromass)
with Lock-Spray dual ion source. The LCT-spectrometer was
coupled with a Waters Alliance 2695 HPLC unit. All solvents
used were of reagent grade and were further dried. Reactions
were monitored by thin layer chromatography (TLC) on silica
gel 60 F254 (E. Merck, Darmstadt) and spots were detected either
by UV-absorption or by charring with KMnO4/NaOH in water.
Preparative column chromatography was performed on silica gel
60 (E. Merck, Darmstadt). Reagents were purified and dried by
standard techniques.
80 ◦C overnight. The solidified mass was rinsed with cyclohexane
using a Soxhlet-extractor. At this point the monolithic materials
were separated from the polymer powder, which was formed as a
by-product. Typically, 200 mg of polymer could be incorporated
into one glass rod. The preparation of the casing for the glass rods
has been described elsewhere,6c as well as the morphology of the
polymer phase formed by precipitation polymerisation.
IR (Golden Gate/ATR): 2919, 2324, 1681, 1487, 1445, 1421,
1265, 833, 797, 708 cm−1.
Amine 2 (184 mg, 0.6 mmol) and sodium tert-butylate (73 mg,
0.74 mmol) were dissolved in N-methylpyrrolidinone (15 mL) and
pumped at 60 ◦C in cycles through a PASSflow reactor containing
the resin (approx. 200 mg, 1 mmol chloride) described above. After
12 h the reactor was washed (H2O, MeOH, CH2Cl2, MeOH, H2O,
MeOH and CH2Cl2; 10 mL each) and dried under reduced pressure
to yield resin 3. The loading was determined gravimetrically and
by IR-spectroscopy (45 mg, 0.15 mmol).
Methyl 6-amino-2-(bis(methoxycarbonylmethyl)amino)hexanoate
(2)
IR (Golden Gate/ATR): 2921, 1733, 1661, 1445, 1224, 1154,
1021, 797, 704 cm−1.
Bromoacetic acid (1.4 g, 10.0 mmol) was dissolved in NaOH
For ester saponification, lithium hydroxide (99 mg, 4.1 mmol)
in MeOH–water (1 : 1, 20 mL) was pumped through the reactor at
50 ◦C for 60 h, after which time the reactor was washed with
HCl (1 M), H2O, MeOH, CH2Cl2, MeOH, H2O, HCl (1 M),
H2O, MeOH, CH2Cl2 (10 mL each). Loading of 4 was determined
gravimetrically and by IR spectroscopy (33 mg, 0.14 mmol).
IR (Golden Gate/ATR): 2918, 2359, 1729, 1667, 1607, 1443,
1269, 1153, 1017, 800, 707 cm−1.
(10 mL, 2 M). A solution of x-N-Cbz L-lysine (981 mg, 3.50 mmol)
◦
in NaOH (15 mL, 2 M) was added dropwise at 0 C and stirred
for 2 h. Stirring was continued for 12 h at r.t. The mixture was
adjusted to pH 1.0 with aqueous HCl (10%). The crude product
was filtered, purified by recrystallisation in HCl at pH 1.0 and
dried under vacuum to yield bisalkylated x-N-Cbz lysine as a
white solid (1.21 g, 3.05 mmol, 87%).
To a solution of the alkylated lysine (1.2 g, 3.03 mmol) in
MeOH (10 mL) was added H2SO4 (1 mL), and the mixture was
heated under reflux for 12 h. After cooling to r.t. the reaction
mixture was neutralised with saturated NaHCO3 and extracted
with dichloromethane (3×). The combined organic layers were
dried (MgSO4) and concentrated under reduced pressure. Without
further purification the product was dissolved in MeOH. After
addition of Pd/C (400 mg) the mixture was stirred at r.t. under
an H2 atmosphere. After 12 h the solution was filtered through a
pad of CeliteTM and the filtrate was concentrated under reduced
pressure to yield the title compound 2 (903 mg, 2.97 mmol, 84%).
1H NMR (400 MHz, d6-DMSO, d5-DMSO = 2.50 ppm):
d = 3.71 (d, J = 5.3 Hz, 2 H, NCHHꢀCO2CH3), 3.69 (s, 3 H,
CO2CH3) 3.68 (s, 6 H, NCH2CO2CH3), 3.66 (d, J = 5.3 Hz, 2 H,
NCHHꢀCO2CH3), 3.47 (t, J = 7.5 Hz, 1 H, H-2), 2.95–2.90 (m,
2 H, H-6), 1.65–1.47 (m, 2 H, H-3), 1.44–1.21 (m, 4 H, H-4, H-
5) ppm. 13C NMR (100 MHz, d6-DMSO, d6-DMSO = 39.5 ppm):
d = 174.4, 173.7, 65.4, 53.5, 52.1, 52.1, 52.0, 40.7, 30.5, 28.2,
23.7 ppm. HRMS (ESI): m/z calcd for C13H25N2O6: 305.1115,
found: 305.1110.
Treatment with dichlorodimethylsilane (10 mL, 10% in
dichloromethane, 12 h at 20 ◦C) followed by NiCl2 (10 mL,
saturated, in water–DMSO 10 : 1) and water (10 mL) yielded
Ni-NTA PASSflow reactor 5.
(Methoxycarbonyl)methyl
2-(bis(methoxycarbonylmethyl)-L-
amino)-3-(4-hydroxyphenyl)propionate (7). A solution of O-
benzyl-L-tyrosine (2.0 g, 7.4 mmol) in 3-pentanone (15 mL)
was treated with methyl bromoacetate (8.5 g, 55.5 mmol),
K2CO3 (5.5 g, 55.5 mmol) and tetra-N-butylammonium iodide
(20 mg, 54 lmol), and was stirred for 16 h at 108 ◦C. After
cooling the suspension to r.t., water was added until all the
K2CO3 was dissolved. The aqueous layer was extracted with
ethyl acetate and the combined organic phases were dried
(MgSO4). After concentration under reduced pressure, the
crude product was purified by flash column chromatography
(petroleum ether–EtOAc 5 : 1) to yield (methoxycarbonyl)-
methyl 2-(bis(methoxycarbonylmethyl)-L-amino)-3-(4-(benzyloxy)-
phenyl)propionate (1.99 g, 4.08 mmol, 55%).
1H NMR (400 MHz, CDCl3, CHCl3 = 7.26 ppm): d = 7.42–
7.30 (m, 5 H, OBn), 7.15 (d, J = 8.5 Hz, 2 H, H-5), 6.88
(d, J = 8.5 Hz, 2 H, H-6), 5.01 (s, 2 H, OBn), 4.59 (d, J =
15.8 Hz, 2 H, OCHHꢀCO2CH3), 4.50 (d, J = 15.8 Hz, 2 H,
OCHHꢀCO2CH3), 3.80 (dd, J = 8.7 and 6.3 Hz, 1 H, H-2), 3.73
(s, 4 H, NCH2CO2CH3), 3.70 (s, 3 H, OCH2CO2CH3), 3.67 (s,
6 H, NCH2CO2CH3), 3.05 (dd, J = 13.6 and 8.7 Hz, 1H, H-
3a), 3.00 (dd, J = 13.6 and 6.3 Hz, 1H, H-3b) ppm. 13C NMR
(100 MHz, CDCl3, CDCl3 = 77.0 ppm): d = 171.5, 170.8, 167.9,
137.0, 130.2, 129.4, 128.4, 127.8, 127.3, 114.6, 69.8, 66.4, 60.4,
52.5, 52.0, 51.6, 35.9 ppm. HRMS-ESI: m/z [M + H]+ calcd for
C25H30NO9: 488.1921, found: 488.1925.
Preparation of polar Merrifield Ni-NTA phase (5)
1,4-Divinylbenzene (3.41 g, 26.2 mmol), p-vinyl benzyl chlo-
ride (48.09 g, 315.1 mmol) and N-vinylpyrrolidinone (20.42 g,
183.8 mmol) were mixed and the volume was made up to
630 mL with C14–C17 n-paraffin. After dissolution of azo-
bis(isobutyronitrile) (788 mg, 4.8 mmol), porous glass rods
(5.3 mm diameter; 110 mm length) and Raschig rings (6 mm inner
diameter, 9 mm outer diameter, 9 mm length) were immersed in this
solution. To remove air bubbles from the pore volume, a vacuum
was applied for several minutes. Then, the mixture was heated to
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 3657–3664 | 3661
©

View Article Online
Table 1
Pd/C (10%, 600 mg) was added to a solution of (methoxy-
carbonyl)methyl 2-(bis(methoxycarbonylmethyl-L-amino)-3-(4-
benzyloxyphenyl)propionate (1.99 g, 4.08 mmol) in MeOH, and
the mixture was stirred at r.t. under an H2 atmosphere. After 12 h
the solution was filtered through a short pad of CeliteTM and the
eluant was concentrated under reduced pressure to yield the title
compound 7 (1.61 g, 4.06 mmol, 99%).
Amount of benzaldehyde
Reaction time Isolated yield of (R)-benzoin
15 lL, 16 mg, 0.15 mmol
95 min
15 mg, 0.07 mmol, 93.0%
29 mg, 0.13 mmol, 99.9%
102 mg, 0.48 mmol, 98.0%
205 mg, 0.97 mmol, 98.0%
30 lL, 29 mg, 0.27 mmol 158 min
100 lL, 104 mg, 0.98 mmol 245 min
200 lL, 209 mg, 1.97 mmol 405 min
IR (Golden Gate/ATR): 2954, 1733, 1516, 1437, 1205, 1142,
1012, 829 cm−1. 1H NMR (400 MHz, CDCl3, CHCl3 = 7.26 ppm):
d = 7.07 (d, J = 8.5 Hz, 2 H, H-5), 6.73 (d, J = 8.5 Hz, 2 H,
H-6), 4.60 (d, J = 15.7 Hz, 2 H, OCHHꢀCO2CH3), 4.53 (d, J =
15.7 Hz, 2 H, OCHHꢀCO2CH3), 3.80 (dd, J = 8.6 and 6.6 Hz, 1 H,
H-2), 3.73 (s, 7 H, NCH2CO2CH3, OCH2CO2CH3), 3.69 (s, 6 H,
NCH2CO2CH3), 3.03 (dd, J = 13.6 and 8.6 Hz, 1H, H-3a), 2.98
(dd, J = 13.6 and 6.6 Hz, 1H, H-3b) ppm. 13C NMR (100 MHz,
CDCl3, CDCl3 = 77.0): d = 171.8, 171.1, 168.1, 154.7, 130.4, 128.7,
115.3, 66.6, 60.6, 52.7, 52.3, 51.7, 36.0 ppm. HRMS-ESI: m/z [M +
Na]+ calcd for C18H23NO9Na: 420.1271; found: 420.1271.
reaction buffer) was added. The mixture was shaken (80 rpm) at
30 ◦C. For kinetic analysis (Fig. 4), a sample (200 lL) was taken
every 30 min, extracted with EtOAc (200 lL) and analyzed by
GC. For measuring the overall yield of benzoin production, the
reaction was repeated under identical conditions, with the results
described in Table 1.
Synthesis of (R)-benzoin (11) using BAL on Ni-NTA resin (9)
Centrifuged lysate (250 lL, 50× concentrated) of E. coli (BAL,
EC 4.1.2.38) was diluted with one aliquot lysis buffer [sodium
phosphate (50 mM, pH 8.0), NaCl (300 mM), imidazole (10 mM),
glycerol (10%)], and 100 mg of Ni2+ NTA-resin 9 was added. After
one minute the resin was filtered and washed with lysis buffer (2 ×
1 mL), washing buffer (4 × 1 mL) [sodium phosphate (50 mM,
pH 8.0), NaCl (300 mM), imidazole (20 mM), glycerol (10%)]
and reaction buffer (2 × 1 mL) [sodium phosphate (125 mM,
pH 7.0), MgSO4 (2.5 mM), thiamine diphosphate (0.25 mM),
DMSO (25%)]. Then, benzaldehyde (15 lL to 200 lL in 5 mL
reaction buffer) was added. The mixture was shaken (80 rpm) at
30 ◦C. For kinetic analysis (Fig. 5), a sample (200 lL) was taken
every 30 min, extracted with EtOAc (200 lL) and analyzed by
GC. For measuring the overall yield of benzoin production, the
reaction was repeated under identical conditions, with the results
described in Table 2.
Preparation of a polar Merrifield Ni-NTA phase (9)
Tyrosine-based NTA-linker 7 (227 mg, 0.60 mmol) was dissolved
in anhydrous DMF (15 mL). K2CO3 (166 mg, 1.2 mmol) and CsI
(130 mg, 0.5 mmol) were added and the mixture was pumped at
60 ◦C in a circular mode through a PASSflow reactor containing
the resin described above (approx. 200 mg; 1 mmol chloride). After
72 h the reactor was washed with H2O, MeOH, CH2Cl2, MeOH,
H2O, MeOH and CH2Cl2 (10 mL each), and dried under reduced
pressure to yield resin 8. The loading was verified gravimetrically
and by IR spectroscopy (110 mg, 0.30 mmol, 30% of theoretical
binding sites).
IR (Golden Gate/ATR): 2919, 2324, 1741, 1680, 1512, 1441,
1266, 1215, 1155, 1017, 796, 709 cm−1.
For ester saponification, lithium hydroxide (99 mg, 4.1 mmol)
in MeOH–water (1 : 1, 20 mL) was pumped through the reactor at
50 ◦C for 60 h, after which time the reactor was washed with
HCl (1 N), H2O, MeOH, CH2Cl2, MeOH, H2O, HCl (1 N),
H2O, MeOH, dichloromethane (10 mL each). The loading of 9
was determined gravimetrically and by IR-spectroscopy (110 mg,
0.29 mmol, 97%).
IR (Golden Gate/ATR): 2918, 2324, 1731, 1679, 1608, 1511,
1442, 1220, 1176, 1016, 797, 707 cm−1.
Treatment with dichlorodimethylsilane (10 mL, 10% in
dichloromethane, 12 h at 20 ◦C in cycles) followed by NiCl2 (10 mL,
saturated in water–DMSO 10 :1) and water (10 mL) yielded Ni-
NTA PASSflow reactor 9.
Flow-through synthesis of (R)-benzoin (11)
BAL (EC 4.1.2.38)11 containing cell lysate [750 lL, 25× concen-
trated in 50 mM potassium phosphate pH 8.0, E. coli SG13009
transformed with pKK233-2/BALHis6 and induced with IPTG
(0.1 mmol); cell lysis by ultrasonification on ice] was centrifuged
and diluted with the same volume of lysis buffer (50 mM sodium
phosphate pH 8.0, 300 mM NaCl, 10 mM imidazole, 10%
glycerol) and pumped through the Ni2+-NTA PASSflow reactor
9 at 0.5 mL min−1. The column was washed with 3 mL lysis
buffer and with 8 mL washing buffer (50 mM sodium phosphate
pH 8.0, 300 mM NaCl, 20 mM imidazole, 10% glycerol). The
reactor was equilibrated with 5 mL reaction buffer [125 mM
sodium phosphate pH 7.0, containing DMSO (25%), MgSO4
(2.5 mM) and thiamine diphosphate (0.25 mM)]. Benzaldehyde
(30 lL, 0.3 mmol) in reaction buffer (5 mL, see above) was
Synthesis of (R)-benzoin (11) using BAL on Ni-NTA resin (5)
◦
pumped through the reactor (37 C, 1 mL min−1) followed by
Centrifuged lysate (250 lL, 50× concentrated) of E. coli (BAL,
EC 4.1.2.38) was diluted with one aliquot lysis buffer [sodium
phosphate (50 mM, pH 8.0), NaCl (300 mM), imidazole (10 mM),
glycerol (10%)], and 100 mg of Ni2+ NTA-resin 5 was added. After
one minute the resin was filtered and washed with lysis buffer (2 ×
1 mL), washing buffer (4 × 1 mL) [sodium phosphate (50 mM,
pH 8.0), NaCl (300 mM), imidazole (20 mM), glycerol (10%)]
and reaction buffer (2 × 1 mL) [sodium phosphate (125 mM,
pH 7.0), MgSO4 (2.5 mM), thiamine diphosphate (0.25 mM),
DMSO (25%)]. Then, benzaldehyde (15 lL to 200 lL in 5 mL
Table 2
Amount of benzaldehyde
Reaction time Isolated yield of (R)-benzoin
15 lL, 16 mg, 0.15 mmol
30 lL, 29 mg, 0.27 mmol
100 lL, 104 mg, 0.98 mmol 90 min
200 lL, 209 mg, 1.97 mmol 150 min
500 lL, 522 mg, 4.97 mmol 540 min
85 min
90 min
14 mg, 0.07 mmol, 93.0%
29 mg, 0.13 mmol, 99.9%
98 mg, 0.46 mmol, 94.0%
200 mg, 0.94 mmol, 96.0%
386 mg, 1.82 mmol, 74.0%
3662 | Org. Biomol. Chem., 2007, 5, 3657–3664
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
reaction buffer (5 mL). No benzaldehyde could be detected in
the eluate. The column was washed with EtOAc (20 mL) and the
aqueous buffer was extracted with EtOAc. The combined extracts
were dried (MgSO4). After concentration under reduced pressure,
the crude product was purified by flash column chromatography
(petroleum ether–EtOAc 5 : 1) to yield the title compound 11
(31.5 mg, 0.15 mmol, 99%). The NMR data were found to be
identical with those reported in the literature.15
1 H, H-6b), 4.16 (br m, 1H, H-3), 3.93 (ddd, J = 10.2, 5.8 and
2.4 Hz, 1 H, H-5), 3.62 (br m, 1 H, H-4), 2.94 (s, 1 H, -OH),
2.08 (s, 3 H, OAc) ppm. 13C NMR (200 MHz, d6-acetone, d6-
acetone = 29.8 ppm): d = 171.0, 143.7, 105.2, 77.2, 70.6, 69.9, 63.8,
20.6 ppm.
Acknowledgements
[a]2D0 −141.2 (c 0.99 in acetone) 98% ee; Rf: 0.35 (CH2Cl2).
HRMS-ESI: m/z [M + Na]+ calcd for C14H11O2Na: 235.0784,
found: 235.0779.
This work was supported by the Deutsche Forschungsgemein-
schaft (DFG; grants Ki 397/8-1 and excellence cluster REBIRTH)
and the Fonds der Chemischen Industrie. We thank Prof. Martina
Pohl (University of Du¨sseldorf) for the expression vector pKK233-
2/BALHis6 and Prof. Uwe T. Bornscheuer (University of Greif-
swald) for a sample of BsubpNBE.
Flow-through synthesis of (R)-2-hydroxy-1-phenylpropan-1-one
(13)
BAL (EC 4.1.2.38)11 containing cell lysate [750 lL, 25× concen-
trated in 50 mM potassium phosphate pH 8.0, E. coli SG13009
transformed with pKK233-2/BALHis6 and induced with IPTG
(0.1 mmol); cell lysis by ultrasonification on ice] was centrifuged
and diluted with the same volume of lysis buffer (50 mM sodium
phosphate pH 8.0, 300 mM NaCl, 10 mM imidazole, 10%
glycerol) and pumped through the Ni2+-NTA PASSflow reactor
9 at 0.5 mL min−1. The column was washed with 3 mL lysis buffer
and with 8 mL washing buffer (50 mM sodium phosphate pH 8.0,
300 mM NaCl, 20 mM imidazole, 10% glycerol). The reactor
was equilibrated with 5 mL reaction buffer [125 mM sodium
phosphate pH 7.0, containing DMSO (25%), MgSO4 (2.5 mM)
and thiamine diphosphate (0.25 mM)]. A mixture of benzaldehyde
(30 lL, 0.3 mmol) and acetaldehyde (85 lL, 1.5 mmol) in reaction
buffer (6 mL, see above) was pumped through the reactor in
a circular mode. After 3 h at 37 ◦C no benzaldehyde could be
detected in the eluate by GC, and the reaction was stopped. The
column was washed with ethyl acetate (20 mL) and the aqueous
buffer was extracted with ethyl acetate. The combined extracts
were dried (MgSO4). After concentration under reduced pressure,
the crude product was purified by flash column chromatography
(petroleum ether–EtOAc 5 : 1) to yield the title compound 13
(42 mg, 0.28 mmol, 92%). The NMR data were found to be
identical with those reported in the literature.15
References and notes
1 N. End and K.-U. Scho¨ning, Top. Curr. Chem., 2004, 242, 273–317.
2 U. T. Bornscheuer, Angew. Chem., 2003, 115, 3458–3459; U. T.
Bornscheuer, Angew. Chem., Int. Ed., 2003, 42, 3336–3337; A. M.
Klibanov, Anal. Biochem., 1979, 79, 1–25; H. R. Luckarift, J. C. Spain,
R. R. Naik and M. O. Stone, Nat. Biotechnol., 2004, 22, 211–213; P. T.
Vasudevan, N. Lo´pez-Corte´s, H. Caswell, D. Reyes-Duarte, F. J. Plou,
A. Ballesteros, K. Como and T. Thomson, Biotechnol. Lett., 2004,
26, 473–477; D. S. Wentorth, D. Skonberg, D. W. Donahue and A.
Ghanem, J. Appl. Polym. Sci., 2004, 91, 1294–1299; A. Basso, L. De
Martin, L. Gardossi, G. Margetts, I. Brazendale, A. Y. Bosma, R. V.
Ulijn and S. L. Flitsch, Chem. Commun., 2003, 1296–1297; R. Torres,
C. Mateo, G. Fernandez-Lorente, C. Ortiz, M. Fuentes, J. M. Palomo,
J. M. Guisan and R. Fernandez-Lafuente, Biotechnol. Prog., 2003, 19,
1056–1060; A. Heilmann, N. Teuscher, A. Kiesow, D. Janasek and U.
Spohn, J. Nanosci. Nanotechnol., 2003, 3, 375–379; X. Gao, K. M. K.
Yu, K. Y. Tam and S. C. Tsang, Chem. Commun., 2003, 2998–2999; C.
Lei, Y. Shin, J. Liu and E. J. Ackerman, J. Am. Chem. Soc., 2002, 124,
11242–11243; H. A. Sousa, C. Rodrigues, E. Klein, C. A. M. Afonso
and J. G. Crespo, Enzyme Microb. Technol., 2001, 29, 625–634.
3 T. Honda, M. Miyazaki, H. Nakamura and H. Maeda, Adv. Synth.
Catal., 2006, 348, 2163–2171.
4 W. Solodenko, K. Mennecke and A. Kirschning, Chem.–Eur. J., 2006,
12, 5972–5990.
5 G. Jas and A. Kirschning, Chem.–Eur. J., 2003, 9, 5708–5723.
6 A. Kirschning, C. Altwicker, G. Dra¨ger, J. Harders, N. Hoffmann,
U. Hoffmann, H. Scho¨nfeld, W. Solodenko and U. Kunz, Angew.
Chem., 2001, 113, 4118–4120; A. Kirschning, C. Altwicker, G. Dra¨ger,
J. Harders, N. Hoffmann, U. Hoffmann, H. Scho¨nfeld, W. Solodenko
and U. Kunz, Angew. Chem., Int. Ed., 2001, 40, 3995–3998; U. Kunz,
A. Kirschning, C. Altwicker and W. Solodenko, J. Chromatogr., A,
2003, 1006, 241–249; U. Kunz, H. Scho¨nfeld, W. Solodenko, G. Jas
and A. Kirschning, Ind. Eng. Chem. Res., 2005, 44, 8458–8467; U.
Kunz, A. Kirschning, H.-L. Wen, W. Solodenko, R. Cecillia, C. O.
Kappe and T. Turek, Catal. Today, 2005, 105, 318–324; A. Kirschning,
K. Mennecke, U. Kunz, A. Michrowska and K. Grela, J. Am. Chem.
Soc., 2006, 128, 13261–13267; W. Solodenko, G. Jas, U. Kunz and A.
Kirschning, Synthesis, 2007, 583–589.
Flow-through synthesis of 6-O-acetyl-D-glucal (17)
BsubpNBE (5 mg) in loading buffer (1 mL, 50 mM sodium
phosphate pH 8.0, 300 mM NaCl, 10% glycerol) was pumped
though the Ni2+-NTA PASSflow reactor 9 at 0.5 mL min−1. The
column was washed with 3 mL loading buffer and 7 mL reaction
buffer (50 mM sodium phosphate pH 7.0). A solution of tri-O-
acetyl-D-glucal (16, 100 mg, 0.38 mmol) in reaction buffer (45 mL)
and MTBE (5 mL) was pumped through the reactor in a circular
mode. After 60 h at 37 ◦C no starting material could be detected
by TLC, and the column was washed with MTBE (20 mL). The
combined aqueous and organic extracts were dried and purified by
flash column chromatography (EtOAc) to yield the title compound
17 (57 mg, 0.30 mmol, 80%) and D-glucal 18 (7 mg 0.05 mmol,
12%).
7 C. Auge´, A. Malleron, H. Tahrat, A. Marc, J.-J. Goergen, M. Cerutti,
W.-F. Steelant, P. Delannoy and A. Lubineau, Chem. Commun., 2000,
2017–2018; J. Nahalka, Z. Liu, X. Chen and P. G. Wang, Chem.–Eur. J.,
2003, 2, 373–377.
8 B. C. Roy and S. Mallik, J. Org. Chem., 1999, 64, 2969–2974; S. Gritsch,
K. Neumaler, L. Schmitt and R. Tampe, Biosens. Bioelectron., 1995,
10, 805–812.
9 S. Mallik and I. Mallik, Synlett, 1996, 734–736.
10 For both linker systems we prepared five different polymeric samples
(DVB/VBC/vinylpyrrolidinone vol %): sample 1 = 10 : 50 : 40;
sample 2 = 5 : 95 : 0; sample 3 = 10/70/20; sample 4 = 20 : 60 :
20 and sample 5 = 5 : 60 : 35. Sample 3 gave the highest loading of
tyrosine linker (30%) and protein immobilisation.
1H NMR (400 MHz, d6-acetone, d5-acetone = 2.05 ppm): d =
6.31 (dd, J = 6.0 and 1.5 Hz, 1 H, H-1), 4.72 (dd, J = 6.0 and
2.2 Hz, 1 H, H-2), 4.58 (br d, J = 4.4 Hz, 1 H, OH), 4.43 (dd,
J = 12.0 and 2.4 Hz, 1 H, H-6a), 4.32 (dd, J = 12.0 and 5.8 Hz,
11 M. Sanchez-Gonzalez and J. P. M. Rosazza, Adv. Synth. Catal.,
¨
2003, 345, 819–824; A. S. Demir, O. Sesenoglu, P. Du¨nkelmann and
M. Mu¨ller, Org. Lett., 2003, 5, 2047–2050; A. S. Demir, O. Sesenoglu,
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 3657–3664 | 3663
©

View Article Online
E. Eren, B. Horsrik, M. Pohl, E. Janzen, D. Kolter, R. Feldmann,
P. Du¨nkelmann and M. Mu¨ller, Adv. Synth. Catal., 2002, 334, 96–103;
T. Du¨nnwald, A. S. Demir, P. Siegert, M. Pohl and M. Mu¨ller,
Eur. J. Org. Chem., 2000, 2161–2170.
glass only had to be done once, even if several consecutive reactions
were carried out in the PASSflow reactor.
18 P. Du¨nkelmann, D. Kolter-Jung, A. Nitsche, A. S. Demir, P. Siegert,
B. Lingen, M. Baumann, M. Pohl and M. Mu¨ller, J. Am. Chem. Soc.,
2002, 124, 12084–12085.
19 J. Zock, C. Cantwell, J. Swartling, R. Hodges, T. Pohl, K. Sutton, P.
Rosteck, Jr., D. McGilvray and S. Queener, Gene, 1994, 151, 37–43; E.
Henke, J. Pleiss and U. T. Bornscheuer, Angew. Chem., Int. Ed., 2002,
41, 3211–3213.
20 I. Fotakopoulou, E. Barbayianni, V. Constantinou-Kokotou, U. T.
Bornscheuer and G. Kokotos, J. Org. Chem., 2007, 72, 782–786; E.
Barbayianni, I. Fotakopoulou, M. Schmidt, V. Constantinou-Kokotou,
U. T. Bornscheuer and G. Kokotos, J. Org. Chem., 2005, 70, 8730–
8733; M. Schmidt, E. Barbayianni, I. Fotakopoulou, M. Ho¨hne, V.
Constantinou-Kokotou, U. T. Bornscheuer and G. Kokotos, J. Org.
Chem., 2005, 70, 3737–3740.
12 M. Pohl, G. A. Sprenger and M. Mu¨ller, Curr. Opin. Biotechnol., 2004,
15, 335–342.
13 M. A. Bradford, Anal. Biochem., 1976, 72, 248–254.
14 U. K. Laemmli, Nature, 1970, 227, 680–685.
15 With a DMSO content of 0% no transformation was observed, while a
DMSO content of 15% and higher led to reduced reaction rates at pH 7.
For additional details on BAL refer to: P. D. De Maria, T. Stillger, M.
Pohl, S. Wallert, K.-H. Drauz, H. Gro¨ger, H. Trautwein and A. Liese,
J. Mol. Catal. B: Enzym., 2006, 38, 43–47.
16 Elution of the protein after termination of the reaction expectedly gave
a positive Bradford assay.
17 Silylation and prevention of non-specific binding of proteins to the
3664 | Org. Biomol. Chem., 2007, 5, 3657–3664
This journal is
The Royal Society of Chemistry 2007
©