anti-MUC1纳米抗体在E.coil中的高效表达和纯化

ca o n d ca ic Received 24 February 2005, and in revised form 12 April 2005 Available online 11 May 2005 Abstract In contrast to the murine and human VHs, camels’ single domain antibodies (sdAb) have suYcient solubility. These antigen-spe- ciWc fragments are expressed well in Escherichia coli. Here, we report high expression and puriWcation of sdAbs against MUC1 mucin. MUC1 is a high molecular weight glycoprotein with an aberrant expression proWle in various malignancies. The sdAb genes were sub-cloned into a pET32a+ vector to overexpress the protein coupled with fusion tags in E. coli BL21(DE3). The expressed sin- gle domain antibodies were puriWed by immobilized metal aYnity chromatography and antigen aYnity chromatography. Analysis by SDS–PAGE and Western blotting demonstrated the integrity of the sdAbs-tags, while ELISA results conWrm that the activity of these molecules compare favorably with that of the parent recombinant antibodies. Enterokinase treated sdAb showed a band at the molecular weight around 12 kDa which demonstrated the naked protein in its natural structure with activities comparable to that of native protein. The high binding activity to MUC1 antigen puriWed from ascitic Xuid (of patients with small-cell lung aggressive car- cinoma and metastasis to peritoneum) and the very close similarity of these molecules to human VHs illustrated the potential appli- cation of these novel products as an immunodiagnostic and immunotherapeutic reagent.  2005 Elsevier Inc. All rights reserved. Keywords: Single domain antibody; MUC1; Expression; PuriWcation The molecular nature of antibody molecules allows for an almost unlimited number of domain rearrange- ments. Antibody engineering allows the researcher to design and use a variety of binding and eVector domains [1]. Number of industrial applications for functional- ized antibody fragments can be envisaged including waste water treatment, industrial scale separation pro- cesses (e.g., of chiral molecules), abzymes, and as an ingredient in novel consumer goods with new or Although signiWcant progress has been made with respect to the production of single chain Fv (scFv)1 fragments by the Escherichia coli, bulk production 1 Abbreviations used: BSA, bovine serum albumin; CH1, constant heavy-chain domain; D-HMFG, deglycosylated human milk fat glob- ule membrane; ELISA, enzyme linked immunosorbant assay; Fvs, non-covalently associated heterodimers of VH and VL domains; HM- FG, human milk fat globule membrane; HRP, horseradish peroxidase; Ig, immunoglobulin(s); NSB, non-speciWc binding; OD, optical densi- Protein Expression and PuriW High expression and puriWcati anti-MUC1 single domain a Fatemeh Rahbarizadeh a, M Mehdi Forouzandeh-Mogha a Department of Clinical Biochemistry, Faculty of Medi b Department of Medical Biotechnology, Faculty of Med 1046-5928/$ - see front matter  2005 Elsevier Inc. All rights reserved. doi:10.1016/j.pep.2005.04.008 improved functionalities [1–4]. These applications require large amount of inexpensive molecules. * Corresponding author. Fax: +98 21 8006544. E-mail address: Rasaee_m@modares.ac.ir (M.J. Rasaee). tion 44 (2005) 32–38 www.elsevier.com/locate/yprep n of the recombinant camelid tibodies in Escherichia coli ohammad Javad Rasaee a,¤, am b, Abdol-Amir Allameh a l Sciences, Tarbiat Modarres University, Tehran, Iran al Sciences, Tarbiat Modarres University, Tehran, Iran ty; PBS, phosphate-buVered saline; scFv, single chain variable domain fragment; sdAb, single domain antibody; SDS–PAGE, sodium dodecyl sulfate–polyacrylamide gel electrophoresis; VH, variable domain from the heavy chain of conventional antibody; VHH, the VH fragment from a camel heavy-chain antibody; VL, variable domain from the light chain of conventional antibody. ss F. Rahbarizadeh et al. / Protein Expre (>100 kg) at low costs is still not feasible [5]. The major reason for this is the structural and functional proper- ties of these fragments. The conventional Fv containing the two variable domains (VH + VL) of immunoglobu- lins and the engineered fragment (e.g., scFv) have been considered as the smallest antibody fragments with retention of the full antigen-binding capacity. PCR and the development of powerful panning techniques led to the generation of large libraries of scFvs from which several speciWc binders could be selected successfully [6,7]. This strategy was a major breakthrough in molec- ular biology. However, the application of the technique is not straight forward. The cloning of the two correctly spliced gene fragments (VH + VL) is a diYcult step, gen- eration of a representative library is tedious, and the genetic constructs are often unstable in the bacterial host. Also, the expression yield, stability, and function- ality of scFv often turn out to be problematic [8]. The work of some investigators indicated that iso- lated VH domains are expected to bind antigen in absence of VL domains [9]. This matter led to attempts to obtain an even smaller antigen-binding unit in a VH format. Unfortunately, the poor solubility, the reduced aYnity for the antigen and the irreproducible outcome of manipulation of the murine and human VHs showed that additional protein engineering would be required to successfully generate single domain antibody fragments. By serendipity, it was discovered that this engineering is already performed continuously in nature. The discovery of camelidae heavy-chain antibodies naturally devoid of light chains opened up a new opportunity to develop novel sdAbs with improved solution properties [10]. These unique antibody isotypes interact with antigen by virtue of only one single variable domain, referred to as VHH. Despite the absence of the VH–VL combinatorial diversity, the VHH antibody fragments are expressed well in E. coli, extremely stable, highly soluble, and react speciWcally and with high aYnity to the antigens. VHH antibody fragments and some VH fragments derived from camels’ conventional antibodies are highly soluble and stable in solution [11,12]. Several sdAbs have been raised against diVerent hap- ten and protein antigens [13,14]. Previously, we have generated and reported two sdAb libraries (from Came- lus dromedarius and Camelus bactrianus) displayed on phage particles [15]. We used the MUC1-related peptide to evaluate the possibility of obtaining antigen-speciWc sdAb fragments against peptide antigens. The MUC1 membrane mucin has an extensive extracellular domain composed of variable numbers of a highly conserved 20 amino acid repeat sequence (PDTRPAPGSTAPPAH- GVTSA), which is abundant in O-glycosylated regions rich in serine, threonine, and proline [16]. The dominant feature of epitopes within the MUC1 protein core is the presence, in full or part, of the hydrophilic sequence of PDTRPAP [16,17]. ion and PuriWcation 44 (2005) 32–38 33 There are some reports on the generation of recombinant antibody fragments speciWc for tandem repeat region of MUC1 [15,18,19]. However, transgenic protein levels were relatively low, making future improvements necessary. Although many expression systems are available, the T7 promoter- driven system is among the most successful, due in large part to its ability to stringently control basal expression levels. In this study, the anti-MUC1 VHs genes were cloned into a pET32a+ vector to overex- press the protein coupled with fusion tags in E. coli BL21(DE3). Here, we report our eVorts for the construction of two anti-MUC1 sdAbs, their overexpression, puriWcation, and characterization. Materials and methods Materials Synthetic mucin peptide (TSA-P1-24-TSAPDTRPAP GST APP AHGVTSA PDTR), corresponding to the mucin core protein, which was chemically conjugated to bovine serum albumin (BSA) by reaction with glutaral- dehyde, was purchased from Q-BIO-GENE (Kayser- berg, France). All other reagents used in this study were at least of analytical grade and purchased from Sigma Chemical (St. Louis, MO). Methods PuriWcation of MUC1 from various sources Fresh human milk samples from healthy mother were collected. BrieXy, Xoated cream from the milk was obtained, disrupted with a homogenizer, and the membranes were pelleted at 80,000g for 90 min at 5 °C [20]. The pellets were resuspended in 0.3 mol/L sucrose, 70 mmol/L KCl, 2 mmol/L MgCl2, and 10 mmol/L Tris–HCl buVer, pH 7.4. The cream fraction was extracted twice with two volumes of chlo- roform and twice with one volume of ether, and stored at ¡70 °C. Preparation of chemically deglycosylated HMFG (D- HMFG), which closely resembles cancerous MUC1, was performed by incubation of the extensively dried sample (1 mg) in triXuoromethanesulfonic acid for 2 h at 0 °C, followed by neutralization with pyridine/water (3:2) at ¡20 °C, and dialysis against phosphate buVer (10 mM, pH 7.2) [21]. The native cancerous MUC1 was puriWed from ascitic Xuid of a patient with aggressive small-cell lung carcinoma and metastasis to peritoneum, by an anti- body–Sepharose aYnity column as described before [15]. ss 34 F. Rahbarizadeh et al. / Protein Expre Preparation of rabbit anti-camel labeled-HRP as a tracer Anti-camel IgG fractions were prepared in four rab- bits and puriWed by protein A–Sepharose aYnity chro- matography (Amersham Pharmacia Biotech, Vienna, Austria) as described by the manufacturer. Immuno- globulin fractions puriWed from antisera were conju- gated to HRP following a simpliWed NaIO4 method. A suitable concentration of enzyme–conjugate was selected based on a titration assay [22]. Media composition The Luria–Bertani (LB) medium contained 1% tryp- tone, 1% NaCl, and 0.5% yeast extract (plus 1.5% agar in plates) that was supplemented with 85�g/ml ampicillin for the selection of transformants. The LB medium supplemented with 0.01 M MgCl2 and 0.02 M glucose was used as electroporation medium. The TerriWc Broth (TB) was prepared with 1.2% tryp- tone, 2.4% yeast extract, 0.4% glycerol, 0.23% KH2PO4, 1.25% K2HPO4, and 70�g/ml ampicillin. Strains and vectors Escherichia coli Top10 strain (Invitrogen, San Diego, CA) was used as a host for plasmid manipulations and cultures in LB medium. The pET32a+ vector, and E. coli BL21(DE3) strain (Novagen, Madison, WI) were used for sdAb overexpression. Construction of plasmid Construction and screening of anti-MUC1 sdAb libraries have been described in detail elsewhere [15]. Selected clones (RR-EB of C. bactrianus and RR-ED of C. dromedarius) were sequenced and characterized. These two sdAbs isolated in this manner were conven- tional antibody VH sequences that represented a minor subpopulation in the library. The VHs encoding sequences in these clones were modiWed by PCR to include Xanking EcoRI and NotI restriction sites. Polymerase chain reaction (PCR) was carried out using the Pfu DNA polymerase (MBI, Fermantase, Opelstr., Germany), and cloned into the pET32a+ expression vector. A pair of PCR primers, 5�-AGCGGCCGCCT AGTGAGGAG-3� and 5�-GAATTCGTGCAGCT GCAGCAGCTGCAGCAGTC-3�, was designed to generate products with vector cohesive overhangs. The ampliWcation protocol consisted of a 10 min denatur- ation at 94 °C followed by 30 cycles of denaturation at 94 °C for 20 s, annealing at 55 °C for 30 s and 72 °C for 3 min. The ampliWed sdAb genes were gel-puriWed from agarose by high pure PCR product puriWcation kit (Roche, Mannheim, Germany) digested with EcoRI and NotI, and puriWed again. These single domain antibody fragments coding regions were cloned into pET32a+ vector. This vector is designed for expression of recombi- nant protein fused to the 109 amino acid thioredoxin (11.7 kDa), a 6 amino acid His-tag, and 15 amino acid ion and PuriWcation 44 (2005) 32–38 S-tag sequences upstream of the cloning site. The fusion tags together can be removed from the recombinant tar- get protein by protease cleavage using enterokinase. Electroporation Top10 E. coli strains were transformed by electropor- ation. Preparation of electrocompetent Top10 was per- formed as per supplier’s instructions. Eighty microliters of electrocompetent cells was mixed with 1–3�g of ligated plasmid in a 0.2 cm electroporation cuvette, incu- bated on ice for 2 min, and electroporated in an Eppen- dorf Multiporatore (Germany) with settings of 2500 V and 5 ms. After pulsing, 1.0 ml of ice-cold medium (LB medium supplemented with 0.01 M MgCl2 and 0.02 M glucose) was added immediately to the cuvette and the cells were transferred to a sterile 15 ml culture tube. The tubes were incubated at 37 °C without shaking for 10 min, then 2.0 ml supplemented LB medium was added to the tube, and the cells were allowed to recover for 1 h at 37 °C at 250 rpm. Transformants were plated (10 or 100�l) on LB plates containing 85�g/ml ampicillin and grown at 37 °C to isolate ampicillin-resistant transfor- mants. The selected transformants were checked by PCR and digestion using restriction enzymes (EcoRI and NotI). E. coli BL21(DE3) cells were transformed with positive recombinant plasmid and used for protein expression. Cultivation conditions Escherichia coli BL21(DE3) cells were transformed with selected recombined plasmids, RR-EB-ET and RR- ED-ET of C. bactrianus and C. dromedarius, respec- tively. The transformants were culture in 15 ml LB medium containing 70�g/ml ampicillin and grown over- night at 37 °C and 250 rpm. These pre-inocula were then transferred to 100 ml TB medium containing ampicillin at the same concentration. The cultures were grown at 37 °C and 250 rpm until OD600 of 0.7 was achieved. These cultures were aliqouted (2 ml) in sterile 45 ml cul- ture tubes. The tubes were induced with diVerent concen- trations of IPTG (0.5, 1, and 2 mM) in various conditions of induction time and temperature (2, 4, and 8 h and 25, 30, and 37 °C). Preparation of cell lysates and comparison of various induction conditions The cells were harvested by centrifugation at 6000g and 4 °C for 20 min. The pellets containing the bacteria of each aliquot were suspended in 200�l of 50 mM Tris– HCl buVer, pH 8.0, containing 100 mM NaCl and 1 mM EDTA (adsorption buVer). The protease inhibitor phen- ylmethylsulfonyl Xuoride (PMSF) and lysozyme were added to Wnal concentration of 1.0 mM and 1.0 mg/ml, respectively. The suspensions were incubated for 20 min at 4 °C with stirring, and then 0.04 mg of deoxycholic acid (Sigma) and 1�l of 1 mg/ml DNase (Roche) were ss F. Rahbarizadeh et al. / Protein Expre added. The suspensions were incubated at 37 °C with stirring for 30 min, then Triton X-100 and RNase were added to the tube to a Wnal concentration of 1% and 5�g/ml, respectively, and continued the incubation with rocking for another 10 min at 4 °C. The lysed material was clariWed by centrifugation at 14,000g for 30 min at 4 °C and the supernatant was collected. To conWrm and compare the quality of recombinant protein expression, SDS–PAGE electrophoresis of the proteins was per- formed as described by Laemmli [23], using 12% acryl- amide gels followed by staining with Coomassie brilliant blue and scanned on a densitometer. Biological activity of recombinant sdAb-tags and sdAbs was compared by ELISA. PuriWcation The target proteins were puriWed by loading the clari- Wed supernatant onto a 1£ 5 cm column, packed with 1.0 ml nickel–nitrilo-triacetic acid (Ni+–NTA) resin (Qiagen, Valencia, CA).The resin was washed with 10 column volumes (CV) of adsorption buVer (see above). The adsorbed proteins were eluted from the column using an imidazole gradient (from 0 to 250 mM in 10 CV) in adsorption buVer. Fractions containing 1.0 CV of volume were collected and a Xow rate of 0.5 ml/min was used during all the chromatographic steps. Fractions were assayed for total protein concentration according to the method presented by Bradford [24], and analyzed by SDS–PAGE [23]. The puriWed samples were dialyzed against adsorption buVer. Alternatively, the target proteins were puriWed by immuno-aYnity chromatography. This column was pre- pared by linking TSA-P1-24-BSA to CNBr-Sepharose 4B (Amersham Pharmacia Biotech) as described by the manufacturer. The cleared cell lysates were loaded onto these peptide–BSA–Sepharose column, equilibrated with phosphate-buVered saline (PBS) (0.39 g/L NaH2PO4, 0.89 g/L Na2HPO4, and 8.9 g/L NaCl, pH 7.0). After washing with PBS, the bound recombinant tagged sdAb- tags were eluted with 100 mM glycine–HCl buVer, pH 2.7. Fractions of 1 ml were collected, immediately neu- tralized with Tris buVer (1 M, pH 9.5), and dialyzed against phosphate buVer (10 mM, pH 7.0). The amounts of puriWed proteins were determined [24] and then ana- lyzed by SDS–PAGE [23]. The recombinant sdAb was separated from the fusion tags by enterokinase proteolysis (1 U/�l per 10�g of recombinant protein). The hydrolysis was performed at 25 °C for 14 h according to the protease supplier’s instructions. The protease was inactivated by PMSF at 1 mM and the proteolysis products were analyzed by SDS–PAGE. The sample was dialyzed against adsorp- tion buVer containing 10–20 mM imidazole to prepare for the next puriWcation step. The Wnal puriWcation step was performed using the same column resin mentioned above. The sample con- ion and PuriWcation 44 (2005) 32–38 35 taining the sdAb and the cleaved fusion protein was loaded onto the column equilibrated with the same dial- ysis buVer. The Xow through proteins (sdAb) were col- lected and washed out with 10 CV of the same buVer. Bound proteins were eluted (0.5 ml/min) with equilibra- tion buVer containing 200 mM imidazole. The fractions containing 1 CV were collected and analyzed by SDS– PAGE, and their total soluble protein concentration was determined. Reactivity of recombinant sdAb towards synthetic peptide or puriWed MUC1 of various sources The synthetic peptide–BSA (50–1000 ng/well), D- HMFG (50–1000 ng/well) or puriWed native cancerous MUC1 (10–200 ng/well) were coated onto the wells of microtiter plates at 37 °C overnight. The same concen- tration of BSA or a 14 amino acid irrelevant peptide (LEEKKGNVVTDHC) conjugated to BSA was used as a negative control. The plates were washed and blocked with a 2% solution of BSA in PBS for 1 h at 37 °C. At the end of incubation time, wells were washed and added with diluted cell lysates, puriWed recombinant sdAb-tags, or sdAbs. In this experiment, cell lysate of BL21(DE3) cell transformed with pET32a+ (without insert) was used as non-speciWc binding (NSB). The contents of the wells were incubated at 37 °C for 2 h, washed, added with rab- bit anti-camel conjugated to HRP, and incubated at 37 °C for 1 h. The rest of experiment was performed as explained before [15]. Results Construction and transformation of the vectors The sequence of the anti-MUC1 sdAbs was modiWed by PCR at the 5�-end by introduction of an EcoRI site. The 0.4kb inserts were ligated into the multiple cloning site region downstream of the Trx-Tag/His-Tag/S-Tag of pET32a+ vector using the EcoRI/NotI restriction sites. The resulting plasmids (RR-EB-ET and RR-ED-ET) were transformed into the E. coli strain Top10. The pET32a+ plasmid contains the ampicillin resistance gene for selec- tion in E. coli. Plasmids carrying the insert were selected on LB plates containing ampicillin. The selected transfor- mants were then conWrmed by PCR and digestion using restriction enzymes (EcoRI and NotI) E. coli BL21(DE3). Comparison of cultivation conditions Dose dependence, temperature, and time course stud- ies of the induction of the recombinant protein expres- sion, analyzed by SDS–PAGE led to a IPTG concentration of 1 mM and induction time of 8 h at 30 °C. The expression of the fusion protein containing ss 36 F. Rahbarizadeh et al. / Protein Expre the sdAb (sdAb-tags) was considered 48 and 19% of total soluble protein of RR-EB-ET and RR-ED-ET bac- terial cell lysate, respectively. PuriWcation of sdAb antibody fragments High quantities of relatively pure recombinant pro- tein containing the Tr