A three-enzyme cascade reaction through positional assembly of enzymes in a polymersome nanoreactor

Article abstract of DOI:10.1002/chem.200802114

Porous polymersomes based on block copolymers of isocyanopeptides and styrene have been used to anchor enzymes at three different locations, namely, in their lumen (glucose oxidase, GOx), in their bilayer membrane (Candida antarctica lipase B, CalB) and o

Full text of DOI:10.1002/chem.200802114

FULL PAPER
DOI: 10.1002/chem.200802114
A Three-Enzyme Cascade Reaction through Positional Assembly of Enzymes
in a Polymersome Nanoreactor
Stijn F. M. van Dongen,[a] Madhavan Nallani,[b] Jeroen J. L. M. Cornelissen,[a]
Roeland J. M. Nolte,[a] and Jan C. M. van Hest*[a]
Abstract: Porous polymersomes based
on block copolymers of isocyanopepti-
des and styrene have been used to
anchor enzymes at three different loca-
tions, namely, in their lumen (glucose
oxidase, GOx), in their bilayer mem-
brane (Candida antarctica lipase B,
CalB) and on their surface (horseradish
peroxidase, HRP). The surface cou-
pling was achieved by click chemistry
between acetylene-functionalised an-
chors on the surface of the polymer-
somes and azido functions of HRP,
residues of the enzyme. To determine
the encapsulation and conjugation effi-
ciency of the enzymes, they were deco-
rated with metal-ion labels and ana-
lysed by mass spectrometry. This re-
vealed an almost quantitative immobi-
lisation efficiency of HRP on the sur-
face of the polymersomes and a more
than statistical incorporation efficiency
for CalB in the membrane and for
GOx in the aqueous compartment. The
enzyme-decorated polymersomes were
studied as nanoreactors in which glu-
cose acetate was converted by CalB to
glucose, which was oxidised by GOx to
gluconolactone in a second step. The
hydrogen peroxide produced was used
by HRP to oxidise 2,2’-azinobis(3-eth-
ylbenzothiazoline-6-sulfonic
acid)
(ABTS) to ABTS +. Kinetic analysis
revealed that the reaction step cata-
lysed by HRP is the fastest in the cas-
cade reaction.
C
Keywords:
cascade reactions
·
enzyme immobilization
·
macro-
molecular chemistry
· nanostruc-
which were introduced by using
a
tures · polymers
direct diazo transfer reaction to lysine
Introduction
bination thereof can be found, for instance, in mitochondria
for the enzymes involved in the citric acid cycle.
Compartmentalisation is one of the techniques that cells
adopt to enable a high level of control over (bio-)chemical
processes, for instance, the order in which enzymes react. In
many cases, the compartment also serves to protect the cell
from the action of its degrading contents, as is the case with
lysosomes. It furthermore serves as a scaffold for the precise
positional assembly of enzymes that work together in a mul-
tistep cascade reaction. Positioning enzymes on the surface
of a compartment, in its interior, its membrane or any com-
In an effort to mimic these complex enzyme systems,
many studies concerning enzyme encapsulation or assembly
have been reported in the literature.[1–5] The focus of this re-
search initially was on phospholipid liposomes,[6,7] the bilay-
er membranes of which are rather similar to those of natu-
rally occurring cells. However, the relative fragility of lipo-
somes limits their potential applicability. Increased mechani-
cal and thermodynamic stability has been achieved by pre-
paring vesicles with layer-by-layer deposition methods.[8,9]
Compartmentalisation based on sol–gel chemistry has also
been reported,[1,3] although this approach parts with the
notion of discrete vesicular objects.
[a] S. F. M. van Dongen, Dr. J. J. L. M. Cornelissen,
Prof. Dr. R. J. M. Nolte, Prof. Dr. J. C. M. van Hest
Department of Organic Chemistry
Like liposomes, polymersomes are spherical aggregates
that contain a bilayer architecture. They are formed by the
self-assembly of amphiphilic block copolymers in an aque-
ous environment.[10] Their polymeric bilayer shows a greater
stability, mainly due to the lower critical aggregation con-
centration of amphiphilic macromolecules.[10] The large
chemical versatility possible in block copolymer synthesis
Institute for Molecules and Materials
Radboud University Nijmegen, Heyendaalseweg 135
6525 AJ Nijmegen (The Netherlands)
Fax : (+31)24-36-53393
E-mail: J.vanHest@science.ru.nl
[b] Dr. M. Nallani
allows one to tune the properties of polymersomes.[11–13]
A
Institute of Materials Research & Engineering (IMRE)
Research Link 3, Singapore 117602 (Singapore)
drawback of polymersome membranes is their low permea-
Chem. Eur. J. 2009, 15, 1107 – 1114
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1107

bility, even to water, which hampers their application as
nanoreactors.[10] To overcome this problem, the diffusion of
solvent and substrate molecules has been enabled by the in-
clusion of channel proteins[14,15] or proton pumps[16] in the
polymeric bilayer.
A more convenient method to prepare permeable poly-
mersomes is the use of a block copolymer that gives an in-
trinsically porous bilayer when self assembled. One such co-
polymer is polystyrene40-b-poly(l-isocyanoalanine(2-thio-
phen-3-yl-ethyl)amide)50 (PS-PIAT). It is a rod–coil type of
amphiphilic copolymer consisting of a rigid polyisocyanide
block and a flexible polystyrene block.[17–19] On dispersal in
water it forms porous polymersomes that possess a relatively
high degree of diffusion. Small molecules can move across
their membranes, whereas larger molecules, such as proteins,
cannot.[20,17]
Scheme 1. Structures of block copolymer “anchor” 1, diazo transfer re-
agent 2, Ru-labelling compound 3 and glucose acetate 4.
Previously we described the use of PS-PIAT polymer-
somes as scaffolds for the positional assembly of two differ-
ent enzymes.[21] Glucose oxidase (GOx) was encapsulated in
the lumen of the polymersome, whereas horseradish perox-
idase (HRP) was embedded in its bilayer. The activity of
these two enzymes was studied as part of a three-enzyme
cascade system, which also involved the enzyme Candida
antarctica lipase B (CalB). CalB was added separately to the
polymersome dispersion. In this cascade system, 1,2,3,4-
tetra-O-acetyl-b-glucopyranose (GAc4) was deprotected by
CalB to produce free glucose, which was subsequently oxi-
dised by GOx. This in turn produced hydrogen peroxide,
which was the substrate for HRP, which was used to convert
brane porosity, that is, the ability of the polymersome to let
small molecules diffuse through its membrane. We have
now used this anchoring approach to develop an enzyme
cascade system in which all enzymes involved are associated
with one single polymersome through a controlled spatial
positioning procedure, as shown in Figure 1. This system,
therefore, extends our ability to create complex cascade bio-
mimetic systems. A practical advantage of this approach is
that all catalytic enzyme species can be removed from solu-
tion by a single filtration step. The system also features a
more controlled spatial positioning of the enzymes. Herein
the construction of the three-enzyme polymersome reactor
is described (Figure 2), together with the quantification of
the efficiency of enzyme incorporation into the polymer-
some membrane, the lumen and the efficiency of surface
conjugation.
2,2’-azinobis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS)
+
C
into ABTS
.
In the system described above, only two out of three en-
zymes were linked to the polymersome. While a variety of
tools is available for the preparation of polymer/enzyme bi-
ohybrids,[22] we have recently developed methodologies to
specifically anchor an enzyme to the surface of a polymer-
some,[23,24] one of which is based on the 1,3-dipolar cycload-
dition between an azide and an alkyne. To this end, a block
copolymer “anchor” with an acetylene-functionalised hydro-
philic terminus (1, Scheme 1) was admixed with PS-PIAT to
introduce surface functionalities while maintaining mem-
Finally, to be able to prepare this three-step reactor, azido
groups needed to be introduced into HRP. To realise this,
we applied a diazo transfer reaction directly on the amines
of the lysine residues and the amino terminus of HRP.[25]
This facile diazo transfer reaction uses imidazole-1-sulfonyl
azide hydrochloride (2, Scheme 1).[26]
Figure 1. Positional assembly of enzymes in a polymersome. A mixture of PS-PIAT and anchor 1 is lyophilised with CalB and then dissolved in THF.
This mixture is then injected into an aqueous buffer containing GOx, encapsulating it in the inner compartment and subsequently trapping CalB in the
polymeric bilayer. A third enzyme, HRP, is immobilised on the polymersomal perimeter though a covalent linkage to anchor 1, creating an outer shell of
enzymes.
1108
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1107 – 1114

A Three-Enzyme Cascade Reaction
FULL PAPER
presence of anchor 1. To this
end, PS-PIAT with 10 wt%
anchor 1 was dissolved in THF
and injected into an aqueous
solution of CalB (2 mgmLꢁ1),
after which the dispersion was
immediately flash frozen in
liquid nitrogen and lyophilised.
The resulting CalB–polymer
hybrid[21] was then dissolved in
THF and gently dripped into a
phosphate buffer containing
GOx (0.25 mgmLꢁ1). After half
an hour of equilibration, non-
encapsulated enzymes and THF
were removed by filtration, and
the resulting mixture was ana-
lysed by electron microscopy
(Figure 3), revealing the pres-
ence of polymersomes with di-
ameters ranging from 90 to
180 nm. This shows that the two
enzymes and anchor 1 do not
prevent the PS-PIAT polymersomes from being formed.
Figure 2. Schematic representation of the multistep reaction. 1) Monoacetylated Glucose (4) is deprotected by
CalB, which is embedded in the polymersome membrane. 2) In the inner aqueous compartment, GOx oxidises
glucose to gluconolactone, providing a molecule of hydrogen peroxide. 3) Hydrogen peroxide is used by HRP
to convert ABTS to ABTS +. HRP is tethered to the polymersome surface.
C
Results and Discussion
Design of the three-enzyme cascade system: The objective
of this study was to position enzymes at three specific loca-
tions within or on a polymersome: GOx in its lumen, CalB
in its membrane and HRP on its surface (see Figure 2).
Since GOx is a rather large tetrameric enzyme (ꢀ160 kDa),
it would disrupt PS-PIAT membranes if it were embedded
in them;[21] hence it was decided to include this enzyme
inside the more spacious lumen. HRP and CalB have lower
molecular weights (ꢀ43 kDa and ꢀ33 kDa, respectively)
and both enzymes have previously been successfully incor-
porated in a PS-PIAT membrane.[20,21] Our choice to embed
CalB in the bilayer of the vesicles was primarily motivated
by the fact that it is more hydrophobic than HRP. To enable
the effective use of anchor 1, any enzyme that was to be
positioned on the surface of the polymersome needed to be
functionalised with azido moieties. To realise this, we decid-
ed to apply a diazo transfer reaction directly on the amines
of the lysine residues and the amino terminus of the
enzyme.[25,26] Given our intent to modify the primary amines,
the lower density of amines present in HRP, as opposed to
CalB, would reduce the risk of polymersome aggregation
due to cross-linking through bridging enzymes. Furthermore,
the risk of decreasing enzymatic activity is lowered when a
modified enzyme more closely resembles its native state.
These considerations led us to position GOx in the lumen,
CalB in the bilayer and HRP on the outer surface, necessi-
tating the synthesis of azido-HRP.
Preparation of azido-HRP: To expand this two-enzyme
system into a three-enzyme one, HRP needed to be
Figure 3. Electron micrographs of the biohybrid polymersomes. All mi-
crographs shown are of PS-PIAT vesicles containing an admixed 10 wt%
of 1. The scale bar represents 500 nm. Their enzymatic content is as fol-
lows: A) GOx in the lumen (TEM); B) GOx in the lumen, HRP on the
surface (SEM); C) GOx in the lumen, CalB in the membrane, HRP on
the surface (TEM); D) GOx in the lumen, CalB in the membrane, HRP
on the surface (SEM); E) 3-labelled GOx in the lumen (TEM); F) 3-la-
belled CalB in the membrane.
Polymersomes loaded with two enzymes: As a first step to-
wards the three-enzyme cascade system, polymersomes
loaded with CalB in their membranes and GOx in their
aqueous compartments were prepared. It was also investi-
gated whether this system would be compatible with the
Chem. Eur. J. 2009, 15, 1107 – 1114
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1109

J. C. M. van Hest et al.
equipped with azide moieties. Although azido-modified
amino acids can be introduced by single- or multisite protein
engineering replacement strategies,[27] the level of control
provided by these laborious methods is not needed for
merely immobilising enzyme molecules. This led us to apply
the mild and generally facile diazo transfer reaction to
amines, carried out with 2 (Scheme 1) as the reagent and
catalysed by CuII,[26] which has been shown to be suitable for
protein modification in aqueous solution.[25] To this end, an
aqueous solution of HRP, K2CO3 and CuIISO4 was treated
with water-soluble 2 for twelve hours, after which time the
enzyme was purified by using centrifugal filter devices. ESI-
TOF analysis of the treated enzyme showed that an average
of four amines were transformed into azides. This is a satis-
factory result, given the fact that in principle only one azide
per enzyme is required for surface immobilisation.
tegrity of the polymersomes remained unperturbed, as is
shown in Figure 3.
Polymersomes loaded with three enzymes: In a next step,
azido-HRP was conjugated to the acetylene anchors that
were part of the above-mentioned two-enzyme polymer-
somes with GOx in their aqueous compartment and CalB in
their bilayer membrane. After conjugation, the vesicles were
washed by using a cutoff filter that allowed passage of un-
connected enzymes while retaining intact polymersomes in
the supernatant. The transmission electron micrographs
shown in Figure 3 revealed the unaltered spherical morphol-
ogy of the polymersomes.
Little has previously been reported on the actual efficien-
cy of these incorporation processes. Rameez et al. deter-
mined the efficiency of haemoglobin (Hb) encapsulation in
the polymersome lumen by lysis of the vesicles, followed by
UV/Vis detection of released Hb with an Hb-specific proto-
col.[28] For a more general approach, we chose to employ a
commercially available ruthenium complex with an isothio-
cyanato moiety (3), which is easily covalently linked to any
molecules containing free amines, as most proteins do in
their lysine residues or N termini. After encapsulation or
immobilisation, the use of inductively coupled plasma–mass
spectrometry (ICP–MS), allowed the quantitative detection
of ruthenium, and hence, of proteins.
Subsequently, PS-PIAT polymersomes containing GOx in
their inner aqueous compartments and 10% acetylene
anchor 1 in their membranes were prepared. After treat-
ment of these aggregates with azido-HRP and CuI, they
were washed through filtration techniques until no HRP ac-
tivity could be detected in the washings. The resulting poly-
mersomes, containing GOx in their lumens and HRP on
their surfaces, were dispersed in phosphate buffer and stud-
ied as a two-enzyme cascade system, in which glucose is
converted to gluconolactone and the resulting hydrogen per-
oxide used to oxidise ABTS to the coloured radical cation
Each of the three different enzymes was labelled and sep-
arately incorporated into the polymersomes or conjugated
to them in the same way as shown above for the unlabelled
enzymes (polymersomes containing the labelled enzymes
are shown in Figure 3). Results for CalB and GOx incorpo-
ration, as well as HRP-conjugation efficiency, are given in
Table 1. From these results, it appears that the efficiency of
ABTS +. As can be seen in Figure 4, the increasing absorb-
C
+
C
ance of ABTS clearly indicates that the two-enzyme cas-
cade system is working, which demonstrates both the suc-
cessful conjugation of HRP to the functionalised polymer-
some nanoreactor and the viability of more complex
enzyme positioning methods such as this. The structural in-
Table 1. Encapsulation and conjugation efficiencies of enzymes in PS-
PIAT polymersomes.
Labelled
enzyme
Location
Quantity
[nmol]
Quantity in bio-
AHCTUNGTREGUNhNN ybrid [nmol]
Efficiency
[%]
CalB
polymeric
bilayer
lumen
5.61
0.80–1.13
17.2ꢂ2.96
GOx
HRP
3.91
12.50
0.90–1.06
2.77–3.61
25.0ꢂ2.09
surface
25.5ꢂ3.37[a]
[a] Measured: 25.5ꢂ3.37%; the theoretical maximum for HRP is dictat-
ed by the amount of acetylenes introduced via anchor 1.
conjugating azido-HRP to the polymersome surface is very
high. By applying 10 wt% of 1, 7 nmol of the alkyne moiet-
ies became incorporated in the periphery of the polymer-
some. Assuming an equal distribution of the anchor mole-
cules over both membrane layers, it can be calculated that
around 3.5 nmol of acetylene functions are present on the
polymersome outer surface. The average value, as deter-
mined by ICP–MS, of 3.2 nmol HRP after conjugation im-
plies that more than 90% of all available acetylene sites
were occupied by HRP.
Figure 4. Progress curve for GOx polymersomes with surface-conjugated
HRP. Two types of polymersomes are incubated with glucose and ABTS,
+
C
and the appearance of ABTS is measured as a function of time. After a
short initial incubation period, polymersomes that only contain GOx do
not show any increase in absorbance (grey dots). Polymersomes in which
both enzymes are present do show an increasing concentration of
+
C
ABTS (black dots).
1110
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1107 – 1114

A Three-Enzyme Cascade Reaction
FULL PAPER
For GOx and CalB, the incorporation efficiency was much
higher than 0.1%, which would be expected based on statis-
tical inclusion.[21] This implies a mechanism of polymersome
formation that is somehow influenced by the presence of
the enzymes, perhaps through nucleation of block-copoly-
mer aggregates around transient protein aggregates.
it to convert any H2O2 molecule relatively quickly after its
generation, eliminating HRP from the rate equation. Its
method of immobilisation also renders it the enzyme that is
least hindered by its crowded environment. Furthermore, it
is also the enzyme that has not been suffering from the ef-
fects of THF because it never came into contact with this
solvent. For comparison, a reaction with a polymersome
system lacking CalB in the membrane was performed
(Figure 5). These polymersomes do not have the ability to
catalytically deprotect 4; the progress of the reaction is de-
pendent on the hydrolysis of 4 by water. As can be seen in
Figure 5, this system displays a significantly lower rate of re-
action. The similarity between the curves for the two-
enzyme and the three-enzyme system shows that CalB is not
the fastest of the three enzymes, again pointing towards
HRP as taking this role. Unfortunately, our analysis using
Equation (1) does not allow us to determine the slowest
step of the three-enzyme cascade reaction. Measurement of
a solution that was filtered to remove the polymersomes
showed no conversion of ABTS, indicating that all activity
that was measured resulted from enzymes associated with
the polymersomes.
Three-enzyme cascade catalysis: As a substrate for the
three-enzyme cascade reaction of CalB, GOx and HRP, an
acetate-protected glucose was chosen. Previously, the tetra-
acetate GAc4 had been used for this purpose.[21] For the
present study, however, an increased solubility of the sub-
strate in aqueous buffers was desired to achieve a more ho-
mogeneous reaction mixture. The orthogonally protected
2,3,4,6-tetra-O-benzyl-d-glucopyranose was acetylated at its
anomeric position and subsequently hydrogenated to pro-
duce 4 (Scheme 1). Besides its greater solubility, substrate 4
is also more rapidly deprotected by CalB, which enhances
the overall reaction rate of the cascade system.
The multistep nanoreactors were incubated with 4 and
ABTS as described in the Experimental Section and the in-
+
C
crease in ABTS concentration as a function of time was
measured by UV/Vis spectroscopy. The results are presented
in Figure 5. The curve is S shaped and the first two thirds of
The relatively high concentration of enzyme molecules
measured by ICP–MS does not translate to a higher total ac-
tivity. A possible explanation is that diffusion of substrate or
product through the membrane is rate limiting, or that the
confinement of the enzyme molecules in a small space is
detrimental to their activity. The temporary exposure to
THF that some enzymes undergo may also adversely influ-
ence their activity. Further studies are required to obtain
more information regarding this issue.
Conclusion
We have constructed biohybrid polymersome nanoreactors
in which three different enzymes are spatially positioned,
and precisely ordered. These enzymes were incorporated in
the membrane (CalB), encapsulated in the inner aqueous
compartment (GOx) and attached to the surface of the pol-
ymersome (HRP). The conjugation of HRP to the polymer-
somes was realised through a CuI-catalysed [3+2] Huisgen
cycloaddition. To this end, HRP was provided with azide
functions following a diazo transfer reaction on lysine resi-
dues or the N terminus, carried out in aqueous buffer. The
vesicular morphology of the resulting three-step nanoreactor
is unaffected by this decoration. More than 90% of the
available handles on the polymersome surface are occupied
by an azido-HRP molecule, and the lumen incorporates ap-
proximately 25% of the added GOx enzymes, which is more
than expected for a purely statistical encapsulation process.
The CalB enzymes are incorporated in the bilayer mem-
brane with an efficiency of 17%. The nanoreactor is capable
of performing a three-step cascade reaction and can be re-
moved from the solution by a single filtration step. The
progress curve of the reaction fits to a two-enzyme reaction
model, suggesting that one of the enzymes, that is, HRP,
Figure 5. Progress curve for the three-enzyme cascade reaction (see
Figure 2). The black dotted line shows the increasing concentration of
the final product of the cascade, ABTS +. After filtration of the solution
C
to remove polymersomes, no further enzymatic activity could be detected
(open dots). The system that lacked CalB still showed activity due to
chemical hydrolysis of 4 (grey dots).
the data points can be fitted to Equation (1), which de-
scribes a two-enzyme reaction (R2 =0.9956; Cp denotes the
concentration of the final product, ABTS +; [4]0 is the initial
C
concentration of the substrate; data were fit by using the
GraphPad Prism 5.0a software for Mac OS X.)
½4ꢃ0
1
2
ð1Þ
CpðtÞ ¼
ðꢁk2eꢁk tþk1eꢁk tÞþ½4ꢃ0
k2ꢁk1
This suggests that one of the three enzymes has an activity
high enough not to influence the total kinetics. We propose
that this is HRP. Its spatial position relative to GOx enables
Chem. Eur. J. 2009, 15, 1107 – 1114
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1111

J. C. M. van Hest et al.
Mass spectrometry: ESI-TOF measurements were performed by using an
AccuTOF-CS (Jeol) instrument. Samples were prepared in ultrapure
water containing 0.5% (v/v) formic acid with a final concentration of
2 mgmLꢁ1. ESI ion trap spectra were obtained by using an LCQ advant-
age max (Thermo Finnigan, Thermo scientific) instrument on samples in
does not influence the overall kinetics, probably as a result
of its location on the surface of the polymersome. Although
the macromolecular assembly of the model enzyme cascade
reaction presented herein does not provide an inherent cata-
lytic advantage over a mixture of the same enzymes when
freely dissolved, it clearly illustrates the viability of ad-
vanced enzyme positioning in polymersomes.
MeOH with a final concentration of 1 mgmLꢁ1
.
Preparation of polymersomes: PS-PIAT (0.5 mg) was dissolved in THF
(0.5 mL) containing the appropriate wt% of anchor 1. Subsequently, it
was gently dripped into a phosphate buffer (2.5 mL, 20 mm, pH 7.4) and
left to self-assemble for 30 min. The suspension was then transferred to
an Amicon Ultra Free-MC centrifugal filter with a cutoff of 100 kDa and
centrifuged to dryness. The polymersomes were redispersed in phosphate
buffer (600 mL, 20 mm, pH 7.4) and then centrifuged again. This step was
repeated six times. The resulting vesicles were redispersed in phosphate
buffer (1 mL, 20 mm, pH 7.4). TEM images of all types of polymersomes
used in this study are shown in Figure 6.
Experimental Section
Materials: PS-b-PEG-acetylene
1 was prepared as previously de-
scribed.[23] Diazo transfer reagent 2 was prepared as described else-
where.[25] Deuterated chloroform (CDCl3, 99.8%), deuterated methanol
(CD3OD, 99.8%) and heavy water (D2O, 99.9%) were purchased from
Aldrich. CuSO4 (Merck); sodium ascorbate (Fluka, >99%); 4,7-diphen-
yl-1,10-bathophenanthroline disulfonic acid disodium salt (Sigma-Al-
drich); anhydrous sodium sulfate (Fluka, 99%); K2CO3 (Fisher); magne-
sium sulfate dihydrate (Fluka, 99%); NaN3 (Acros, 99%); 3 (BioChe-
mika, Aldrich); 2,3,4,6-tetra-O-benzyl-d-Glucopyranose (Aldrich);
Celite; acetic anhydride (Ac2O); NaHCO3, Na2CO3 and Na2HPO4
(Merck) and 2,2’-azinobis(3-ethyl-benzothiazoline-6-sulfonic acid) dia-
mmonium salt (ABTS) (Fluka, 99%) were all used as received. Char-
coal-supported Pd catalyst (Pd/C), HNO3, HCl, MeOH, EtOAc, NaCl,
AgOAc, toluene, 1,4-dioxane, DMF and Et2O were of technical grade
(Baker) and used as received. Styrene (Aldrich) was distilled before use.
THF (Acros, >99%) was distilled from sodium and benzophenone and
CH2Cl2 (Baker) was distilled from calcium hydride and Et3N (Baker) was
distilled from CaCl prior to use. Ultrapure water was purified by using a
Encapsulation of enzymes in polymersomes: For vesicles containing GOx
in their lumen, 250 mL of the buffer was replaced by an equal volume of
a GOx stock solution in the same buffer (2.5 mgmLꢁ1, 15.6 mm). The re-
maining procedure was unchanged. For polymersomes containing CalB
in their membranes, the solution containing the block-copolymers in
THF was first injected into 100 mL of a CalB stock solution (2 mgmLꢁ1
,
56.1 mm) in ultrapure water. This dispersion was lyophilised and redis-
solved in THF (0.5 mL) and then used as described above. TEM images
of all types of polymersomes used in this study are shown in Figure 6.
Diazo transfer to HRP: A solution of HRP in ultrapure water (200 mL,
2.5 mgmLꢁ1) was treated with K2CO3 (100 mL of an aqueous solution,
2 mgmLꢁ1), along with a solution of CuIISO4·5H2O in ultrapure water
WaterPro PS polisher (Labconco, Kansas City, MO) set to 18.2mWcmꢁ1
.
Candida antarctica Lipase B, recombinant from Aspergillus oryzae (E.C.
3.1.1.3), horseradish peroxidase (E.C. 1.11.1.7) type VI and glucose ox-
idase (E.C. 1.1.3.4) type X-S from Aspergillus niger were purchased from
Sigma (BioChemika). PS40-PIAT50 was purchased from Encapson B.V.
(Nijmegen, The Netherlands).
1H NMR spectroscopy: spectra were recorded on a Varian inova400 in-
1
strument at room temperature. H NMR spectra are reported in ppm (d)
relative to tetramethylsilane (d=0.00) when measured in CDCl3. In
CD3OD the solvent residual peak is used as a reference. Multiplicities
are reported as follows: s (singlet), d (doublet), t (triplet), m (multiplet)
or br (broad). The number of protons for a given resonance is based on
spectral integration values.
Transmission electron microscopy (TEM): TEM images were obtained
by using a JEOL JEM 1010 microscope (60 kV) equipped with a CCD
camera. Samples were prepared by placing 10 mL of a solution on a
carbon-coated copper grid for 15 min, after which time the liquid was re-
moved. The grid was washed with 10 mL ultrapure water, which was sub-
sequently removed, after which the grid was dried in vacuo. The struc-
tures were visualised without further treatment.
Scanning electron microscopy (SEM): SEM images were obtained by
using a JEOL JSM T300 Scanning microscope (30 kV). Samples were
prepared by placing 10 mL of a solution on a carbon-coated copper grid
for 15 min, after which time the liquid was removed. The grid was
washed with 10 mL ultrapure water, which was subsequently removed,
after which the grid was dried in vacuo. Grids were subsequently coated
in 1.5 nm Pt/Au by using a BALZERS sputter machine. The structures
were visualised without further treatment.
UV spectroscopy: UV absorption spectra were measured by using a
Wallac Multilabel Counter 1420 (Victor Wallac) with 96-well microtiter
plates. All samples were freshly prepared and measurements were started
immediately after mixing.
Figure 6. Transmission electron micrographs of polymersomes. All micro-
graphs are TEM images of PS-PIAT polymersomes containing 10 wt%
of anchor 1. The black scale bar denotes 200 nm. A), C) and E) are poly-
mersomes containing GOx in their lumen and CalB in their membranes.
B), D) and F) are the same polymersomes, but now with HRP conjugat-
ed to their surfaces. In A) and B) HRP is labelled with Ru by using 3. In
C) and D) CalB is labelled with Ru by using 3. In E) and F) GOx is la-
belled with Ru by using 3.
Inductively coupled plasma–mass spectrometry (ICP–MS): ICP–MS
measurements were performed by using an Xseries I quadropole machine
(Thermo Fisher Scientific) with 5 mL samples containing 0.5 mgmLꢁ1
AgOAc as an internal standard.
1112
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1107 – 1114

A Three-Enzyme Cascade Reaction
FULL PAPER
(25 mL, 1 mgmLꢁ1). After mixing, compound 2 was added as a solution in
ultrapure water (15 mL, 2 mgmLꢁ1, 1.75 equiv) and the reaction was left
on a roller bank overnight. The reaction mixture was transferred to an
Amicon UltraFree-MC centrifugal filter with a 3 kDa cutoff and centri-
fuged to dryness. The supernatant was redissolved in ultrapure water
(600 mL) and centrifuged again. This procedure was repeated for a total
of five such washings. Finally, the product was redissolved in ultrapure
water (200 mL). It was analysed by ESI-TOF. MS: m/z: 43282.00 [M]
(calcd (for four transfers based on 43178.00 found for unreacted HRP):
43281.96).
belled azido-HRP was reacted with 3 (0.1 equiv) prior to diazo transfer
as described above.
ICP–MS analysis of biohybrid polymersomes: Dispersions of polymer-
somes in ultrapure water containing Ru-labelled enzymes were lyophi-
lised. The dry vesicles were then destructed in concentrated nitric acid
(0.5 mL) at 808C for 3 h. The samples were cooled to room temperature
and AgOAc was added as an internal standard (2 mgmLꢁ1 in ultrapure
water, 1.25 mL). The total volume of each sample was then brought to
5.0 mL by using ultrapure water prior to measurement. The resulting
ppm values were expressed as molarities by standardising Ru counts on
Ag counts and comparing these results to samples containing known
amounts of labelled enzyme.
Conjugation of azido-HRP to polymersome surfaces: An aqueous solu-
tion of azido-functionalised HRP (33 mL, 75 mm, 2 equiv relative to 1) was
added to a dispersion of acetylene-functionalised polymersomes in phos-
phate buffer (200 mL, 20 mm, pH 7.4). Aqueous solutions of
CuIISO4·5H2O containing sodium ascorbate (10 mm each, 33 mL) and
bathophenanthroline ligand (10 mm, 33 mL) were pre-mixed and subse-
quently added to the dispersion, which was left at 48C for 60 h. The mix-
ture was then transferred to an Amicon UltraFree-MC centrifugal filter
with a 100 kDa cutoff and centrifuged to dryness. The supernatant poly-
mersomes were redispersed in phosphate buffer (600 mL, 20 mm, pH 7.4)
and centrifuged again. This step was repeated until no enzyme activity
could be detected in the filtrate. The resulting biohybrid was redispersed
in phosphate buffer (200 mL, 20 mm, pH 7.4).
Curve fitting: Curves were fitted by using Prism 5.0a for Mac OS X. All
fits were least-squares fits using one thousand iterations. For Equa-
tion (1), initial values were set as follows: [4]0 =550, k1 =0.02, k2 =0.002.
Only the first two thirds of the data points were fitted, leading to a curve
with an R2 value of 0.9956. The last forty percent of the data points could
ACTHNUTRGNEUNG
be fitted to a sigmoidal curve of the shape Cp(t)=[4]0 ꢁth/(Km+th), in
which Km represents the Michaelis–Menten constant. This led to a curve
+
with an R2 value of 0.9997, which suggests that the decay of ABTS is
C
responsible for the departure from Equation (1). In the measurement of
polymersomes that do not have CalB in their membranes, water can still
hydrolyse 4. The absence of CalB can be translated into Equation (1) by
reducing k1 to a very small number relative to its original value. The fol-
lowing initial values were used for the fitting procedure: [4]0 =550, k1 =
2ꢁ10ꢁ9, k2 =0.002. The resulting curve, while less convincing with an R2
value of 0.9682, still suggests that the overall shape of the progress curve
is unaltered by the drastic reduction of k1, indicating that CalB is not the
slowest enzyme in the triad.
Synthesis of 1-O-acetyl-2,3,4,6-tetra-O-benzyl-d-glucopyranose (5):
2,3,4,6-Tetra-O-benzyl-d-glucopyranose (990 mg, 1.8 mmol), was dis-
solved in dry CH2Cl2 (30 mL). Et3N (240 mL, 1.9 mmol, 1.05 equiv) was
added, followed by acetic anhydride (180 mL, 1.9 mmol, 1.05 equiv).
After stirring for 2 h, a further portion of Et3N (240 mL, 1.9 mmol,
1.05 equiv) was added, followed by more Ac2O (180 mL, 1.9 mmol,
1.05 equiv), which was left to stir for 2 h. The crude reaction mixture was
washed with 1m aqueous HCl (3ꢁ), a 5% aqueous solution of NaHCO3
(2ꢁ), ultrapure water and brine. Compound 5 (860 mg, 82%) was ob-
tained after flash chromatography (CH2Cl2). 1H NMR (400 MHz,
CDCl3): d=2.05 (s, 3H; Ac), 3.59 (m, 2H; C6H2), 3.73 (m, 4H; C2H,
C3H, C4H, C5H), 4.50–4.91 (m, 8H; benzylic), 5.60 (d, 1H; anomeric),
7.14–7.32 ppm (m, 20H; Ar); MS: m/z: 605.30 [M+Na].
Acknowledgements
We thank Alexander B. C. Deutman for helpful discussions and
Ruud J. R. W. Peters for technical assistance. This work was funded by
the National Research School Combination Catalysis (NRSCC).
Synthesis of 1-O-acetyl-d-glucopyranose (4): Compound
5 (500 mg,
0.86 mmol) was dissolved in MeOH/EtOAc (25 mL, 2:1, v/v). Pd/C
(10 mg) was added to this solution. The solution was shaken under 3 bar
H2 pressure for 90 min by using a Parr apparatus. The Pd/C was removed
by filtration over Celite and the solution was concentrated to yield 4 as a
clear, waxy solid (176 mg, 92%). 1H NMR (400 MHz, CD3OD): d=2.02
(s, 3H; Ac), 3.04 (br, 2H; C3H and C5H), 3.13 (br, 2H; C6H2), 3.38 (m,
1H; C2H), 3.58 (m, 1H; C5H), 4.55 (t, 1H; CH2OH), 4.99 (br, 1H;
C2HOH), 5.08 (br, 1H; C3HOH), 5.25 (br, 1H; C4HOH), 5.29 ppm (d,
1H; anomeric H); MS: m/z: 245.05 [M+Na].
[1] D. Avnir, T. Coradin, O. Lev, J. Livage, J. Mater. Chem. 2006, 16,
1013–1030.
[2] P. Walde, S. Ichikawa, Biomol. Eng. 2001, 18, 143–177.
[3] V. B. Kandimalla, V. S. Tripathi, H. X. Ju, Crit. Rev. Anal. Chem.
2006, 36, 73–106.
[4] D. M. Vriezema, M. C. Aragones, J. A. A. W. Elemans, J. J. L. M.
Cornelissen, A. E. Rowan, R. J. M. Nolte, Chem. Rev. 2005, 105,
1445–1489.
[5] G. Murtas, Y. Kuruma, P. Bianchini, A. Diaspro, P. L. Luisi, Bio-
chem. Biophys. Res. Commun. 2007, 363, 12–17.
[6] G. Weissman, G. Sessa, M. Standish, A. D. Bangham, J. Clin. Invest.
1965, 44, 1109.
[7] T. Oberholzer, K. H. Nierhaus, P. L. Luisi, Biochem. Biophys. Res.
Commun. 1999, 261, 238–241.
[8] O. Kreft, A. G. Skirtach, G. B. Sukhorukov, H. Mohwald, Adv.
Mater. 2007, 19, 3142–3145.
[9] F. Caruso, D. Trau, H. Mohwald, R. Renneberg, Langmuir 2000, 16,
1485–1488.
[10] D. E. Discher, A. Eisenberg, Science 2002, 297, 967–973.
[11] L. B. Luo, A. Eisenberg, Langmuir 2001, 17, 6804–6811.
[12] F. Meng, C. Hiemstra, G. H. M. Engbers, J. Feijen, Macromolecules
2003, 36, 3004–3006.
Activity assay for biohybrid polymersomes: A stock solution of 4 (1m in
20 mm phosphate buffer, pH 7.4) was freshly prepared before each series
of measurements, as was a stock solution of ABTS (4 mm in 20 mm phos-
phate buffer, pH 7.4). A dispersion of polymersomes (100 mL) or an ali-
quot of control solution (100 mL) was placed in a single well of a 96-well
microtiter plate, followed by the stock solutions of glucose acetate
(40 mL) and ABTS (20 mL). Monitoring the formation of the radical
cation of ABTS by its absorption at 405 nm was started immediately
after mixing.
Ru-labelling of enzymes: An aqueous solution of the desired enzyme in
ultrapure water was added to a weighed quantity of 3 in such an amount
that 0.5 equiv of the metal complex were present for every amine in the
protein, counting only its lysine residues and N terminus. Then, 10 vol%
of a solution of Na2CO3 in ultrapure water (1 mgmLꢁ1) was added and
the reaction mixture was left at 48C for 14 h. Thereafter, it was filtered
by using an Amicon UltraFree-MC centrifugal filter with a 3 kDa cutoff.
The supernatant was redissolved in ultrapure water (600 mL) and centri-
fuged again. This procedure was repeated for a total of five such wash-
ings. Finally, the product was redissolved in an aliquot of ultrapure water
equal to that of the enzymatic solution initially used. Reactions were
verified by inductively coupled plasma mass spectrometry (ICP–MS). La-
[13] B. M. Discher, H. Bermudez, D. A. Hammer, D. E. Discher, Y. Y.
Won, F. S. Bates, J. Phys. Chem. B 2002, 106, 2848–2854.
[14] C. Nardin, S. Thoeni, J. Widmer, M. Winterhalter, W. Meier, Chem.
Commun. 2000, 1433–1434.
[15] M. Nallani, S. Benito, O. Onaca, A. Graff, M. Lindemann, M. Win-
terhalter, W. Meier, U. Schwaneberg, J. Biotechnol. 2006, 123, 50–
59.
Chem. Eur. J. 2009, 15, 1107 – 1114
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1113

J. C. M. van Hest et al.
[16] H. J. Choi, C. D. Montemagno, Nano Lett. 2005, 5, 2538–2542.
[17] D. M. Vriezema, J. Hoogboom, K. Velonia, K. Takazawa, P. C. M.
Christianen, J. C. Maan, A. E. Rowan, R. J. M. Nolte, Angew. Chem.
2003, 115, 796–800; Angew. Chem. Int. Ed. 2003, 42, 772–776.
[18] D. M. Vriezema, A. Kros, R. de Gelder, J. J. L. M. Cornelissen,
A. E. Rowan, R. J. M. Nolte, Macromolecules 2004, 37, 4736–4739.
[19] H.-P. M. de Hoog, D. M. Vriezema, M. Nallani, S. Kuiper, J. J. L. M.
Cornelissen, A. E. Rowan, R. J. M. Nolte, Soft Matter 2008, 4, 1003–
1010.
[22] J. F. Lutz, H. G. Borner, Prog. Polym. Sci. 2008, 33, 1–39.
[23] S. F. M. Van Dongen, M. Nallani, S. Schoffelen, J. J. L. M. Cornelis-
sen, R. J. M. Nolte, J. C. M. van Hest, Macromol. Rapid Commun.
2008, 29, 321–325.
[24] M. Felici, M. Marzꢂ-Pꢃrez, N. S. Hatzakis, R. J. M. Nolte, M. C. Feit-
ers, Chem. Eur. J. 2008, 14, 9914–9920.
[25] S. F. M. Van Dongen, M. Nallani, R. L. M. Teeuwen, S. S. van Ber-
kel, J. J. L. M. Cornelissen, R. J. M. Nolte, J. C. M. van Hest, Biocon-
jugate Chem. 2008, DOI: 10.1021/bc8004304.
[20] M. Nallani, H.-P. M. de Hoog, J. Cornelissen, A. R. A. Palmans,
J. C. M. van Hest, R. J. M. Nolte, Biomacromolecules 2007, 8, 3723–
3728.
[26] E. D. Goddard-Borger, R. V. Stick, Org. Lett. 2007, 9, 3797–3800.
[27] S. Schoffelen, M. H. L. Lambermon, M. B. van Eldijk, J. C. M. van
Hest, Bioconjugate Chem. 2008, 19, 1127–1131.
[21] D. M. Vriezema, P. M. L. Garcia, N. Sancho Oltra, N. S. Hatzakis,
S. M. Kuiper, R. J. M. Nolte, A. E. Rowan, J. C. M. van Hest,
Angew. Chem. 2007, 119, 7522–7526; Angew. Chem. Int. Ed. 2007,
46, 7378–7382.
[28] S. Rameez, H. Alosta, A. F. Palmer, Bioconjugate Chem. 2008, 19,
1025–1032.
Received: October 13, 2008
Published online: December 15, 2008
1114
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1107 – 1114