Enzyme and Microbial Technology 52 (2013) 279–285
Contents lists available at SciVerse ScienceDirect
Enzyme and Microbial Technology
journal homepage: www.elsevier.com/locate/emt
Glycosylation site-targeted PEGylation of glucose oxidase retains native
enzymatic activity
Dustin W. Ritter a , Jason R. Roberts a , Michael J. McShane a,b,∗
a
b
Department of Biomedical Engineering, Texas A&M University, College Station, TX 77843, USA
Materials Science and Engineering Program, Texas A&M University, College Station, TX 77843, USA
a r t i c l e
i n f o
Article history:
Received 26 September 2012
Received in revised form 4 January 2013
Accepted 9 January 2013
Keywords:
Glucose oxidase
PEGylation
PEG-hydrazide
Operational stability
Glucose sensing
a b s t r a c t
Targeted PEGylation of glucose oxidase at its glycosylation sites was investigated to determine the effect
on enzymatic activity, as well as the bioconjugate’s potential in an optical biosensing assay. Methoxypoly(ethylene glycol)-hydrazide (4.5 kDa) was covalently coupled to periodate-oxidized glycosylation
sites of glucose oxidase from Aspergillus niger. The bioconjugate was characterized using gel electrophoresis, liquid chromatography, mass spectrometry, and dynamic light scattering. Gel electrophoresis data
showed that the PEGylation protocol resulted in a drastic increase (ca. 100 kDa) in the apparent molecular mass of the protein subunit, with complete conversion to the bioconjugate; liquid chromatography
data corroborated this large increase in molecular size. Mass spectrometry data proved that the extent
of PEGylation was six poly(ethylene glycol) chains per glucose oxidase dimer. Dynamic light scattering
data indicated the absence of higher-order oligomers in the PEGylated GOx sample. To assess stability, enzymatic activity assays were performed in triplicate at multiple time points over the course of
29 days in the absence of glucose, as well as before and after exposure to 5% w/v glucose for 24 h. At
a confidence level of 95%, the bioconjugate’s performance was statistically equivalent to native glucose
oxidase in terms of activity retention over the 29 day time period, as well as following the 24 h glucose
exposure. Finally, the bioconjugate was entrapped within a poly(2-hydroxyethyl methacrylate) hydrogel containing an oxygen-sensitive phosphor, and the construct was shown to respond approximately
linearly with a 220 ± 73% signal change (n = 4, 95% confidence interval) over the physiologically-relevant
glucose range (i.e., 0–400 mg/dL); to our knowledge, this represents the first demonstration of PEGylated
glucose oxidase incorporated into an optical biosensing assay.
© 2013 Elsevier Inc. All rights reserved.
1. Introduction
Protein PEGylation—the covalent attachment of poly(ethylene
glycol) (PEG) to a protein or peptide—has been widely employed for
therapeutic purposes since its introduction in 1977 by Abuchowski
et al. [1]. A number of reviews outline the benefits that PEGylation
can impart upon therapeutic proteins, such as enhanced circulation half-life in vivo and decreased immunogenicity [2–5]. Until
recently, modification of the ε-amino group of superficial lysine
residues with an amine-reactive PEG has been most commonly
employed due to the large number of these reactive groups (lysine
residues comprise 10% of a typical protein [6]); however, conjugates
prepared using this technique are typically heterogeneous and
Abbreviations: PEG, poly(ethylene glycol); PEG-Hz, poly(ethylene glycol)hydrazide; GOx, glucose oxidase; pHEMA, poly(2-hydroxyethyl methacrylate).
∗ Corresponding author. Tel.: +1 979 845 7941; fax: +1 979 845 4450.
E-mail addresses: [email protected] (D.W. Ritter), [email protected]
(J.R. Roberts), [email protected] (M.J. McShane).
0141-0229/$ – see front matter © 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.enzmictec.2013.01.004
often require purification to isolate the preferred isoform [4,5,7].
Furthermore, in enzymes, activity loss is an issue when random
multi-site PEGylation is applied, which has been partially attributed
to lower substrate binding affinity due to steric hinderance of the
binding site and disruption of the protein tertiary structure [8–11].
To overcome this issue, Zalipsky and co-workers designed a PEGhydrazide (PEG-Hz) derivative [12] that could be used to target
oligosaccharides on glycoproteins, allowing for PEGylation without
affecting the primary structure of the enzyme [2,13].
Glucose oxidase (GOx) is a dimeric enzyme that is used widely
in the food industry to produce gluconic acid, act as a food preservative, and determine the glucose content in foodstuffs [14]. Many
of the properties that make GOx an ideal choice for glucose sensors in the food industry also make it the most suitable choice
for incorporation into glucose biosensors for biomedical applications [15]. Our lab has developed optical glucose biosensors based
on GOx [16–18], and recent work has focused on extending the
operating lifetime of these biosensors [19]. Loss of enzymatic activity of GOx can result in undesirable changes in sensor response
characteristics (requiring recalibration and decreasing operational
280
D.W. Ritter et al. / Enzyme and Microbial Technology 52 (2013) 279–285
a fume hood, 20 ␮L of 5 M sodium cyanoborohydride in 1 N sodium hydroxide was
added. Caution: sodium cyanoborohydride is extremely toxic; as such, all operations should be performed with care in a fume hood. The sodium cyanoborohydride
is reacted with the PEGylated GOx for 30 min at room temperature under gentle
agitation to yield 4. Unreacted aldehyde sites were blocked by addition of 100 ␮L of
1 M ethanolamine (pH 9.6) and reaction for 30 min at room temperature under gentle agitation to yield 5. The PEGylated GOx was purified from low-molecular-weight
contaminants using a desalting column equilibrated with 10 mM sodium phosphate
containing 154 mM sodium chloride (pH 7.2).
2.3. Gel electrophoresis
Reducing sodium dodecyl sulfate polyacrylamide gel electrophoresis was performed to test for an increase in the hydrodynamic size of GOx as a result of the
PEGylation process. Protein samples were combined with sample buffer (containing 54 mg/mL DL-dithiothreitol), vigorously agitated, and loaded onto a 10-well,
10% precast polyacrylamide gel; all samples were duplicated symmetrically on the
gel (i.e., the first sample was loaded onto lane 1 and lane 10, etc.). Following electrophoresis, the gel was rinsed three times with deionized water and cut between
lanes 5 and 6 so half of the lanes could be stained for protein and the other half for
PEG. To stain for protein, one half of the gel was placed in 30 mL of the Coomassie
staining solution for 1 h, followed by rinsing with deionized water overnight. To
stain for PEG, the other half of the gel was placed in 30 mL of perchloric acid for
15 min, and then 10 mL of 5% w/v barium chloride and 4 mL of 0.1 N iodine were
added. The gel was stained for 10 min, followed by extensive rinsing with deionized water [23]. Both stained halves of the gel were imaged separately using a gel
imaging system (Bio-Rad model 170-8270).
2.4. Liquid chromatography
Scheme 1. Glycosylation site-targeted PEGylation of GOx.
lifetime); therefore, we hypothesized that PEGylation of GOx would
provide stability against denaturation or hydrolytic cleavage. Moreover, GOx is highly glycosylated, with a total carbohydrate content
of 18.8 ± 0.6% of its molecular mass [20], so PEG-Hz is an attractive option to avoid blocking the binding site or affecting the
protein conformation. Therefore, in this work, we PEGylated GOx
using PEG-Hz to target glycosylation sites, compared the native and
PEGylated GOx enzymatic activities, and demonstrated the bioconjugate’s ability to function as a glucose sensor. To our knowledge,
this is the first reported study to assess the enzymatic activity of
GOx modified with PEG-Hz and to incorporate PEGylated GOx into
a hydrogel and demonstrate its function as an optical biosensor.
Gel-filtration chromatography was performed to separate the PEGylated GOx
from unattached PEG-Hz (i.e., purify the conjugate), but also served as an independent secondary characterization tool to confirm the gel electrophoresis data.
The samples were injected into a liquid chromatography system (GE Healthcare Life
Sciences model ÄKTAexplorer 10) equipped with a gel-filtration column (GE Healthcare Life Sciences model HiLoad Superdex 200 PG) equilibrated with 10 mM sodium
phosphate containing 154 mM sodium chloride (pH 7.2). Absorbance at 280 nm was
monitored and 2 mL fractions were collected.
2.5. Mass spectrometry
Mass spectra were acquired with a matrix-assisted laser desorption/ionization
time-of-flight mass spectrometer (Shimadzu model Axima-CFR) operating in linear mode to determine the extent of PEGylation (i.e., the number of PEG chains
attached per GOx). Protein samples (3 mg/ml) in 10 mM sodium phosphate containing 154 mM sodium chloride (pH 7.2) were concentrated and desalted (Millipore
model ZipTipC4 Pipette Tips); the eluate was spotted directly onto a steel sample
plate, where it was combined with an equal volume of saturated sinapinic acid
solution and air-dried.
2. Materials and methods
2.6. Dynamic light scattering
2.1. Materials
GOx from Aspergillus niger (type VII, 168.8 U/mg solid, 80% protein by biuret)
and peroxidase from Amoracia rusticana (type II, 188 U/mg solid) were obtained
from Sigma. Methoxy-poly(ethylene glycol)-hydrazide (PEG-Hz, 4.5 kDa by gel permeation chromatography) was obtained from Laysan Bio.
2.2. Preparation of PEGylated GOx
A modification of Zalipsky’s PEGylation protocol was used (Scheme 1) [21]. GOx
(6 mg, 1) was dissolved in 1.8 mL of 10 mM sodium phosphate containing 154 mM
sodium chloride (pH 7.2). Separately, 8.6 mg of sodium periodate was dissolved in
200 ␮L of deionized water and protected from light. The sodium periodate solution
was immediately added to the GOx solution, and the sample was slowly agitated. The
mixture was reacted in the dark for 1 h at room temperature to yield 2. It is important to note that proteins exposed to oxidants such as periodate have been reported
to form higher-order oligomers in certain cases due to intermolecular crosslinking;
however, Nakamura et al. exposed GOx from A. niger to a five-fold higher concentration of periodate for 5 h and found that the size and shape of the periodate-oxidized
GOx was essentially the same as the native GOx [22]. The reaction was quenched by
the addition of 2.5 ␮L of glycerol, corresponding to a 20-fold molar excess of glycerol
to sodium periodate. The oxidized GOx was purified using a desalting column equilibrated with 100 mM sodium phosphate containing 154 mM sodium chloride (pH
6.0). PEG-Hz (33.8 mg) was added to the oxidized GOx solution, yielding a 200-fold
molar excess of PEG-Hz to GOx. The extremely low pKa of the hydrazide reactive
group (pKa = 3), paired with its smaller size and large molar excess as compared to
the GOx, makes attachment of PEG more favorable than intermolecular crosslinking
between oxidized sugars and superficial amines on GOx. The reaction solution was
reacted in the dark for 2 h at room temperature under gentle agitation to yield 3. In
A photon correlation spectrometer (Malvern model Zetasizer Nano ZS) was used
to acquire size distributions of the GOx and PEGylated GOx samples. This was necessary to determine the change in size after modification, as well as to gauge the extent
of oligomerization during oxidation or subsequent PEGylation. Disposable 3.5 mL
cuvettes were filled with protein samples (0.6 mg/mL) in 10 mM sodium phosphate
containing 154 mM sodium chloride (pH 7.2). In all cases, the derived count rate
exceeded 300 kcps.
2.7. Enzymatic activity experiments
Enzymatic assays of PEGylated GOx and native GOx were performed to
determine activity [24], and in all cases, enzymatic activity measurements were
performed in triplicate at pH 5.1 and 35 ◦ C. All enzyme concentrations were determined using a UV/Vis spectrophotometer (PerkinElmer model LAMBDA 45), with
a molar extinction coefficient ( = 280 nm) of 2.672 × 105 M−1 cm−1 and a molecular mass of 160 kDa for the GOx dimer (based on the manufacturer’s datasheet). To
observe the effect PEGylation has on the spontaneous denaturation of GOx, enzymatic activity was assayed over the course of 29 days stored at 37 ◦ C (elevated
temperature to accelerate deactivation) in the absence of glucose. To compare the
rates of deactivation for PEGylated GOx and native GOx under operating conditions, 0.25 mL of each protein solution (0.25 mg/mL) was injected into a dialysis
cassette (10 kDa molecular-weight cutoff, 0.5 mL capacity) and placed into 1 L of
10 mM sodium phosphate containing 154 mM sodium chloride and 5% w/v glucose
(pH 7.2). Both solutions were stirred and air-equilibrated by bubbling air through a
gas diffuser. After 24 h at room temperature, the dialysis cassettes were transferred
into 1 L of 10 mM sodium phosphate containing 154 mM sodium chloride (pH 7.2).
The dialysate was tested for the presence of glucose using a biochemical analyzer
(YSI Life Sciences model 2700 SELECT Biochemistry Analyzer) and exchanged for
D.W. Ritter et al. / Enzyme and Microbial Technology 52 (2013) 279–285
281
Fig. 1. Gel electrophoresis of PEGylated GOx and native GOx. Protein samples were combined with sample buffer containing a reducing agent to break the GOx dimer into
monomeric subunits, and loaded onto a 10% polyacrylamide gel. Lane 1: Prestained molecular weight marker; Lane 2: Native GOx; Lane 3: Blank; Lane 4: PEGylated GOx.
fresh buffer until glucose levels were undetectable. At that point, protein samples
were removed from the dialysis cassettes and assayed for enzymatic activity; PEGylated GOx and native GOx samples not exposed to glucose were also assayed for
enzymatic activity.
2.8. Sensor response with PEGylated GOx
We have previously reported on the encapsulation and testing of native GOx
in a poly(2-hydroxyethyl methacrylate) (pHEMA) hydrogel [25], and the same
general approach was used herein. Briefly, PEGylated GOx and Pd(II) meso-tetra(4carboxyphenyl) porphine, an oxygen-sensitive phosphor, were combined with a
pHEMA precursor solution at final concentrations of 20 ␮M and 100 ␮M, respectively. The pHEMA precursor solution comprised 2-hydroxyethyl methacrylate
(monomer), tetra(ethyleneglycol) diacrylate (crosslinker), and 2,2-dimethoxy2-phenylacetophenone (photoinitiator) dissolved in an ethylene glycol/water
mixture, and the molar ratio of monomer/crosslinker was 98:2. Microscope slides
were silanized with (trimethoxysilyl)propyl methacrylate to facilitate binding of the
pHEMA to the glass. The pHEMA precursor solution containing the PEGylated GOx
and phosphor was then cast into a mold placed directly on the microscope slide and
photopolymerized; four pHEMA hydrogels (ca. 2.5 mm in diameter and 1 mm thick)
on four separate microscope slides were prepared in this manner from the same
stock precursor solution.
To test each pHEMA hydrogel for a glucose response, the microscope slide was
placed into a custom-built dynamic testing apparatus (Supplementary Fig. 1) and
equilibrated in phosphate-buffered saline until the baseline phosphorescence signal was stable. The hydrogel was then exposed to random glucose concentrations
(2 h per concentration) within the physiological range (i.e., 0–400 mg/dL) by flowing phosphate-buffered saline doped with glucose across the pHEMA hydrogel at a
volumetric flow rate of 4 mL/min. A modulated green light-emitting diode (523 nm)
was coupled into a bifurcated fiber optic bundle to illuminate the hydrogel, and
phosphorescence from the oxygen-sensitive phosphor was captured through the
same fiber bundle; luminescence lifetime measurements were collected using a
frequency-domain lifetime system (TauTheta model MFPF-100).
3. Results
3.1. PEGylated GOx characterization
Our data—collected using gel electrophoresis, liquid chromatography, and mass spectrometry—prove that GOx can be PEGylated
using the protocol described in the Experimental Procedures
section. Fig. 1 shows a composite image of the gel on which PEGylated GOx and native GOx were electrophoresed. The left half of
Fig. 1, which depicts the half of the gel that was stained and imaged
for the presence of protein, clearly shows an increase in the apparent monomer molecular mass upon PEGylation of GOx; lane 2
containing native GOx displays a band at 86.4 kDa, while lane 4 containing the PEGylated GOx displays a band at 182.2 kDa. The sample
bands corresponding to the native and the PEGylated GOx are quite
diffuse. However, this is to be expected given that (1) GOx is a glycoprotein and therefore has a distribution of molecular masses;
(2) glycoproteins run anomalously in SDS-PAGE, typically with an
apparent molecular mass that is higher than the actual molecular
mass due to suppressed binding of sodium dodecyl sulfate [26,27];
and (3) the PEG has its own molecular mass distribution, which
serves to further broaden the PEGylated GOx band. Upon inspection of the right half of Fig. 1, which depicts the half of the gel
that was stained and imaged for the presence of PEG, one finds the
absence of a band in close proximity to the protein band in lane
2, while lane 4 containing the PEGylated GOx displays a faint band
nearly coincident (apparent molecular mass of 175.7 kDa) with the
protein band.
Fig. 2 shows the overlaid chromatograms of PEGylated GOx
(black line) and native GOx (gray line) subjected to gel-filtration
chromatography. A calibration curve (Supplementary Fig. 2) was
compiled using seven proteins over a wide range of molecular
masses within the column’s fractionation range (i.e., 10–600 kDa)
and was used to convert elution volume to molecular mass. Native
GOx (elution volume of 67.52 mL) is estimated to be 140 kDa,
while the apparent molecular mass of PEGylated GOx (elution volume of 56.33 mL) could not be estimated because it eluted before
the largest calibrant (i.e., thyroglobulin with a molecular mass of
669 kDa and an elution volume of 60.3 mL). Therefore, while the
column that was chosen for gel-filtration chromatography was
appropriate to separate the bioconjugate from the native GOx, the
only conclusion that can be made about the bioconjugate’s apparent
molecular mass is that it is larger than thryoglobulin.
282
D.W. Ritter et al. / Enzyme and Microbial Technology 52 (2013) 279–285
Fig. 2. Overlaid chromatograms of PEGylated GOx (black line) and native GOx (gray
line). Native GOx eluted at 67.52 mL and is estimated to be 140 kDa, while PEGylated
GOx eluted at 56.33 mL (molecular mass cannot be estimated as it is out of the
calibration range).
Fig. 3 shows the mass spectra of PEGylated GOx (black line) and
native GOx (gray line) collected using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. The spectrum
collected from the native GOx sample depicts a broad peak centered at approximately 72 kDa, while the spectrum collected from
the PEGylated GOx sample depicts an even broader peak centered
at approximately 85 kDa; this shift corresponds to a mass change of
approximately 13 kDa per monomer as a result of GOx PEGylation.
Fig. 4 contains the size distributions of PEGylated GOx (black
bars) and native GOx (gray bars). The PEGylated GOx has a mean
hydrodynamic diameter of 17.13 nm, while the native GOx has a
mean hydrodynamic diameter of 11.48 nm, corresponding to an
approximate 50% increase in the hydrodynamic size upon PEGylation. The size distributions have approximately equivalent widths
(as determined by measuring the full width at half of the maximum).
3.2. Enzymatic activity experiments
Enzymatic activity measurements of PEGylated GOx and native
GOx were performed at multiple time points over the course of 29
days in the absence of glucose, as well as before and after exposure
to 5% w/v glucose for 24 h. Fig. 5 depicts the loss in specific activity
of PEGylated GOx (black filled circles) and native GOx (gray filled
Fig. 3. Mass spectra of PEGylated GOx (black line) and native GOx (gray line). The
peak for the native GOx sample corresponds to 72 kDa, while the peak for the PEGylated GOx sample corresponds to 85 kDa.
Fig. 4. Size distributions of PEGylated GOx (black bars) and native GOx (gray bars).
Native GOx has a mean hydrodynamic diameter of 11.48 nm, while PEGylated GOx
has a mean hydrodynamic diameter of 17.13 nm.
circles) in the absence of glucose and at 37 ◦ C. Data points represent
the average specific activities at each time point, and error bars represent 95% confidence intervals (n = 3). The data were fitted with
a first-order exponential decay, with the black line corresponding to the PEGylated GOx data and the gray line corresponding
to the native GOx data. Although the specific activities for PEGylated GOx and native GOx are statistically different at intermediate
time points, they are not statistically distinct at the initial two time
points and the final time point. From Day 1 to Day 29, the activity retention for PEGylated GOx and native GOx is 44.0 ± 6.64%
and 37.2 ± 2.29%, respectively. Half-life values are given by t1/2 = (ln
2)/, where t1/2 is the half-life and is the decay constant from
the fitted exponential decay. Half-life values for PEGylated GOx
and native GOx kept at storage conditions range from 16.8 to 28.6
days and from 9.54 to 20.1 days, respectively (95% confidence intervals). Table 1 compares the activity retention of PEGylated GOx and
native GOx following glucose exposure for 24 h; data represent 95%
confidence intervals (n = 3). Although GOx apparently has greater
specific activity before and after glucose exposure as compared
to PEGylated GOx before and after glucose exposure, the activity
retention of GOx and PEGylated GOx following glucose exposure
for 24 h are statistically equivalent. Half-life values for PEGylated
Fig. 5. Loss in specific activity of PEGylated GOx (black filled circles) and native
GOx (gray filled circles) in the absence of glucose and at 37 ◦ C. Data points represent the average specific activities at each time point, and error bars represent
95% confidence intervals (n = 3). The data were fitted with a first-order exponential
decay, with the black line corresponding to the PEGylated GOx data and the gray
line corresponding to the native GOx data.
D.W. Ritter et al. / Enzyme and Microbial Technology 52 (2013) 279–285
Table 1
Activity retention following glucose exposure for 24 h.a
Specific activity (U/mg)
Sample
Pre-glucose
exposure
Post-glucose
exposure
Activity
retentionb (%)
GOx
PEGylated GOx
237 ± 12
195 ± 4.1
159 ± 15
122 ± 5.8
67.3 ± 7.3
62.6 ± 3.2
a
b
Data are expressed as 95% confidence intervals (n = 3).
Activity retention = (post-glucose exposure) ÷ (pre-glucose exposure) × 100.
GOx and native GOx exposed to glucose are estimated to range from
30.7 to 40.4 h and from 35.7 to 48.5 h, respectively (95% confidence
intervals).
3.3. Sensor response with PEGylated GOx
Fig. 6 shows the sensor response that was obtained by entrapping PEGylated GOx and an oxygen-sensitive phosphor in a pHEMA
hydrogel. Data points represent the average luminescence lifetimes
at each glucose concentration, and error bars represent 95% confidence intervals (n = 4). The sensor response is approximately linear,
and luminescence lifetime values increase by 220 ± 73% over the
physiologically-relevant glucose range (i.e., 0–400 mg/dL).
4. Discussion
4.1. PEGylated GOx characterization
It is important to note that, of the various analytical techniques
that were used to characterize the PEGylated GOx, mass spectrometry is the only one expected to provide the actual molecular mass
of the conjugate; both gel electrophoresis and liquid chromatography can only provide apparent molecular mass of the PEGylated
GOx (based on calibration curves constructed using non-PEGylated
protein molecular mass standards), which can differ greatly from
the actual molecular mass.
Gel electrophoresis and liquid chromatography data
(Figs. 1 and 2) indicate complete modification of GOx, as evidenced
by the lack of a band in the gel image or a peak in the chromatogram
corresponding to the presence of unmodified GOx in the PEGylated
GOx sample. Furthermore, the presence of a single bioconjugated
specie suggests that the chosen PEG:GOx molar ratio of 200:1
results in saturation of the available conjugation sites. To confirm
283
that a PEG:GOx molar ratio of 200:1 results in saturation, GOx
was PEGylated using higher PEG:GOx molar ratios (i.e., 500:1 and
2000:1) and the bioconjugate was subjected to gel-filtration chromatography equipped with a column with a fractionation range
of 5 kDa to 5 MDa (GE Healthcare Life Sciences model Superose 6
HR); the three bioconjugates, which are within the fractionation
range of the column, have similar elution volumes (Supplementary
Fig. 3), indicating that the lowest PEG:GOx molar ratio of 200:1 is
sufficient to saturate the available conjugation sites.
The peaks in the mass spectra (Fig. 3) are relatively broad; as
explained in Section 3.1, this is expected from a glycoprotein due
to the existence of multiple glycoforms with various molecular
masses. The peak from the PEGylated GOx sample is broader than
the native GOx peak, which is also expected as the PEG’s molecular
mass distribution will introduce further broadening in the molecular mass distribution of the bioconjugate; this effect has been
reported by other groups as well [28]. The monomeric molecular
mass for native GOx by mass spectrometry matches well with the
dimeric molecular mass for native GOx estimated by liquid chromatography, while the monomeric molecular mass for native GOx
estimated by gel electrophoresis is slightly higher (likely due to
glycosylation; see Section 3.1). The mass change of approximately
13 kDa per monomer as a result of GOx PEGylation corresponds to
approximately three 4.5 kDa PEG chains; therefore, our PEGylation
protocol is believed to result in the attachment of approximately
six PEG chains per GOx dimer.
In 2004, Fee and Van Alstine proposed a model that was shown
to accurately predict the viscosity radius of various PEGylated
proteins, which was related to the bioconjugate’s behavior in gelfiltration chromatography [29]. By applying this model to our
system, we find that despite the modest increase in molecular
mass upon GOx PEGylation (ca. 26 kDa), the viscosity diameter of
the bioconjugate is estimated to increase from 8.6 nm (unmodified
dimeric GOx; 144 kDa by mass spectrometry) to 14.3 nm (144 kDa
GOx + 26 kDa PEG). This 66% increase in viscosity diameter results in
an approximate five-fold increase in the apparent molecular mass
upon GOx PEGylation (from 144 kDa to 667 kDa), which agrees with
our earlier observation that the bioconjugate’s apparent molecular
mass exceeds the column’s exclusion limit of 600 kDa. Furthermore, a 66% increase in viscosity diameter agrees well with the
50% increase in hydrodynamic size that was observed by dynamic
light scattering.
The dynamic light scattering data do not support the presence of
higher-order oligomers in the PEGylated GOx sample, which would
broaden the peak or present as slightly larger secondary peaks (corresponding to dimers, trimers, etc.). Although a secondary peak
exists at 526 nm (not visible on graph as shown), it represents only
0.3% of the sample by volume (the primary peak accounts for the
remaining 99.7% by volume), and is much larger than one would
expect for an oligomer formed by intermolecular crosslinking of
GOx during oxidation or PEGylation; it likely represents a small
fraction of aggregated denatured protein or a contaminant.
4.2. Enzymatic activity experiments
Fig. 6. Sensor response obtained by entrapping PEGylated GOx and an oxygensensitive phosphor in a pHEMA hydrogel. Data points represent the average
luminescence lifetimes at each glucose concentration, and error bars represent 95%
confidence intervals (n = 4).
Enzymatic activity assays of PEGylated GOx and native GOx
were performed in triplicate at multiple time points over
the course of 29 days in the absence of glucose, as well as
before and after exposure to 5% w/v glucose for 24 h. Although
half-life values are dependent upon many variables (e.g., temperature, pH, substrate concentration, enzyme concentration, enzyme
immobilization), our measurements are consistent with previous reports in the literature [30–34]. For example, Krishnaswamy
and Kittrell reported that solution-phase GOx (citrate-phosphate
buffer, pH 5.5, air-equilibrated) exposed to 10 mM glucose has a
half-life of 23.1 h [30], which is slightly lower than what we
284
D.W. Ritter et al. / Enzyme and Microbial Technology 52 (2013) 279–285
observed for native GOx; similarly, Fortier and Belanger showed
that GOx entrapped in a polypyrrole matrix and exposed to 250 mM
glucose (ca. 5% w/v glucose) at room temperature had a half-life of
approximately 18.7 h. The same group demonstrated that solutionphase GOx has a half-life of 6 h at 50 ◦ C in the absence of glucose
[32], which is significantly lower than what we observed at 37 ◦ C.
However, Tse and Gough reported a much higher half-life (i.e.,
87.2 days) for immobilized GOx stored in the absence of glucose at
37 ◦ C in 0.1 M phosphate buffer (pH 7.3) [31]. At a confidence level
of 95%, the bioconjugate’s performance was statistically equivalent to native GOx in terms of activity retention over the 29 day
time period and following the 24 h glucose exposure. These findings agree well with recently published work by Untrweger et al.
[28]; site-specific PEGylation of a mutated L-lactate oxidase with
one or two maleimide-activated PEGs did not alter the enzymatic
activity of the mutant. Conversely, Slavica et al. [35] showed that
orthogonal maleimide-thiol coupling of d-amino acid oxidase with
three maleimide-activated PEGs resulted in a marked decrease in
substrate catalytic efficiency for the dioxygen-dependent reaction.
Finally, there have been mixed reports concerning post-PEGylation
activity when amine-reactive PEGs are employed, with some studies demonstrating complete retention of enzymatic activity [36,37]
and others indicating partial loss of enzymatic activity [8–11].
5. Conclusion
PEG-Hz was covalently coupled to periodate-oxidized glycosylation sites of GOx from A. niger. Gel electrophoresis and
liquid chromatography data indicated that the PEGylation protocol resulted in a drastic increase in the apparent molecular mass
of GOx, with complete conversion to the bioconjugate. Mass spectrometry data proved that the extent of PEGylation was three PEG
chains per GOx subunit (i.e., six PEG chains per GOx dimer). Enzymatic activity assays of PEGylated GOx and native GOx revealed
that the bioconjugate’s performance was statistically equivalent
to native GOx in terms of activity retention over the 29 day time
period and following the 24 h glucose exposure. Therefore, PEGylation of GOx with 4.5 kDa PEG-Hz does not provide stability against
denaturation or hydrolytic cleavage as was hypothesized; however, larger PEGs or PEGs of various geometries exist and might
provide better protection of the GOx. Furthermore, coupling of
other potentially stabilizing molecules (e.g., other synthetic or natural polymers, proteins, etc.) to glucose oxidase via its glycosylation
sites can now be explored as it is likely that targeted attachment to
these sites will not have a detrimental effect on enzymatic activity.
Nevertheless, it is expected that future applications of this work
will utilize the bioconjugate’s drastically increased size and nativelike enzymatic activity. Many polymeric matrices, including the
pHEMA used in this work, have a mesh size that is on the same
order as the size of the protein; therefore, increases in the hydrodynamic radius of the protein could translate to better entrapment.
Future work in our lab will explore this possibility. Finally, the
bioconjugate was entrapped within a pHEMA hydrogel containing an oxygen-sensitive phosphor, and the construct was shown to
respond approximately linearly and with high sensitivity over the
physiologically-relevant glucose range.
Author contributions
DWR, JRR, and MJM conceived and designed the experiments.
DWR and JRR performed the experiments and analyzed the data.
MJM contributed materials and analysis tools. DWR, JRR, and
MJM wrote and edited the manuscript. All authors have read and
approved the manuscript
Acknowledgments
This work was supported, in part, by the National Institutes
of Health (1R43DK093139) and the Texas Engineering Experiment Station. Dustin W. Ritter thanks the Dwight Look College
of Engineering at Texas A&M University and the National Science Foundation for financial support through the Barclay/Willson
National Excellence Fellowship and the Graduate Research Fellowship, respectively. The authors also gratefully acknowledge the
Protein Chemistry Laboratory at Texas A&M University for performing gel-filtration chromatography and matrix-assisted laser
desorption/ionization time-of-flight mass spectrometry, and for
providing technical expertise, comments, and suggestions. Jared
M. Newton is acknowledged for his assistance in collecting preliminary data for this work.
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.enzmictec.
2013.01.004.
References
[1] Abuchowski A, Van Es T, Palczuk NC, Davis FF. Alteration of immunological
properties of bovine serum albumin by covalent attachment of polyethylene
glycol. Journal of Biological Chemistry 1977;252:3578–81.
[2] Zalipsky S. Chemistry of polyethylene glycol conjugates with biologically active
molecules. Advanced Drug Delivery Reviews 1995;16:157–82.
[3] Veronese FM. Peptide and protein PEGylation: a review of problems and solutions. Biomaterials 2001;22:405–17.
[4] Jevševar S, Kunstelj M, Porekar VG. PEGylation of therapeutic proteins. Biotechnology Journal 2010;5:113–28.
[5] Pasut G, Veronese FM. State of the art in PEGylation: the great versatility achieved after forty years of research. Journal of Controlled Release
2012;161:461–72.
[6] Fee CJ, Van Alstine JM. PEG-proteins: reaction engineering and separation
issues. Chemical Engineering Science 2006;61:924–39.
[7] Bailon P, Berthold W. Polyethylene glycol-conjugated pharmaceutical proteins.
Pharmaceutical Science & Technology Today 1998;1:352–6.
[8] Matsushima A, Kodera Y, Hiroto M, Nishimura H, Inada Y. Bioconjugates of
proteins and polyethylene glycol: potent tools in biotechnological processes.
Journal of Molecular Catalysis B: Enzymatic 1996;2:1–17.
[9] Veronese FM, Caliceti P, Schiavon O, Sartore L. Preparation and properties of
monomethoxypoly(ethylene glycol)-modified enzymes for therapeutic applications. In: Harris JM, editor. Poly(ethylene glycol) Chemistry: Biotechnical and
Biomedical Applications. New York: Plenum Press; 1992, 127–137.
[10] Rodriguez-Martinez JA, Rivera-Rivera I, Sola RJ, Griebenow K. Enzymatic activity and thermal stability of PEG-alpha-chymotrypsin conjugates. Biotechnology
Letters 2009;31:883–7.
[11] Lopez-Cruz JI, Viniegra-Gonzalez G, Hernandez-Arana A. Thermostability of
native and pegylated Myceliophthora thermophila laccase in aqueous and mixed
solvents. Bioconjugate Chemistry 2006;17:1093–8.
[12] Zalipsky S, Lee C, Menon-Rudolph S. Hydrazine containing conjugates of
polypeptides and glycopolypeptides with polymers. 1992.
[13] Zalipsky S. Functionalized poly(ethylene glycols) for preparation of biologically
relevant conjugates. Bioconjugate Chemistry 1995;6:150–65.
[14] Crueger A, Crueger W. Glucose transforming enzymes. In: Fogarty WM, Kelly
CT, editors. Microbial enzymes and biotechnology. 2nd edn. London: Elsevier
Applied Science; 1990. p. 177–226.
[15] Wilson R, Turner APF. Glucose oxidase: an ideal enzyme. Biosensors and Bioelectronics 1992;7:165–85.
[16] Brown JQ, Srivastava R, McShane MJ. Encapsulation of glucose oxidase and
an oxygen-quenched fluorophore in polyelectrolyte-coated calcium alginate
microspheres as optical glucose sensor systems. Biosensors and Bioelectronics
2005;21:212–6.
[17] Stein EW, Grant PS, Zhu H, McShane MJ. Microscale enzymatic optical biosensors using mass transport limiting nanofilms 1. Fabrication and
characterization using glucose as a model analyte. Analytical Chemistry
2007;79:1339–48.
[18] Stein EW, Singh S, McShane MJ. Microscale enzymatic optical biosensors using
mass transport limiting nanofilms 2. Response modulation by varying analyte
transport properties. Analytical Chemistry 2008;80:1408–17.
[19] Singh S, McShane M. Enhancing the longevity of microparticle-based glucose
sensors towards 1 month continuous operation. Biosensors and Bioelectronics
2010;25:1075–81.
[20] Tsuge H, Natsuaki O, Ohashi K. Purification, properties, and molecular features of glucose oxidase from Aspergillus niger. Journal of Biochemistry
1975;78:835–43.
D.W. Ritter et al. / Enzyme and Microbial Technology 52 (2013) 279–285
[21] Zalipsky S, Menon-Rudolph S. Hydrazide derivatives of poly(ethylene glycol)
and their bioconjugates. American Chemical Society 1997:318-41.
[22] Nakamura S, Hayashi S, Koga K. Effect of periodate oxidation on the structure and properties of glucose oxidase. Biochimica Et Biophysica Acta, Enzymol
1976;445:294–308.
[23] Skoog B. Determination of polyethylene glycols 4000 and 6000 in plasma protein preparations. Vox Sanguinis 1979;37:345–9.
[24] Bergmeyer HU, Gawehn K, Grassl M. Methods of enzymatic analysis. 2nd edn.
New York: Academic Press; 1974.
[25] Roberts JR, Park J, Helton K, Wisniewski N, McShane MJ. Biofouling of polymer
hydrogel materials and its effect on diffusion and enzyme-based lumniescent
glucose sensor functional characteristics. Journal of Diabetes Science and Technology 2012;6:1267–75.
[26] Skerrow CJ, Matoltsy AG. Chemical characterization of isolated epidermal
desmosomes. Journal of Cell Biology 1974;63:524–30.
[27] Hartshorne RP, Catterall WA. The sodium channel from rat brain—purification
and subunit composition. Journal of Biological Chemistry 1984;259:1667–75.
[28] Untrweger B, Stoisser T, Leitgeb S, Birner-Grünberger R, Nidetzky B. Engineering of Aerococcus viridans L-lactate oxidase for site-specific PEGylation:
characterization and selective bioorthogonal modification of a S218C mutant.
Bioconjugate Chemistry 2012.
[29] Fee CJ, Van Alstine JM. Prediction of the viscosity radius and the size exclusion chromatography behavior of PEGylated proteins. Bioconjugate Chemistry
2004;15:1304–13.
285
[30] Krishnaswamy S, Kittrell JR. Deactivation studies of immobilized glucose oxidase. Biotechnology and Bioengineering 1978;20:821–35.
[31] Tse PHS, Gough DA. Time-dependent inactivation of immobilized glucose oxidase and catalase. Biotechnology and Bioengineering 1987;29:
705–13.
[32] Fortier G, Belanger D. Characterization of the biochemical behavior of glucoseoxidase entrapped in a polypyrrole film. Biotechnology and Bioengineering
1991;37:854–8.
[33] Chen Q, Kenausis GL, Heller A. Stability of oxidases immobilized in silica gels.
Journal of the American Chemical Society 1998;120:4582–5.
[34] Gouda MD, Singh SA, Rao AGA, Thakur MS, Karanth NG. Thermal inactivation
of glucose oxidase—mechanism and stabilization using additives. Journal of
Biological Chemistry 2003;278:24324–33.
[35] Slavica A, Dib I, Nidetzky B. Selective modification of surface-exposed thiol
groups in Trigonopsis variabilis d-amino acid oxidase using poly(ethylene
glycol) maleimide and its effect on activity and stability of the enzyme. Biotechnology and Bioengineering 2007;96:9–17.
[36] Seymour SL, Klinman JP. Comparison of rates and kinetic isotope effects using
PEG-modified variants and glycoforms of glucose oxidase: the relationship of
modification of the protein envelope to C H activation and tunneling. Biochemistry 2002;41:8747–58.
[37] Hadley KB, Sato PH. Catalytic activity of administered gulonolactone oxidase
polyethylene glycol conjugates. Enzyme 1989;42:225–34.