Generation and analysis of the improved human HAL9/10 antibody phage display libraries
- Jonas Kügler1, 2,
- Sonja Wilke2,
- Doris Meier1,
- Florian Tomszak1,
- André Frenzel1, 3,
- Thomas Schirrmann1, 3,
- Stefan Dübel1,
- Henk Garritsen4, 5,
- Björn Hock6,
- Lars Toleikis6,
- Mark Schütte†6 and
- Michael Hust†1Email author
© Kügler et al.; licensee BioMed Central. 2015
Received: 11 November 2014
Accepted: 9 February 2015
Published: 19 February 2015
Antibody phage display is a proven key technology that allows the generation of human antibodies for diagnostics and therapy. From naive antibody gene libraries - in theory - antibodies against any target can be selected. Here we describe the design, construction and characterization of an optimized antibody phage display library.
The naive antibody gene libraries HAL9 and HAL10, with a combined theoretical diversity of 1.5×1010 independent clones, were constructed from 98 healthy donors using improved phage display vectors. In detail, most common phagemids employed for antibody phage display are using a combined His/Myc tag for detection and purification. We show that changing the tag order to Myc/His improved the production of soluble antibodies, but did not affect antibody phage display. For several published antibody libraries, the selected number of kappa scFvs were lower compared to lambda scFvs, probably due to a lower kappa scFv or Fab expression rate. Deletion of a phenylalanine at the end of the CL linker sequence in our new phagemid design increased scFv production rate and frequency of selected kappa antibodies significantly. The HAL libraries and 834 antibodies selected against 121 targets were analyzed regarding the used germline V-genes, used V-gene combinations and CDR-H3/-L3 length and composition. The amino acid diversity and distribution in the CDR-H3 of the initial library was retrieved in the CDR-H3 of selected antibodies showing that all CDR-H3 amino acids occurring in the human antibody repertoire can be functionally used and is not biased by E. coli expression or phage selection. Further, the data underline the importance of CDR length variations.
The highly diverse universal antibody gene libraries HAL9/10 were constructed using an optimized scFv phagemid vector design. Analysis of selected antibodies revealed that the complete amino acid diversity in the CDR-H3 was also found in selected scFvs showing the functionality of the naive CDR-H3 diversity.
KeywordsscFv Phage display Antibody engineering Library Panning Screening
Since the inception of antibody technology twenty years ago, phage display is a powerful tool to generate antibodies for proteome research [1-4], diagnostics [5-8] or for therapeutic purposes [9-11]. Therapeutic antibodies are currently one of the fastest developing class of biologicals in the pharmaceutical market . The main indications for therapeutic antibodies are cancer and auto-immune diseases [13,14]. To date, 44 antibodies and antibody conjugates are EMA and/or FDA approved (status autumn 2014) (http://www.imgt.org/mAb-DB/index) and about 350 antibodies and antibody fusion proteins were under development in 2013 . Two major strategies for generating fully human antibodies are: transgenic mice and antibody phage display. In transgenic mice, the chromosome segments encoding antibody gene fragments are replaced with the corresponding human chromosome segments encoding human immunoglobulins. These animals allow the generation of fully human antibodies by hybridoma technology [16-18]. An advantage of transgenic mice is the in vivo affinity maturation of antibodies, but on the other hand, all in vivo antibody generations are restricted by the natural immune system itself: The limitation in antigen processing and presentation and the tolerance against conserved epitopes . Antibody phage display is an alternative or complementing technology to generate human antibody fragments from universal antibody gene libraries as lead candidates for therapeutic development [17,20-22]. Here, the selection is an in vitro process and is not limited by the restrictions of the immune system and selection conditions can be adjusted and controlled, thus allowing to select for properties not achievable by in vivo immune systems . To isolate human antibodies by phage display, two types of antibody gene libraries are used: immune libraries and universal or “single-pot” libraries [24,25]. Immune libraries from patients are suited to select specific antibodies against a disease or pathogen, e.g. cancer [26,27], human immunodeficiency virus or herpes simplex virus . “Single-pot” libraries allow the selection of antibodies - in theory - against any target. The human naive antibody gene libraries HAL4/7/8 are “single-pot” libraries. Antibodies against a panel of different antigens were selected from these HAL libraries and applied for different purposes, e.g. [8,30-35]. Antibody fragments from these libraries can directly be cloned into a selection of compatible expression vectors to produce e.g. in vivo biotinylated antibodies , scFv-Fc  or full IgG (Frenzel et al. unpublished). The scFv-Fc format (Yumab) is an alternative, functionally identical to IgG in most assays. Due to its quicker and easier production, it provides a robust format for screening of large numbers of antibody candidates, and can be converted to full IgG afterwards.
In this work, the scFv phagemid vector design was optimized and the “single-pot” antibody gene libraries HAL9/10 were constructed and analyzed, demonstrating significant improvements over previous designs.
Construction of phage display vectors
The phage display vector pHAL30 was constructed by cloning a DNA fragment encoding His-/C-Myc tag flanked by NotI and BamHI, which was generated with two PCR primers (MHMycHIS-NotI_f:5′ CGCGTGCGGCCGCAGGTTCTGAACAAAAGCTGATCTC 3′; MHMycHisBamHi_r: 5′ CGCGTGGATCCCTAATGATGATGGTGATGATGGG 3′) into pHAL14. ScFv coding sequences from pHAL14 were cloned in pHAL30 using the restriction sites NcoI and NotI. The C-terminal phenylalanine of the VК was deleted from scFv coding sequences by PCR using the primer set MkpelB_f (5′ GCCTACGGCAGCCGCTGG 3′) and MhkappaCLscFv-NotI_r2 (5′ ACCGCCTCCGCGGCCGCGACAGATGGTGCAGCCACAGT 3′). The PCR products were cloned into the vector pHAL14  and pHAL30 between the vector encoded signal peptide and the c-Myc/ His-tag or His/ c-Myc-tag sequences, respectively, using the restriction sites NcoI and NotI.
Production of soluble antibodies in microtitre plates
96-well MTPs with polypropylene (PP) wells (U96 PP 0,5 mL, Greiner, Frickenhausen, Germany) containing 150 μL 2 × YT-GA  were inoculated with the bacteria bearing scFv expressing phagemids. MTPs were incubated overnight at 37°C at 800 rpm in a MTP shaker (Thermoshaker PST-60HL-4, Lab4You, Berlin, Germany). A volume of 180 μL phosphate-buffered 2 × YT-GA in a PP-MTP well was inoculated with 10 μL of the overnight culture and grown at 37°C and 800 rpm for 2 h. Bacteria were harvested by centrifugation for 10 min at 3,220 × g and 180 μL supernatants were removed. The pellets were resuspended in 180 μL 2 × YT supplemented with 100 μg/mL ampicillin, 100 mM sucrose and 50 μM isopropyl-beta D thiogalacto pyranoside (IPTG) and incubated at 30°C and 800 rpm overnight. Bacteria were pelleted by centrifugation for 10 min at 3,220 × g and 4°C. The scFv-containing supernatant was transferred to a new PP-MTP and stored at 4°C before analysis.
Production of antibody phage
50 mL 2 × YT medium + 100 μg/mL ampicillin + 100 mM glucose were inoculated with an overnight culture to O.D.600 = 0.01. Bacteria were grown to O.D.600 = 0.4–0.5 at 37°C and 250 rpm. 2 mL bacteria (∼1 × 109 bacteria) were infected with 2 × 1010 helper phage Hyperphage [38,39], incubated at 37°C for 30 min without shaking, followed by 30 min at 250 rpm. Infected cells were harvested by centrifugation for 10 min at 3,220 × g and the pellet was resuspended in 30 mL 2 × YT + 100 μg/mL ampicillin + 50 μg/mL kanamycin. Phage were produced at 30°C and 250 rpm for 16 h. Cells were pelleted for 10 min at 3,220 × g. Phage in supernatant were precipitated with 1/5 volume of 20% PEG/2.5 M NaCl solution for 1 h on ice and pelleted by centrifugation 1 h at 3,220 × g at 4°C. Precipitated phage were resuspended in 300 μL phage dilution buffer, and cell debris was pelleted by additional centrifugation for 5 min at 15,400 × g at 20°C. The supernatant containing the scFv phage was stored at 4°C. Phage titration (cfu) was done according to .
100 ng of antigen was coated to 96 well microtitre plates (MaxiSorp, Nunc) in PBS pH 7.4 or 50 mM NaHCO3 pH 9.6 overnight at 4°C. After coating, the wells were washed three times with PBST and blocked with 2%MPBST for 1.5 h at RT, followed by three washing steps with PBST. 108 cfu scFv-phage and helperphage as control were loaded in the first well of each line of the plate. Each sample was diluted 1:2 in PBS along the plate ending with 100 μL proteinaceous solution per well. The loaded plate was incubated for 1.5 h at RT, followed by three PBST washing cycles. Bound scFv phage were detected with mouse anti-M13 HRP conjugate (Amersham Biosciences, Freiburg, Germany) (1:5,000 diluted in 2%MPBST). Visualization was performed using TMB (3,3′,5,5′-tetramethylbenzidine) as a substrate and the staining reaction was stopped by adding 100 μL 1 N sulphuric acid. Absorbance at 450 nm (620 nm reference) was measured by using a SUNRISE™ microtitre plate reader (Tecan, Crailsheim, Germany).
Antigens were analyzed by 12% SDS-PAGE using a Protean II Minigel system (BioRad Inc, München, Germany) according to . Protein gels were stained with coomassie brilliant blue.
Western blotting and immunostaining
Protein samples in SDS polyacrylamide gels were transferred to and immobilized on polyvinylidene difluoride (PVDF) membranes (Roth, Karlsruhe) by a semidry procedure (Biorad, München). The membrane was blocked with 2% (w/v) skimmed milk powder in PBST for 1 h at RT. ScFvs were detected by mouse anti c-myc mAb Myc1-9E10 (1:1,000 diluted in 2% MPBST) (Yumab, Braunschweig) for 1.5 h at RT and goat anti-mouse HRP mAb (A0168, Sigma, Taufkirchen) (1:3,200 diluted in 2%MPBST) for 1.5 h at RT. For visualization Super Signal® WestPico Chemiluminescence Substrate (Thermo Scientific, Bonn) was applied and protein bands were detected by enhanced chemiluminescence (ECL) on a luminometer (Biorad, München). For detection of antibody phage, the minor coat protein pIII was detected with 1:2,000 diluted mouse mAb anti-pIII (Mobitec, Göttingen) for 1.5 h at RT, followed by goat anti-mouse (Fc specific) mAb (Sigma, Taufkirchen) conjugated with AP (1:10,000 diluted). The development was performed by NBT/BCIP.
The study was performed in accordance with the Declaration of Helsinki. The study participants were selected from blood donors of the Institute for Clinical Transfusion medicine, Städtisches Klinikum Braunschweig gGmbH, Braunschweig, Germany. All voluntary donors were informed about the project and gave their informed consent. The use of blood samples for the amplification of antibody gene fragments to develop antibody phage display libraries was approved by the ethical commitee of the Technische Universität Braunschweig (Ethik-Kommission der Fakultät 2 der TU Braunschweig, approval number DM-2014-08).
Lymphocytes were isolated from 54 donors with various ethnical backgrounds using the Lymphoprep Kit (Progen, Heidelberg) according to the manufacturers instructions. Total RNA was isolated using Trizol (Invitrogen, Karlsruhe) and mRNA was isolated using the Oligotex mRNA Minikit (Qiagen, Hilden). Then cDNA was synthesized using Superscript III (Invitrogen), random hexamer oligonucleotide primers and 50–250 ng mRNA.
Production of soluble scFv antibody fragments randomly chosen from the HAL8 antibody gene library
For VL cloning 5 μg vector pHAL30 and 2 μg kappa or lambda PCR products were digested in a volume of 100 μL using 30 U MluI and NotI (NEB, Frankfurt) at 37°C for 3 h. This step was performed multiple times to produce enough material for up to 50 transformations. For quality control each digestion was performed separately with each enzyme in parallel. The digestions were inactivated at 80°C for 20 min, followed by adding 0.5 U calf intestine phosphatase (CIP) (MBI Fermentas, St. Leon-Rot) and incubation at 37°C for 30 min, this step was repeated once. Vector and PCR products were purified using the NucleoSpin Gel and PCR Clean-up Kit, removing short stuffer fragments between MluI and NotI in pHAL30. Each ligation was performed with 1 μg digested vector and 270 ng PCR product using 3 U T4 ligase (Promega, Mannheim) in a volume of 100 μL overnight at 16°C. Ligations were inactivated at 70°C for 10 min and precipitated with 10 μL 3 M sodium acetate pH 5.2 and 250 μL ethanol for 2 min at RT, followed by 5 min centrifugation at 16000 × g and 4°C. Pellet was washed two times with 500 μL 70% (v/v) ethanol. The dried pellet was resuspended in 35 μL dH2O and mixed on ice with 25 μL electrocompetent E. coli XL1-Blue MRF’. Electroporation was performed with 0.1 cm prechilled cuvettes using a MicroPulser electroporator (BioRad, München) at 1.7 kV. Immediately, 1 mL prewarmed 37°C SOC medium pH 7.0 (2% (w/v) tryptone, 0.5% (w/v) yeast extract, 0.05% (w/v) NaCl, 20 mM Mg solution, 20 mM glucose) was added and incubated for 1 h at 600 rpm and 37°C. Transformation was plated out on 25 cm square Petri dishes with 2xYT-GAT agar (1,6% (w/v) Tryptone, 1% (w/v) yeast extract, 0,5% (w/v) NaCl, 100 mM glucose, 100 μg/mL ampiciline, 20 μg/mL tetracycline, 1.5% (w/v) agar-agar) and incubated overnight at 37°C. In parallel 10−6 dilutions of each transformation reaction were plated out to determine the transformation rate. These plates were used also for quality control by colony PCR using the oligonucleotide primers MHLacZ-Pro_f and MHgIII_r to verify the insert size. The colonies were scraped from the 25 cm plates using 40 mL 2xYT medium (1,6% (w/v) Tryptone, 1% (w/v) yeast extract, 0,5% (w/v) NaCl) and 5 mL were directly used for midi plasmid preparation using the Nucleobond Plasmid Midi Kit (Macherey-Nagel).
For VH cloning material from 44 donors of the previous described HAL7/8 libraries  was used in addition to material of 54 new donors. In a first step 5 μg pHAL30-VL sublibraries and 2 μg VH PCR products were digested in a volume of 100 μL using 30 U NcoI-HF and 30 U HindIII-HF (NEB) at 37°C for 3 h. This step was performed several times to produce enough material for 150–200 transformations. The digestions, ligations and transformations were performed as described for VL cloning with following modifications. Ligation was performed with 250 ng VH PCR product. After transformation of XL1-Blue MRF’, overnight incubation on 25 cm 2 × YT-GAT plates and resuspension in 25 mL 2 × YT medium, glycerol stocks of each library were made using 800 μL bacteria solution (~1010 bacteria) and 200 μL glycerol (Roth, Karlsruhe) and stored at −80°C. For the final library one glycerol stock of each sublibrary was thawed, mixed together, aliquoted (1 mL) and stored at −80°C.
The antibody gene libraries HAL9/10 were packaged separately inoculating 400 mL 2 × YT-GA (2 × YT containing 100 mg/mL ampicillin and 100 mM glucose) with a library glycerol stock. The bacteria were grown to an optical density at 600 nm (OD600) of 0.4–0.5 at 37°C and 250 rpm. 25 mL bacterial culture (~1.25 × 1010 cells) were infected with 2.5 × 1011 Hyperphage particles , incubated at 37°C for 30 min without shaking followed by 30 min at 250 rpm. Infected cells were harvested by centrifugation for 10 min at 3220xg. The pellet was resuspended in 400 mL 2 × YT-AK (2 × YT containing 100 mg/mL ampicillin and 50 mg/mL kanamycin). Phage were produced at 30°C and 250 rpm over night. The bacteria were centrifuged for 20 min at 10,000 × g. Phage particles in the supernatant were precipitated with 1/5 volume of 20% (w/v) polyethylene glycol (PEG)/2.5 M NaCl solution for one hour on ice with gentle shaking and pelleted by centrifugation for one hour at 10,000 × g at 4°C. The precipitated phage were resuspended in 10 mL phage dilution buffer (10 mM TrisHCl pH 7.5, 20 mM NaCl, 2 mM EDTA), 1/5 volume of PEG solution was added and incubated on ice for 20 min, followed by centrifugation for 30 min at 10,000 × g and 4°C. The pellet was resupended in 1 mL phage dilution buffer. Residual bacteria and cell debris were removed by additional centrifugation for 5 min at 16000 × g at 20°C. Supernatants containing antibody phage were stored at 4°C. Phage titration was done as described before . The scFv display rates of the packaged libraries were analyzed by 10% SDS-PAGE, Western blot and anti-pIII immunostain (mouse anti-pIII 1:2000, goat anti-mouse IgG AP conjugate 1:10000). Wildtype pIII has a calculated molecular mass of 42.5 kDa, but it runs at an apparent molecular mass of 65 kDa in SDS-PAGE . Accordingly, the scFv::pIII fusion protein runs at about 95 kDa.
Antibody selection (panning and screening)
Antibody selections were performed against proteins or peptides as described before . The elution with trypsin - pHAL30 and the former vector pHAL14 contain a trypsin site between the tags and gIII - is more efficient as elution by pH shift (Schirrmann, unpublished results).
Analysis of the new phage display vector pHAL30
Next, four different pHAL14 and pHAL30 phagemids were packaged with Hyperphage. The resulting phage titers were comparable among each other. They ranged from 6 × 109 to 6 × 1010 cfu/ mL culture medium. The scFv display level on phage was identical, as confirmed by Western immunoblot (data not shown). Approximately 90% of g3p was fused to the scFv antibody fragments.
A functional comparison of pHAL14 and pHAL30 scFv-phage regarding antigen binding was performed by titration antigen ELISA (Figure 2B). The titration curves were similar for three different scFv-phage, further illustrating similar display level. For TM43-E10 more functional scFv were displayed on phage. Taken together, both phagemids are almost identical in phage display level and phage binding assays, but the pHAL30 phagemid provided a significant increase in production of soluble antibody fragments for screening which may improve the hit rate in initial screening assays.
Influence of the phenylalanine residue in kappa CL on scFv phage display
Target independent, it was observed that scFv antibody fragments bearing a VК light chain were less frequently isolated from HAL7/8 than those with a Vλ light chain [31,43]. A possible explanation could be a lower expression of antibody fragments with a VК light chain. Therefore, the phagemid was further optimized by deleting the bulky C-terminal phenylalanine of the VК light chain. First, the soluble production level of five known scFv fragments in pHAL14 and six known kappa scFv fragments in pHAL30 was analyzed and compared by Western immunoblot (data not shown). The deletion of the phenylalanine in the VК light chain resulted in a higher production level for all tested scFv. Further expression analysis by Western immunblot included the comparison of 32 randomly chosen scFv fragments from the kappa HAL8 antibody gene library and their phenylalanine deletion version (Table 1). The expression levels were higher with the phenylalanine deletion mutants in 14 cases, slightly lower for three clones and in the same range for the remaining 15 clones.
Construction of HAL9 and HAL10
Using blood samples obtained from 98 Caucasian, African, Indian and Chinese non-immunized human donors, the new antibody libraries HAL9 and HAL10 were constructed using the improved phagemid backbone pHAL30. In detail, blood lymphocytes from 54 Caucasian, African, Indian and Chinese non-immunized human donors were isolated for library construction. Lymphocyte mRNA was reverse transcribed to cDNA and the antibody gene repertoire was amplified in two steps. In the first step, VH and the full length light chains were amplified using a set of framework 1 forward primers and IgM, kappa and lambda constant domain reverse primers . In the second step, specific restriction sites were incorporated for cloning into the pHAL30 phagemid vector. For the second amplification step of VH and lambda subfamilies the DNA antibody repertoire (1st PCR material) from 44 donors of the previously constructed HAL7 was added for construction of the new HAL9 and HAL10 libraries. Both libraries contain the same VH repertoire (from 98 donors) but differ in their light chain repertoire. HAL9 includes all lambda subfamilies (from 98 donors), HAL10 all kappa subfamilies (from 54 donors) except the pseudogene encoding Vκ7. Cloning of the antibody genes was done in two steps. First, light chain genes were cloned with 1 × 109 independent clones for each library. Then heavy chain genes were cloned in the light chain containing phagemid leading to a final repertoire size of the HAL9 and HAL10 scFv libraries of 1.04 × 1010 and 4.45 × 109, respectively.
Analysis of V-gene subfamily distribution
24 of the 33 functional Vλ genes were found in the HAL9 library sample (Figure 4). One Vλ gene (IGLV1-50) was found in the unselected HAL9 as well in selected scFvs that is predicted to be not functional . As seen for the heavy chain repertoire, the frequency of several Vλ germline genes changed during the selection process. Usage of the Vλ1 subfamily was increased after selection mostly due to a higher abundance of the IGLV1-47 germline gene, increasing from 5.8% in the unselected library to 14.5% after selection and a higher abundance of IGLV1-44, increasing from 12.3% to 15.5% (Figure 3C and Figure 4). In the Vλ2 subfamily, the frequency of the IGLV2-14 germline gene showed a slight increase from 5.3% to 8.4% after selection but the overall abundance of Vλ2 genes decreased due to a lower usage of the remaining germline genes. Three of nine different Vλ3 germline genes were retrieved in high frequency before and after selection: IGLV3-1, IGLV3-19 and IGLV2-21. The findings reflect the in vivo subfamily distribution, where IGLV3-1 and IGLV2-21 account for ~80% of Vλ3 . After selection the frequency of IGLV3-1 was reduced from 9.2% to 5.3% and increased for IGLV3-21 from 9.4% to 15.1%. Usage of IGLV3-19 remained constant. The abundance of the most prevalent Vλ gene IGLV6-57 in the unselected HAL9 was decreased from 16.3% to 11.1% after selection.
In the HAL10 library 29 of 38 functional Vκ germline genes were observed (Figure 5). Two Vκ genes (IGKV1-NL1 and IGKV3-NL5) found in the unselected library were described to be only potentially functional . The most abundant Vκ germline gene before selection was IGKV4-1 with a frequency of ~25%. This is a more than twofold higher prevalence than found in vivo  and usage of IGKV4-1 after selection was reduced to ~4%. The frequency of Vκ1 subfamily genes was raised from 37% in the unselected library to 48% after selection (Figure 3B and Figure 5). Three genes IGKV1-12, IGKV1-D33 and IGKV1-D39 increased from 4.4 to 9.2%, 3.3 to 9.2% and from 12.9 to 13.8%, respectively.
VH and VL germline gene pairings
The abundance of heavy- and light-chain pairings that were observed in the unselected libraries and within selected antibodies are illustrated in Figure 4 for VH-Vλ combinations and in Figure 5 for VH-Vκ combinations. In the unselected libraries no particular VH-VL combination is dominating (Figure 4 and 5). The pairings are evenly distributed depending on their total abundance. VH-Vλ pairings were observed in frequencies between 0.1% and 1.8% in the unselected HAL9 with VH3-30/Vλ6-57 being the most abundant combination. In the unselected HAL10 pairing frequencies up to 2.8% were observed for VH-Vκ due to the high abundance of the IGKV4-1 germline gene. However, after 121 individual selections a clear preference for particular VH-VL pairings was detected. For VH-Vλ pairings the most prevalent combination was VH3-30/Vλ3-21 which was observed in 6.1% of the selected antibodies. The λ3-21 gene showed a preference for pairing with VH3 germline genes. The VH3-30 germline gene was by far the most abundant VH gene after selection and paired with nearly all light chain genes in high frequency. Interestingly, the light chain gene Vλ8-61 paired almost exclusively with VH3-30. The remaining Vλ germline genes showed preferential pairing with the VH1 and VH3 germline genes and with lower frequency also to the VH6-1 germline gene. A second VH3 subfamily gene, VH3-23, was also observed in high frequency after selection. Similar to VH3-30 no clear preference in pairing with any Vλ genes segment was seen. For three VH1 genes a more distinct preference for single Vλ genes was observed. The combinations VH1-46/Vλ1-47 and VH1-69/Vλ1-44 were enriched to 2.3% and the combination VH1-18/Vλ1-47 was enriched to 2.6%. The five most common VH-Vλ combinations (VH3-30/Vλ3-21, VH3-30/Vλ1-44, VH3-30/Vλ6-57, VH3-30/Vλ2-14 and VH3-30/Vλ3-14) account for nearly 20% of all selected antibodies.
An enrichment of the VH3-30 germline gene after selection was also observed for the VH-Vκ gene pairing. Furthermore a high frequency of Vκ1 and Vκ2 genes occurred after selection, but due to the relatively small sample size of selected antibodies bearing a kappa light chain no significant conclusion can be made about enrichment of single VH-Vκ combinations.
CDR3 length and amino acid distribution of light-chain and heavy-chain
For the CDR3-L length distribution of HAL9 (lambda light chain) before and after selection 776 sequences of the unselected library and 685 sequences of selected antibodies were analyzed (Figure 6B). In the unselected library the CDR3 length ranges from 8 to 14 aa with a median length of about 11 aa. The most common CDR3 length for lambda light chains are 9, 10 or 11 aa that account together for more than 80% of all found CDR3-L lengths. After selection, CDRs with a length of 11 aa were found more frequently and the occurrence of CDRs with a length of 9 aa was decreased. CDR3-L length distribution of HAL10 (kappa light chain) was analyzed with 461 sequences of the unselected library and 126 sequences of selected antibodies (Figure 6C). In the unselected library the CDR3 length ranges from 7 to 11 aa. Most kappa light chains are bearing a CDR3 with a length of 9 aa (~70%) and the length distribution did not change significantly after selection.
Since the development of phagemids for antibody phage display by Breitling et al. , the technology was continuously further developed.
The most used phage display vectors have the tag order His-Myc when using both tags, e.g. pCANTAB3his , pCES , pHEN2 , pHAL14 , pIT2 , pMod1  or pMID21 . To our knowledge, there is only one phage display vector using the opposite orientation: pSD3 . Comparing both tag orders, the Myc-His order (pHAL30) results in improved production of functional scFv compared to His-Myc (pHAL14) whereas the display of functional scFv on phage is not effected.
Using a kappa (HAL4 or HAL8) and a lambda (HAL7) antibody gene library in the same selection, we have observed that the output of selected kappa antibodies was always lower compared to lambda antibodies [31,43]. Løset et al. and Tiller et al. also describe a lower production yield of kappa scFvs or Fabs compared to lambda antibodies that confirms our observation with the HAL libraries [45,56]. The scFv design in pHAL14 included the first six amino acids of CH1 as part of the linker and the first 16 amino acids of CL lambda, respectively the first nine amino acids of CL kappa. The last amino acid in CL kappa used in the scFv design is a phenylalanine. Phenylalanine is encoded by only two codons which may lead to a reduced translation and has a bulky hydrophobic benzol side group which may hamper solubility. In projects with macaque immune libraries using also pHAL14 and derived pHAL vectors, the phenylalanine was omitted. The selected kappa scFv were produced well, leading to a generation of kappa scFv against different targets [8,57-62]. For human scFv, deletion of phenylalanine improved the expression rate of known kappa antibodies and randomly chosen kappa clones. This modification increased the rate of selected kappa antibodies from 4% using HAL4/7/8 to 15.6% when using HAL9/10.
Using the novel vector pHAL30 and scFv design, two new universal human naive antibody gene libraries were constructed. The theoretical diversity of the HAL9/10 libraries is with 1.5 × 1010 independent clones in the same range like other published scFv or Fab libraries with a theoretical diversity in the range of 1x1010-1x1011, e.g. the naive McCafferty scFv library (1.1x1010 independent clones) , the naive Pfizer scFv library (3.1 × 1010) , the naive CAT2.0 scFv library (1.29 × 1011) , the naive/synthetic Dyax Fab (3.5 × 1010)  library, the synthetic Ylanthia Fab library (1.3 × 1011)  or very recently published the PHILODiamond scFv library (4 × 1010) .
The representation of antibody subfamilies in the HAL9/10 libraries is comparable with the HAL7/8 library [31,43], it differs only in some minor points, e.g. VH4, Vκ4 and Vλ4 are more abundant in HAL9/10. In general, VH gene usage is dominated by VH3 and VH1. Dominant VL are usually Vκ1, Vκ3, Vκ4 and Vκ2 and Vλ by Vλ3, Vλ1, Vλ2 and Vλ6. In the new libraries, the distribution of V genes is comparable with their frequency in vivo  but with following exceptions. The subfamily VH4 is more abundant in vivo whereas Vκ4 and Vλ6 are more frequent in HAL antibody gene libraries and rare in vivo. The main difference is the frequency of Vλ1 > Vλ2 > Vλ3 in vivo compared to Vλ3 > Vλ1 > Vλ2 in HAL9/10. The Ylanthia library  fits perfectly with the in vivo distribution. The distribution of V genes for VH and Vκ of HAL9/10 is in accordance with the Pfizer library . A comparison of Vλ with HAL9 is not possible, because the Pfizer library lacks most Vλ genes. Compared to HAL9/10, the distribution of V genes in the CAT2.0 library  differs more from in vivo situation. Here, VH is dominated by VH1 and Vκ is dominated by one Vκ1 gene. With ~14 amino acids, the average HAL9/10 CDR-H3 length is longer than in the naive Pfizer  or synthetic Ylanthia  libraries (~11 amino acids), but slightly shorter compared to the Dyax library with ~15 amino acids . In addition more very long CDR-H3 (>20 amino acids) are found in HAL9/10 compared to the Pfizer library. The average length of CDR-L3 for λ is eleven amino acids and therefore two amino acids longer as the average length of CDR-L3 for κ. In HAL9/10, 19 or 20 different aa are occurring at each CDR-H3 position whereas for Ylanthia 13 to 17 different aa are possible at each position, planned by rational design . In Ylanthia, the CDR-H3 is dominated by serine, tyrosine and glycine (~40% frequency). In the naive HAL9/10 libraries, the amino acid repertoire in CDR-H3 is more balanced, but the frequency of tyrosine increases from N’terminus to C’terminus as well (aa 95 - 100i).
Significantly, within the group of selected scFvs, the representation of VH antibody subfamilies is largely in accordance with the McCafferty library  and Pfizer library , but not with the CAT2.0 library , where VH is dominated by VH1 > VH3 compared to VH3 > VH1 for McCafferty, HAL9/10 or Pfizer library. For Vκ, more Vκ1 than Vκ3 were selected from HAL10 or McCafferty, whereas from Pfizer or CAT2.0 mostly Vκ3 were selected. Regarding Vλ, here Vλ1-3 and 6 are mainly isolated from McCaffery and HAL9, whereas Vλ6 is rarely selected from CAT2.0. The frequency of selected VH4, Vκ4 and Vλ7 antibodies from HAL9/10 is very low compared with the frequency in the initial library. The same was observed for HAL7/8  and CAT2.0 . The VH4-34 antibody gene is described as toxic for B-cells . As consequence, two VH4-34 and VH4-59 genes were excluded when constructing the synthetic Ylanthia library, despite the fact, that these VH4 genes are very frequent in the natural human antibody repertoire . Remarkably, a majority of antibodies selected from macaque immune antibody gene libraries against botulinum A toxin are VH4 antibodies, especially VH4-59 . The macaque antibody genes are very similar to their human counterparts. Therefore, the VH4 gene family may be preadapted to some targets, or at least best suited to structurally complement the respective epitopes.
The average CDR length of CDR-H3 is only very slightly shorter after selection compared with the initial library, presumably due to a better growth of bacteria producing shorter antibody genes during antibody selection. The length of the CDR-L3 does not significantly differ in selected antibodies. The CDR composition in the CDR-L3 is completely different between λ and κ light chains, giving a hint why two different kinds of light chains have been evolved. Very importantly, the frequency of amino acids at each CDR3 position does not significantly change after selection, clearly demonstrating that any bacterial expression bias is minor, and all amino acids used in the naive human gene repertoire can be found in selected, functional antibodies.
Some VH-genes are preferred after selection, e.g. VH1-18, VH1-46, VH3-23, VH3-30. These are also V-genes often used in Ylanthia  and CAT2.0 . Notably, the very common VH3-23 gene results in a low IgG1 expression rate, but is most frequent in the natural human antibody repertoire . This V-gene is a very stable and was used also for semisynthetic libraries, e.g. Tomlinson Library . For κ light chain, IGKV1-12, IGKV1D-33, IGKV1D-39 and IGKV3-20 are selected frequently. IGKV3-20 is the most used κ gene in the natural repertoire and is used for Ylanthia . The λ V-genes IGLV1-44, IGLV1-47, IGLV2-14, IGLV3-19, IGLV3-21 and IGLV6-57 were most frequently found after HAL selection. IGLV2-14 is the most used λ V-gene in the natural antibody repertoire and is used for Ylanthia, IGLV1-44 and IGLV1-47 are also often used in nature, but IGLV1-44 was excluded from the Ylanthia design .
The naive antibody gene libraries HAL9/10 have unique features compared to the libraries published earlier. The libraries have a slightly greater maximum diversity than the McCafferty library, they comprise more lambda V-genes and have longer CDR-H3 compared to the Pfizer library, the distribution of V-genes resembles more the natural antibody repertoire compared to CAT2.0 and they show more CDR-H3 diversity than Ylanthia. The successful generation of >800 antibodies against >100 targets with overwhelming success and in short time with miniaturized systems illustrates the huge advances in human library construction and use achieved since the first description of this method.
We gratefully acknowledge the financial support by the FP7 collaborative projects AffinityProteome (contract 222635) and AFFINOMICS (contract 241481) for some parts of this work. The new HAL libraries are dedicated to HAL9000 (2001: a space odyssey).
- Colwill K, Gräslund S. A roadmap to generate renewable protein binders to the human proteome. Nat Methods. 2011;8:551–8.View ArticleGoogle Scholar
- Dübel S, Stoevesandt O, Taussig MJ, Hust M. Generating recombinant antibodies to the complete human proteome. Trends Biotechnol. 2010;28:333–9.View ArticleGoogle Scholar
- Liu B, Huang L, Sihlbom C, Burlingame A, Marks JD. Towards proteome-wide production of monoclonal antibody by phage display. J Mol Biol. 2002;315:1063–73.View ArticleGoogle Scholar
- Strassberger V, Gutbrodt KL, Krall N, Roesli C, Takizawa H, Manz MG, et al. A comprehensive surface proteome analysis of myeloid leukemia cell lines for therapeutic antibody development. J Proteomics. 2014;99:138–51.View ArticleGoogle Scholar
- Leow CH, Jones M, Cheng Q, Mahler S, McCarthy J. Production and characterization of specific monoclonal antibodies binding the Plasmodium falciparum diagnostic biomarker, histidine-rich protein 2. Malar J. 2014;13:277.View ArticleGoogle Scholar
- Meyer T, Stratmann-Selke J, Meens J, Schirrmann T, Gerlach GF, Frank R, et al. Isolation of scFv fragments specific to OmpD of Salmonella Typhimurium. Vet Microbiol. 2011;147:162–9.View ArticleGoogle Scholar
- Pitaksajjakul P, Lekcharoensuk P, Upragarin N, Barbas CF, Ibrahim MS, Ikuta K, et al. Fab MAbs specific to HA of influenza virus with H5N1 neutralizing activity selected from immunized chicken phage library. Biochem Biophys Res Commun. 2010;395:496–501.View ArticleGoogle Scholar
- Schütte M, Thullier P, Pelat T, Wezler X, Rosenstock P, Hinz D, et al. Identification of a putative Crf splice variant and generation of recombinant antibodies for the specific detection of Aspergillus fumigatus. PLoS One. 2009;4:e6625.View ArticleGoogle Scholar
- Baker KP, Edwards BM, Main SH, Choi GH, Wager RE, Halpern WG, et al. Generation and characterization of LymphoStat-B, a human monoclonal antibody that antagonizes the bioactivities of B lymphocyte stimulator. Arthritis Rheum. 2003;48:3253–65.View ArticleGoogle Scholar
- Lima XT, Abuabara K, Kimball AB, Lima HC. Briakinumab. Expert Opin Biol Ther. 2009;9:1107–13.View ArticleGoogle Scholar
- Mazumdar S. Raxibacumab. MAbs. 2009;1:531–8.View ArticleGoogle Scholar
- Reichert JM. Antibody-based therapeutics to watch in 2011. MAbs. 2011;3:76–99.View ArticleGoogle Scholar
- Reichert JM. Antibodies to watch in 2014. MAbs. 2014;6:5–14.View ArticleGoogle Scholar
- Sliwkowski MX, Mellman I. Antibody therapeutics in cancer. Science. 2013;341:1192–8.View ArticleGoogle Scholar
- Reichert JM. Which are the antibodies to watch in 2013? MAbs. 2013;5:1–4.View ArticleGoogle Scholar
- Fishwild DM, O’Donnell SL, Bengoechea T, Hudson DV, Harding F, Bernhard SL, et al. High-avidity human IgG kappa monoclonal antibodies from a novel strain of minilocus transgenic mice. Nat Biotechnol. 1996;14:845–51.View ArticleGoogle Scholar
- Nelson AL, Dhimolea E, Reichert JM. Development trends for human monoclonal antibody therapeutics. Nat Rev Drug Discov. 2010;9:767–74.View ArticleGoogle Scholar
- Osborn MJ, Ma B, Avis S, Binnie A, Dilley J, Yang X, et al. High-affinity IgG antibodies develop naturally in Ig-knockout rats carrying germline human IgH/Igκ/Igλ loci bearing the rat CH region. J Immunol Baltim Md 1950. 2013;190:1481–90.Google Scholar
- Winter G, Milstein C. Man-made antibodies. Nature. 1991;349:293–9.View ArticleGoogle Scholar
- Hoet RM, Cohen EH, Kent RB, Rookey K, Schoonbroodt S, Hogan S, et al. Generation of high-affinity human antibodies by combining donor-derived and synthetic complementarity-determining-region diversity. Nat Biotechnol. 2005;23:344–8.View ArticleGoogle Scholar
- Schirrmann T, Meyer T, Schütte M, Frenzel A, Hust M. Phage display for the generation of antibodies for proteome research, diagnostics and therapy. Mol Basel Switz. 2011;16:412–26.View ArticleGoogle Scholar
- Thie H, Meyer T, Schirrmann T, Hust M, Dübel S. Phage display derived therapeutic antibodies. Curr Pharm Biotechnol. 2008;9:439–46.View ArticleGoogle Scholar
- Bradbury ARM, Sidhu S, Dübel S, McCafferty J. Beyond natural antibodies: the power of in vitro display technologies. Nat Biotechnol. 2011;29:245–54.View ArticleGoogle Scholar
- Hust M, Dübel S. Mating antibody phage display with proteomics. Trends Biotechnol. 2004;22:8–14.View ArticleGoogle Scholar
- Winter G, Griffiths AD, Hawkins RE, Hoogenboom HR. Making antibodies by phage display technology. Annu Rev Immunol. 1994;12:433–55.View ArticleGoogle Scholar
- Ayat H, Burrone OR, Sadghizadeh M, Jahanzad E, Rastgou N, Moghadasi S, et al. Isolation of scFv antibody fragments against HER2 and CEA tumor antigens from combinatorial antibody libraries derived from cancer patients. Biol J Int Assoc Biol Stand. 2013;41:345–54.Google Scholar
- Thie H, Toleikis L, Li J, von Wasielewski R, Bastert G, Schirrmann T, et al. Rise and fall of an anti-MUC1 specific antibody. PLoS One. 2011;6:e15921.View ArticleGoogle Scholar
- Trott M, Weiβ S, Antoni S, Koch J, von Briesen H, Hust M, et al. Functional characterization of two scFv-Fc antibodies from an HIV controller selected on soluble HIV-1 Env complexes: a neutralizing V3- and a trimer-specific gp41 antibody. PLoS One. 2014;9:e97478.View ArticleGoogle Scholar
- Diebolder P, Keller A, Haase S, Schlegelmilch A, Kiefer JD, Karimi T, et al. Generation of “LYmph Node Derived Antibody Libraries” (LYNDAL) for selecting fully human antibody fragments with therapeutic potential. MAbs. 2014;6:130–42.View ArticleGoogle Scholar
- Fuchs M, Kämpfer S, Helmsing S, Spallek R, Oehlmann W, Prilop W, et al. Novel human recombinant antibodies against Mycobacterium tuberculosis antigen 85B. BMC Biotechnol. 2014;14:68.View ArticleGoogle Scholar
- Hust M, Meyer T, Voedisch B, Rülker T, Thie H, El-Ghezal A, et al. A human scFv antibody generation pipeline for proteome research. J Biotechnol. 2011;152:159–70.View ArticleGoogle Scholar
- Kirsch M, Hülseweh B, Nacke C, Rülker T, Schirrmann T, Marschall H-J, et al. Development of human antibody fragments using antibody phage display for the detection and diagnosis of Venezuelan equine encephalitis virus (VEEV). BMC Biotechnol. 2008;8:66.View ArticleGoogle Scholar
- Mersmann M, Meier D, Mersmann J, Helmsing S, Nilsson P, Gräslund S, et al. Towards proteome scale antibody selections using phage display. New Biotechnol. 2010;27:118–28.View ArticleGoogle Scholar
- Meyer T, Schirrmann T, Frenzel A, Miethe S, Stratmann-Selke J, Gerlach GF, et al. Identification of immunogenic proteins and generation of antibodies against Salmonella Typhimurium using phage display. BMC Biotechnol. 2012;12:29.View ArticleGoogle Scholar
- Wezler X, Hust M, Helmsing S, Schirrmann T, Dübel S. Human antibodies targeting CD30+ lymphomas. Hum Antibodies. 2012;21:13–28.Google Scholar
- Jäger V, Büssow K, Wagner A, Weber S, Hust M, Frenzel A, et al. High level transient production of recombinant antibodies and antibody fusion proteins in HEK293 cells. BMC Biotechnol. 2013;13:52.View ArticleGoogle Scholar
- Sambrook J, Russell D. Molecular Cloning: A Laboratory Manual. 3rd ed. New York: Cold Spring Harbor Laboratory Press; 2001.Google Scholar
- Rondot S, Koch J, Breitling F, Dübel S. A helper phage to improve single-chain antibody presentation in phage display. Nat Biotechnol. 2001;19:75–8.View ArticleGoogle Scholar
- Soltes G, Hust M, Ng KKY, Bansal A, Field J, Stewart DIH, et al. On the influence of vector design on antibody phage display. J Biotechnol. 2007;127:626–37.View ArticleGoogle Scholar
- Frenzel A, Kügler J, Wilke S, Schirrmann T, Hust M. Construction of human antibody gene libraries and selection of antibodies by phage display. Methods Mol Biol Clifton NJ. 2014;1060:215–43.View ArticleGoogle Scholar
- Hust M, Dübel S, Schirrmann T. Selection of recombinant antibodies from antibody gene libraries. Methods Mol Biol. 2007;408:243–55.View ArticleGoogle Scholar
- Goldsmith ME, Konigsberg WH. Adsorption protein of the bacteriophage fd: isolation, molecular properties, and location in the virus. Biochemistry (Mosc). 1977;16:2686–94.View ArticleGoogle Scholar
- Frenzel A, Fröde D, Meyer T, Schirrmann T, Hust M. Generating recombinant antibodies for research, diagnostics and therapy using phage display. Curr Biotechnol. 2012;1:33–41.View ArticleGoogle Scholar
- Mollova S, Retter I, Hust M, Dübel S, Müller W. Analysis of Single Chain Antibody Sequences Using the VBASE2 Fab Analysis Tool. In: Antibody Engineering. 2nd ed. Heidelberg/New York: Springer Verlag; 2010. p. 3–10.View ArticleGoogle Scholar
- Tiller T, Schuster I, Deppe D, Siegers K, Strohner R, Herrmann T, Berenguer M, Poujol D, Stehle J, Stark Y, Heßling M, Daubert D, Felderer K, Kaden S, Kölln J, Enzelberger M, Urlinger S: A fully synthetic human Fab antibody library based on fixed VH/VL framework pairings with favorable biophysical properties. mAbs 2013, 5.Google Scholar
- Pallarès N, Lefebvre S, Contet V, Matsuda F, Lefranc MP. The human immunoglobulin heavy variable genes. Exp Clin Immunogenet. 1999;16:36–60.View ArticleGoogle Scholar
- Pallarès N, Frippiat JP, Giudicelli V, Lefranc MP. The human immunoglobulin lambda variable (IGLV) genes and joining (IGLJ) segments. Exp Clin Immunogenet. 1998;15:8–18.View ArticleGoogle Scholar
- Barbié V, Lefranc MP. The human immunoglobulin kappa variable (IGKV) genes and joining (IGKJ) segments. Exp Clin Immunogenet. 1998;15:171–83.View ArticleGoogle Scholar
- Breitling F, Dübel S, Seehaus T, Klewinghaus I, Little M. A surface expression vector for antibody screening. Gene. 1991;104:147–53.View ArticleGoogle Scholar
- McCafferty J, Fitzgerald KJ, Earnshaw J, Chiswell DJ, Link J, Smith R, et al. Selection and rapid purification of murine antibody fragments that bind a transition-state analog by phage display. Appl Biochem Biotechnol. 1994;47:157–71. discussion 171–3.View ArticleGoogle Scholar
- De Haard HJ, van Neer N, Reurs A, Hufton SE, Roovers RC, Henderikx P, et al. A large non-immunized human Fab fragment phage library that permits rapid isolation and kinetic analysis of high affinity antibodies. J Biol Chem. 1999;274:18218–30.View ArticleGoogle Scholar
- Goffinet M, Chinestra P, Lajoie-Mazenc I, Medale-Giamarchi C, Favre G, Faye J-C. Identification of a GTP-bound Rho specific scFv molecular sensor by phage display selection. BMC Biotechnol. 2008;8:34.View ArticleGoogle Scholar
- Goletz S, Christensen PA, Kristensen P, Blohm D, Tomlinson I, Winter G, et al. Selection of large diversities of antiidiotypic antibody fragments by phage display. J Mol Biol. 2002;315:1087–97.View ArticleGoogle Scholar
- Pansri P, Jaruseranee N, Rangnoi K, Kristensen P, Yamabhai M. A compact phage display human scFv library for selection of antibodies to a wide variety of antigens. BMC Biotechnol. 2009;9:6.View ArticleGoogle Scholar
- Li Y, Cockburn W, Kilpatrick JB, Whitelam GC. High affinity ScFvs from a single rabbit immunized with multiple haptens. Biochem Biophys Res Commun. 2000;268:398–404.View ArticleGoogle Scholar
- Løset GA, Løbersli I, Kavlie A, Stacy JE, Borgen T, Kausmally L, et al. Construction, evaluation and refinement of a large human antibody phage library based on the IgD and IgM variable gene repertoire. J Immunol Methods. 2005;299:47–62.View ArticleGoogle Scholar
- Chahboun S, Hust M, Liu Y, Pelat T, Miethe S, Helmsing S, et al. Isolation of a nanomolar scFv inhibiting the endopeptidase activity of botulinum toxin A, by single-round panning of an immune phage-displayed library of macaque origin. BMC Biotechnol. 2011;11:113.View ArticleGoogle Scholar
- Hülseweh B, Rülker T, Pelat T, Langermann C, Frenzel A, Schirrmann T, et al. Human-like antibodies neutralizing Western equine encephalitis virus. MAbs. 2014;6:717–26.View ArticleGoogle Scholar
- Miethe S, Rasetti-Escargueil C, Liu Y, Chahboun S, Pelat T, Avril A, et al. Development of neutralizing scFv-Fc against botulinum neurotoxin A light chain from a macaque immune library. MAbs. 2014;6:446–59.View ArticleGoogle Scholar
- Pelat T, Hust M, Laffly E, Condemine F, Bottex C, Vidal D, et al. High-affinity, human antibody-like antibody fragment (single-chain variable fragment) neutralizing the lethal factor (LF) of Bacillus anthracis by inhibiting protective antigen-LF complex formation. Antimicrob Agents Chemother. 2007;51:2758–64.View ArticleGoogle Scholar
- Pelat T, Hust M, Hale M, Lefranc M-P, Dübel S, Thullier P. Isolation of a human-like antibody fragment (scFv) that neutralizes ricin biological activity. BMC Biotechnol. 2009;9:60.View ArticleGoogle Scholar
- Rülker T, Voß L, Thullier P, O’ Brien LM, Pelat T, Perkins SD, et al. Isolation and characterisation of a human-like antibody fragment (scFv) that inactivates VEEV in vitro and in vivo. PLoS One. 2012;7:e37242.View ArticleGoogle Scholar
- Schofield DJ, Pope AR, Clementel V, Buckell J, Chapple SD, Clarke KF, et al. Application of phage display to high throughput antibody generation and characterization. Genome Biol. 2007;8:R254.View ArticleGoogle Scholar
- Glanville J, Zhai W, Berka J, Telman D, Huerta G, Mehta GR, et al. Precise determination of the diversity of a combinatorial antibody library gives insight into the human immunoglobulin repertoire. Proc Natl Acad Sci U S A. 2009;106:20216–21.View ArticleGoogle Scholar
- Lloyd C, Lowe D, Edwards B, Welsh F, Dilks T, Hardman C, et al. Modelling the human immune response: performance of a 1011 human antibody repertoire against a broad panel of therapeutically relevant antigens. Protein Eng Des Sel PEDS. 2009;22:159–68.View ArticleGoogle Scholar
- Weber M, Bujak E, Putelli A, Villa A, Matasci M, Gualandi L, et al. A highly functional synthetic phage display library containing over 40 billion human antibody clones. PLoS One. 2014;9:e100000.View ArticleGoogle Scholar
- Bhat NM, Bieber MM, Teng NN. Cytotoxicity of murine B lymphocytes induced by human VH4-34 (VH4.21) gene-encoded monoclonal antibodies. Clin Immunol Immunopathol. 1997;84:283–9.View ArticleGoogle Scholar
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.