This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/authorsrights

Author's personal copy

Yeast surface display for the expression, purification and characterization of wild-type and B11 mutant glucose oxidases

Marija Blazic a, Gordana Kovacevic b, Olivera Prodanovic c, Raluca Ostafe d, Marija Gavrovic-Jankulovic b,Rainer Fischer d, Radivoje Prodanovic b,⇑a Center for Chemistry IHTM, University of Belgrade, Njegoševa 12, 11001 Belgrade, Serbia b Faculty of Chemistry, University of Belgrade, Studentski trg 12, 11000 Belgrade, Serbia c Institute for Multidisciplinary Research, University of Belgrade, Kneza Višeslava 1, 11030 Belgrade, Serbia d Molecular Biotechnology, Faculty of Biology, RWTH Aachen University, Worringerweg 1, 52078 Aachen, Germany

a r t i c l e i n f o

Article history: Received 19 December 2012 and in revised form 4 March 2013 Available online 3 April 2013

Keywords:Saccharomyces cerevisiae Glucose oxidase Directed evolution High throughput screening Chimera

a b s t r a c t

Glucose oxidase (GOx) catalyzes the oxidation of glucose to form gluconic acid and hydrogen peroxide, areaction with important applications in food preservation, the manufacture of cosmetics and pharmaceu- ticals, and the development of glucose monitoring devices and biofuel cells. We expressed Aspergillusniger wild type GOx and the B11 mutant, which has twi ce the activity of the wild type enzyme at pH 5.5, as C-terminal fusions with the Saccharomyces cerevisiae Aga2 protein, allowing the fusion proteins to be displayed on the surface of yeast EBY100 cells. After expression, we extracted the proteins from the yeast cell wall and purified them by ion-exchange chromatography and ultrafiltration. This produced a broad 100–140 kDa band by denaturing SDS–PAGE and a high-molecular-weight band by native PAGE corresponding to the activity band revealed by zymography. The wild type and B11 fusion proteins had kcat values of 33.3 and 61.3 s�1 and Km values for glucose of 33.4 and 27.9 mM, respectively. The pH opti- mum for both enzymes was 5.0. The kinetic properties of the fusion proteins displayed the same ratio as their native counterpart s, confirming that yeast surface display is suitable for the high-throughput direc- ted evolution of GOx using flow cytometry for selection. Aga2–GOx fusion proteins in the yeast cell wall could also be used as immobilized catalysts for the production of gluconic acid.

� 2013 Elsevier Inc. All rights reserved.

Introductio n

Glucose oxidase (GOx)1, also known as b-D-glucose:oxyge n1-oxidoredu ctase (EC 1.1.3.4), catalyzes the oxidation of glucose to form gluconic acid and hydrogen peroxide. Derivatives of GOx from the filamentous fungus Aspergillus niger are widely used in industry. The A. niger enzyme is a homodimeri c flavoprotein with a molecular mass of 160 kDa (10–25% of which (w/w) is repre- sented by glycans) whose crystal structure has been solved [1].GOx has been used in industry for food preservation [2], gluconic acid production and glucose measure ment [3], and for the manu- facture of miniature biofuel cells that should power implanted bio- medical devices using glucose from human blood as a source of energy [4] and for electrochem ical glucose sensors [5]. The effi-ciency of such devices depends on the activity of the enzyme and the electron transfer rate between the electrode surface and the

active site enzyme, which contains FAD/FADH 2. Wild type GOx from A. niger has several limitations as a component of miniature biofuel cells implanted in blood vessels, including (i) its Km valuefor glucose (�30 mM) and pH optimum (�5.0) which result in low- er activity under the physiologica l conditions of human blood (4 mM glucose, pH 7.4); and (ii) the active site is deeply buried, so it is necessary to use redox mediators that compete with oxygen as the natural substrate during the oxidative half reaction, in order to increase the rate of electron transfer from the electrode surface [6]. This has been addressed by expressing the enzyme in Saccha-romyces cerevisiae [7] and Pichia pastoris [8,9] and using directed evolution to reduce the Km value for glucose [10] and to increase the activity with redox mediators at pH 7.4 [11].

Directed evolution involves iterative rounds of diversity generation and screening allowing the selection of enzymes with improved properties [12–14]. The most powerful methods for diversity generation give rise to libraries containing up to 10 12 dif-ferent mutants, but contempor ary screening platforms based on microtiter plate assays are only suitable for libraries with a maxi- mum complexi ty of 10 4 [15]. We have recently addressed this bottlenec k by developing an ultra-high-thr oughput screening plat- form for GOx gene libraries in yeast based on flow cytometry and

1046-5928/$ - see front matter � 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.pep.2013.03.014

⇑ Corresponding author. E-mail address: rprodano@chem.bg.ac.rs (R. Prodanovic).

1 Abbreviations used: GOx, glucose oxidase; GPI, glycosylphos phatidylinositol; HRP, horseradish peroxi dase; SDS–PA GE, so dium dod ec yl sulfate polyacryl ami de gelelectrophoresis.

Protein Expression and Purification 89 (2013) 175–180

Contents lists available at SciVerse ScienceDi rect

Prote in Expressio n and Puri fication

journal homepage: www.elsevier .com/ locate /yprep

Author's personal copy

in vitro compartme ntalization in emulsions, which is compatible with high-complexi ty libraries and therefore much more likely to identify improved enzyme variants [10]. Similar platform was pre- viously developed for enzymes expressed in the bacterium Esche-richia coli or by in vitro translation, allowing the screening of libraries containing up to 10 9 mutants [16–18]. One of the major advantages of this platform is that it maintains a physical connec- tion between the gene encoding the enzyme and the product of the enzymatic reaction by compartme ntalizing cells in single or double emulsions, but this requires either intracellular expression of the enzyme or cell surface display [19].

Yeast surface display was developed for the directed evolution of antibodies in S. cerevisiae , and involves the fusion of antibody variable domains to Aga2p, the adhesion subunit of the yeast agglutinin protein. Aga2p binds via disulfide bonds to the mem- brane protein Aga1p, which is embedde d in the membran e via aglycosylpho sphatidylinosit ol (GPI) anchor [20]. The Aga2-antibodyfusion gene is cloned in the vector pCTCON, whereas the Aga1pgene is integrated into the yeast genome, but both are under the control of galactose-indu cible promoters [20]. This surface display system has been used for the directed evolution of horseradish per- oxidase (HRP) [21] and expression of GOx for applications in bio- fuel cells [22]. One drawback of the system is that the enzyme is not presented in its native form, which may alter its activity and kinetic characterist ics, especiall y in the case of GOx where native enzyme should be in a dimeric form in order to be active [1]. It is often said that in directed evolution experime nts ‘you get what you screen for’ [23] so it is important to establish a correlation be- tween the activity of the surface-displ ayed fusion proteins and their soluble counterparts before using surface display to select en- zymes that will be utilized in their native form.

In order to see if there is a correlation between the activity of the surface-display ed GOx and the native GOx we constructed fu- sion genes for the wild type enzyme and the more active B11 mu- tant fused to the C-terminus of Aga2. We purified the fusion proteins from the yeast cell wall and compared their activities and kinetic properties to those of the correspondi ng soluble en- zymes also expressed in yeast. We found that the ratio of kinetic parameters between the soluble wild type and mutant enzymes was preserved when they were expressed as fusion proteins.

Materials and methods

Introduction of GOx sequences into the pCTCON2 vector

The A. niger NRLL-3 GOx gene (P13006) was synthesized by GenScript USA Inc. (NJ, USA) and the mutant B11 (T30V, I94V)was generated by directed evolution as previously described [10]. The genes were amplified by PCR using forward primer NhEI_fp_GOx (50-ATC GCT AGC AGC AAT GGC ATT GAA GC-3 0)and reverse primer BamHI_rp _GOx (50-ATC GGA TCC TCC CTG CAT GGA AGC-3 0). The products were digested with BamHI and NheI and inserted in vector pCTCON2 (kindly provided by Prof. Dane Wittrup, MIT, MA, USA) which had been linearized with the same enzymes . The recombinant vector was cloned in E. coli strain DH5 a, which was cultivated in Luria–Bertani med- ium supplemented with 100 mg/L ampicillin.

Transformati on of S. cerevisiae EBY 100 cells

S. cerevisiae strain EBY100 (kindly provided by Prof. Dane Witt- rup) is based on strain BJ5465 (MATa ura 3–52 trp 1 leu2D1his3D200 pep4:HIS3 prb 1D1.6R can1 GAL) and contains a stably- integrated Aga1 gene and the URA marker. Yeast cells were trans- formed with the recombinant pCTCON2 vector using the Gietz

protocol [24] with an additional 42 �C heat shock for 2.5 h. Trans- formed cells were plated on YNB-CAA Gal/Raf agar and positive clones were identified by overlaying with 0.1 M sodium acetate buffer (pH 5.5) containing 2 mM 2,2 0-azino-bis(3-ethylbenzothiaz -oline-6-s ulphonic acid) (ABTS), 100 mM glucose and 1 U/mL HRP.

Expression of GOx in liquid culture

Positive colonies identified in the ABTS overlay assay (seeabove) were inoculated into 25 mL YNB-CAA Glc liquid medium and incubated for 48 h at 30 �C, 250 rpm. GOx expression was in- duced by adding 75 mL YNB-CAA Gal/Raf medium and incubating the culture as above for various duration s to optimize the expres- sion protocol.

Purification of GOx

Yeast cells were harvested by centrifugation (3000g, 5 min, 4 �C) and resuspended in 10 mL 0.1 M sodium acetate buffer (pH5.5) containing 1 mM 2-mercaptoethan ol. The suspension was incubate d at 4 �C for 4 h then centrifuged as above and the super- natant was dialyzed against 10 mM sodium phosphat e buffer (pH6.0). The enzymes were purified by ion-exch ange chromatograp hy on a HiTrap DEAE FF 5 mL column using a linear gradient from 10 to 500 mM sodium phosphate buffer (pH 6.0). Fractions containing active GOx were collected and concentrated using Vivaspin ultra- filtration columns with a 50-kDa molecular cutoff.

Quantitati ve GOx assay

GOx activity was measured using a quantitative version of the ABTS agar plate assay described above. The purified enzyme was mixed with 0.1 M sodium acetate buffer (pH 5.5) containing 2 mM ABTS, 100 mM glucose and 1 U/mL HRP. Absorbance was measure d at 420 nm using Mapada UV-3100PC spectrophotom eter

Fig. 1. Agarose gel electrophoresis to confirm the integrity of the expression vectors by diagnostic digestion with EcoRI to yield fragments of 7 and 1.3 kbp. Lanes show empty vector (pCTCON2), size markers (MM), the wild type construct (wt) and the B11 mutant (B11).

176 M. Blazic et al. / Protein Expression and Purification 89 (2013) 175–180

Author's personal copy

based on the extinction coefficient of oxidized ABTS. We definedone unit of GOx activity as the amount of enzyme required to oxi- dize 1 lmol of glucose in 1 min at 25 �C. Protein concentratio n was determined by measuring absorbance at 280 nm.

Polyacrylami de gel electrophoresis and zymography

The molecular mass and homogeneit y of the enzyme prepara- tions were determined by denaturin g sodium dodecylsulfate poly- acrylamide gel electrophor esis (SDS–PAGE) in 8% polyacrylamid egels containing 2-mercapto ethanol. Protein bands were revealed by staining with Coomassie Brilliant Blue R-250 and silver nitrate, and were compare d to molecular weight standards (Thermo Fisher Scientific, MA, USA). Native electrophor esis was carried out under same condition s in 8% polyacrylami de gels lacking SDS and 2- mercaptoethan ol. For zymography, the gels were supplem ented with 0.1 M sodium acetate buffer (pH 5.5) containing 1 U/mL HRP, 9 mM guaiacol and 100 mM glucose.

Characteri zation of enzyme properties

Kinetic constants were determined by measuring enzyme activ- ity in different concentratio ns of glucose (5–100 mM glucose) in 0.1 M sodium acetate buffer (pH 5.5) and fitting Michaeli s–Mentencurves directly using Origin 8.0 software. We determined the pH optima by measuring enzyme activity in 0.1 M sodium phosphate citrate buffer (pH 2–8).

Results and discussion

Cloning and expression of the fusion proteins

The wild type and B11 GOx sequences for mature gene without the native signal sequence were inserted downstream of the Aga2gene in pCTCON2 to create C-terminal fusions by first amplifying the synthetic wild type allele sequence and the B11 allele (housedin vector pYES2) [10] using primers containing BamHI and NheI restrictio n sites. The 1.8-kb PCR products (verified by agarose gel electroph oresis) were digested with BamHI and NheI and inserted into the correspondi ng sites in pCTCON2. The inserts were verifiedby colony PCR and by digesting isolated plasmid DNA with EcoRI, which yielded two diagnost ic bands (Fig. 1).

Competent S. cerevisiae EBY100 cells were transformed with verified constructs and selected on YNB-CAA Gal/Raf plates using an ABTS agar plate assay [11]. Positive clones were picked and transferred to liquid YNB-CAA Glc medium, and cultivated for 2 days before diluting to an OD 600 nm of 0.8 with YNB-CAA Gal medium to induce GOx expression. GOx activity was monitore dto determine the ideal duration of induction to maximize the expression of the wild type and B11 constructs. In both cases, the greatest activity was achieved after 12 h induction, with GOx activ- ity declining if the fermentation was extended any longer (Fig. 2).

Purification of the fusion proteins

After expression, enzyme was extracted from the cell wall, dia- lyzed and loaded onto DEAE Sepharose column. The proteins were eluted using a 10–500 mM of the same buffer (Fig. 3).

Fractions containing GOx activity were concentr ated using 50 kDa cutoff Vivaspin ultrafiltration columns and yield of enzyme

Fig. 2. The activities of the wild type GOx fusion protein (Aga2–wtGOx) and the B11 fusion proteins (Aga2–B11GOx) during fermentation in YNB-CAA Gal liquid medium, showing the peak of activity for both enzymes as 12 h.

Fig. 3. Chromatogram showing the elution of the GOx fusion protein (Aga2–GOx) on 5 mL Sepharose FF columns in a 10–500 mM gradient of sodium phosphate buffer (pH6.0).

M. Blazic et al. / Protein Expression and Purification 89 (2013) 175–180 177

Author's personal copy

purification was determined (see Suppl. Table 1). Purity of the enzyme was determined by native polyacrylamid e gel electropho- resis and zymography. The silver staining of the gels after native electrophor esis revealed very broad protein band on the polyacry l- amide gel that matched the position of the broad activity band de- tected by zymography using glucose, guaiacol and HRP (Fig. 4).This observation and symmetric protein peak during ion exchange chromatograp hy (Fig. 3) that corresponded to the GOx activity were confirming the Aga2–GOx fusion protein was pure. The molecular weight of the fusion proteins was determined by SDS–PAGE, revealing a very broad band between 100 and 140 kDa for both enzymes expresse d in S.cerevisiae and sharp

band of 75 kDa for commercial GOx secreted by A. niger (Fig. 5).These molecular weights are higher than theoretically expected molecula r weights for native GOx of 65 kDa and for AGA2 of 9.5 kDa (74.5 kDa for AGA2–GOx construct), which is a result of glycosyla tion that occurs during secretion both in S. cerevisiae and A. niger . The broad band was also detected by gel filtration(see Suppl. Fig. 2) and it is a well-known fact for secreted proteins in S. cerevisiae that was previously reported for invertase whose molecula r weight of subunit was between 60 kDa and 120 kDa [25] and for GOx whose molecular weight was between 90 and 130 kDa [26]. The broad protein band in all of these cases is a result of microhetero geneity caused by hyperglycos ylation of recombi- nant proteins that occurs in S. cerevisiae during secretion. The molecula r weight of AGA2–GOx we observed is slightly higher to the one previously reported for native GOx expressed extracellu- lary in the same yeast (S. cerevisiae ) [26], probably because of the 9.5 kDa, AGA2 fusion partner.

Kinetic characterizati on of the fusion proteins

We determined the GOx activity of the purified fusion proteins in different glucose concentr ations, allowing the data to be fittedonto a Michaelis–Menten curve to calculate the correspond ing ki- netic paramete rs (Fig. 6). Both fusion proteins had a lower catalytic activity than their corresponding native forms. The kcat of the wild type and B11 fusion proteins were 1.65-fold and 1.30-fold lower than the native enzymes , respectively , and the Km values of the wild type and B11 fusion proteins were 1.52-fold and 1.74-fold higher than the native enzymes, respectively (Table 1).

The lower catalytic activity of the fusion proteins may reflectthe impact of the higher molecular weight on the diffusion coeffi-cient of the enzyme, although because native GOx exists as a di- mer, the ability of Aga2 to form additional disulfide bridges may

Fig. 4. Native 8% polyacrylamide gel electrophoresis of (1) wild type GOx fusion protein Aga2–wtGOx and (2) B11 GOx fusion protein Aga2–B11GOx. (A) Protein bands in the gel after silver staining. (B) Activity bands in the gel after incubation in substrate solution (100 mM glucose, 1 U/mL HRP and 9 mM guaiacol.

Fig. 5. SDS–PAGE of purified fusion protein Aga2–GOx compared to commercial GOx (cGOx) and molecular weight markers (MM).

Fig. 6. Michaelis–Menten curve for wild type GOx fusion protein (Aga2–wtGOx)and B11 fusion protein (Aga2–B11GOx) determined by incubation in different concentrations of glucose. Error bars show standard deviation.

Table 1Kinetic constants of wild type GOx fusion protein (Aga2–wtGOx), the B11 fusion protein (Aga2–B11GOx), native wild type GOx (wtGOx) and native B11 (B11GOx). The data for native enzymes are from our previous research [10].

Aga2–wtGOx Aga2–B11GOx wtGOx B11GOx

kcat (s�1) 33.3 ± 1.6 61.3 ± 2.0 54.8 80.0 Km (mM) 33.4 ± 5.4 27.9 ± 3.1 22.0 16.0 kcat/Km (s�1/mM�1) 0.997 2.20 2.49 5.00

178 M. Blazic et al. / Protein Expression and Purification 89 (2013) 175–180

Author's personal copy

result in the oligomerizati on of the Aga2–GOx proteins, reducing the catalytic activity and increasing the Km value as often observed for enzymes that are immobilized by cross-link ing [27].

Even so, the catalytic activity of Aga2-B11 remained higher than that of the wild type fusion protein, mirroring the comparative behavior of the native enzymes. We therefore compared the kcat/Km ratios of the fusion proteins and their native counterparts , find-ing that the kcat/Km ratio for Aga2-B11 was 2.21 times higher than that of the wild type fusion protein, which was similar to the rela- tive values for the native proteins expresse d in S. cerevisiae , where the kcat/Km ratio for native B11 was found to be twice the value for wild type GOx [10]. These results show that the detection of an im- proved GOx variant by surface display in yeast is likely to reflect agenuine improvement in the propertie s of the correspondi ng na- tive enzyme rather than an artifact of the surface display platform.

In our previous directed evolution experime nts we have sought to increase the activity of GOx at pH 7.4, so that the improved en- zymes can be used for the developmen t of miniature biofuel cells. We therefore determined the pH optima of the fusion proteins in comparison to the native enzymes , and found that the presence of the Aga2 fusion partner had no significant impact on pH sensi- tivity, with the optimum remaining at pH 5.0 for both variants (Fig. 7). This means that yeast surface display is suitable for the di- rected evolution of GOx with the aim of identifying variants with particular pH optima.

GOx–Aga2 fusion proteins are also suitable for immobilization due to the presence of two additional cysteine residues that would enable crosslinking to substrate s such as the gold surfaces often used in biosensors without interfering with substrate access to the enzyme [28]. We found that after lysing yeast cells with tolu- ene and washing the cell wall residues, GOx activity remained bound to the cell wall residues, which would therefore be ideal for further immobilizat ion studies and for gluconic acid produc- tion, as previously reported for invertase [29].

Conclusion s

We expresse d wild type GOx and the more active B11 mutant as C-terminal fusions with the yeast membrane- associated protein Aga2, allowing the enzymes to be characterized by yeast surface display to determine whether this expression is suitable for the developmen t of an ultra-high-thr oughput screening platform for the directed evolution of GOx. The purified fusion proteins were

less active than their native counterparts , but the loss of activity af- fected both variants to a similar degree, and other characteri stics such as the pH optimum remained unchanged. Therefore, despite the lower absolute activities of the enzymes , their relative activi- ties remained unchanged by the fusion protein format, meaning that yeast surface display is suitable for the developmen t of screening platforms for the selection of improved GOx variants. The two additional sulfhydryl groups provided by the Aga2 fusion partner could facilitate the immobilizat ion of GOx by cross-linkin gto substrate s with sulfhydryl groups or directly onto gold surfaces, which are often used in biosensors. The B11 GOx mutant cova- lently bound to the surface of yeast cells could also be used as awhole-cel l catalyst, either in suspension or immobilized to a solid phase, for the efficient production of gluconic acid.

Acknowled gments

This work was supported by Grants on 172049 and on 173017 funded by the Ministry of Education and Science, Republic of Serbia.

Appendi x A. Supplementar y data

Supplement ary data associate d with this article can be found, in the online version, at http://dx.doi.o rg/10.1016/ j.pep.2013.03.014 .

References

[1] B.E.P. Swoboda, V. Massey, Purification and properties of glucose oxidase from Aspergillus niger , J. Biol. Chem. 240 (1965) 2209–2215.

[2] D. Kirstein, W. Kuhn, Glucose-oxidase – properties and application in the food- industry, Lebensmittel Industrie 28 (1981) 205–208.

[3] C.M. Wong, K.H. Wong, X.D. Chen, Glucose oxidase: natural occurrence, function, properties and industrial applications, Appl. Microbiol. Biol. 78 (2008) 927–938.

[4] A. Cavalcanti, B. Shirinzadeh, L.C. Kretly, Medical nanorobotics for diabetes control, Nanomed. Nanotechnol. Biol. Med. 4 (2008) 127–138.

[5] S.V. Dzyadevych, V.N. Arkhypova, A.P. Soldatkin, A.V. El’skaya, C. Martelet, N. Jaffrezic-Renault, Amperometric enzyme biosensors: past, present and future, IRBM 29 (2008) 171–180.

[6] Q. Liu, X.H. Xu, G.L. Ren, W. Wang, Enzymatic biofuel cells, Progr. Chem. 18 (2006) 1530–1537.

[7] D.F. Malherbe, M. du Toit, R.R.C. Otero, P. van Rensburg, I.S. Pretorius, Expression of the Aspergillus niger glucose oxidase gene in Saccharomycescerevisiae and its potential applications in wine production, Appl. Microbiol. Biol. 61 (2003) 502–511.

[8] Y. Guo, F.X. Lu, H.Z. Zhao, Y.C. Tang, Z.X. Lu, Cloning and heterologous expression of glucose oxidase gene from Aspergillus niger Z-25 in Pichiapastoris, Appl. Biochem. Biotechnol. 162 (2010) 498–509.

[9] M. Yamaguchi, Y. Tahara, A. Nakano, T. Taniyama, Secretory and continuous expression of Aspergillus niger glucose oxidase gene in Pichia pastoris , Protein Expr. Purif. 55 (2007) 273–278.

[10] R. Prodanovic, R. Ostafe, A. Scacioc, U. Schwaneberg, Ultrahigh-throughput screening system for directed glucose oxidase evolution in yeast cells, Comb. Chem. High Throughput Screen 14 (2011) 55–60.

[11] Z. Zhu, M. Wang, A. Gautam, J. Nazor, C. Momeu, R. Prodanovic, U. Schwaneberg, Directed evolution of glucose oxidase from Aspergillus niger forferrocenemethanol-mediated electron transfer, Biotechnol. J. 2 (2007) 241–248.

[12] C. O’Fagain, Engineering protein stability, Methods Mol. Biol. 681 (2011) 103–136.

[13] T. Matsuura, T. Yomo, In vitro evolution of proteins, J. Biosci. Bioeng. 101 (2006) 449–456.

[14] N.J. Turner, Directed evolution of enzymes for applied biocatalysis, Trends Biotechnol. 21 (2003) 474–478.

[15] V. Taly, B.T. Kelly, A.D. Griffiths, Droplets as microreactors for high-throughput biology, ChemBioChem 8 (2007) 263–272.

[16] A. Aharoni, G. Amitai, K. Bernath, S. Magdassi, D.S. Tawfik, High-throughput screening of enzyme libraries: thiolactonases evolved by fluorescence-activated sorting of single cells in emulsion compartments, Chem. Biol. 12 (2005) 1281–1289.

[17] K. Bernath, M. Hai, E. Mastrobattista, A.D. Griffiths, S. Magdassi, D.S. Tawfik, In vitro compartmentalization by double emulsions: sorting and gene enrichment by fluorescence activated cell sorting, Anal. Biochem. 325 (2004)151–157.

[18] E. Mastrobattista, V. Taly, E. Chanudet, P. Treacy, B.T. Kelly, A.D. Griffiths, High- throughput screening of enzyme libraries: in vitro evolution of a beta-

Fig. 7. Activity of the wild type GOx fusion protein (WT YSD) and the B11 fusion proteins (B11 YSD) compared to commercial GOx (cGOx) at different pH values, confirming the pH optimum in each case remains �pH 5.0.

M. Blazic et al. / Protein Expression and Purification 89 (2013) 175–180 179

Author's personal copy

galactosidase by fluorescence-activated sorting of double emulsions, Chem. Biol. 12 (2005) 1291–1300.

[19] E.T. Boder, K.D. Wittrup, Yeast surface display for directed evolution of protein expression, affinity, and stability, Methods Enzymol. 328 (2000) 430–444.

[20] G. Chao, W.L. Lau, B.J. Hackel, S.L. Sazinsky, S.M. Lippow, K.D. Wittrup, Isolating and engineering human antibodies using yeast surface display, Nat. Protoc. 1(2006) 755–768.

[21] E. Antipov, A.E. Cho, K.D. Wittrup, A.M. Klibanov, Highly L and D

enantioselective variants of horseradish peroxidase discovered by an ultrahigh-throughput selection method, Proc. Nat. Acad. Sci. USA 105 (2008)17694–17699.

[22] S. Fishilevich, L. Amir, Y. Fridman, A. Aharoni, L. Alfonta, Surface display of redox enzymes in microbial fuel cells, J. Am. Chem. Soc. 131 (2009). 12052–+.

[23] P.A. Romero, F.H. Arnold, Exploring protein fitness landscapes by directed evolution, Nat. Rev. Mol. Cell Biol. 10 (2009) 866–876.

[24] R.D. Gietz, R.H. Schiestl, High-efficiency yeast transformation using the LiAc/SS carrier DNA/PEG method, Nat. Protoc. 2 (2007) 31–34.

[25] F.K. Chu, F. Maley, The effect of glucose on the synthesis and glycosylation of the polypeptide moiety of yeast external invertase, J. Biol. Chem. 255 (1980)6392–6397.

[26] I.C. Momeu, Improving Glucose Oxidase Properties by Directed Evolution, School of Engineering and Science, Jacobs University Bremen, Bremen, 2007. p. 109.

[27] R.A. Sheldon, Enzyme immobilization: the quest for optimum performance, Adv. Synth. Catal. 349 (2007) 1289–1307.

[28] K. Saha, S.S. Agasti, C. Kim, X.N. Li, V.M. Rotello, Gold nanoparticles in chemical and biological sensing, Chem. Rev. 112 (2012) 2739–2779.

[29] Z. Vujcic, Z. Miloradovic, A. Milovanovic, N. Bozic, Cell wall invertase immobilisation within gelatin gel, Food Chem. 126 (2011) 236–240.

180 M. Blazic et al. / Protein Expression and Purification 89 (2013) 175–180