- Email: [email protected]
Functional expression and display of an antibody Fab fragment in Escherichia coli: study of vector designs and culture conditions
Protein Expression and Purification 34 (2004) 270–279 www.elsevier.com/locate/yprep
Functional expression and display of an antibody Fab fragment in Escherichia coli: study of vector designs and culture conditions Susanne Corisdeoa,b and Baiyang Wanga,* a
Department of Immunotherapeutics, Purdue Pharma LP, 6 Cedar Brook Drive, Cranbury, NJ 08512, USA b Department of Microbiology, Thomas Jefferson University, Philadelphia, PA 19107, USA Received 9 October 2003, and in revised form 25 November 2003
Abstract Several different vector designs are currently being used to display and express Fab molecules in Escherichia coli, but their relative efficiency in phage display and protein expression cannot be compared from the published data. We systematically investigated which vector design most effectively displays and expresses Fab molecules in E. coli using, as a model system, a human Fab against tetanus toxoid (tt). Three different vectors were used in this study: pFab1 where the VL-CL and VH-CH1 genes were driven by two promoters in two separate expression cassettes, and pFab2 and pFab3 that both contain one dicistronic expression cassette with two translation initiation sites and either VH-CH1 before VL-CL or VL-CL before VH-CH1, respectively. The display of tt-Fab on the surface of phage and the expression of tt-Fab protein in E. coli were compared for the aforementioned vectors. Our results showed that the pFab3 vector was most effective in Fab display. A 10-fold increase in the expression of secreted Fab was observed in pFab3 when compared with vectors pFab1 and pFab2. Further experiments were conducted using pFab3 to optimize expression levels using different strains of E. coli and various culture conditions. The highest expression of tt-Fab was obtained using the pFab3 vector in host strain JM105 with an induction temperature at 37 °C and IPTG concentration of 0.1 mM. Ó 2003 Elsevier Inc. All rights reserved. Keywords: Fab, expression and phage display; Vector, design; Condition, culture, E. coli strain
The fragment antigen binding (Fab) fragment of an antibody molecule is a heterodimer of VH-CH1 and VL-CL linked together through a disulfide bond. When simultaneously expressed with signal sequences in Escherichia coli, the two peptides (VH-CH1 and VL-CL) are secreted into the periplasm and form functional Fab molecules [1]. The Fab can also be expressed as a gIII fusion protein and displayed on the surface of filamentous phage (Fab phage) [2,3]. This Fab phage can be selected through a panning procedure that allows the isolation of target-specific Fabs from Fab libraries [2,3]. The combination of Fab expression and display has revolutionized the isolation and engineering of antibodies. Some of the most successful applications include de novo isolation of high affinity human antibodies from na€ıve [4,5] and immunized B-cell repertoire libraries [6] and affinity maturation of existing antibodies [7]. * Corresponding author. Fax: 1-609-409-6922. E-mail address: [email protected] (B. Wang).
1046-5928/$ - see front matter Ó 2003 Elsevier Inc. All rights reserved. doi:10.1016/j.pep.2003.11.020
One of the keys for the successful expression and display of Fab in E. coli is the expression vector design. Three different versions of phagemid vectors are currently used to display and express Fab molecules in E. coli: (a) VL-CL and VH-CH1 genes driven by two promoters in two separate expression cassettes [8,9]; (b) one dicistronic expression cassette with two translation initiation sites with VH-CH1 before VL-CL [10,11]; and (c) one dicistronic expression cassette with two translation initiation sites with VL-CL before VH-CH1 [12,13]. Many different Fabs have been expressed at varying levels using different vector designs and bacterial strains. Because the level of Fab expression is affected not only by vector design and bacterial strains but also by the sequences of the Fabs [14], the effects of vector design on the relative efficiency of Fab display and expression are difficult to assess using the published data. We report here a systematic study to investigate which vector design most effectively displays and expresses Fab molecules in E. coli. In this study,
S. Corisdeo, B. Wang / Protein Expression and Purification 34 (2004) 270–279
we constructed and compared all three configurations of phagemid vectors for their ability to express and display Fab in E. coli. A human Fab that binds tetanus toxoid (tt) [3] was used as a model Fab in this study. We investigated the level of tt-Fab expression in different cellular locations in E. coli and on the surface of phage, and also compared its expression under different culture conditions.
Materials and methods Reagents, strains, and media All enzymes were purchased from New England Biolabs (Beverly, MA) unless otherwise stated and used according to manufacturerÕs directions. All oligonucleotide primers (Table 1) were synthesized at Genosys (The Woodlands, TX). SurfZap cloning kit, bacterial strains BL21, and XL1-Blue competent cells were purchased from Stratagene (La Jolla, CA); Top10F0 , JM105, and DM1 cells were purchased from Invitrogen (Carlsbad, CA).
Construction of Fab expression and display vectors The parental vector for all vectors (Fig. 1) in this report is the pSurfscript SK()) vector (Stratagene). Among the individual components of the vector, CH1 and CLj fragments were amplified from human antibody sequences, and restriction sites, Shine and Dalgarno (SD), His6, PelB, and OmpA sequences were either incorporated into the vectors by PCR oligonucleotides or constructed as individual units with synthetic oligonucleotides. Recombinant DNA techniques were performed following standard protocols using modifications recommended by manufacturers. Whenever commercial kits were used, instructions were followed without modification. All PCR products were TA cloned and their DNA sequences were confirmed. Construction of the pFab1 For the construction of pFab1 (Fig. 1A), PCR primer A (a detailed description of all primers is in Table 1) and primer B were used to amplify a DNA fragment containing pelB from pSurfscript SK). The resulting PCR product was amplified again with primer A and primer
Table 1 Oligonucleotide primers Primer
Oligonucleotide sequence
A
Annealing upstream of the SapI site AGCGAGTCAGTGAGCGAGGAAG Annealing to PelB sequence without NotI site GCTAAGAGGAGGAGACCTGCGGCAGCCGTAGGCAATAGGTATTTC Annealing to PelB sequence with NotI site ACGAGAAAGCGGCCGCCATTGCTGGTTGTGCTGCTAAGAGGAGGAGACCTGC Annealing to 50 of CH with NotI/XhoI sites AGCAATGGCGGCCGCTTAATAAATCTCGAGTGCTAGCACCAAG Annealing to 30 of CH with SpeI site ACCTTGCACTAGTCTCACCAACTTTCTTGTCCAC Annealing to LacP sequence with 50 XbaI/his6/stop elements ACCTTGCTCTAGACATCACCATCACCATCACTAATAGCAACGCAATTAATGTGAGTTAGC Annealing to PelB sequence with SfiI/BamHI sites ACCTTGCGGATCCCATCATGGCCGGTTGGGCCGCTAAGAGGAGGAGACCTGC Annealing to 50 of CL with BamHI/SfiI sites ACCTTGCGGATCCCATCATGGCCGGCTCGGCCACTGTGGCTGCACCATCTGTC Annealing to 30 of CL containing stop codon and BglII/NheI/XhoI sites ACCTTGCCTCGAGTGCTAGCGAGATCTTACTAACACTCTCC Annealing to 50 untranslated region upstream of pelB containing the first 39 bases of the OmpA signal AGCCAGTGCCACTGCAATCGCGATAGCTGTCTTTTTCATTATGACTGTCTCCTTGGCGAGTAG Annealing to OmpA sequence containing the last 24 bases of the OmpA signal sequence and NotI site TTATTAGCGGCCGCCTGCGCTACGGTAGCGAAACCAGCCAGTGCCACTGCAATCGC Annealing to 50 of CL with SfiI/BglII sites TTAGCGGCCCAACCGGCCATGGCCAGATCTGTGGCTGCACCATCTGTCTTC Annealing to 30 of CL containing stop codon and BamHI site TCTCGTTGGATCCTTACTAACACTCTCCCCTGTT Annealing to 50 of pelB sequence containing XbaI/his6/stop/SD elements ACCTTGCTCTAGACATCACCATCACCATCACTAATAGATTCTTAACTACTCGCCAAGGAGA Annealing to 50 of the OmpA signal sequence containing BamHI site and SD CTTGAAGGATCCAATTCTTAACTACTCGCCAAGG XbaI/BamHI linker containing His6/stop codon 50 -CTAGACATCACCATCACCATCACTAATGAG-30 30 -TGTAGTGGTAGTGGTAGTGATTACTCCTAG-50
B C D E F G H I J K L M N O Oligolinker A
271
272
S. Corisdeo, B. Wang / Protein Expression and Purification 34 (2004) 270–279
Fig. 1. Construction of expression vectors.
C to generate a DNA fragment between SapI and NotI of the pFab1. This operation moved the original NotI site from the middle of the pelB signal peptide to the end of it without altering the amino acid sequence of pelB. Primer D and primer E were used to obtain the CH1 fragment between NotI and SpeI of the pFab1 from a heavy chain of a human IgG1 antibody cDNA. These two fragments were linked together and cloned into
pSurfscript SK()) that created vector pFab1a (containing part of the pFab1 between SapI and XbaI). PCR primer F and primer B were used to amplify a DNA fragment containing pelB from pSurfscript SK). The resulting PCR product was amplified further with primer F and primer G to generate a DNA fragment between XbaI and the first SfiI sites of the pFab1b. This operation removed the original NotI site from the
S. Corisdeo, B. Wang / Protein Expression and Purification 34 (2004) 270–279
273
Fig. 1. (continued)
middle of the pelB signal peptide and added an SfiI site at the 30 of pelB without altering its amino acid sequence. The fragment between the second SfiI and BglII sites was amplified with primer H and primer I from a light chain of a human IgG1 antibody cDNA. These two fragments were ligated together with an XbaI/XhoI digested pBluescript SK) vector to create vector pFab1b (containing part of pFab1 between the XbaI and BglII). The ligation of the two different parts of pFab1 resulted in the complete pFab1 vector.
Construction of the pFab2 For the construction of pFab2 (Fig. 1B), we changed the first pelB signal sequence of pFab1 to an OmpA signal sequence by a two-step PCR using two pairs of oligonucleotide primers. The pFab1 vector was first amplified with primer A and primer J. A second PCR amplification was performed on the first PCR product with primer A and primer K. The PCR product was digested with SapI/NotI and cloned into the SapI/NotI site of pFab1 to create vector pFab2a.
274
S. Corisdeo, B. Wang / Protein Expression and Purification 34 (2004) 270–279
Fig. 1. (continued)
We then eliminated the BamHI site and replaced the second SfiI site with a BglII site in the Fab2a vector. The CL fragment was PCR amplified with primer L and primer M. The resulting PCR product was digested with SfiI/BamHI and cloned into the SfiI/BglII sites of the pFab2a to create the pFab2b vector. The second LacP of the pFab2b was changed to an SD sequence by PCR using primers N and primer G. This PCR product was
digested with XbaI/SfiI and cloned into corresponding sites in the pFab2b to create vector pFab2. Construction of the pFab3 To construct pFab3 (Fig. 1C), pFab2a was first cut with XbaI/BglII and ligated with the Oligolinker A to create pFab3a vector. An SD/ompA fragment was created by PCR amplification of the pFab2 using primer O
S. Corisdeo, B. Wang / Protein Expression and Purification 34 (2004) 270–279
and primer K. This PCR product was digested with BamHI and NotI. Another fragment containing LacP/ SD/pelB/CL was PCR amplified from pFab2b using primer F and primer M. This PCR fragment was digested with EcoRI and BamHI. The pFab3 vector was constructed through ligation of the BamHI/NotI digested SD/ompA fragment and EcoRI/BamHI digested LacP/ SD/pelB/CL fragment with the EcoRI/NotI digested pFab3a. Single-strand rescue for preparation of phage particle Phagemids were transformed into E. coli XL1-Blue cells and single clones were selected and grown overnight in LB medium containing ampicillin (50 lg/ml), tetracycline (12.5 lg/ml), and 1% glucose at 37 °C with shaking. The overnight cultures were diluted to OD600 ¼ 0:05 with 10 ml Amp+/Tet/LB medium and incubation continued until OD600 > 0:2. The cells were collected by centrifugation. The 8 108 cells were resuspended in 1 ml fresh medium and infected with VCSM13 helper phage (phage to cell ratio or multiplicity of infection, MOI ¼ 5:1) at 37 °C for 15 min. The infected cells were diluted to O.D.600 about 0.1 by addition of 10 ml fresh medium and incubation was continued until OD600 reached about 0.9–1.0. Bacteriophage was isolated from the cultures after centrifugation and filtration through a 0.2 lM syringe filter. Phage titration The phage titer was determined by mixing 100 ll logarithmic phase XL1-Blue cells with 1 ll of 104 , 105 , and 106 dilutions of the medium containing the phage particles. The mixtures were incubated for 15 min at 37 °C and plated on LB plates containing 100 lg/ml ampicillin. The total number of colony-forming units recovered was calculated as follows: CFU ¼
number of colonies dilution factor total volume in microliters : volume plated in microliters
Phage-ELISA Multiwell plates (Nalge Nunc International) were coated with either 0.1 lg/well of tetanus toxoid antigen or BSA diluted in PBS and incubated overnight at 4 °C. The wells were then washed with PBS four times and blocked for 1 h with 5% non-fat milk in PBS. 100 ll/well of phage samples was allowed to bind to the immobilized antigen at 25 °C for 1 h. Unbound phage was washed away with PBS and bound phage was detected with an anti-M13 peroxidase conjugate (Amersham Pharmacia Biotech, Piscataway, NJ) and subsequent colorimetric detection with ABTS substrate (Vector laboratories, Burlingame, CA). The absorbance at
275
405 nm was measured using a ThermoMax microplate reader (Molecular Devices). Fab expression in Escherichia coli To generate soluble Fab, the gIII fragment was removed from the phagemid vectors by SpeI and XbaI digestion. E. coli cells transformed with phagemid DNA carrying gIII-deleted tt-Fab gene were inoculated into 40 ml LB (100 lg/ml ampicillin and 1% glucose) with overnight culture prepared from a single colony. Cells were grown until OD600 ¼ 1:0 and then pelleted at room temperature for 10 min at 3000g. The cell pellet was resuspended in 40 ml pre-warmed LB (100 lg/ml ampicillin and different concentrations of IPTG) and cells were grown at 37 °C with shaking for various times. Preparation of soluble and secreted protein After incubation, cells were harvested from liquid culture by centrifugation at 5000g for 10 min. For preparation of soluble protein, the cell pellet was resuspended in BugBuster Protein Extraction reagent (Novagen) using 1/10th of the original bacterial culture volume. The cell suspension was incubated on a shaking platform for 20 min at room temperature. Insoluble cell debris was removed by centrifugation at 16,000g for 20 min at 4 °C. Protein concentration was determined using the Bradford assay with the Bio-Rad Protein Assay Kit. For the isolation of the secreted protein from the culture medium, the culture supernatant was concentrated 10 times using Millipore Ultrafree-4 centrifugal filter units. SDS–PAGE, Western immunoblot analysis, and Fab quantitation Western immunoblot analysis was performed according to standard procedures. Samples were resuspended in equal volumes of Bio-RadÕs Laemmli sample buffer without b-mercaptoethanol, heated for 5 min at 95 °C, and loaded on a 4–15% SDS–polyacrylamide gel. Following electrophoresis, the samples were transferred to nitrocellulose by electroblotting. The membranes were immunoblotted with either a HRP-conjugated goat anti-human kappa (0.5 lg/ml, Southern Biotechnology Associates, Birmingham, AL) for free light chain and total Fab detection or a HRP-conjugated goat anti-human IgG gamma (gamma chain specific) (0.1 lg/ml, Southern Biotechnology Associates, Birmingham, AL) for free heavy chain and total Fab detection. Fab expression was quantitated using a Kodak image station 440. X-ray films of Western blots were converted to a digital image and the density of the bands was analyzed with Kodak Digital Science 1D V.3.0.2. A commercial human Fab (Rockland, Gilbertsville, PA) of
276
S. Corisdeo, B. Wang / Protein Expression and Purification 34 (2004) 270–279
known concentration was used to establish the standard curve for quantitation of unknown samples.
Results and discussion Vector construction Three different vectors, pFab1, pFab2, and pFab3, were constructed to express and display a human Fab against tetanus toxoid (tt) (Fig. 1). Transcription of the Fab gene was driven by the lac promoter, and the PelB and ompA leader sequences were used to direct the secretion of the fusion protein to the periplasmic space of E. coli. In all three vectors, VH-CH1 was produced as a fusion to the gIII protein. The resulting Fab molecules were incorporated onto the surface of phage in the presence of helper phage in E. coli. To achieve expression of the soluble Fab, the gIII sequence was removed by SpeI and XbaI digestion. These two restriction enzymes generated compatible ends that could be ligated to put a stop codon inframe with VH-CH1. The VH and VL of a human Fab that binds to tetanus toxoid were cloned into the vectors that allowed the heavy and light chains to be coexpressed and displayed. pFab1 contains VL-CL and VH-CH1-gIII genes driven by two Lac promoters in two separate expression cassettes. pFab2 contains one dicistronic expression cassette with two translation initiation sites with VHCH1-gIII before VL-CL. pFab3 was constructed from pFab2 with the VL-CL before the VH-CH1-gIII gene. Effect of vector design on Fab expression and display The relative efficiency of the three different expression vectors to display functional tetanus toxoid binding Fab on the surface of phage was compared by phage-ELISA analysis. Phagemid particles were rescued from equal numbers of XL1-Blue cells containing the aforementioned vectors by co-infecting with VCSM13 helper phage. Equal numbers of phage were incubated with either immobilized tetanus toxoid antigen or BSA. Binding phage was detected using an anti-M13 antibody and absorbance was measured using a colorimetric assay. Three independent experiments were conducted, each with samples in quadruplicate. Within an individual experiment, the means of the ELISA readings for tt and BSA were calculated (data not shown). The mean BSA reading was subtracted from the mean tt reading to obtain a background-corrected tt reading. The corrected readings were proportional to the number of phage that displays tt-Fab. The background-corrected tt readings for pFab2tt and pFab3tt were divided by the background-corrected tt reading for pFab1tt to calculate pFab1tt-relative absorbances.
Fig. 2. Comparison of tt-antigen binding by tt-displayed phage produced from three different vectors using phage-ELISA analysis. Data are displayed as absorbance relative to pFab1tt. A non-related angioFab was used as negative control.
Fig. 2 shows the ELISA-relative absorbances of the different vectors in relation to pFab1tt. Each data point represents the average of quadruplicate samples. The ELISA-relative absorbance of pFab3tt was, on average, twofold higher than pFab1tt and pFab2tt. These data suggest that the pFab3 vector is more effective than the pFab1 and pFab2 vectors in phage display of Fabs. The efficiency of Fab display is a critical factor in Fab identification since it directly affects not only the numbers of phage that will have Fab molecules on their surface but also the copy number of Fab per phage particle. The improved display efficiency of pFab3 phagemid would be expected to facilitate the isolation of target-specific Fab binders using phage display, including those Fab molecules that bind with lower affinity. We compared all three vectors, pFab1, pFab2, and pFab3, for levels of tt-Fab protein production. XL1-Blue cells were transformed with the expression plasmids, single colonies were selected and grown at 37 °C, and Fab expression was induced with IPTG. Fab produced inside cells and soluble in salt buffer was identified as lysate Fab. Soluble Fab that was secreted into the culture medium was identified as secreted Fab. Equal amounts of soluble cellular lysates (10 lg) and culture media that was adjusted to the cell paste weight were resolved by SDS– PAGE, transferred to a nitrocellulose membrane, and probed with anti-kappa antibody. The recombinant protein was visualized as bands of 49.4 kDa (assembled Fab) and 23.4 kDa (free light chain) (Fig. 3A). Protein loading was monitored by Coomassie blue staining (Fig. 3B). The highest level of assembled Fab was produced in secreted form using the pFab3 vector and showed an 8-fold increase in the production level as measured by analysis of the Western blot signals (data not shown). Using tt-Fab as a model molecule, we demonstrated that pFab3 vector most effectively expressed and
S. Corisdeo, B. Wang / Protein Expression and Purification 34 (2004) 270–279
277
Fig. 3. Comparison of the effect of vector design on the production levels of tt-Fab as secreted and cell lysate products in E. coli. Three different vectors (pFab1, pFab2, and pFab3) were constructed and tested for tt-Fab production in XL1-Blue cells with IPTG induction. Ten micrograms of soluble cell lysate (lys) and concentrated medium (sec) adjusted to the cell paste weight was separated by non-reducing SDS–PAGE. (A) Western blot detection of the assembled Fab molecule (49.4 kDa) and free light chain (23.4 kDa) by HRP-labeled anti-human kappa chain antibody. (B) Coomassie stained gel demonstrating protein loading.
displayed Fab in E. coli. Since we employed the same lacZ promoter in all constructs and expressed and displayed the same tt-Fab in the same XL1-Blue E. coli strain under the same culture conditions, the transcription efficiency and protein stability should not be contributing factors. The difference in expression and display is likely due to the variable efficiency in protein translation reflecting the different secondary structures of transcribed RNAs due to the differences in sequences. Other factors influencing tt-Fab expression Cell strain and culture temperature Separate fragments of heavy and light chains were produced in the cytoplasm of E. coli and secreted into the periplasm. The heavy and light chain assembly takes place in the oxidizing environment of the periplasm yielding the heterodimer Fab. Fig. 4 shows that a large portion of heavy and light chains are not assembled. The yield of Fab expression could potentially be increased through the enhancement of Fab assembly either in vivo or in vitro.
Fig. 4. Western blot analysis of free light chain, free heavy chain, and assembled tt-Fab produced using the pFab4 vector in XL1-Blue cells. Ten micrograms of soluble cell lysate and concentrated medium adjusted to the cell paste weight was separated by non-reducing SDS–PAGE. (A) Detection using anti-gamma antibody of free heavy chain (26.1 kDa) and assembled Fab (49.4 kDa). (B) Detection using anti-kappa antibody of free light chain (23.4 kDa) and assembled Fab (49.4 kDa).
Fig. 5. Comparison of Fab levels produced as soluble cell lysate and secreted products in different bacterial strains. pFab3 vector was used to produce tt-Fab in E. coli strains JM10, DM1, TOP10F0 , XL1-Blue, and BL21. Ten micrograms of soluble cell lysate (lanes 7–12) and concentrated medium adjusted to the cell paste weight (lanes 1–6) was separated by non-reducing SDS–PAGE. (A) Western blot analysis of assembled Fab molecule (49.4 kDa) by anti-kappa antibody. (B) Coomassie stained gel demonstrating protein loading.
The efficiency of Fab production also varies between different cell strains and under different growth condition. Five strains of E. coli (JM105, DM1, Top10F, XL1-Blue, and BL21) were transformed with the pFab3tt plasmid; their total cell lysates (10 lg) and culture supernatants (adjusted to cell paste weight) were assayed by Western blot analysis (Fig. 5A). The Kodak image station analysis showed 12- to 13-fold more assembled Fab produced in the cell lysates of both the JM105 and BL21 bacterial strains over the other three strains (data not shown). The highest levels of Fab produced were recovered from the supernatant of JM105 and BL21 strains. A bacterial protein has recently been identified that is thought to facilitate the
278
S. Corisdeo, B. Wang / Protein Expression and Purification 34 (2004) 270–279
Fig. 6. Comparison of the effect of different growth temperatures during induction on Fab produced in JM105 bacterial cells. IPTG was added at 0.1 mM concentration to cultures grown overnight at 16, 25 or 37 °C. Ten micrograms of soluble cell lysate and concentrated medium adjusted to cell paste weight was separated by non-reducing SDS–PAGE. (A) Western blot detection of assembled Fab molecule (49.4 kDa) using anti-kappa antibody. (B) Coomassie stained gel demonstrating protein loading.
expression of antibody fragments in E. coli [15]. This endogenous protein may be expressed at different levels in different E. coli strains and may be one possible explanation for strain-specific differences. Lowering the bacterial growth temperature has been shown to decrease periplasmic aggregation and increase the yield of soluble antibody protein [16]. To test this, we incubated bacteria in shake-flasks with LB medium with 0.1 mM IPTG at 16, 25, and 37 °C. Ten micrograms of soluble cell lysate and concentrated medium adjusted to cell paste weight was separated by SDS–PAGE and Western blot detection was performed using anti-kappa antibody (Fig. 6A). Under these growth conditions, no significant effect was found on the production of lysate Fab and much more secreted Fab was produced at higher temperature. The majority of assembled Fab protein was recovered in the medium of 37 °C culture rather than the cell lysate. Culture conditions have been shown to modulate the relative distribution of antibody fragments between the periplasmic space and the medium [17], and conservative amino acid substitutions within the recombinant antibody fragments have resulted in the secretion of the
antibody rather than its retention in the periplasm [10]. We noticed severe bacterial lysis when incubating cultures expressing tt Fab at 37 °C after IPTG induction. Coomassie staining also showed a lysate-like staining pattern in the 37 °C lane of the secreted material even when equal volumes of culture were loaded (Fig. 6B). Our data suggested that, for certain Fabs, high level of Fab production at high temperature is toxic to cells and causes a significant degree of Fab release into the culture medium due to cell lysis. The highest production can be achieved by growing cells to high density in the presence of repressor and inducing protein expression at high temperature. IPTG Concentration Increasing concentration of IPTG (from 0.02 to 2 mM) had little effect on Fab production in our system (Fig. 7A). All inductions were performed at 37 °C using the pFab3 vector to produce tt-Fab in JM105 cells. Protein was loaded as previously described and Western blot detection was performed with the anti-kappa antibody. Expression using 0.02 mM IPTG induction yielded similar levels of Fab as that obtained using 2 mM IPTG.
Fig. 7. Comparison of Fab production under different conditions of induction. All inductions were performed at 37 °C using the pFab4 vector to produce tt-Fab in JM105 cells. Ten micrograms of soluble cell lysate and concentrated medium adjusted to cell paste weight was separated by nonreducing SDS–PAGE. (A) Western blot detection of assembled Fab molecule (49.4 kDa) using anti-kappa antibody. (B) Coomassie stained gel demonstrating protein loading.
S. Corisdeo, B. Wang / Protein Expression and Purification 34 (2004) 270–279
279
References
Fig. 8. Yield of secreted tt-Fab produced under optimized conditions: high density of JM105 cells using expression vector pFab3 in LB medium induced with 0.1 mM IPTG for 16 h at 37 °C. (A) Western blot detection of known concentrations of commercial Fab and dilutions of culture sample using anti-kappa antibody. (B) Standard curve representing the relationship between the quantity of Fab standard and densitometry reading. The densitometry reading of the experimental sample was located on the standard curve and protein quantity was determined.
[1] A. Skerra, A. Pluckthun, Assembly of a functional immunoglobulin Fv fragment in Escherichia coli, Science 240 (1988) 1038– 1041. [2] H.R. Hoogenboom, A.D. Griffiths, K.S. Johnson, P. Hudson, G. Winter, Multisubunit proteins on the surface of filamentous phage: methodologies for displaying antibody (Fab) heavy and light chains, Nucleic Acids Res. 19 (1991) 4133–4137. [3] C.F. Barbas III, A.S. Kang, R.A. Lerner, S.J. Benkovic, Assembly of combinatorial antibody libraries on phage surfaces: the gene III site, Proc. Natl. Acad. Sci. USA 213 (1991) 201–212. [4] H. Gram, L. Marconi, C.F. Barbas III, T.A. Collet, R.A. Lerner, A.S. Kang, In vitro selection and affinity maturation of antibodies from na€ıve combinatorial immunoglobulin library, Proc. Natl. Acad. Sci. USA 89 (1992) 3576–3580. [5] P. Chames, S.E. Hofton, P.G. Coulie, B. Uchanska-Ziegler, H.R. Hoogenboom, Direct selection of a human antibody fragment directed against the tumor T-cell epitope HLA-A1-MAGE-A1 from a nonimmunized phage-Fab library, Proc. Natl. Acad. Sci. USA 97 (2000) 7969–7974. [6] I. Moreno de Alboran, C. Martinez-alonso, C.F. Barbas III, D.R. Burton, H.J. Ditzel, Human monoclonal Fab fragments specific for viral antigens from combinatorial IgA libraries, Immunotechnology 1 (1995) 21–28. [7] W.P. Yang, K. Green, S. Pinz-Sweeney, A.T. Briones, D.R. Burton, C.F. Barbas III, CDR walking mutagenesis for the affinity maturation of a potent human anti-HIV-1 antibody into the picomolar range, J. Mol. Biol. 254 (1995) 392–403. [8] R. Burioni, P. Plaisant, F. Bugli, V.D. Carri, M. Clementi, G. Fadda, A vector for the expression of recombinant monoclonal Fab fragments in bacteria, J. Immunol. Methods 217 (1998) 195– 199. [9] M. Foti, F. Granucci, P. Ricciardi-Castagnoli, A. Spreafico, M. Ackermann, M. Suter, Rabbit monoclonal Fab derived from a phage display library, J. Immunol. Methods 213 (1998) 201–212. [10] G. Forsberg, M. Forsgren, M. Jaki, M. Norin, C. Sterky, A.
Yield of Fab production under optimized conditions We identified optimized conditions for tt-Fab production through testing different vector designs, bacterial strains, and culture and induction conditions. The optimized conditions are as follows: high density of JM105 cells containing the pFab3tt vector in LB medium plus 0.1 mM IPTG induction for 16 h at 37 °C. The Fab production under these conditions was estimated by Western blot analysis using known concentrations of commercial Fab as a quantitation standard (Fig. 8A). The assembled Fab was detected using anti-kappa antibody and the Western blot signals were quantitated using a Kodak image station. The quantity of the assembled Fab standard was adjusted from the known protein quantity that was loaded on the gel using the densitometry reading of all bands within the lane (calculation not shown). A standard curve was established between the quantity of the assembled Fab standard and the densitometry reading (Fig. 8B). The densitometry reading of the experimental sample was located on the standard curve and its corresponding protein quantity was determined. Under optimized conditions, the production level of assembled tt-Fab was approximately 16 mg/L.
[11]
[12]
[13]
[14]
[15]
[16]
[17]
Enhorning, K. Larsson, M. Ericsson, P. Bj€ ork, Identification of framework residues in a secreted recombinant antibody fragment that control production level and localization in Escherichia coli, J. Biol. Chem. 272 (1997) 12430–12436. R. Levy, R. Weiss, G. Chen, B.L. Iverson, G. Georgiou, Production of correctly folded Fab antibody fragment in the cytoplasm of Escherichia coli trxB gor mutants via the coexpression of molecular chaperones, Protein Expr. Purif 23 (2001) 338– 347. P.M. OÕBrien, G. Maxwell, M.S. Campo, Bacterial expression and purification of recombinant bovine Fab fragments, Protein Express. Purif. 24 (2002) 43–50. S. Kumar, J. Kalsi, C.T. Ravirajan, A. Rahman, D. Athwal, D.S. Latchman, D.A. Isenber, L.H. Pearl, Molecular cloning and expression of the Fabs of human autoantibodies in Escherichia coli, J. Biol. Chem. 275 (2000) 35129–35136. D.S. Gill, Y.W. Wong, M.N. Margolies, Differences in sequencespecific expression of two anti-arsonate Fabs in E. coli, Biotechnol. Prog. 13 (1997) 692–694. H. Bothmann, A. Pluckthun, Selection for periplasmic factor improving phage display and functional periplasmic expression, Nat. Biotechnol. 16 (1998) 376–380. A. Skerra, A. Pluckthum, Secretion and in vivo folding of the Fab fragment of the antibody McPC603 in Escherichia coli: influence of disulfides and cis-prolines, Protein Eng. 4 (1991) 971–976. S.M. Kipriyanov, G. Moldenhauer, M. Little, High-level production of soluble single chain antibodies in small-scale Escherichia coli cultures, J. Immunol. Methods 200 (1997) 69–77.
Functional expression and display of an antibody Fab fragment in Escherichia coli: study of vector designs and culture conditions Susanne Corisdeoa,b and Baiyang Wanga,* a
Department of Immunotherapeutics, Purdue Pharma LP, 6 Cedar Brook Drive, Cranbury, NJ 08512, USA b Department of Microbiology, Thomas Jefferson University, Philadelphia, PA 19107, USA Received 9 October 2003, and in revised form 25 November 2003
Abstract Several different vector designs are currently being used to display and express Fab molecules in Escherichia coli, but their relative efficiency in phage display and protein expression cannot be compared from the published data. We systematically investigated which vector design most effectively displays and expresses Fab molecules in E. coli using, as a model system, a human Fab against tetanus toxoid (tt). Three different vectors were used in this study: pFab1 where the VL-CL and VH-CH1 genes were driven by two promoters in two separate expression cassettes, and pFab2 and pFab3 that both contain one dicistronic expression cassette with two translation initiation sites and either VH-CH1 before VL-CL or VL-CL before VH-CH1, respectively. The display of tt-Fab on the surface of phage and the expression of tt-Fab protein in E. coli were compared for the aforementioned vectors. Our results showed that the pFab3 vector was most effective in Fab display. A 10-fold increase in the expression of secreted Fab was observed in pFab3 when compared with vectors pFab1 and pFab2. Further experiments were conducted using pFab3 to optimize expression levels using different strains of E. coli and various culture conditions. The highest expression of tt-Fab was obtained using the pFab3 vector in host strain JM105 with an induction temperature at 37 °C and IPTG concentration of 0.1 mM. Ó 2003 Elsevier Inc. All rights reserved. Keywords: Fab, expression and phage display; Vector, design; Condition, culture, E. coli strain
The fragment antigen binding (Fab) fragment of an antibody molecule is a heterodimer of VH-CH1 and VL-CL linked together through a disulfide bond. When simultaneously expressed with signal sequences in Escherichia coli, the two peptides (VH-CH1 and VL-CL) are secreted into the periplasm and form functional Fab molecules [1]. The Fab can also be expressed as a gIII fusion protein and displayed on the surface of filamentous phage (Fab phage) [2,3]. This Fab phage can be selected through a panning procedure that allows the isolation of target-specific Fabs from Fab libraries [2,3]. The combination of Fab expression and display has revolutionized the isolation and engineering of antibodies. Some of the most successful applications include de novo isolation of high affinity human antibodies from na€ıve [4,5] and immunized B-cell repertoire libraries [6] and affinity maturation of existing antibodies [7]. * Corresponding author. Fax: 1-609-409-6922. E-mail address: [email protected] (B. Wang).
1046-5928/$ - see front matter Ó 2003 Elsevier Inc. All rights reserved. doi:10.1016/j.pep.2003.11.020
One of the keys for the successful expression and display of Fab in E. coli is the expression vector design. Three different versions of phagemid vectors are currently used to display and express Fab molecules in E. coli: (a) VL-CL and VH-CH1 genes driven by two promoters in two separate expression cassettes [8,9]; (b) one dicistronic expression cassette with two translation initiation sites with VH-CH1 before VL-CL [10,11]; and (c) one dicistronic expression cassette with two translation initiation sites with VL-CL before VH-CH1 [12,13]. Many different Fabs have been expressed at varying levels using different vector designs and bacterial strains. Because the level of Fab expression is affected not only by vector design and bacterial strains but also by the sequences of the Fabs [14], the effects of vector design on the relative efficiency of Fab display and expression are difficult to assess using the published data. We report here a systematic study to investigate which vector design most effectively displays and expresses Fab molecules in E. coli. In this study,
S. Corisdeo, B. Wang / Protein Expression and Purification 34 (2004) 270–279
we constructed and compared all three configurations of phagemid vectors for their ability to express and display Fab in E. coli. A human Fab that binds tetanus toxoid (tt) [3] was used as a model Fab in this study. We investigated the level of tt-Fab expression in different cellular locations in E. coli and on the surface of phage, and also compared its expression under different culture conditions.
Materials and methods Reagents, strains, and media All enzymes were purchased from New England Biolabs (Beverly, MA) unless otherwise stated and used according to manufacturerÕs directions. All oligonucleotide primers (Table 1) were synthesized at Genosys (The Woodlands, TX). SurfZap cloning kit, bacterial strains BL21, and XL1-Blue competent cells were purchased from Stratagene (La Jolla, CA); Top10F0 , JM105, and DM1 cells were purchased from Invitrogen (Carlsbad, CA).
Construction of Fab expression and display vectors The parental vector for all vectors (Fig. 1) in this report is the pSurfscript SK()) vector (Stratagene). Among the individual components of the vector, CH1 and CLj fragments were amplified from human antibody sequences, and restriction sites, Shine and Dalgarno (SD), His6, PelB, and OmpA sequences were either incorporated into the vectors by PCR oligonucleotides or constructed as individual units with synthetic oligonucleotides. Recombinant DNA techniques were performed following standard protocols using modifications recommended by manufacturers. Whenever commercial kits were used, instructions were followed without modification. All PCR products were TA cloned and their DNA sequences were confirmed. Construction of the pFab1 For the construction of pFab1 (Fig. 1A), PCR primer A (a detailed description of all primers is in Table 1) and primer B were used to amplify a DNA fragment containing pelB from pSurfscript SK). The resulting PCR product was amplified again with primer A and primer
Table 1 Oligonucleotide primers Primer
Oligonucleotide sequence
A
Annealing upstream of the SapI site AGCGAGTCAGTGAGCGAGGAAG Annealing to PelB sequence without NotI site GCTAAGAGGAGGAGACCTGCGGCAGCCGTAGGCAATAGGTATTTC Annealing to PelB sequence with NotI site ACGAGAAAGCGGCCGCCATTGCTGGTTGTGCTGCTAAGAGGAGGAGACCTGC Annealing to 50 of CH with NotI/XhoI sites AGCAATGGCGGCCGCTTAATAAATCTCGAGTGCTAGCACCAAG Annealing to 30 of CH with SpeI site ACCTTGCACTAGTCTCACCAACTTTCTTGTCCAC Annealing to LacP sequence with 50 XbaI/his6/stop elements ACCTTGCTCTAGACATCACCATCACCATCACTAATAGCAACGCAATTAATGTGAGTTAGC Annealing to PelB sequence with SfiI/BamHI sites ACCTTGCGGATCCCATCATGGCCGGTTGGGCCGCTAAGAGGAGGAGACCTGC Annealing to 50 of CL with BamHI/SfiI sites ACCTTGCGGATCCCATCATGGCCGGCTCGGCCACTGTGGCTGCACCATCTGTC Annealing to 30 of CL containing stop codon and BglII/NheI/XhoI sites ACCTTGCCTCGAGTGCTAGCGAGATCTTACTAACACTCTCC Annealing to 50 untranslated region upstream of pelB containing the first 39 bases of the OmpA signal AGCCAGTGCCACTGCAATCGCGATAGCTGTCTTTTTCATTATGACTGTCTCCTTGGCGAGTAG Annealing to OmpA sequence containing the last 24 bases of the OmpA signal sequence and NotI site TTATTAGCGGCCGCCTGCGCTACGGTAGCGAAACCAGCCAGTGCCACTGCAATCGC Annealing to 50 of CL with SfiI/BglII sites TTAGCGGCCCAACCGGCCATGGCCAGATCTGTGGCTGCACCATCTGTCTTC Annealing to 30 of CL containing stop codon and BamHI site TCTCGTTGGATCCTTACTAACACTCTCCCCTGTT Annealing to 50 of pelB sequence containing XbaI/his6/stop/SD elements ACCTTGCTCTAGACATCACCATCACCATCACTAATAGATTCTTAACTACTCGCCAAGGAGA Annealing to 50 of the OmpA signal sequence containing BamHI site and SD CTTGAAGGATCCAATTCTTAACTACTCGCCAAGG XbaI/BamHI linker containing His6/stop codon 50 -CTAGACATCACCATCACCATCACTAATGAG-30 30 -TGTAGTGGTAGTGGTAGTGATTACTCCTAG-50
B C D E F G H I J K L M N O Oligolinker A
271
272
S. Corisdeo, B. Wang / Protein Expression and Purification 34 (2004) 270–279
Fig. 1. Construction of expression vectors.
C to generate a DNA fragment between SapI and NotI of the pFab1. This operation moved the original NotI site from the middle of the pelB signal peptide to the end of it without altering the amino acid sequence of pelB. Primer D and primer E were used to obtain the CH1 fragment between NotI and SpeI of the pFab1 from a heavy chain of a human IgG1 antibody cDNA. These two fragments were linked together and cloned into
pSurfscript SK()) that created vector pFab1a (containing part of the pFab1 between SapI and XbaI). PCR primer F and primer B were used to amplify a DNA fragment containing pelB from pSurfscript SK). The resulting PCR product was amplified further with primer F and primer G to generate a DNA fragment between XbaI and the first SfiI sites of the pFab1b. This operation removed the original NotI site from the
S. Corisdeo, B. Wang / Protein Expression and Purification 34 (2004) 270–279
273
Fig. 1. (continued)
middle of the pelB signal peptide and added an SfiI site at the 30 of pelB without altering its amino acid sequence. The fragment between the second SfiI and BglII sites was amplified with primer H and primer I from a light chain of a human IgG1 antibody cDNA. These two fragments were ligated together with an XbaI/XhoI digested pBluescript SK) vector to create vector pFab1b (containing part of pFab1 between the XbaI and BglII). The ligation of the two different parts of pFab1 resulted in the complete pFab1 vector.
Construction of the pFab2 For the construction of pFab2 (Fig. 1B), we changed the first pelB signal sequence of pFab1 to an OmpA signal sequence by a two-step PCR using two pairs of oligonucleotide primers. The pFab1 vector was first amplified with primer A and primer J. A second PCR amplification was performed on the first PCR product with primer A and primer K. The PCR product was digested with SapI/NotI and cloned into the SapI/NotI site of pFab1 to create vector pFab2a.
274
S. Corisdeo, B. Wang / Protein Expression and Purification 34 (2004) 270–279
Fig. 1. (continued)
We then eliminated the BamHI site and replaced the second SfiI site with a BglII site in the Fab2a vector. The CL fragment was PCR amplified with primer L and primer M. The resulting PCR product was digested with SfiI/BamHI and cloned into the SfiI/BglII sites of the pFab2a to create the pFab2b vector. The second LacP of the pFab2b was changed to an SD sequence by PCR using primers N and primer G. This PCR product was
digested with XbaI/SfiI and cloned into corresponding sites in the pFab2b to create vector pFab2. Construction of the pFab3 To construct pFab3 (Fig. 1C), pFab2a was first cut with XbaI/BglII and ligated with the Oligolinker A to create pFab3a vector. An SD/ompA fragment was created by PCR amplification of the pFab2 using primer O
S. Corisdeo, B. Wang / Protein Expression and Purification 34 (2004) 270–279
and primer K. This PCR product was digested with BamHI and NotI. Another fragment containing LacP/ SD/pelB/CL was PCR amplified from pFab2b using primer F and primer M. This PCR fragment was digested with EcoRI and BamHI. The pFab3 vector was constructed through ligation of the BamHI/NotI digested SD/ompA fragment and EcoRI/BamHI digested LacP/ SD/pelB/CL fragment with the EcoRI/NotI digested pFab3a. Single-strand rescue for preparation of phage particle Phagemids were transformed into E. coli XL1-Blue cells and single clones were selected and grown overnight in LB medium containing ampicillin (50 lg/ml), tetracycline (12.5 lg/ml), and 1% glucose at 37 °C with shaking. The overnight cultures were diluted to OD600 ¼ 0:05 with 10 ml Amp+/Tet/LB medium and incubation continued until OD600 > 0:2. The cells were collected by centrifugation. The 8 108 cells were resuspended in 1 ml fresh medium and infected with VCSM13 helper phage (phage to cell ratio or multiplicity of infection, MOI ¼ 5:1) at 37 °C for 15 min. The infected cells were diluted to O.D.600 about 0.1 by addition of 10 ml fresh medium and incubation was continued until OD600 reached about 0.9–1.0. Bacteriophage was isolated from the cultures after centrifugation and filtration through a 0.2 lM syringe filter. Phage titration The phage titer was determined by mixing 100 ll logarithmic phase XL1-Blue cells with 1 ll of 104 , 105 , and 106 dilutions of the medium containing the phage particles. The mixtures were incubated for 15 min at 37 °C and plated on LB plates containing 100 lg/ml ampicillin. The total number of colony-forming units recovered was calculated as follows: CFU ¼
number of colonies dilution factor total volume in microliters : volume plated in microliters
Phage-ELISA Multiwell plates (Nalge Nunc International) were coated with either 0.1 lg/well of tetanus toxoid antigen or BSA diluted in PBS and incubated overnight at 4 °C. The wells were then washed with PBS four times and blocked for 1 h with 5% non-fat milk in PBS. 100 ll/well of phage samples was allowed to bind to the immobilized antigen at 25 °C for 1 h. Unbound phage was washed away with PBS and bound phage was detected with an anti-M13 peroxidase conjugate (Amersham Pharmacia Biotech, Piscataway, NJ) and subsequent colorimetric detection with ABTS substrate (Vector laboratories, Burlingame, CA). The absorbance at
275
405 nm was measured using a ThermoMax microplate reader (Molecular Devices). Fab expression in Escherichia coli To generate soluble Fab, the gIII fragment was removed from the phagemid vectors by SpeI and XbaI digestion. E. coli cells transformed with phagemid DNA carrying gIII-deleted tt-Fab gene were inoculated into 40 ml LB (100 lg/ml ampicillin and 1% glucose) with overnight culture prepared from a single colony. Cells were grown until OD600 ¼ 1:0 and then pelleted at room temperature for 10 min at 3000g. The cell pellet was resuspended in 40 ml pre-warmed LB (100 lg/ml ampicillin and different concentrations of IPTG) and cells were grown at 37 °C with shaking for various times. Preparation of soluble and secreted protein After incubation, cells were harvested from liquid culture by centrifugation at 5000g for 10 min. For preparation of soluble protein, the cell pellet was resuspended in BugBuster Protein Extraction reagent (Novagen) using 1/10th of the original bacterial culture volume. The cell suspension was incubated on a shaking platform for 20 min at room temperature. Insoluble cell debris was removed by centrifugation at 16,000g for 20 min at 4 °C. Protein concentration was determined using the Bradford assay with the Bio-Rad Protein Assay Kit. For the isolation of the secreted protein from the culture medium, the culture supernatant was concentrated 10 times using Millipore Ultrafree-4 centrifugal filter units. SDS–PAGE, Western immunoblot analysis, and Fab quantitation Western immunoblot analysis was performed according to standard procedures. Samples were resuspended in equal volumes of Bio-RadÕs Laemmli sample buffer without b-mercaptoethanol, heated for 5 min at 95 °C, and loaded on a 4–15% SDS–polyacrylamide gel. Following electrophoresis, the samples were transferred to nitrocellulose by electroblotting. The membranes were immunoblotted with either a HRP-conjugated goat anti-human kappa (0.5 lg/ml, Southern Biotechnology Associates, Birmingham, AL) for free light chain and total Fab detection or a HRP-conjugated goat anti-human IgG gamma (gamma chain specific) (0.1 lg/ml, Southern Biotechnology Associates, Birmingham, AL) for free heavy chain and total Fab detection. Fab expression was quantitated using a Kodak image station 440. X-ray films of Western blots were converted to a digital image and the density of the bands was analyzed with Kodak Digital Science 1D V.3.0.2. A commercial human Fab (Rockland, Gilbertsville, PA) of
276
S. Corisdeo, B. Wang / Protein Expression and Purification 34 (2004) 270–279
known concentration was used to establish the standard curve for quantitation of unknown samples.
Results and discussion Vector construction Three different vectors, pFab1, pFab2, and pFab3, were constructed to express and display a human Fab against tetanus toxoid (tt) (Fig. 1). Transcription of the Fab gene was driven by the lac promoter, and the PelB and ompA leader sequences were used to direct the secretion of the fusion protein to the periplasmic space of E. coli. In all three vectors, VH-CH1 was produced as a fusion to the gIII protein. The resulting Fab molecules were incorporated onto the surface of phage in the presence of helper phage in E. coli. To achieve expression of the soluble Fab, the gIII sequence was removed by SpeI and XbaI digestion. These two restriction enzymes generated compatible ends that could be ligated to put a stop codon inframe with VH-CH1. The VH and VL of a human Fab that binds to tetanus toxoid were cloned into the vectors that allowed the heavy and light chains to be coexpressed and displayed. pFab1 contains VL-CL and VH-CH1-gIII genes driven by two Lac promoters in two separate expression cassettes. pFab2 contains one dicistronic expression cassette with two translation initiation sites with VHCH1-gIII before VL-CL. pFab3 was constructed from pFab2 with the VL-CL before the VH-CH1-gIII gene. Effect of vector design on Fab expression and display The relative efficiency of the three different expression vectors to display functional tetanus toxoid binding Fab on the surface of phage was compared by phage-ELISA analysis. Phagemid particles were rescued from equal numbers of XL1-Blue cells containing the aforementioned vectors by co-infecting with VCSM13 helper phage. Equal numbers of phage were incubated with either immobilized tetanus toxoid antigen or BSA. Binding phage was detected using an anti-M13 antibody and absorbance was measured using a colorimetric assay. Three independent experiments were conducted, each with samples in quadruplicate. Within an individual experiment, the means of the ELISA readings for tt and BSA were calculated (data not shown). The mean BSA reading was subtracted from the mean tt reading to obtain a background-corrected tt reading. The corrected readings were proportional to the number of phage that displays tt-Fab. The background-corrected tt readings for pFab2tt and pFab3tt were divided by the background-corrected tt reading for pFab1tt to calculate pFab1tt-relative absorbances.
Fig. 2. Comparison of tt-antigen binding by tt-displayed phage produced from three different vectors using phage-ELISA analysis. Data are displayed as absorbance relative to pFab1tt. A non-related angioFab was used as negative control.
Fig. 2 shows the ELISA-relative absorbances of the different vectors in relation to pFab1tt. Each data point represents the average of quadruplicate samples. The ELISA-relative absorbance of pFab3tt was, on average, twofold higher than pFab1tt and pFab2tt. These data suggest that the pFab3 vector is more effective than the pFab1 and pFab2 vectors in phage display of Fabs. The efficiency of Fab display is a critical factor in Fab identification since it directly affects not only the numbers of phage that will have Fab molecules on their surface but also the copy number of Fab per phage particle. The improved display efficiency of pFab3 phagemid would be expected to facilitate the isolation of target-specific Fab binders using phage display, including those Fab molecules that bind with lower affinity. We compared all three vectors, pFab1, pFab2, and pFab3, for levels of tt-Fab protein production. XL1-Blue cells were transformed with the expression plasmids, single colonies were selected and grown at 37 °C, and Fab expression was induced with IPTG. Fab produced inside cells and soluble in salt buffer was identified as lysate Fab. Soluble Fab that was secreted into the culture medium was identified as secreted Fab. Equal amounts of soluble cellular lysates (10 lg) and culture media that was adjusted to the cell paste weight were resolved by SDS– PAGE, transferred to a nitrocellulose membrane, and probed with anti-kappa antibody. The recombinant protein was visualized as bands of 49.4 kDa (assembled Fab) and 23.4 kDa (free light chain) (Fig. 3A). Protein loading was monitored by Coomassie blue staining (Fig. 3B). The highest level of assembled Fab was produced in secreted form using the pFab3 vector and showed an 8-fold increase in the production level as measured by analysis of the Western blot signals (data not shown). Using tt-Fab as a model molecule, we demonstrated that pFab3 vector most effectively expressed and
S. Corisdeo, B. Wang / Protein Expression and Purification 34 (2004) 270–279
277
Fig. 3. Comparison of the effect of vector design on the production levels of tt-Fab as secreted and cell lysate products in E. coli. Three different vectors (pFab1, pFab2, and pFab3) were constructed and tested for tt-Fab production in XL1-Blue cells with IPTG induction. Ten micrograms of soluble cell lysate (lys) and concentrated medium (sec) adjusted to the cell paste weight was separated by non-reducing SDS–PAGE. (A) Western blot detection of the assembled Fab molecule (49.4 kDa) and free light chain (23.4 kDa) by HRP-labeled anti-human kappa chain antibody. (B) Coomassie stained gel demonstrating protein loading.
displayed Fab in E. coli. Since we employed the same lacZ promoter in all constructs and expressed and displayed the same tt-Fab in the same XL1-Blue E. coli strain under the same culture conditions, the transcription efficiency and protein stability should not be contributing factors. The difference in expression and display is likely due to the variable efficiency in protein translation reflecting the different secondary structures of transcribed RNAs due to the differences in sequences. Other factors influencing tt-Fab expression Cell strain and culture temperature Separate fragments of heavy and light chains were produced in the cytoplasm of E. coli and secreted into the periplasm. The heavy and light chain assembly takes place in the oxidizing environment of the periplasm yielding the heterodimer Fab. Fig. 4 shows that a large portion of heavy and light chains are not assembled. The yield of Fab expression could potentially be increased through the enhancement of Fab assembly either in vivo or in vitro.
Fig. 4. Western blot analysis of free light chain, free heavy chain, and assembled tt-Fab produced using the pFab4 vector in XL1-Blue cells. Ten micrograms of soluble cell lysate and concentrated medium adjusted to the cell paste weight was separated by non-reducing SDS–PAGE. (A) Detection using anti-gamma antibody of free heavy chain (26.1 kDa) and assembled Fab (49.4 kDa). (B) Detection using anti-kappa antibody of free light chain (23.4 kDa) and assembled Fab (49.4 kDa).
Fig. 5. Comparison of Fab levels produced as soluble cell lysate and secreted products in different bacterial strains. pFab3 vector was used to produce tt-Fab in E. coli strains JM10, DM1, TOP10F0 , XL1-Blue, and BL21. Ten micrograms of soluble cell lysate (lanes 7–12) and concentrated medium adjusted to the cell paste weight (lanes 1–6) was separated by non-reducing SDS–PAGE. (A) Western blot analysis of assembled Fab molecule (49.4 kDa) by anti-kappa antibody. (B) Coomassie stained gel demonstrating protein loading.
The efficiency of Fab production also varies between different cell strains and under different growth condition. Five strains of E. coli (JM105, DM1, Top10F, XL1-Blue, and BL21) were transformed with the pFab3tt plasmid; their total cell lysates (10 lg) and culture supernatants (adjusted to cell paste weight) were assayed by Western blot analysis (Fig. 5A). The Kodak image station analysis showed 12- to 13-fold more assembled Fab produced in the cell lysates of both the JM105 and BL21 bacterial strains over the other three strains (data not shown). The highest levels of Fab produced were recovered from the supernatant of JM105 and BL21 strains. A bacterial protein has recently been identified that is thought to facilitate the
278
S. Corisdeo, B. Wang / Protein Expression and Purification 34 (2004) 270–279
Fig. 6. Comparison of the effect of different growth temperatures during induction on Fab produced in JM105 bacterial cells. IPTG was added at 0.1 mM concentration to cultures grown overnight at 16, 25 or 37 °C. Ten micrograms of soluble cell lysate and concentrated medium adjusted to cell paste weight was separated by non-reducing SDS–PAGE. (A) Western blot detection of assembled Fab molecule (49.4 kDa) using anti-kappa antibody. (B) Coomassie stained gel demonstrating protein loading.
expression of antibody fragments in E. coli [15]. This endogenous protein may be expressed at different levels in different E. coli strains and may be one possible explanation for strain-specific differences. Lowering the bacterial growth temperature has been shown to decrease periplasmic aggregation and increase the yield of soluble antibody protein [16]. To test this, we incubated bacteria in shake-flasks with LB medium with 0.1 mM IPTG at 16, 25, and 37 °C. Ten micrograms of soluble cell lysate and concentrated medium adjusted to cell paste weight was separated by SDS–PAGE and Western blot detection was performed using anti-kappa antibody (Fig. 6A). Under these growth conditions, no significant effect was found on the production of lysate Fab and much more secreted Fab was produced at higher temperature. The majority of assembled Fab protein was recovered in the medium of 37 °C culture rather than the cell lysate. Culture conditions have been shown to modulate the relative distribution of antibody fragments between the periplasmic space and the medium [17], and conservative amino acid substitutions within the recombinant antibody fragments have resulted in the secretion of the
antibody rather than its retention in the periplasm [10]. We noticed severe bacterial lysis when incubating cultures expressing tt Fab at 37 °C after IPTG induction. Coomassie staining also showed a lysate-like staining pattern in the 37 °C lane of the secreted material even when equal volumes of culture were loaded (Fig. 6B). Our data suggested that, for certain Fabs, high level of Fab production at high temperature is toxic to cells and causes a significant degree of Fab release into the culture medium due to cell lysis. The highest production can be achieved by growing cells to high density in the presence of repressor and inducing protein expression at high temperature. IPTG Concentration Increasing concentration of IPTG (from 0.02 to 2 mM) had little effect on Fab production in our system (Fig. 7A). All inductions were performed at 37 °C using the pFab3 vector to produce tt-Fab in JM105 cells. Protein was loaded as previously described and Western blot detection was performed with the anti-kappa antibody. Expression using 0.02 mM IPTG induction yielded similar levels of Fab as that obtained using 2 mM IPTG.
Fig. 7. Comparison of Fab production under different conditions of induction. All inductions were performed at 37 °C using the pFab4 vector to produce tt-Fab in JM105 cells. Ten micrograms of soluble cell lysate and concentrated medium adjusted to cell paste weight was separated by nonreducing SDS–PAGE. (A) Western blot detection of assembled Fab molecule (49.4 kDa) using anti-kappa antibody. (B) Coomassie stained gel demonstrating protein loading.
S. Corisdeo, B. Wang / Protein Expression and Purification 34 (2004) 270–279
279
References
Fig. 8. Yield of secreted tt-Fab produced under optimized conditions: high density of JM105 cells using expression vector pFab3 in LB medium induced with 0.1 mM IPTG for 16 h at 37 °C. (A) Western blot detection of known concentrations of commercial Fab and dilutions of culture sample using anti-kappa antibody. (B) Standard curve representing the relationship between the quantity of Fab standard and densitometry reading. The densitometry reading of the experimental sample was located on the standard curve and protein quantity was determined.
[1] A. Skerra, A. Pluckthun, Assembly of a functional immunoglobulin Fv fragment in Escherichia coli, Science 240 (1988) 1038– 1041. [2] H.R. Hoogenboom, A.D. Griffiths, K.S. Johnson, P. Hudson, G. Winter, Multisubunit proteins on the surface of filamentous phage: methodologies for displaying antibody (Fab) heavy and light chains, Nucleic Acids Res. 19 (1991) 4133–4137. [3] C.F. Barbas III, A.S. Kang, R.A. Lerner, S.J. Benkovic, Assembly of combinatorial antibody libraries on phage surfaces: the gene III site, Proc. Natl. Acad. Sci. USA 213 (1991) 201–212. [4] H. Gram, L. Marconi, C.F. Barbas III, T.A. Collet, R.A. Lerner, A.S. Kang, In vitro selection and affinity maturation of antibodies from na€ıve combinatorial immunoglobulin library, Proc. Natl. Acad. Sci. USA 89 (1992) 3576–3580. [5] P. Chames, S.E. Hofton, P.G. Coulie, B. Uchanska-Ziegler, H.R. Hoogenboom, Direct selection of a human antibody fragment directed against the tumor T-cell epitope HLA-A1-MAGE-A1 from a nonimmunized phage-Fab library, Proc. Natl. Acad. Sci. USA 97 (2000) 7969–7974. [6] I. Moreno de Alboran, C. Martinez-alonso, C.F. Barbas III, D.R. Burton, H.J. Ditzel, Human monoclonal Fab fragments specific for viral antigens from combinatorial IgA libraries, Immunotechnology 1 (1995) 21–28. [7] W.P. Yang, K. Green, S. Pinz-Sweeney, A.T. Briones, D.R. Burton, C.F. Barbas III, CDR walking mutagenesis for the affinity maturation of a potent human anti-HIV-1 antibody into the picomolar range, J. Mol. Biol. 254 (1995) 392–403. [8] R. Burioni, P. Plaisant, F. Bugli, V.D. Carri, M. Clementi, G. Fadda, A vector for the expression of recombinant monoclonal Fab fragments in bacteria, J. Immunol. Methods 217 (1998) 195– 199. [9] M. Foti, F. Granucci, P. Ricciardi-Castagnoli, A. Spreafico, M. Ackermann, M. Suter, Rabbit monoclonal Fab derived from a phage display library, J. Immunol. Methods 213 (1998) 201–212. [10] G. Forsberg, M. Forsgren, M. Jaki, M. Norin, C. Sterky, A.
Yield of Fab production under optimized conditions We identified optimized conditions for tt-Fab production through testing different vector designs, bacterial strains, and culture and induction conditions. The optimized conditions are as follows: high density of JM105 cells containing the pFab3tt vector in LB medium plus 0.1 mM IPTG induction for 16 h at 37 °C. The Fab production under these conditions was estimated by Western blot analysis using known concentrations of commercial Fab as a quantitation standard (Fig. 8A). The assembled Fab was detected using anti-kappa antibody and the Western blot signals were quantitated using a Kodak image station. The quantity of the assembled Fab standard was adjusted from the known protein quantity that was loaded on the gel using the densitometry reading of all bands within the lane (calculation not shown). A standard curve was established between the quantity of the assembled Fab standard and the densitometry reading (Fig. 8B). The densitometry reading of the experimental sample was located on the standard curve and its corresponding protein quantity was determined. Under optimized conditions, the production level of assembled tt-Fab was approximately 16 mg/L.
[11]
[12]
[13]
[14]
[15]
[16]
[17]
Enhorning, K. Larsson, M. Ericsson, P. Bj€ ork, Identification of framework residues in a secreted recombinant antibody fragment that control production level and localization in Escherichia coli, J. Biol. Chem. 272 (1997) 12430–12436. R. Levy, R. Weiss, G. Chen, B.L. Iverson, G. Georgiou, Production of correctly folded Fab antibody fragment in the cytoplasm of Escherichia coli trxB gor mutants via the coexpression of molecular chaperones, Protein Expr. Purif 23 (2001) 338– 347. P.M. OÕBrien, G. Maxwell, M.S. Campo, Bacterial expression and purification of recombinant bovine Fab fragments, Protein Express. Purif. 24 (2002) 43–50. S. Kumar, J. Kalsi, C.T. Ravirajan, A. Rahman, D. Athwal, D.S. Latchman, D.A. Isenber, L.H. Pearl, Molecular cloning and expression of the Fabs of human autoantibodies in Escherichia coli, J. Biol. Chem. 275 (2000) 35129–35136. D.S. Gill, Y.W. Wong, M.N. Margolies, Differences in sequencespecific expression of two anti-arsonate Fabs in E. coli, Biotechnol. Prog. 13 (1997) 692–694. H. Bothmann, A. Pluckthun, Selection for periplasmic factor improving phage display and functional periplasmic expression, Nat. Biotechnol. 16 (1998) 376–380. A. Skerra, A. Pluckthum, Secretion and in vivo folding of the Fab fragment of the antibody McPC603 in Escherichia coli: influence of disulfides and cis-prolines, Protein Eng. 4 (1991) 971–976. S.M. Kipriyanov, G. Moldenhauer, M. Little, High-level production of soluble single chain antibodies in small-scale Escherichia coli cultures, J. Immunol. Methods 200 (1997) 69–77.