Protein Expression and Purification 93 (2014) 87–92

Contents lists available at ScienceDirect

Protein Expression and Purification

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

Improved extracellular expression and purification of recombinantStaphylococcus aureus protein A

1046-5928/$ - see front matter � 2013 Elsevier Inc. All rights reserved.http://dx.doi.org/10.1016/j.pep.2013.10.013

⇑ Corresponding author. Tel.: +49 89 289 15 750; fax: +49 89 289 15 766.E-mail address: s.berensmeier@tum.de (S. Berensmeier).

1 Abbreviations used: Immunoglobulin G, IgG; Staphylococcus aureus protein A, SpA;pectate lyase subunit B, pelB; isopropyl-D-thiogalactopyranoside, IPTG; 2-amino-2-hydroxymethyl-propane-1,3-diol, Tris; benzene-1,2-diamine, OPD; ethylenediamine-tetraacetic acid, EDTA.

Matthias Freiherr von Roman, Anja Koller, Daniel von Rüden, Sonja Berensmeier ⇑Bioseparation Engineering Group, Technische Universität München, Boltzmannstr. 15, 85748 Garching, Germany

a r t i c l e i n f o

Article history:Received 11 June 2013and in revised form 21 October 2013Available online 30 October 2013

Keywords:Recombinant protein AExtracellular expressionChemical permeabilizationpelB

a b s t r a c t

Protein A from Staphylococcus aureus plays one key role as an immobilized affinity ligand for the purifi-cation of antibodies. A simple method for its extracellular expression in Escherichia coli and subsequentpurification is reported herein. The N-terminus of the gene coding for the five IgG binding domains wasfused to a pelB signal peptide which is responsible for periplasmic localization and which is removedafter translocation into the periplasmic space of E. coli. Different additives, which were added at the sametime with the induction of the protein expression by IPTG, were tested in order to facilitate the release ofthe target protein. With help of this optimized release protocol, more than 380 mg L�1 of protein A wereobtained when Tris–HCl pH 8.5 was added up to a final concentration of 180 mM in shaking flask exper-iments. Based on these observations, a protocol was developed for the extracellular production of SpA in astirred tank bioreactor yielding 5.5 g L�1 of the secreted target protein. After cell removal by centrifuga-tion, the protein A-containing supernatant was concentrated and dialyzed by tangential flow filtration.The target protein was subsequently purified by anion exchange chromatography with a total processyield of 90% and a final purity of P95% (RP HPLC) was achieved.

� 2013 Elsevier Inc. All rights reserved.

Introduction

Due to its high affinity for the Fc domain of several classes ofImmunoglobulin G (IgG)1 and its high stability against cleaningagents such as acids and urea, immobilized Staphylococcus aureusprotein A (SpA) has been extensively used as a platform technologyfor the purification of antibodies. Protein A is a type I membrane pro-tein which consists of three different functional regions. Next to fivehomologous domains (A–E) that are responsible for the interactionwith different IgG subclasses, SpA consists of a secretion signal se-quence (S) and a cell wall linkage region (XM) [1]. To date, onlyfew reports can be found addressing the purification of SpA, whichis in clear contrast to its widespread use in research and pharmaceu-tical production. These reports were either using IgG affinity chro-matography [2] or a sequence of different chromatographictechniques for the purification of SpA which was expressed intothe cytosol of Escherichia coli [3].

Purification of secreted proteins is advantageous as only aminority of E. coli host cell proteins are able to leach into the med-ium. Nevertheless, the successful secretion of heterologous ex-

pressed proteins is restricted to only a few examples and titers ofthe target protein were rather low. In order to improve secretionefficiency, different signal peptides such as pectate lyase subunitB (pelB), outer membrane protein OmpA and many more havebeen developed in order to target proteins to the periplasmicspace. Other reports indicate that some of the IgG binding domainsof SpA in combination with the native signal sequence lead to aperiplasmic localization of proteins which are fused to the N-ter-minus [4,5]. Proteins attached to these bacterial signal peptidesare translocated via different secretion pathways such as the Secpathway and the signal peptide is removed by a signal peptidase[6]. Secretion of periplasmic proteins into the culture supernatantis generally obtained by diffusion across the outer cell membraneof E. coli, which can be increased by leaky strains showing muta-tions in the coding region of outer membrane lipoproteins or thecoexpression of the bacteriocin release proteins (BRP) [7]. Further-more, chemical permeabilization processes using glycine, deter-gents [8] or glycol ethers [9] have been developed recently inorder to help proteins to pass the outer lipopolysaccharide mem-brane of E. coli. Using this approach, even heterologous expressedproteins were secreted from the cytosol of E. coli by adding thesechemicals in combination with a heat shock treatment [10].

A new SpA purification process was developed by taking advan-tage of both methods using a combination of the pelB signal se-quence which is cleaved after successful processing and achemical release approach.

88 M. Freiherr von Roman et al. / Protein Expression and Purification 93 (2014) 87–92

Materials and methods

Chemicals

All chemicals were obtained from Carl Roth GmbH (Karlsruhe,Germany) or AppliChem (Darmstadt, Germany). Restriction andother DNA modifying enzymes were obtained from New EnglandBiolabs GmbH (Frankfurt, Germany). Solvents for HPLC were allHPLC grade (J.T. Baker, Center Valley, USA).

Strains and plasmids

E. coli strains DH5a (Invitrogen, Carlsbad, USA) and BL21 (DE3)(Novagen, Madison, USA) were used for cloning and overexpres-sion experiments. The expression vectors pET20b and pET28a wereobtained from Novagen (Madison, USA).

Construction of the SpA expression vector pSPA

The gene sequence coding for the five IgG binding domains(A–E) of protein A from S. aureus was codon optimized for heterol-ogous expression in E. coli and was obtained from Geneart (LifeTechnologies, Paisley, UK). The pMA Vector containing the geneof interest flanked by NcoI and EcoRI restriction sites was digestedwith the corresponding enzymes and the gene of interest was in-serted into the multiple cloning site of the pET20b and pET28a vec-tors which were linearized with the corresponding restrictionenzymes in advance. The obtained expression vectors (pSPA) wereamplified in DH5a cells and the correctness of the sequence wasverified by analytical restriction digest and DNA sequencing (Euro-fins MWG Operon, Ebersberg, Germany). Isolated plasmids weresubsequently transformed into BL21(DE3) cells for target proteinexpression.

Media and cultivation conditions

For all expression studies, 5 mL of LB medium (10 g L�1 tryp-tone, 5 g L�1 yeast extract, 10 g L�1 NaCl) supplemented with100 lg mL�1 ampicillin was inoculated from a single bacterial col-ony and was incubated for 8 h at 37 �C with shaking at 220 rpm.For the secreting expression of SpA, cells were grown in a syntheticM9 minimal media (12.8 g L�1 Na2HPO4, 3.0 g L�1 KH2PO4, 0.5 g L�1

NaCl, 2.0 g L�1 NH4Cl, 20 g L�1 glucose, 0.1 mM CaCl2, 1.0 mMMgSO4, and 10 lM FeCl3, pH 7.4) [11]. Therefore a 1 L flask contain-ing 200 mL M9 medium was inoculated with 4 mL of the seed cul-ture and cells were grown for 16 h at 37�C up to a final opticaldensity (absorption at 600 nm) OD600 of 3.5 on a rotary shaker.Protein expression and secretion was induced by the addition ofsterile 1 mM IPTG and Tris–HCl was added up to a final concentra-tion of 180 mM. Cultivation temperature was lowered to 30 �C andprotein expression was allowed to proceed for further 24 h.

Fermentation process

The fermentation process for the extracellular expression of SpAwas developed on a BioStat B Plus Twin fermenter (Satorius StedimBiotech, Göttingen, German) controlled by the Biostat MFCS DAprocess recording software. All experiments were performed inglass vessels with a maximum working volume of 1 L equippedwith an Easyferm K8 160 pH probe and a Visiferm DO 160 pO2

probe (Hamilton, Bonaduz, Switzerland). All fermentation experi-ments were performed as triplicates. Modified M9 medium con-taining 12.8 g L�1 Na2HPO4, 3.0 g L�1 KH2PO4, 0.5 g L�1 NaCl,2.0 g L�1 NH4Cl, 20 g L�1 glucose, 0.1 mM CaCl2, 1.0 mM MgSO4,and 10 lM FeCl3 [11] supplemented with 100 lg mL�1 ampicillin

was used as batch medium. Medium were inoculated with seedculture and cells were grown at a constant pH of 6.4 until thedepletion of glucose. Next to this, cell growth was maintained bythe continuous addition of 40% glucose (w/v) for further 12 h. Sub-sequent to the induction of the protein expression with 1 mM IPTGthe maximum concentration of target protein was achieved within24 h.

Separation and purification of SpA

Cells were removed by centrifugation 24 h after induction withIPTG, and the protein containing supernatant (total of 1.1 L) wasrecovered. The solution was 10-fold concentrated and the bufferwas exchanged for binding buffer (20 mM L-histidine pH 6.5,20 mM NaCl) with the help of a Slice 200 (Satorius Stedim Biotech,Göttingen, Germany) ultrafiltration device. For this purpose, two10 kDa cut-off Hydrosart� membranes (total filtration area of0.04 m2) were used. Diafiltration was performed at room tempera-ture with a constant transmembrane pressure of 0.7 bar. SpA wassubsequently purified by ion exchange chromatography on an ÄK-TAexplorer 100 (GE Healthcare life science, Freiburg, Germany).Therefore, the protein containing material was loaded onto aXK16 column (1.6 � 40 cm) packed with 53 mL Unosphere Q (BioRad, München, Germany), which was equilibrated with bindingbuffer in advance. Impurities were removed by washing with astep gradient to 80 mM NaCl and the purified target protein waseluted by increasing the salt concentration to 280 mM NaCl. Frac-tions containing the target protein were pooled and dialyzedagainst PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 2 mMKH2PO4, pH 7.4) by diafiltration in order to increase storagestability.

High performance liquid chromatography

Supernatant concentrations of the target protein were deter-mined by analytical reverse phase chromatography (Agilent 1100Series, Santa Clara, USA). Cells were removed by centrifugationand the supernatant was filtrated with a 0.2 lm syringe filter(Satorius Stedim Biotech, Göttingen, Germany). 10 lL of the ob-tained solution was applied onto a BioBasic C4 column(1 � 150 mm, Thermo Fisher Scientific, Waltham, USA) and the tar-get protein was eluted with a linear acetonitrile gradient (7–90%,0.1% TFA). Protein concentration was determined by peak areacomparison using purified SpA as a standard. Protein concentrationof purified SpA samples was determined by a BCA assay accordingto the manufacturer’s instruction using BSA as reference protein.

ELISA

Binding characteristics of the secreted proteins were evaluatedby a sandwich ELISA technique in 96-well microtiter plates (Brand,Wertheim, Germany) according to Warnes et al. [12]. Wells werecoated with 100 lL hIgG (10 lg mL�1 in PBS, Octapharma, Des-sau-Rosslau, Germany) for 2 h and unbound proteins were subse-quently removed by washing with 300 lL PBST (0.1% Tween 20).Plates were blocked overnight with blocking buffer (PBST contain-ing 5% skim milk powder) at 4 �C. Several dilutions of the targetprotein (0.1–600 ng mL�1) were added to the washed wells andbound SpA was detected by an HRP conjugated anti-SpA IgY (Gal-lus Immunotech, USA). In order to determine the concentrations ofunbound SpA per well, a second ELISA was performed by transfer-ring the supernatant containing unreacted SpA solution onto a sec-ond IgG coated plate and performing the subsequent detectionsteps. Wells were developed with 50 lL of a OPD solution (0.04%(w/v) OPD in 0.2 M NaH2PO4, 0.1 M citric acid pH 4.0), and the col-orimetric reaction was stopped by the addition of 50 lL 1 M sulfu-

M. Freiherr von Roman et al. / Protein Expression and Purification 93 (2014) 87–92 89

ric acid. Protein concentration was determined by data fitting theresponse curve with 4 parameter logistic function in SigmaPlot(Systat Software Inc., San Jose, CA, USA).

Results and discussion

Effect of additives on the secretion of SpA

Due to an increased stability and a facile purification, the tar-geting of heterologous expressed proteins in E. coli to the peri-plasmic and the extracellular space is highly desired [13].Different signal sequences including the natural secretion signalsequence and various domains of SpA have already been studiedfor periplasmic targeting which is needed for an efficient secre-tion. Based on these reports, SpA was targeted into the culturemedium using a N-terminal pelB sequence in order to find a sim-plified process for the purification of SpA. All shaking flask exper-iments were conducted using a defined M9 minimal mediasupplemented with 2% of an appropriate carbon source and dif-ferent trace elements, as the high content of hydrolyzed proteinsin complex media which could not be removed by diafiltrationinterfered with subsequent purification steps. Protein expressionwas induced by the addition of IPTG in the late log phase of cellgrowth as earlier reports proved a maximized expression andsecretion efficiency [14].

With a finally achieved level of secreted SpA of around10 mg L�1, the yield of target protein was rather low (Fig. 1). Thiscorresponds well to the yield of secreted SpA subdomains of15–35 mg L�1, which were recovered from an E. coli culture super-natant by former reports [15]. Nevertheless, secretion to the culti-vation medium often suffers from low efficiency when shakingflasks are used for expression studies [16]. This problem was ad-dressed by the use of permeabilizing additives such as glycine, gly-col ethers and nonionic detergents such as Triton X-100 whichresulted in the release of target proteins from the cytoplasm ofE. coli with high yields. As these chemicals were reported to inter-act with the inner membrane of E. coli, a higher degree of host cellproteins was delivered into the supernatant which could only bereduced by a further heat precipitation step [9,10].

Therefore, a new approach was investigated in order to increasethe release of pelB fused SpA. Different additives which are knownto modulate the permeability of the outer membrane of E. coli were

Fig. 1. Impact of different additives on the secretion of protein A. E. coli BL21DE3 cells cgrowth was achieved. Protein expression was induced by the addition of 1 mM IPTG andadded with the desired concentration. Supernatant concentration of SpA was monitoredpelB signal sequence (-pelB) and by pH adjustment to 7.5.

supplemented to the cells simultaneously with the induction of theprotein expression with 1 mM IPTG [17] during the stationarygrowth phase of the bacteria (here: optical densities about 4). Incontrast to earlier reports, detergents such as Triton were not in-cluded into the study due to possible interference with a subse-quent purification process. The effects of the supplementedadditives glycine, EDTA, and Tris on SpA secretion and cell growthare shown in Figs. 1 and 2.

Compared to previous reports extracellular concentrations ofSpA and cell growth did not remarkably increase after inductionwhen EDTA or glycine was added to the cultures with a final con-centration of 34 mM (1% w/v) or 130 mM (2% w/v) [8]. In contrastto these observations, the addition of a Tris–HCl to a final concen-tration of 180 mM leads as well to a dramatic increase in the re-lease of SpA as to further cell growth. The target protein iscontinuously delivered into the cultivation media and reaches a fi-nal concentration of about 380 mg L�1 24 h after the inductionwith IPTG, which is almost 40 times higher than in cultures whichwere not supplemented with additives (see Fig. 1). Additionally180 mM Tris–HCl leads to a further growth of the bacteria to opti-cal density of about 7.0. This is almost double of the amount of cul-tures which were only supplemented with water as control (seeFig. 2). Increased secretion of the target protein may be explainedby the readjustment of the pH due to the addition of Tris–HCl. Afterit’s addition, the pH of the broth increased from 4.7 to 7.5 in case ofthe addition of 180 mM Tris–HCl and again decreased to 6.5 duringthe phase of protein expression. In contrast cultures supplementedwith a lower final concentration of Tris–HCl released no SpA andpossess a slightly lower pH (below 6.6) before expression whichfurther decreases to 5.5 during the expression of SpA. In order tovalidate the influence of the pH on the SpA secretion and cellgrowth, the pH of the shake flask cultures was adjusted to pH 7.5by the addition 12.5% ammonium hydroxide parallel to the induc-tion of the target protein expression by IPTG. As expected pHadjustment had a positive effect for both SpA secretion and cellgrowth. Despite an even increased optical density of 10, the yieldof target protein is slightly lower than for cultures supplementedwith Tris–HCl. To sum up, the protein release by viable cells duringthe cultivation without any additives yields in higher volumetricproductivities and a subsequent simple purification protocol (seeTable 1). In contrast the addition of EDTA or detergents such asTriton X-100 in combination with denaturants such as guanidine

ontaining the plasmid pSpA were grown for 16 h until the stationary phase of cellthe appropriate additive or increasing concentrations of Tris-HCl were immediatelyby RP-HPLC. Control experiments were conducted using a SpA construct lacking the

Fig. 2. The effect of the addition of different additives on the growth of E. coli BL21DE3 cells after the induction with 1 mM IPTG was monitored by measuring the absorptionat 600 nm. Control experiments were conducted by pH adjustment to 7.5.

Table 1Purification scheme for extracellular secreted SpA. Yields and purity were determinedby RP-HPLC and refer to the starting material after the cultivation process.

Fraction Total amount(mgSpA)

Step yield(%)

Total yield(%)

Purity(%)

Supernatant 464 80Concentrate 432 93 93 92IEX Q 419 97 90 96

Fig. 3. Fermentation process for the extracellular expression of SpA in a pHcntrolled stirred tank bioreactor. E. coli BL21DE3 cells containing the plasmid pSpAwere grown for 22 h in a modified M9 media. Protein expression was induced bythe addition of 1 mM IPTG and was allowed to proceed for further 24 h.

90 M. Freiherr von Roman et al. / Protein Expression and Purification 93 (2014) 87–92

resulted in an arrested cell growth of E. coli [10,18]. Furthermore,previous reports have shown an alteration of the E. coli morphol-ogy and a partial cell lysis upon the addition of high glycine con-centrations [8].

Based on reports that the D domain of SpA itself can be used as asignal peptide for periplasmic localization of proteins in E. coli, thesecretion ability of an SpA construct lacking the pelB leader pep-tide was investigated [5]. Therefore, the DNA sequence encodingfor SpA was inserted into the multiple cloning site of pET28a viaNcoI/EcoRI sites, and the resulting construct was inserted intoE. coli BL21DE3 cells for expression studies. Despite the presenceof 180 mM Tris–HCl in the media, no secretion of the target proteininto the culture supernatant was detectable even using the opti-mized conditions which were detected in advance (Fig. 1). Theseresults indicate the influence of the cellular localization on thesecretion capability of SpA.

Process transfer to a stirred tank bioreactor

Based on these results, a new pH-controlled fermentation pro-tocol for the production of SpA in a stirred tank bioreactor wasdeveloped. This multi-step process is based on a first growth phasewhere cells are grown for 22 h to a final optical cell density of 90 bythe continuous addition of glucose (Fig. 3). Protein expression is in-duced by the addition of IPTG up to a final concentration of 1 mMat a controlled pH of 6.4 resulting in a final yield of 4.5 g L�1 targetprotein after 48 h of total process time. The final yield of SpA couldbe furthermore improved up to a concentration of 5.5 g L�1 2 hafter the addition of 180 mM Tris to the resulting fermentationbroth (data not shown). Again, no SpA was detected in the culturesupernatant despite comparable cell growth behavior when SpAconstructs lacking the pelB sequence were used in the fermenta-tion process. In contrast, the extracellular expression of SpA inPichia pastoris was recently reported by Hao et al. [19]. High prod-uct yields of about 8.8 g L�1 were achieved by growing those cells

to a OD600 above 350 and an extended production phase of over68 h, threefold longer than the here discussed E. coli process.Therefore it can be stated that a more productive extracellularexpression system was developed for the production of SpA.

Purification of the recombinant ligand

The recombinant ligand was captured from the crude fermenta-tion broth by anion exchange chromatography. First of all, cellswere removed by centrifugation and the suspension was concen-trated by crossflow filtration as described in the methods sectionwith a final yield of 93%. During buffer exchange with help of a10 kDa cut-off membrane, low molecular weight contaminantssuch as glucose or pigments were removed and the target proteinwas obtained in the appropriate buffer for the subsequent ion ex-change purification step.

Based on the amino acid sequence of the five IgG binding sub-units of SpA lacking the signal sequence, a pI of 4.7 was calculatedfor the target protein (ExPASy; http://web.expasy.org/). In order toensure an appropriate charge of the target protein which is neededfor a complete binding, a pH of 6.5 was chosen for this chromato-graphic step.

Fig. 4. 12% SDS PAGE analysis of the purification process. Lane 1, supernatant ofcultivation 24 h after induction; lane 2, tangential flow concentrate after bufferexchange; lane 3, flow through after passing the ion exchange column; lane 4,eluted protein.

Fig. 5. Evaluation of the binding activity of the purified IgG binding domains bysandwich ELISA. IgG coated polystyrene plates were incubated with solutions withincreasing SpA concentration. Supernatant SpA concentrations were detected withhelp of a second ELISA using an anti SpA-IgY-HRP conjugate and the amount ofbound SpA per well (q) was determined thereof.

M. Freiherr von Roman et al. / Protein Expression and Purification 93 (2014) 87–92 91

The performance of the overall purification process can be seenin Table 1. SpA with a purity of more than 95% was obtained thanksto the highly specific secretion of the target protein. Using bufferswith lowered pH did not result in an increased purity of the finalproduct whereas salt concentration in the washing buffer seemedto be the critical factor for the yield and purity of the anion ex-change chromatographic step. The use of 80 mM NaCl in the wash-ing buffer was chosen to be the optimal concentration for theremoval of host cell protein contaminants, as product lost is re-duced to 97%. Compared to these results, the use of purificationsteps such as heat precipitation, gel filtration, ion exchange andhydrophobic interaction chromatography which have been usedin former reports concerning the purification of SpA resulted in aproduct loss of over 40% [3]. Therefore, a high overall process yieldof more than 90% was achieved whereas the use of IgG affinitychromatography [2] was avoided. This is especially important asthis type of chromatography resin is both expensive and has alow volumetric binding capacity. In addition, the possibility ofleaching affinity ligand can be excluded, which facilitates the puri-fication process. The high product purity of over 95% was further-more confirmed by SDS–PAGE with coomassie blue staining(Fig. 4).

IgG binding activity

IgG binding activity was evaluated using a sandwich ELISA andcompared to commercially available SpA which was obtained fromRepliGen (Waltham, USA). Despite the approach of extracellularexpressing of the ligand, no detectable alterations in the bindingof hIgG could be observed (see Fig. 5).

Affinity constants were evaluated with help of the Scatchardanalysis. The dissociation constant of purified and commerciallyavailable SpA were therefore determined to be 185 and 225 nM,respectively, which is in accordance with data which has been pub-lished before [20]. Therefore it can be assumed that secreted SpAhas retained its functionality and can be used for furtherapplications.

Conclusion

In this study we describe an efficient protocol for the expressionand extracellular secretion of recombinant SpA in E. coli. With helpof a signal sequence responsible for the periplasmic localization ofSpA, the target protein was secreted into the supernatant with highyields. Based on these results, a new fermentation process for theextracellular production of SpA with a final yield of 4.5 g L�1 wasdeveloped. Furthermore, SpA was captured from the growth mediaby tangential flow filtration and was purified by anion exchangechromatography. Due to the simplified two step purification proto-col, product loss was minimized whereas the IgG binding activityremained unaffected by the purification process.

Acknowledgments

The authors acknowledge the financial support of the TUMGraduate School and the Leonhard Lorenz foundation of the Tech-nische Universität München. Chromatographic material was a kindgift of BioRad.

References

[1] S. Hober, K. Nord, M. Linhult, Protein A chromatography for antibodypurification, J. Chromatogr. B 848 (2007) 40–47.

[2] S. Löfdahl, B. Guss, M. Uhlen, L. Philipson, M. Lindberg, Gene for Staphylococcalprotein A, Proc. Natl. Acad. Sci. U.S.A. 80 (1983) 697–701.

[3] P.M. Hammond, K.A. Philip, R.J. Hinton, G.W. Jack, Recombinant protein A fromEscherichia coli JM83, Ann. N.Y. Acad. Sci. 613 (1990) 863–867.

[4] M. Hansson, S. Stahl, R. Hjorth, M. Uhlen, T. Moks, Single-step recovery of asecreted recombinant protein by expanded bed adsorption, Nat. Biotechnol. 12(1994) 285–288.

[5] T. Heel, M. Paal, R. Schneider, B. Auer, Dissection of an old protein reveals anovel application: domain D of Staphylococcus aureus Protein A (sSpAD) as asecretion – tag, Microb. Cell Fact. 9 (2010) 1–7.

[6] S. Tan, W. Wu, J. Liu, Y. Kong, Y. Pu, R. Yuan, Efficient expression and secretionof recombinant hirudin III in E. coli using the l-asparaginase II signal sequence,Protein Expr. Purif. 25 (2002) 430–436.

[7] S.H. Yoon, S.K. Kim, J.F. Kim, Secretory production of recombinant proteins inEscherichia coli, Recent Pat. Biotechnol. 4 (2010) 23–29.

[8] J. Yang, T. Moyana, S. MacKenzie, Q. Xia, J. Xiang, One hundred seventy-foldincrease in excretion of an Fv fragment-Tumor Necrosis Factor Alpha fusionprotein (sFV/TNF-a) from Escherichia coli caused by the synergistic effects ofglycine and Triton X-100, Appl. Environ. Microbiol. 64 (1998) 2869–2874.

[9] J.R. Allen, A.Y. Patkar, T.C. Frank, F.A. Donate, Y.C. Chiu, J.E. Shields, M.E.Gustafson, Use of glycol ethers for selective release of periplasmic proteinsfrom gram negative bacteria, Biotechnol. Prog. 23 (2007) 1163–1170.

[10] F. Xiang-Yang, Extracellular accumulation of recombinant protein byEscherichia coli in a defined medium, Appl. Microbiol. Biotechnol. 88 (2010)75–86.

[11] J. Sambrook, Molecular Cloning: A Laboratory Manual, vol. 3, Cold SpringHarbor Laboratory Press, 2001.

[12] A. Warnes, A. Walkland, J.R. Stephenson, Development of an enzyme-linkedimmunosorbent assay for Staphylococcal protein A produced in Escherichia coliby pUC8 based plasmids containing the Staphylococcus aureus cowan I proteinA gene, J. Immunol. Methods 93 (1986) 63–70.

[13] F.J.M. Mergulhão, D.K. Summers, G.A. Monteiro, Recombinant protein secretionin Escherichia coli, Biotechnol. Adv. 23 (2005) 177–202.

92 M. Freiherr von Roman et al. / Protein Expression and Purification 93 (2014) 87–92

[14] A. Khushoo, Y. Pal, B.N. Singh, K.J. Mukherjee, Extracellular expression andsingle step purification of recombinant Escherichia coli l-asparaginase II,Protein Expr. Purif. 38 (2004) 29–36.

[15] C. Ljungquist, B. Jansson, T. Moks, M. UhlÉN, Thiol-directed immobilizationof recombinant IgG-binding receptors, Eur. J. Biochem. 186 (1989)557–561.

[16] B. Nilsson, L. Abrahmsén, Fusions to Staphylococcal protein A, in: V.G. David(Ed.), Methods in Enzymology, vol. 185, Academic Press, 1990, pp. 144–161.

[17] R.E.W. Hancock, Alterations in outer membrane permeability, Annu. Rev.Microbiol. 38 (1984) 237–264.

[18] J.G. Voss, Effects of organic cations on the gram-negative cell wall and theirbactericidal activity with ethylenediaminetetra-acetate and surface activeagents, J. Gen. Microbiol. 48 (1967) 391–400.

[19] J. Hao, L. Xu, H. He, X. Du, L. Jia, High-level expression of Staphylococcal ProteinA in Pichia pastoris and purification and characterization of the recombinantprotein, Protein Expr. Purif. 90 (2013) 178–185.

[20] U.K. Ljungberg, B. Jansson, U. Niss, R. Nilsson, B.E.B. Sandberg, B. Nilsson, Theinteraction between different domains of Staphylococcal protein A and humanpolyclonal IgG, IgA, IgM and F(ab’)2: separation of affinity from specificity,Mol. Immunol. 30 (1993) 1279–1285.