Protein Expression and Purification 95 (2014) 195–203

Contents lists available at ScienceDirect

Protein Expression and Purification journal homepage: www.elsevier.com/locate/yprep

Facile purification of Escherichia coli expressed tag-free recombinant human tumor necrosis factor alpha from supernatant Chun Zhang a,b, Yongdong Liu b,⇑, Dawei Zhao b, Xiunan Li b, Rong Yu a,⇑, Zhiguo Su b a b

Key Laboratory of Drug Targeting and Drug Delivery System, Ministry of Education, West China School of Pharmacy, Sichuan University, Chengdu 610041, China State Key Laboratory of Biochemical Engineering, Institute of Process Engineering, Chinese Academy of Sciences, Beijing 100190, China

a r t i c l e

i n f o

Article history: Received 17 November 2013 and in revised form 24 December 2013 Available online 8 January 2014 Keywords: Tag-free rhTNF-a Purification E. coli expression Supernatant Two-step ion exchange chromatography

a b s t r a c t Fusing affinity tag at N-terminus was reported to decrease the biological activity of the recombinant human tumor necrosis factor alpha. Although preparation of tag-free rhTNF-a has already been achieved, the processes were yet laborious, especially in large scale. In this paper, tag-free rhTNF-a was almost equally synthesized by Escherichia coli in both soluble and insoluble forms. A two-step ion exchange chromatography, DEAE-Sepharose combined with CM-Sepharose, was developed to purify the soluble specie from supernatant after cell lysis. Native PAGE and HP-SEC showed the rhTNF-a extracted from supernatant existed in a homogeneous form. HP-SAX and SDS–PAGE analysis demonstrated the purity of the final fraction was over 98% with a very high recovery of 75%. Circular dichroism spectrum demonstrated that b-sheet structure was dominant and fluorescence analysis suggested no dramatic exposure of aromatic amino acid residues on the protein surface. Bioassay indicated that purified rhTNF-a was biologically active with a specific activity of approximately 2.0  107 U/mg. All these results suggested that this two-step ion exchange chromatography is efficient for preparation of biologically active tag-free rhTNF-a from supernatant. Ó 2014 Elsevier Inc. All rights reserved.

Introduction In virtue of its high expression level and no need of expensive culture medium, Escherichia coli (E. coli) expression system is widely applied to producing a variety of recombinant proteins, including cytokines, hormones and enzymes [1]. Moreover, chemical modification, like pegylation, as an alternative strategy for substituting the glycan of glycosylated protein produced by prokaryotic expression system has aroused wide concerns [2], which may further expand the application of E. coli expression system. Unlike eukaryotic and mammalian system, recombinant expressed by E. coli could be soluble or insoluble. Although the insoluble form, named inclusion bodies (IBs) could be easily separated from most formidable impurities of bacteria simply by washing and centrifugation, these aggregated proteins are always misfolded and thus biologically inactive. Solubilization and refolding of IBs are cumbersome, low yield and time-consuming, especially for largescale production [3,4]. Fusing affinity tag to protein of interest, exemplified by 6 his-tag and GST-tag, is convenient for subsequent purification. However, the tag would probably cause a series of negative problems, including loss of biological activity [5,6], disturbance of protein conformation [7,8], alteration of innate

⇑ Corresponding authors. Tel./fax: +86 010 82545028 (Y. Liu). E-mail addresses: [email protected] (Y. Liu), [email protected] (R. Yu). 1046-5928/$ - see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.pep.2013.12.012

immunogenicity [9], augmentation of original molecular size [9] and perturbation of subunit assembly [10]. Although those affinity tags could be removed from the fusion proteins by some special enzymes and chemical reagents, the process is always troublesome and costly [11]. From this point of view, establishment of a simple and efficient strategy for purification of tag-free recombinant protein is economical and practical. Tumor necrosis factor alpha (TNF-a), known for its potent cytotoxicity to solid tumor cell whereas no effect to normal counterparts, is a pleiotropic cytokine [12,13]. Naturally produced human TNF-a exists in both transmembrane and soluble forms. TNF is initially generated in transmembrane form (26 kDa) and further converted to soluble form by proteolytic action [14,15]. Human soluble TNF-a is derived from the topologic domain of extracellular with 157 AA and approximately 17.35 kDa. Biologically active TNF-a always exists as a compact homotrimer [16]. It is considered that TNF-a is one of the most promising candidates for treating many kinds of tumors, especially in recent advances in targeting design of protein-based drugs [17]. Additionally, new advances in understanding of TNF-a during the progress of chronic inflammatory and autoimmunity provided an alternative insight for application of TNF-a [18–21]. Benefited from the advances of biotechnology, large-scale production of TNF-a becomes realizable. To achieve high purity of TNF-a with efficient downstream processing, fusing affinity tag

196

C. Zhang et al. / Protein Expression and Purification 95 (2014) 195–203

was frequently adopted in the upstream construction. However, it has been reported that fusing affinity ligands, like histidine clusters, at the N-terminus of TNF-a could lead to an increased steric hindrance which probably disturbs the interaction with its receptors and further results in a dramatic reduction of biological activity [6]. Analogously, an increased steric hindrance at the N-terminus of TNF-a arisen from covalently conjugated polyethylene glycol (PEG) led to a dramatic loss of bioactivity, which further validated that the N-terminal domain of TNF-a plays a certain role in mediation of physiological effects [22]. Although elimination of fusion tag by enzymatic action could be achieved and the biological activity could be regained, the low efficacy of cleavage and the unfavorable cost may restrict its further application in practice [6]. A well-known strategy for production of TNF-a is through expressing in the form of IBs, and the route from the preparation and refolding of inclusion bodies to purification of TNF-a is available [23]. However, refolding by dilution was restricted by low protein concentration and huge volume of the refolding solution, especially for large scale production. Furthermore, TNF-a is not allowed to be exposed to solution at room temperature for its poor stability [24]. Although preparation of TNF-a from supernatant of cell disruption in laboratory scale was reported in several literatures, the bureaucratic process for purification and the unfavorable recovery limited its further application when compared to the route of inclusion bodies [25–28]. Thus, a protocol for extracting tag-free TNF-a from supernatant with high recovery, lower cost and easy-manipulating is economically necessary. In this work, BL21 E. coli strain harboring recombinant human TNF-a (rhTNF-a)1 plasmid was induced and the target protein without any fusion tags was highly expressed in forms of both soluble and inclusion bodies, almost half to half in ratio. A two-step ion exchange chromatography, DEAE-Sepharose combination with CMSepharose was developed for the purification of tag-free rhTNF-a in gram-scale from supernatant, and the feasibility and efficacy were further evaluated in detail. Materials and methods Materials E. coli BL21 (DE3) strain with pET21a-rhTNF-a plasmid was constructed in our laboratory and stored at 70 °C with 20% glycerol before use. Tryptone and yeast extract were purchased from Oxoid Ltd. (England); Sodium dodecyl sulfate (SDS, Sigma); Phenylmethylsulfonyl fluoride (PMSF, Sigma) and b-mercaptoethanol (Lawn, NJ, USA). Actinomycin D (Sigma), Cell counting kit (CCK-8, Beyotime China); RPMI1640 culture (Hyclone), L-Glutamine (Beyotime, China) Fetal bovine serum (FBS, YHJM, China), Trypsin (Sigma); IPTG (Sigma), ampicillin (Sigma), Penicillin–Streptomycin Solution (Solarbio, China), other reagents used in experiment were of analytical grade and prepared in-house. Fermentation E. coli BL21 (DE3) strain harboring plasmid with a recombinant human soluble TNF-a gene was initially activated in Luria broth (LB) medium (0.5% yeast extract, 1% tryptone and 1% sodium 1 Abbreviations used: rhTNF-a, recombinant human tumor necrosis factor alpha; HPLC, high performance liquid chromatography; HP-SEC, high performance size exclusion chromatography; HP-SAX, high performance strong anion exchange; SDS– PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis; CD, circular dichroism; IPTG, isopropyl b-D-thiogalactoside; FBS, fetal bovine serum; IBs, inclusion bodies; Q-Sepharose, quaternary amine-Sepharose; DEAE-Sepharose, diethylaminoethyl-Sepharose; CM-Sepharose, carboxymethyl-Sepharose; SP-Sepharose, sulfopropyl-Sepharose; FF, fast flow.

chloride) supplemented with 50 lg/ml ampicillin overnight at 30 °C. 5.0 ml activated strain was transferred into two 2.0-L conical flasks with 500 ml (1:100) LB medium supplemented by 100 lg/ml ampicillin and these flasks were vigorously shaken in a speed of 220 rpm at 37 °C for until the OD600 reached to 2.0. The resulted 1000 ml culture was switched to a 20-L fermentor (BioFlo 415, New Brunswick Scientific, Germany) with 13.5 L 2 LB medium at a constant temperature of 37 °C and the dissolved oxygen was kept to approximately 10% and the pH value at the range of 6.95 ± 0.05 by setting agitation, gas-flow and other parameters. Once the OD600 reached to 8.0, IPTG was injected into fermentor with a final concentration of 1.0 mM. After induction by IPTG for 3.5 h, 200 ml 10 LB medium supplemented with 20% glycerol was added at once and 800 ml was added into the fermentor at a speed of 20 ml/min. The fermentation was terminated after induction for 4.5 h, and the broth was centrifuged at 4000 rpm for 30 min under 4 °C (Thermo, USA). The expression level of rhTNFa was analyzed by 12% SDS–PAGE and the cell pellets were collected and stored at 20 °C before use. Purification of rhTNF-a from supernatant The harvested cell pellets (about 300 g, wet weight) were resuspended in 2000 ml 1.0 mM EDTA, 20 mM Tris–HCl (pH 8.0) and disintegrated by high-pressure homogenizer (APV-2000, SPX Flow Technology, Germany) with a pressure of 950–1000 bar for three cycles. 1.0 mM PMSF was added to inhibit enzymatic degradation. The suspension was further centrifuged by high-speed centrifuge (Hitachi, Japan) at 12,000 rpm for 30 min. The supernatant and precipitate were analyzed by 12% SDS–PAGE, the precipitate was washed and stored at 20 °C, and the supernatant was accumulated and kept 4 °C before subsequent purification. Two DEAE-Sepharose Fast Flow (GE Healthcare, USA) 50  300 mm columns with 400 ml of column volume (CV) were equilibrated by buffer A (20 mM Tris–HCl, pH 8.0) with 5 CV, about 4000 ml supernatant (double dilution, total protein concentration approximately 7.5 mg/ml) was then separately loaded onto the two DEAE-Sepharose columns at a speed of 20 ml/min (ÄKTA purifier, GE Healthcare). The columns were washed with buffer A until the baseline became smooth and eluted with a 0–30% linear gradient of buffer B (buffer A + 1.0 M NaCl, pH 8.0) with 3.5 CV. Fractions were collected respectively and further analyzed by 12% SDS– PAGE. The fractions containing the intended protein were pooled and then diluted with buffer C (20 mM Na2HPO4/KH2PO4, pH 6.0) until the solution conductivity was less than 5.0 mS/cm and the pH was carefully adjusted to 6.0 with 10 mM HCl solution. The processed fraction was loaded onto a CM-Sepharose Fast Flow (GE Healthcare, USA) 50  300 mm column (CV, 400 ml) which had been equilibrated with 5 CV of buffer C at a speed of 20 ml/min. The column was then washed with buffer C until the baseline tended to be stable around zero, and the intended protein was eluted with a 0–40% gradient of buffer D (buffer C + 1.0 M NaCl) with 4 CV. The unique peak was analyzed by 10% native PAGE and 12% SDS–PAGE, and the protein solution was dialyzed against 50 mM Tris–HCl pH 8.0 for 6 h. The final protein solution was dispensed and stored at 20 °C. SDS–PAGE and native-PAGE analysis of rhTNF-a 12% sodium dodecyl sulfate (SDS) polyacrylamide gels were prepared as described by Laemmli [29]. The 10% native polyacrylamide gels were also prepared according to that except adding surfactant and reductant. Both PAGEs were initially performed at 80 V for 30 min and further carried out at 120 V until the bromophenol blue reached the bottom of gels. Gel-separated proteins were finally stained with Coomassie brilliant blue R250.

C. Zhang et al. / Protein Expression and Purification 95 (2014) 195–203

HPLC analysis of rhTNF-a Analysis of rhTNF-a was carried out with a strong anion exchange column, 6.0  150 mm TSK-GEL SAX (TOSOH, Japan) using HPLC (Agilent 1100, USA). The mobile phases were 50 mM Tris– HCl, pH 8.0 (solution A) and 50 mM Tris–HCl with 1.0 M sodium chloride, pH 8.0 (solution B). All samples were initially desalted (HiTrap Desalting 5.0 ml, GE Healthcare) and then appropriate amount of each sample was drawn in by auto-sampler at a constant flow rate of 0.5 ml/min. Upon sample injection, a wash with solution A for 5 min was performed, followed by a linear gradient of solution B, of which the concentration was raised from 0% to 50% in 40 min and then changed to 100% at once. After elution with 100% solution B for 10 min, the column was equilibrated with 100% solution A for another 10 min. The outlet elution was detected with a diode array detector. The signals were recorded at 280 nm and the reference wavelength was set at 360 nm. N-terminus amino acid sequencing N-terminus amino acid sequencing was accomplished by State Key Laboratory of Protein and Plant Gene Research (Peking University, Beijing). Circular dichroism (CD) spectroscopy To investigate the secondary structure of rhTNF-a obtained from supernatant by the two-step ion exchange chromatography, far-UV CD analysis of rhTNF-a was performed on J-810 spectrometer (Jasco, Japan) at room temperature using a 1.0 mm path length quartz cuvette. The concentration of the sample was 0.3 mg/ml in 50 mM Na2HPO4/KH2PO4 pH 8.0. The signal was recorded from 260 nm to 195 nm at a scanning rate of 1000 nm/min with a bandwidth of 1.0 nm, and each spectrum represented the average of 10 scans with a response of 0.1 s. The final signals were adjusted by subtracting the buffer signal. Fluorescence spectroscopy To further validate the folding, the intrinsic emission fluorescence spectra of rhTNF-a in 50 mM Na2HPO4/KH2PO4, pH 8.0 was measured by F-4500 fluorescence spectrophotometer (Hitachi, Japan). The protein solution was diluted to a final concentration of 0.1 mg/ml. The emission spectra were recorded from 290 nm to 450 nm using 1.0 cm path length cuvette with a constant excitation wavelength of 280 nm. Characterization of rhTNF-a by high performance size exclusion chromatography (HP-SEC) Superdex 75 (GL 10/300 mm, GE Healthcare, USA) was utilized to analyze the quaternary structure of rhTNF-a on ÄKTA purifier with elution buffer of 50 mM Na2HPO4/KH2PO4 0.15 M Na2SO4, pH 7.0. Samples were loaded onto the column by 500 ll loop and the elution was detected at 280 nm with a constant flow rate of 0.6 ml/min.

197

(Invitrogen, USA). 100 ll cell solution was seeded in each well of a 9  12 plate and incubated at 37 °C with 5% CO2 for 6 h (SANYO, Japan). 20 ll rhTNF-a solution and rhTNF-a standard protein (expressed in yeast, Sigma) with appropriate dilution by RPMI-1640 culture containing 6 lg/ml of actinomycin D was added to corresponding well with rhTNF-a concentration ranged from 0.1 pg/ml to 1000 ng/ml. Each sample had three parallel wells. After incubation at 37 °C with 5% CO2 for 20 h, 12 ll CCK-8 work solution was added to each well. The plate was further incubated at 37 °C for another 2 h and the final OD450 was measured by microplate reader (Bio-Rad, USA) with a reference of OD630. One unit of the TNF-a activity was defined as the amount giving a survival ratio of 50% L929 cells in the presence of 1.0 lg/ml actinomycin D [31]. Optimizing the two-step purification process To obtain an optimal process, we investigated the purification efficacy of several analogous separation media used in both steps. During the first step, we compared two types of anion exchange chromatography, including DEAE-Sepharsoe FF (HiTrap 1.0 ml) and Q-Sepharose FF (HiTrap 1.0 ml). The same amounts of supernatant (4.0 ml, 5.0 mg/ml of total protein) were loaded separately onto both columns at a speed of 0.5 ml/min and the absorbed protein was finally eluted by 0~0.4 M sodium chloride with 10 CV. The intended fraction was collected and the total protein was measured by Coomassie Blue G250. The purity was determined by HPLC and the recovery was calculated according to the following formula: recovery ¼ 100%  ðcollected proteinÞ=ðtotal loaded proteinÞ. In the second step, once the DEAE-Sepharose FF elution was sufficiently dialyzed, the same amounts of total protein (5.0 ml, 1.0 mg/ml) were separately loaded on CM-Sepahrsoe FF (HiTrap 1.0 ml) and SP-Sepharose FF (HiTrap 1.0 ml) column and the absorbed protein was washed out by a gradient concentration of 0–0.4 M NaCl with 4 CV. The single peak was collected and the total protein was measured by Coomassie Blue G250. Additionally, the same amount of total protein solution (5.0 mg) was initially adjusted by saturated ammonium sulfate solution to a finally concentration of 1.0 M, further loaded on phenyl-Sepharsoe FF (HiTrap 1.0 ml) and finally washed out with 4 CV. The purity and recovery of intended protein were calculated as mentioned above. Protein concentration assay Protein concentration in each experiment was measured with Coomassie Blue G250 according to Bradford method and bovine serum albumin (BSA) was used as standard [32]. Statistical analysis The data were analyzed by GraphPad Prism 3.0 software. Results were expressed as means ± (SD) standard deviations and the differences were considered significant at p < 0.05. Result Fermentation and purification of rhTNF-a

Cytotoxicity assay The bioactivity of the purified rhTNF-a in our study was evaluated by lysis assay of L929 cell with modification as described by Aggarwal et al. [30]. Briefly, mouse fibroblast L929 cell was cultured in RPMI-1640 supplemented with 10% FBS at 37 °C with 5% CO2. The exponentially growing cells were digested by 0.2% trypsin solution with 1.0 mM EDTA for 1.0 min and the suspended cells were adjusted to 2  105 cells/ml measured by cell counting

Approximately 300 g cells (wet weight) were obtained after fermentation. SDS–PAGE analysis showed that the expression level of rhTNF-a was over 30% of total protein estimated by gel analysis software (BandScan V5.0). After lysis of cells, the intended protein was found almost half in supernatant and half in inclusion bodies. About 3.6 g rhTNF-a was obtained from the 300 g cells (Table 1). In the DEAE-Sepharose FF purification process, rhTNF-a was absorbed on the resins and much of the impurity passed through by carefully

198

C. Zhang et al. / Protein Expression and Purification 95 (2014) 195–203

Table 1 Purification of rhTNF-a from about 300 g E. coli cell pellets by the two-step ion exchange chromatography strategy. Process

Total protein (g)

Target protein (g)

Supernatant

Purity (%)

Recovery (rhTNF-a,%)

Condition (Buffer) 20 mM Tris/HCl, pH8.0

30.0

4.5

15.0

100.0

DEAE-Sepharose

5.1

3.8

74.5

84.5

A: 20 mM Tris/HCl pH8.0 B: buffer A + 1.0 M NaCl

CM-Sepharose

3.6

3.6

>98.0

80.0

A: 20 mM Na2HPO4/KH2PO4 pH6.0 B: buffer A + 1.0 M NaCl

controlling the buffer pH value. Then the absorbed proteins were eluted by about 0.15 M sodium chloride with a recovery of approximately 85%. The purity of rhTNF-a in the eluted peak was about 74.5% estimated by gel density and HLPC (Fig. 1). In the CM-Sepharose FF elution profile, only one symmetrical peak was eluted. Intended protein was eluted by approximately 0.3 M sodium chloride. 12% SDS–PAGE showed that the purity was over 98% estimated by densitometric scanning (Gel imaging analysis system, Bio-Rad). In addition, 10% native PAGE indicated that rhTNF-a existed in a homogeneous form (Fig. 2). HPLC analysis of rhTNF-a The purity of intended protein obtained by the two-step ion exchange chromatography was estimated by HPLC. A single major peak was eluted by approximately 15% solution B at about 15.6 min. The purity of intended protein calculated by the peak area was approximately 98% (Fig. 3). Identification of the N-terminus amino acid of rhTNF-a N-terminal amino acid sequencing analysis of the rhTNF-a demonstrated that the last five amino acids were Val, Arg, Ser, Ser, Ser, indicating that the methionine was cut off from the recombinant protein during expression.

Secondary structure analysis For secondary structure analysis of the purified rhTNF-a, far-UV CD spectrum from 260 nm to 195 nm was obtained. It showed a negative peak at 219 nm and a positive peak at 198 nm (Fig. 4). K2d analysis indicated that the rhTNF-a derived from supernatant mainly existed in b-sheet form and little in a-helix form [28,33]. Fluorescence spectroscopy Fluorescence spectra analysis of rhTNF-a in 50 mM Na2HPO4/ KH2PO4 pH 8.0 buffer was shown in Fig. 5. The maximum emission wavelength of rhTNF-a was approximately 341 nm, a dramatic blue-shift of 7 nm from that of free tryptophan (348 nm), indicating that the aromatic amino acids were mostly buried inside. Characterization of rhTNF-a by HP-SEC In order to determine the quaternary structure of rhTNF-a from supernatant, the purified protein was then subjected to Superdex 75 column and a major peak was eluted at about 10.6 ml. According to the elution profile of protein marker (GE Healthcare, USA), protein eluted at this position should have a molecular weight of approximately 42 kDa (Fig. 6), which was less than the theoretical molecular weight of trimetric TNF-a (52 kDa).

Fig. 1. Chromatographic profile for purifying rhTNF-a from supernatant by DEAE-Sepharose. The insert is the SDS–PAGE (12%) analysis of each fraction from DEAE-Sepharose process. Lane 1, the supernatant; lane 2, pellet of inclusion bodies corresponding to the amount cell of lane 1; lane 3, peak 0 (flow-through); lane 4, peak 1 (the shaded part in elution profile).

C. Zhang et al. / Protein Expression and Purification 95 (2014) 195–203

199

Fig. 2. Elution profile of purification of rhTNF-a by CM-Sepharose and 12% SDS–PAGE (A) & 10% native-PAGE (B) analysis, lane 1, sample from peak 0 in the elution profile (flow-through); lane 2, sample from peak 1 in the elution profile; lane 3, bovine serum albumin; lane 4, sample from peak upon desalting.

Fig. 3. Characterization of rhTNF-a by HPLC. The samples from CM-Sepharose FF elution were firstly desalted and then loaded on a 6.0  150 mm TSK-GEL SAX column. Elution gradient: 0~5 min, 100% solution A; 5–45 min, gradually increased solution B from 0% to 50%.

L929 cells Cytotoxicity L929 cell assay was carried out to analyze the bioactivity of purified rhTNF-a. The dose–response curve ranging from 0.1 pg/ ml to 1000 ng/ml showed an EC50-value of 47.1 ± 5.5 pg/ml, indicating that the purified rhTNF-a in our experiment was biologically active with activity of approximately 2.0  107 U/mg (Fig. 7).

Optimizing the two-step purification process We designed a strategy arming to achieve a high purity of rhTNF-a with an optimal recovery. The main progress achieved in our optimization was summarized in Table 2. In the first round of optimization, strong cationic resin Q-Sepharose FF showed a similar recovery of rhTNF-a when compared with DEAE-Sepharose

FF, whereas the purity was much lower than the later (7.1%). More unrelated proteins were absorbed onto the SP-Sepharose resin. In the second round of optimization, both CM-Sepharose FF and SPSepharose FF resins showed high recovery of rhTNF-a (both above 90%), and the purity of both was as high as approximately 95%, especially for the CM-Sepharose (about 98%). Additionally, rhTNF-a showed strong interaction with phenyl-Sepharose resin in the presence of 1.0 M ammonium sulfate, whereas the selectivity and recovery to rhTNF-a were relatively low compared to both cationic resins.

Discussion Ion-exchange chromatography is widely used to roughly purify recombinant and natural proteins for its low cost and

200

C. Zhang et al. / Protein Expression and Purification 95 (2014) 195–203

Fig. 4. Far-UV CD analysis of rhTNF-a. The spectrum was recorded at room temperature in a 1.0 mm path length quartz cuvette with a concentration of 0.3 mg/ml.

Fig. 5. Intrinsic fluorescence analysis of rhTNF-a. The emission spectrum was recorded from 290 nm to 450 nm with a constant excitation wavelength of 280 nm at room temperature. The path length of sample cuvette is 1.0 cm and sample concentration is about 0.1 mg/ml.

easy-manipulating. Most of the prokaryotic-based proteins are produced by intracellular expression for which allows to achieve a high yield of intended protein [34]. This always results in abundant impurities which arise from bacteria itself and thus lead to laborious separation and purification work. Conventional strategies available for preparation of tag-free TNF-a from supernatant always involve an unmanageable procedure of ammonium sulfate precipitation which mainly responsible for the low recovery of TNF-a [28,35]. PEG was reported to help improve the recovery of

TNF-a [36], whereas it had to add further measures for getting rid of PEG molecules. DEAE-based separation media was commonly adopted to extract tag-free TNF-a from crude protein solution because it shows high efficient for eliminating most impurities with high recovery of intended protein. This has been validated by several literatures. However, the strategies of the subsequent purification for high purity of TNF-a were divergent, including a timeconsuming procedure of size-exclusion chromatography, Blue Dye affinity chromatography and Mono Q column etc. [28,37]. We

C. Zhang et al. / Protein Expression and Purification 95 (2014) 195–203

201

Fig. 6. Superdex 75 analysis of the rhTNF-a. Short dash line represented the protein marker, conalbumin, 75 kDa; ovalbumin, 43 kDa; carbonic anhydrase, 29 kDa; ribonuclease A, 13.7 kDa.

Fig. 7. L929 cells cytotoxicity of rhTNF-a. For the rhTNF-a prepared by the two-step ion exchange chromatography, an EC50-value of 47.1 ± 5.5 pg/ml was determined (standard rhTNF-a derived from yeast expression).

developed a strategy of cation exchange chromatography at pH 6.0 for obtaining high purity of tag-free rhTNF-a. It was reported that the pH value could irreversibly influence the biological activity of TNF-a by undermining its structure. CD spectrum was used to investigate the effects of pH values to the secondary structure of rhTNF-a and demonstrated that the secondary structure was perturbed only when pretreated below pH 5.0 (data not shown). Given that the theoretical isoelectric point of rhTNF-a is approximately 7.0 roughly estimated by ExPASy on-line tool (soluble form, 77– 233 AA) and that rhTNF-a tends to precipitate when the pH value

is close to 6.5, we chose pH 6.0 as optimum for cation exchange chromatography. CD was also used to investigate the secondary structure of rhTNF-a. The feature that TNF-a consists of high content of b-sheets but little a-helix [38], was also observed in rhTNF-a derived from E. coli supernatant. SEC analysis showed a smaller molecular weight than the theoretical value of trimeric TNF-a, and it was inclined to be a dimeric or trimeric form. Trimeric form of TNF-a was reported to be a requirement for its biological activity [16,39]. It was indicated that the rhTNF-a existed as a

202

C. Zhang et al. / Protein Expression and Purification 95 (2014) 195–203

Table 2 Summary for the optimization of the two-step purification process. Process

Media

Total protein (mg)

Supernatant

/

20.0

First step

DEAESepharose Q-Sepharose

3.1 3.5

Second step

CM-Sepharose SP-Sepharose PhenylSepharose

3.8 4.0 4.0

Target protein (mg)

Purity (%)

Recovery (rhTNFa,%)

3.0

15.0

100.0

20 mM Tris/HCl, pH8.0

/

2.5

81.5

83.6

A: 20 mM Tris/HCl pH8.0

AEX

2.6

74.4

85.7

B: buffer A + 1.0 M NaCl

3.7 3.8 3.6

97.9 94.3 88.7

91.6 92.4 87.8

A: 20 mM Na2HPO4/KH2PO4 pH6.0 B: buffer A + 1.0 M NaCl A: 20 mM Na2HPO4/KH2PO4 pH6.0, + 1.0 M ammonium sulfate B: 20 mM Na2HPO4/KH2PO4 pH6.0

Condition (buffer)

Note

CEX HIC

Note: 5.0 mg crude proteins from DEAE-Sepharose elution were used in the second step analysis. AEX, anion exchange chromatography; CEX, cation exchange chromatography; HIC, hydrophobic interaction chromatography.

trimeric form according to the L929 cell cytotoxicity result. Although Superdex 75 column was reported to show moderate hydrophobicity, we did not think the retarded elution of trimeric rhTNF-a was as a result of hydrophobic interaction with the matrix of media, because rhTNF-a was quite soluble at pH 7.0 solution. Besides, 0.15 M Na2SO4 had been used to avoid nonspecific absorption. Ordinarily, association of subunits into an oligomer (dimer, trimer) would increase the overall asymmetric structure of a protein. However, this time, we believed that the compact form and spherical structure of trimeric TNF-a could contribute to the smaller apparent molecular weight measured by gel filtration analysis. Similar scenario was also encountered in several literatures when concerned the quaternary structure of trimeric TNF-a [39,40]. Native-PAGE showed the rhTNF-a almost existed in a homogeneous form, indicating that the trimeric rhTNF-a was hydrodynamically stable. Compared with weak exchanger resins, strong exchanger counterparts showed more intense interaction with both rhTNFa and unrelated impurities, which led to low purity of product during our experiment. Thus, to achieve a higher yield and purity of rhTNF-a, it was highly recommended that weak cation exchanger resin should be adopted in the second round of purification. Additionally, phenyl-Sepharose would provide an alternative step for purification of rhTNF-a though moderate precipitation induced by ammonium sulfate probably resulted in a relatively low recovery. In summary, the two-step ion exchange chromatography, DEAE-Sepharose combined with CM-Sepharose, was demonstrated to be a simple and efficient strategy for large-scale production of tag-free rhTNF-a from supernatant with high recovery of over 75% and the whole process would be finished within 24 h. The final purity of rhTNF-a prepared according to this strategy is over 98%. Acknowledgments This research was financially supported by the National Natural Science Foundation of China (Grant No. 20976178, 20820102036) and Doctoral Fund of the Ministry of Education (Grant Nos. 20120181110036). References [1] M. Kamionka, Engineering of therapeutic proteins production in Escherichia coli, Curr. Pharm. Biotechnol. 12 (2011) 268–274. [2] Y.J. Wang, S.J. Hao, Y.D. Liu, T. Hu, G.F. Zhang, X. Zhang, Q.S. Qi, G.H. Ma, Z.G. Su, PEGylation markedly enhances the in vivo potency of recombinant human non-glycosylated erythropoietin: a comparison with glycosylated erythropoietin, J. Control. Release 145 (2010) 306–313. [3] J. Chen, Y. Liu, X. Li, Y. Wang, H. Ding, G. Ma, Z. Su, Cooperative effects of urea and L-arginine on protein refolding, Protein Expr. Purif. 66 (2009) 82–90. [4] L. Gao, C. Zhang, L. Li, L. Liang, X. Deng, W. Wu, Z. Su, R. Yu, Construction, expression and refolding of a bifunctional fusion protein consisting of Cterminal 12-residue of hirudin-PA and reteplase, Protein J. 31 (2012) 328–336.

[5] Y. Nomine, T. Ristriani, C. Laurent, J.F. Lefevre, E. Weiss, G. Trave, A strategy for optimizing the monodispersity of fusion proteins: application to purification of recombinant HPV E6 oncoprotein, Protein Eng. 14 (2001) 297–305. [6] I. Fonda, M. Kenig, V. Gaberc-Porekar, P. Pristovaek, V. Menart, Attachment of histidine tags to recombinant tumor necrosis factor-alpha drastically changes its properties, ScientificWorldJournal 2 (2002) 1312–1325. [7] M.H. Bucher, A.G. Evdokimov, D.S. Waugh, Differential effects of short affinity tags on the crystallization of Pyrococcus furiosus maltodextrin-binding protein, Acta Crystallogr. D Biol. Crystallogr. 58 (2002) 392–397. [8] A. Chant, C.M. Kraemer-Pecore, R. Watkin, G.G. Kneale, Attachment of a histidine tag to the minimal zinc finger protein of the Aspergillus nidulans gene regulatory protein AreA causes a conformational change at the DNA-binding site, Protein Expr. Purif. 39 (2005) 152–159. [9] J.P. Bannantine, J.K. Hansen, M.L. Paustian, A. Amonsin, L.L. Li, J.R. Stabel, V. Kapur, Expression and immunogenicity of proteins encoded by sequences specific to Mycobacterium avium subsp. paratuberculosis, J. Clin. Microbiol. 42 (2004) 106–114. [10] W. Kaplan, P. Husler, H. Klump, J. Erhardt, N. Sluis-Cremer, H. Dirr, Conformational stability of pGEX-expressed Schistosoma japonicum glutathione S-transferase: a detoxification enzyme and fusion-protein affinity tag, Protein Sci. 6 (1997) 399–406. [11] J. Arnau, C. Lauritzen, G.E. Petersen, J. Pedersen, Current strategies for the use of affinity tags and tag removal for the purification of recombinant proteins, Protein Expr. Purif. 48 (2006) 1–13. [12] E.A. Carswell, L.J. Old, R.L. Kassel, S. Green, N. Fiore, B. Williamson, An endotoxin-induced serum factor that causes necrosis of tumors, Proc. Natl. Acad. Sci. USA 72 (1975) 3666–3670. [13] F. Balkwill, Tumour necrosis factor and cancer, Nat. Rev. Cancer 9 (2009) 361– 371. [14] R.A. Black, C.T. Rauch, C.J. Kozlosky, J.J. Peschon, J.L. Slack, M.F. Wolfson, B.J. Castner, K.L. Stocking, P. Reddy, S. Srinivasan, N. Nelson, N. Boiani, K.A. Schooley, M. Gerhart, R. Davis, J.N. Fitzner, R.S. Johnson, R.J. Paxton, C.J. March, D.P. Cerretti, A metalloproteinase disintegrin that releases tumour-necrosis factor-alpha from cells, Nature 385 (1997) 729–733. [15] M.L. Moss, S.L.C. Jin, M.E. Milla, W. Burkhart, H.L. Carter, W.-J. Chen, W.C. Clay, J.R. Didsbury, D. Hassler, C.R. Hoffman, T.A. Kost, M.H. Lambert, M.A. Leesnitzer, P. McCauley, G. McGeehan, J. Mitchell, M. Moyer, G. Pahel, W. Rocque, L.K. Overton, F. Schoenen, T. Seaton, J.-L. Su, J. Warner, D. Willard, J.D. Becherer, Cloning of a disintegrin metalloproteinase that processes precursor tumour-necrosis factor-alpha, Nature 385 (1997) 733–736. [16] R.A. Smith, C. Baglioni, The active form of tumor necrosis factor is a trimer, J. Biol. Chem. 262 (1987) 6951–6954. [17] B. Neumann, T. Machleidt, A. Lifka, K. Pfeffer, D. Vestweber, T.W. Mak, B. Holzmann, M. Kronke, Crucial role of 55-kilodalton TNF receptor in TNFinduced adhesion molecule expression and leukocyte organ infiltration, J. Immunol. 156 (1996) 1587–1593. [18] M. Feldmann, R.N. Maini, TNF defined as a therapeutic target for rheumatoid arthritis and other autoimmune diseases, Nat. Med. 9 (2003) 1245–1250. [19] P. Ameloot, W. Declercq, W. Fiers, P. Vandenabeele, P. Brouckaert, Heterotrimers formed by tumor necrosis factors of different species or muteins, J. Biol. Chem. 276 (2001) 27098–27103. [20] J. Zalevsky, T. Secher, S.A. Ezhevsky, L. Janot, P.M. Steed, C. O’Brien, A. Eivazi, J. Kung, D.-H.T. Nguyen, S.K. Doberstein, F. Erard, B. Ryffel, D.E. Szymkowski, Dominant-negative inhibitors of soluble TNF attenuate experimental arthritis without suppressing innate immunity to infection, J. Immunol. 179 (2007) 1872–1883. [21] P.M. Steed, M.G. Tansey, J. Zalevsky, E.A. Zhukovsky, J.R. Desjarlais, D.E. Szymkowski, C. Abbott, D. Carmichael, C. Chan, L. Cherry, P. Cheung, A.J. Chirino, H.H. Chung, S.K. Doberstein, A. Eivazi, A.V. Filikov, S.X. Gao, R.S. Hubert, M. Hwang, L. Hyun, S. Kashi, A. Kim, E. Kim, J. Kung, S.P. Martinez, U.S. Muchhal, D.H. Nguyen, C. O’Brien, D. O’Keefe, K. Singer, O. Vafa, J. Vielmetter, S.C. Yoder, B.I. Dahiyat, Inactivation of TNF signaling by rationally designed dominant-negative TNF variants, Science 301 (2003) 1895–1898. [22] Y. Yoshioka, S. Tsunoda, Y. Tsutsumi, Development of a novel DDS for sitespecific PEGylated proteins, Chem. Cent. J. 5 (2011) 25.

C. Zhang et al. / Protein Expression and Purification 95 (2014) 195–203 [23] Y. Yamamoto, Y. Tsutsumi, Y. Yoshioka, T. Nishibata, K. Kobayashi, T. Okamoto, Y. Mukai, T. Shimizu, S. Nakagawa, S. Nagata, T. Mayumi, Site-specific PEGylation of a lysine-deficient TNF-alpha with full bioactivity, Nat. Biotechnol. 21 (2003) 546–552. [24] C.M. Petersen, A. Nykjaer, B.S. Christiansen, L. Heickendorff, S.C. Mogensen, B. Moller, Bioactive human recombinant tumor necrosis factor-alpha: an unstable dimer, Eur. J. Immunol. 19 (1989) 1887–1894. [25] W. Li, Y. Wang, X. Zhu, M. Li, Z. Su, Preparation and characterization of PEGylated adducts of recombinant human tumor necrosis factor-alpha from Escherichia coli, J. Biotechnol. 92 (2002) 251–258. [26] G.S. Rees, C.K. Gee, H.L. Ward, C. Ball, G.M. Tarrant, S. Poole, A.F. Bristow, Rat tumour necrosis factor-alpha: expression in recombinant Pichia pastoris, purification, characterization and development of a novel ELISA, Eur. Cytokine Netw. 10 (1999) 383–392. [27] H. Wang, Z. Yan, J. Shi, W. Han, Y. Zhang, Expression, purification, and characterization of a neovasculature targeted rmhTNF-alpha in Escherichia coli, Protein Expr. Purif. 45 (2006) 60–65. [28] K. Sreekrishna, L. Nelles, R. Potenz, J. Cruze, P. Mazzaferro, W. Fish, M. Fuke, K. Holden, D. Phelps, P. Wood, High-level expression, purification, and characterization of recombinant human tumor necrosis factor synthesized in the methylotrophic yeast Pichia pastoris, Biochemistry 28 (1989) 4117–4125. [29] U.K. Laemmli, Cleavage of structural proteins during the assembly of the head of bacteriophage T4, Nature 227 (1970) 680–685. [30] B.B. Aggarwal, W.J. Kohr, P.E. Hass, B. Moffat, S.A. Spencer, W.J. Henzel, T.S. Bringman, G.E. Nedwin, D.V. Goeddel, R.N. Harkins, Human tumor necrosis factor. Production, purification, and characterization, J. Biol. Chem. 260 (1985) 2345–2354. [31] F. Squadrito, D. Altavilla, G. Squadrito, G.M. Campo, M. Arlotta, C. Quartarone, A. Saitta, A.P. Caputi, Recombinant human erythropoietin inhibits iNOS

[32]

[33]

[34]

[35]

[36]

[37]

[38] [39] [40]

203

activity and reverts vascular dysfunction in splanchnic artery occlusion shock, Br. J. Pharmacol. 127 (1999) 482–488. M.M. Bradford, A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding, Anal. Biochem. 72 (1976) 248–254. M.A. Andrade, P. Chacón, J.J. Merelo, F. Morán, Evaluation of secondary structure of proteins from UV circular dichroism spectra using an unsupervised learning neural network, Protein Eng. 6 (1993) 383–390. C. Khosla, J.E. Curtis, J. DeModena, U. Rinas, J.E. Bailey, Expression of intracellular hemoglobin improves protein synthesis in oxygen-limited Escherichia coli, Biotechnology 8 (1990) 849–853. D. Zhang, S. Nandi, P. Bryan, S. Pettit, D. Nguyen, M.A. Santos, N. Huang, Expression, purification, and characterization of recombinant human transferrin from rice (Oryza sativa L.), Protein Expr. Purif. 74 (2010) 69–79. X.-L. Feng, Z.-Y. Gu, Y.-T. Jin, Z.-G. Su, Polyethylene glycol improves the purification of recombinant human tumor necrosis factor during ion exchange chromatography, Biotechnol. Tech. 12 (1998) 289–293. A. Hoffmann, M.Q. Muller, M. Gloser, A. Sinz, R. Rudolph, S. Pfeifer, Recombinant production of bioactive human TNF-alpha by SUMO-fusion system – high yields from shake-flask culture, Protein Expr. Purif. 72 (2010) 238–243. E.Y. Jones, D.I. Stuart, N.P.C. Walker, Structure of tumour necrosis factor, Nature 338 (1989) 225–228. P. Wingfield, R.H. Pain, S. Craig, Tumour necrosis factor is a compact trimer, FEBS Lett. 211 (1987) 79–84. L.O. Narhi, T. Arakawa, Dissociation of recombinant tumor necrosis factoralpha studied by gel permeation chromatography, Biochem. Biophys. Res. Commun. 147 (1987) 740–746.