A surface expression vector for antibody screening

Epitopes fused to F-pilin are incorporated into functional recombinant pili

In order to develop a system which allows infection by an epitope-specific phage-antibody via an F-pilus expressing that epitope, a study on the expression of foreign sequences on F-pilin was undertaken. Initially, a plasmid library was constructed with random sequences encoding one to five amino acid residues fused to the C terminus of F-pilin (traA) which was used to complement an F-plasmid with an amber mutation in traA. Functional F-pilin fusions were detected using the filamentous phage, fUSE2, which transduces tetracycline resistance, as well as immunoblots using a monoclonal antiserum specific for the acetylated N terminus of pilin. All the clones selected expressed the pilin-fusions and displayed full sensitivity towards fUSE2 infection, which was indistinguishable from the wild-type F-pilin. The sequences of fUSE2-sensitive clones when compared to randomly selected clones which were not fUSE2-sensitive, revealed no obvious pattern in the amino acid residues fused to the C terminus, except for a preference for a hydrophilic amino acid at position +1. Mutating the C-terminal Leu in wt (wild-type) pilin to Ser blocked pilus assembly and fUSE2 infection; the pilin was correctly processed but the level of acetylation at the N terminus appeared to decrease. Fusing a known epitope (myc) directly to the C terminus blocked processing of F-pilin leading to loss of F-pilus assembly and function. The introduction of random sequences between traA and this epitope yielded fully recombinant, functional F-pili but this appeared to be due to processing of the extension by an unidentified protease leading to loss of the epitope. Surface expression of another epitope (G2-10) was clearly demonstrated by immuno-electron microscopy of pili with a G2-10 monoclonal antibody. A different five amino acid residue spacer between the F-pilin C terminus and the G2-10 epitope produced a system that was transfer-proficient and fUSE2-sensitive, but the pili were barely detectable by immunoblots or by electron microscopy. While the underlying rules that govern successful epitope expression at the C terminus of F-pilin remain elusive, many types of foreign sequences can be displayed with varying degrees of success. Our results also suggest that pilin sequence determines a number of steps in the complex pathway for pilus assembly.

Antibody phage display technology and its applications

In recent years, the use of display vectors and in vitro selection technologies has transformed the way in which we generate ligands, such as antibodies and peptides, for a given target. Using this technology, we are now able to design repertoires of ligands from scratch and use the power of phage selection to select those ligands having the desired (biological) properties. With phage display, tailor-made antibodies may be synthesized and selected to acquire the desired affinity of binding and specificity for in vitro and in vivo diagnosis, or for immunotherapy of human disease. This review addresses recent progress in the construction of, and selection from phage antibody libraries, together with novel approaches for screening phage antibodies. As the quality of large naïve and synthetic antibody repertoires improves and libraries becomes more generally available, new and exciting applications are pioneered such as the identification of novel antigens using differential selection and the generation of receptor a(nta)gonists. A combination of the design and generation of millions to billions of different ligands, together with phage display for the isolation of binding ligands and with functional assays for identifying (and possibly selecting) bio-active ligands, will open even more challenging applications of this inspiring technology, and provide a powerful tool for drug and target discovery well into the next decade.

Prolonged in vivo tumour retention of a human diabody targeting the extracellular domain of human HER2/neu

Single-chain Fv (scFv) molecules exhibit highly specific tumour-targeting properties in tumour-bearing mice. However, because of their smaller size and monovalent binding, the quantities of radiolabelled scFv retained in tumours limit their therapeutic applications. Diabodies are dimeric antibody-based molecules composed of two non-covalently associated scFv that bind to antigen in a divalent manner. In vitro, diabodies produced from the anti-HER2/neu (c-erbB-2) scFv C6.5 displayed approximately 40-fold greater affinity for HER2/neu by surface plasmon resonance biosensor measurements and significantly prolonged association with antigen on the surface of SK-OV-3 cells (t1/2 cell surface retention of > 5 h vs 5 min) compared with C6.5 scFv. In SK-OV-3 tumour-bearing scid mice, radioiodinated C6.5 diabody displayed a highly favourable balance of quantitative tumour retention and specificity. By as early as 4 h after i.v. administration, significantly more diabody was retained in tumour (10 %ID g(-1)) than in blood (6.7 %ID ml(-1)) or normal tissue (liver, 2.8 %ID g(-1); lung, 7.1 %ID g(-1); kidney, 5.2 %ID g(-1)). Over the next 20 h, the quantity present in blood and most tissues dropped approximately tenfold, while the tumour retained 6.5 %ID g(-1) or about two-thirds of its 4-h value. In contrast, the 24-h tumour retention of radioiodinated C6.5 scFv monomer was only 1 %ID g(-1). When diabody retentions were examined over the course of a 72-h study and cumulative area under the curve (AUC) values were determined, the resulting tumor-organ AUC ratios were found to be superior to those previously reported for other monovalent or divalent scFv molecules. In conclusion, the diabody format provides the C6.5 molecule with a distinct in vitro and in vivo targeting advantage and has promise as a delivery vehicle for therapeutic agents.

Strategies for selection of antibodies by phage display

Phage antibody-display is rapidly maturing into a very effective tool for antibody generation. The recent development of large primary antibody libraries enables selection of antibodies against most targets in under two weeks and many of these antibodies have relatively high (nanomolar) affinities. Successful strategies have also been developed to affinity mature these antibodies into the picomolar range if required.

Use of mutator cells as a means for increasing production levels of a recombinant antibody directed against Hepatitis B

A mutation strategy which utilises phage display technology and the Escherichia coli mutator strains, mutD5-FIT and XL1-RED, was applied to a Hepatitis B (HepB) specific single-chain Fv (scFv) to incorporate random mutations throughout the gene. Messenger RNA from a hybridoma producing antibodies against HepB was isolated, reverse transcribed and used as template for the production of scFv. Following production of the scFv protein using an E. coli expression vector (pGC), the scFv gene was recloned into a phage display vector (pHFA). This gene construct was introduced into E. coli mutator cells and the transformed cells were used as an inoculum for liquid cultures. After five cycles of growth at 37 degrees C, each followed by dilution and re-inoculation of fresh media, recombinant phage were recovered. Nucleotide sequence analysis of the scFv gene in phage selected on HBsAg-coated magnetic beads identified amino acid substitutions which produced an increase of greater than 10-fold in apparent production levels. Competitive ELISA studies showed that the selected scFv mutants appeared to have similar affinity to HBsAg as the parent scFv. The apparent increase in production was not the result of improved surface characteristics of regions uniquely exposed in scFvs, as the sites did not correlate with the variable/constant interface of the scFv variable region normally masked in Fabs or IgGs.

Determination of the relative affinities of antibody fragments expressed in Escherichia coli by enzyme-linked immunosorbent assay

We have previously utilized an M13 phage expression system and codon-based mutagenesis to increase the avidity of the tumor-specific antibody, BR96, 65-fold (Yelton et al., 1995, J. Immunol. 155, 1994-2004). Mutants with improved affinity were identified by screening phage-expressed antibodies on a carcinoma cell line. In this study we describe a more broadly applicable assay which permits rapid and quantitative comparison of affinities of related antibodies produced in an M13 phage expression system. BR96 variants displaying a range of affinities were expressed as soluble antibody fragments (Fabs) in the periplasmic space of bacteria and isolated from small-scale cultures grown in a 96-well format, yielding between 142 ng and 1.06 microg of Fab. Although the small-scale cultures expressed variable levels of Fab, the lower quantities were sufficient to saturate microtiter plates coated with a limiting amount of anti-human Fab antibody, resulting in the capture of uniform quantities of the Fab variants. The relative affinities of the variants were then compared by assessing binding to biotinylated antigen followed by detection with streptavidin-alkaline phosphatase conjugates. This approach permitted the direct comparison of the relative affinities of large numbers of antibody variants in a single step without multiple antibody dilutions. The assay is readily adaptable for screening phage-expressed antibody libraries against any biotinylated target and does not require purified antigen. For instance, the biotinylated BR96 antigen utilized in these studies was from a total cell extract prepared by labeling the surface of live tumor cells followed by detergent extraction. Thus, this approach may be applicable to the screening of antibody libraries against other cell surface antigens, such as transmembrane receptors, which are difficult to purify.

Molecular characterization of murine humoral immune response to botulinum neurotoxin type A binding domain as assessed by using phage antibody libraries

To produce antibodies capable of neutralizing botulinum neurotoxin type A (BoNT/A), the murine humoral immune response to BoNT/A binding domain (H(C)) was characterized at the molecular level by using phage antibody libraries. Mice were immunized with BoNT/A H(C), the spleens were harvested, and single-chain Fv (scFv) phage antibody libraries were constructed from the immunoglobulin heavy and light chain variable region genes. Phage expressing BoNT/A binding scFv were isolated by selection on immobilized BoNT/A and BoNT/A H(C). Twenty-eight unique BoNT/A H(C) binding scFv were identified by enzyme-linked immunosorbent assay and DNA sequencing. Epitope mapping using surface plasmon resonance in a BIAcore revealed that the 28 scFv bound to only 4 nonoverlapping epitopes with equilibrium constants (Kd) ranging from 7.3 x 10(-8) to 1.1 x 10(-9) M. In a mouse hemidiaphragm assay, scFv binding epitopes 1 and 2 significantly prolonged the time to neuroparalysis, 1.5- and 2.7-fold, respectively, compared to toxin control. scFv binding to epitopes 3 and 4 showed no protection against neuroparalysis. A combination of scFv binding epitopes 1 and 2 had an additive effect on time to neuroparalysis, which increased to 3.9-fold compared to the control. The results suggest that there are two "productive" receptor binding sites on H(C) which lead to toxin internalization and toxicity. Blockade of these two epitopes with monoclonal antibodies may provide effective immunoprophylaxis or therapy against BoNT/A intoxication.

Primary structure and functional scFv antibody expression of an antibody against the human protooncogen c-myc

The immunoglobulin heavy- and light-chain variable region (Vh and Vl) genes were isolated from Myc1-9E10 hybridoma cells, which secreted monoclonal antibody against human oncogen c-myc. The expression vector pOPE52-c-myc was constructed for the recombinant production in E. coli. A 30 kDa single chain fragment (scFv) expression product was found in the periplasmic space by SDS-PAGE and immunoblotting. A significant fraction was processed correctly as demonstrated with an antiserum recognizing the processed aminoterminus only. The specific binding of the scFv fragment to the peptide epitope of the maternal monoclonal antibody was demonstrated and the primary sequence of the variable regions was determined. Sequence comparison with previously published partial Vh and Vl sequences from this hybridoma cell line revealed a genetic heterogeneity for the light chain variable region. The potential use of this scFv as a new tool for detection and purification of tagged proteins, for adding costimulatory signals to the surface of cancer cells as well as for analyzing c-myc function in the living cell by cytoplasmic expression is discussed.

The antigen-binding domain of a human IgG-anti-F(ab')2 autoantibody

Recent studies revealed an immunoregulatory role of natural IgG-anti-F(ab')2 antibodies in both healthy individuals and patients with certain diseases. The implication of anti-F(ab')2 antibodies in the pathogenesis of diseases prompted us to study the gene segment structure of their antigen-binding domains and their binding characteristics. cDNA was prepared from the lymphocytes of a patient with a high IgG-anti-F(ab')2 serum titer. Variable heavy and light gene segments were amplified by PCR and inserted into a phagemid surface expression vector. Single-chain antibodies displayed on the phage surface were screened for binding to F(ab')2 fragments. The subsequent analysis of 95 single clones demonstrated that they all bound specifically to F(ab')2. Sequence analyses of 12 clones showed that 11 were identical and 1 contained a silent point mutation in the heavy chain and three amino acid exchanges in the light chain. The heavy chains belonged to the V(H)3 and the light chains to the V(kappa)2 gene family. The 11 identical light-chain genes were completely homologous to a germ-line sequence (DPK-15). Binding assays showed that the single-chain antibodies bind to F(ab')2, but not to Fab, Fc, or intact IgG. This binding pattern was confirmed by surface plasmon resonance studies, which revealed a relatively high affinity (Ka = 2.8 x 10(7) M(-1)). The strong binding capacity was further demonstrated by competitive inhibition of the serum anti-IgG antibody's interaction with antigen. The present study defines for the first time to our knowledge the gene segment structure of the antigen-binding domain of two human IgG-anti-F(ab')2 autoantibody clones and describes the binding kinetics of the purified monomeric fragments.

High level production of soluble single chain antibodies in small-scale Escherichia coli cultures

We have investigated the effect of growth and induction conditions on the production of soluble single-chain Fv antibody fragments in Escherichia coli under the control of wt lac promoter. The scFv was directed into the periplasmic space by a pelB leader sequence. Addition of sucrose to the medium gave a 15-25-fold increase in the yield of soluble scFv-phOx (3.0 mg/l) for bacterial shake-tube cultures and an increase of 80-150-fold (16.5 mg/l) for shake-flask cultures. Using flask culture in the presence of 0.4 M sucrose, a significant amount of scFv was released into the medium. We found that the scFv could be made to accumulate in the periplasm or be secreted into the medium by simply changing the incubation conditions and the concentration of the inducer. The ratio between soluble antibody fragments and insoluble scFv aggregates proved to be dependent on the strength of the promoter. Lowering the incubation temperature below 20 degrees C had no effect on the yield of soluble antibody fragments in the periplasm, but they were no longer secreted into the medium. An example of high level production in shake-flask cultures and one-step purification by immobilized metal affinity chromatography (IMAC) is described for a soluble scFv specific for the T cell surface antigen CD3. The biological activity of the purified anti-CD3 scFv was demonstrated by flow cytometry. This method should be especially useful for the functional screening of a large number of clones in small-scale cultures.

Phage-display libraries of murine and human antibody Fab fragments

We provide efficient and detailed procedures for construction, expression, and screening of comprehensive libraries of murine or human antibody Fab fragments displayed on the surface of filamentous phage. In addition, protocols for producing and using ultra-electrocompetent cells, for producing Fab phages from libraries, and for selecting antigen binders by panning are presented. The latter protocol includes a procedure for trypsin elution of bound phage.

Kawasaki disease: update on diagnosis, treatment, and a still controversial etiology

Diagnosis of Kawasaki syndrome still relies solely on clinical criteria because the etiology is unknown. However, the function and structure of different bacterial superantigens as potential pathogens are discussed. In this regard, the recent determination of the crystal structure of the toxic shock syndrome toxin-1 superantigen complexed with major histocompatibility complex class II suggests potential implications for the controversial findings concerning a role of those superantigens in Kawasaki disease. Although a specific therapy is not available, coronary complications can be significantly reduced with the help of intravenous immunoglobulin therapy combined with oral aspirin.

Rapid detection of recombinant antibody fragments directed against cell-surface antigens by flow cytometry

Cloning the correct genes coding for antibody variable domains (especially VL kappa) from hybridomas is often complicated by the presence of several immunoglobulin transcripts, some of them arising from the myeloma cell line. Indeed, four different VL genes were obtained after the amplification of immunoglobulin genes by PCR from the hybridoma HD37, which produces an antibody against the human CD19 B cell differentiation antigen. Most of the variants (eight out of 15) were derived from the kappa chain of the myeloma MOPC-21. For the rapid functional evaluation of recombinant antibody fragments against cell surface antigens, we established an efficient expression and detection system. First, deleted and mutated genes were eliminated by a colony screening procedure. Bacteria from picked colonies were then induced and grown in the presence of 0.4 M sucrose to increase the accumulation of soluble scFv in the periplasm (5-10 micrograms per ml of bacterial shake-tube culture). Finally, the cell-specific binding of scFv in crude periplasmic extracts was detected by flow cytometry. This procedure facilitated the efficient cloning of a functional anti-CD19 VH/VL combination from the hybridoma cDNA.

Affinity maturation of recombinant antibodies using E. coli mutator cells

BACKGROUND

Phage libraries can display repertoires of antibodies which are greater in number than the mammalian immune response. However, the selected antibodies often have low binding affinity to their target antigen or hapten (KD below 10(-6) M), which is characteristic of the primary immune repertoire. There is a need for procedures to mimic somatic hypermutation through antigen driven affinity maturation, thereby increasing the affinity of selected immunoglobulins.

OBJECTIVE

To investigate the effectiveness of mutation and affinity selection of recombinant antibody genes with mutator E. coli cells, incorporating phage-display strategies.

STUDY DESIGN

Unique human scFvs were selected from a naive Fd-phage library. These genes were mutated by propagation in mutD5 mutator E. coli cells (mutD5-FIT) which were competent for Fd (M13) based phagemid transfections and generated point mutations (transversions and transitions) in the scFv genes. Individual phage-displayed scFvs were affinity selected from the mutation library and were assayed as soluble scFvs by ELISA and BIAcore for binding to antigen.

RESULTS

The in vivo mutation of phage-displayed scFvs in E. coli mutD5-FIT, combined with affinity selection against antigen, produced scFv molecules with improved binding activity. The point mutations which resulted in single amino acid substitutions frequently produced ten fold increases in apparent binding affinity. Structural comparisons revealed that these point mutations were in framework regions (adjacent to the CDRs) and within the CDRs. In one case the apparent affinity of an anti-glycophorin scFv after mutation in the VL framework region close to CDR3 increased by 10(3). However, this increase in apparent affinity was accompanied by an increased propensity to dimerise and form aggregates.

CONCLUSIONS

A strategy for the rapid affinity maturation of scFv and Fab antibody fragments has been developed which utilises mutator strains of E. coli and incorporates phage display of antibody repertoires (libraries).

A novel and efficient route for the isolation of antibodies that recognise T cell receptor V alpha(s)

Studies of the T cell repertoire have been hindered by the lack of antibodies that recognise V region families, particularly for V alpha regions. In this report, single chain Fv (scFv) fragments have been isolated that recognise both recombinant V alpha(s) and native V alpha(s) on the surface of T cells. Mice have been immunised with purified soluble T cell receptors (TCRs) and antibody heavy and light chain variable domain (VH and VL, respectively) genes isolated from splenocytes using the polymerase chain reaction (PCR). The VH and VL genes have been assembled as scFv gene libraries and a bacteriophage display system used to isolate scFvs that recognise a soluble V alpha. Five scFvs have been purified and characterised in detail using enzyme-linked immunosorbent assays (ELISAs) and flow cytometry. Three of these five scFvs recognise native V alpha(s) on the surface of T cell hybridomas. This method therefore offers a rapid route to the generation of scFvs that recognise native TCRs and can readily be extended to the production of anti-human TCR antibodies for use in therapy and diagnosis.

Immobilizing proteins on the surface of yeast cells

Yeast has a rigid cell wall comprising an outer layer of glycoproteins and an internal skeletal layer of glucan; heterologous proteins can be targeted to the glycoprotein layer and become covalently linked to the glucan skeleton. Yeast is a eukaryote that has 'generally regarded as safe' (GRAS) status, and is easy to cultivate, so it seems ideally suited for applications including the manufacture of recyclable, immobilized, biocatalysts, whole-cell vaccines, the presentation of peptide or antibody libraries, and the presentation of adhesion or metal-binding proteins.

Identification of functional and structural amino-acid residues by parsimonious mutagenesis

For in vitro evolution of protein function, we previously proposed using parsimonious mutagenesis (PM), a technique where mutagenic oligodeoxynucleotides (oligo) are designed to minimize coding sequence redundancy and limit the number of amino acid (aa) residues which do not retain parental structural features. For this work, PM was used to increase the affinity of C6.5, a human single-chain Fv (scFv) that binds the glycoprotein tumor antigen, c-erbB-2. A phage antibody library was created where 19 aa located in three of the heavy (H) and light (L) chain antigen-binding loops (L1, L3 and H2) were simultaneously mutated. After four rounds of selection, 50% of scFv had a lower dissociation rate constant (koff) than the parental scFv. The Kd of these scFv ranged from twofold (Kd=7.0 x 10(-9) M) to sixfold (Kd=2.4 x 10(-9) M) lower than the parental scFv (Kd=1.6 x 10(-8) M). In higher affinity scFv, substitutions occurred at 10/19 of the positions, with 21/28 substitutions occurring at only four positions, two in H2, and one each in L1 and L3. Only the wild type (wt) aa was observed at 9/19 aa. Based on a model of C6.5, seven of the nine conserved aa have a structural role in the variable domain, either in maintaining the main chain conformation of the loop, or in packing on the H-chain variable domain. Two of the conserved aa are solvent exposed, suggesting they may play a critical role in recognition. Thus, PM identified three types of aa: structural aa, functional aa which modulate affinity, and functional aa, which are critical for recognition. Since the sequence space was not completely sampled, higher affinity scFv could be produced by subjecting functional aa which modulate affinity to a higher rate of mutation. Furthermore, PM could prove useful for modifying function in other proteins that belong to structurally related families.

Characterization of human variable domain antibody fragments against the U1 RNA-associated A protein, selected from a synthetic and patient-derived combinatorial V gene library

This is the first study describing recombinant human antibody fragments directed to the U1 RNA-associated A protein (U1A). Three anti-U1A antibody fragments (Fab) were isolated from a semi-synthetic human Fab library and one anti-U1A single-chain variable fragment (scFv) was isolated from a library which was derived from the IgG-positive splenic lymphocytes of an autoimmune patient. Competition studies with autoantibodies against the U1 small nuclear ribonucleoprotein (snRNP) particle from patients with systemic lupus erythematosus (SLE) and SLE-overlap syndromes revealed that U1A binding of these antibody fragments can be inhibited by about 40% of the patient sera. All antibody fragments recognized the native U1 snRNP in immunoprecipitation assays. Two of three Fab clones as well as the scFv clone derived from the repertoire of an autoimmune patient use the same heavy chain germ-line gene DP-65. Epitope mapping revealed that these three clones appear to recognize an identical epitope domain present on the C-terminal RNP motif of the U1A protein. The DP-65 heavy chain gene is used in less than 1% of the B cells in healthy individuals, while three out of four anti-U1A antibody fragments use this gene. This points to a restricted VH gene usage in the case of U1A, suggesting that the DP-65 heavy chain has a natural shape complementarity to the U1A protein.

An antibody VH domain with a lox-Cre site integrated into its coding region: bacterial recombination within a single polypeptide chain

Bacterial lox-Cre recombination within a single antibody VH domain was achieved through integration of a loxP site into its coding sequence. The 5' half of the VH gene, in which the H2 loop was replaced by a mutant loxP site, was fused to geneIII in an 'acceptor' fd-phage vector containing also a wild type loxP site. With a 'donor' plasmid vector harbouring the 3' half of the VH gene flanked by the same, differing loxP sites it recombined into a full-length VH with the loxP site-H2 loop. This VH was purified from bacterial periplasm, where it folded into a typical immunoglobulin domain. The system allows the generation of large VH repertoires using lox-Cre recombination.

A human pancreatic secretory trypsin inhibitor presenting a hypervariable highly constrained epitope via monovalent phagemid display

Hypervariable gene banks displaying ligands which can be used for affinity optimisation are valuable resources for examining shape space. They have added value if the ligand is small, if there is extensive information on its tertiary structure and if the variable region is highly constrained. These features would be expected to stabilise complexes by reducing the dissociation constants and to facilitate their use as 'lead substances' for the development of synthetic mimetics. The synthesis and characterisation of such phagemid-display banks is described here, in which the variable region is a 7-amino acid (aa) (pSKAN8-HyB/C) or 8-aa (pSKAN8-HyA) extended peptide held between two disulfide bridges at the exposed tip of the human pancreatic secretory trypsin inhibitor (PSTI). A phagemid pSKAN8 was created which contains a fusion between the PSTI and M13 pIII protein-coding genes. Cassettes containing the sequences (NNK)8 [HyA], (NNK)7 [HyB] or (NNK)6GTT [Hy-C] (where K = G or T) were used to randomize the aa coding region in the trypsin-inhibitory loop (aa 17 to 23) of PSTI. Some 31 million individual clones were generated in a mutS Escherichia coli strain kept as frozen cell stocks. Analysis of controls which had not undergone selection showed very low levels of deletion. The quality of the hypervariable region and bias of codon usage was quantified by DNA sequencing. It was estimated from SDS-PAGE that hybrid protein was represented statistically at a frequency of one molecule per two phagemid particles. The functionality and reproducibility of the system was demonstrated by trypsin-binding of the original vector and in selecting novel chymotrypsin inhibitors from the banks.

Human antibody libraries in Escherichia coli

A potentially vast pool of human antibodies with novel specificities for diagnostic and therapeutic purposes can be generated in Escherichia coli. Antibodies to infectious agents have already been isolated by amplifying the heavy and light chain repertoires of donor lymphocytes and they have even been rescued many years after the initial infection from memory cells cultivated in SCID mice. Eventually, however, the creation of extremely large and diverse libraries from the naive antibody repertoire of unactivated B lymphocytes or by gene synthesis using random oligonucleotides for the hypervariable regions could provide a rapid means of obtaining human antibodies to any particular antigen. An important breakthrough for exploiting the potential size and diversity of these libraries has been the development of systems for the surface display of antibodies that are physically linked to their own genes. This allows large numbers of clones to be screened simultaneously and antibodies with affinities of up to 10(8) M-1 have already been obtained using these vectors. It seems quite feasible, therefore, that antibodies with affinities approaching those obtained in the secondary immune response can be obtained by systematically optimizing the strategies for making antibody libraries. Furthermore, it might be possible to establish extremely large antibody repertoires in E. coli by the in vivo recombination of phage and plasmid antibody libraries. The affinity of the selected antibodies could be increased by chain shuffling or random mutagenesis followed by several rounds of selection under increasingly stringent conditions.

A single expression system for the display, purification and conjugation of single-chain antibodies

To facilitate the purification and conjugation of single-chain antibodies (scFv) selected from a phage display library, we have incorporated His6, an amber stop codon and a C-terminal Cys into a surface expression vector. The vector also contains a lacIq gene for improving the efficiency of regulation and a sequence coding for a marker peptide.

Similar binding properties for a neutralizing anti-tetanus toxoid human monoclonal antibody and its bacterially expressed Fab

A high-affinity anti-tenanus toxoid (TT) human monoclonal antibody showing neutralizing activity was isolated from a fusion between mouse myeloma and human splenic cells. Fab fragments from this antibody were obtained using a recombinant phage surface-display expression system. The parental antibody and the corresponding Fab had identical immunological activities, including specificity and affinity. These results confirm the feasibility of developing Escherichia coli expression of monoclonal human Fab from hybridoma cells.

Amino acid sequence based PCR primers for amplification of rearranged human heavy and light chain immunoglobulin variable region genes

Previously described primers for PCR amplification of variable immunoglobulin (Ig) genes were based on gene sequences. To include the large number of amino acid sequences of antibodies whose DNA has not been sequenced and to ensure a maximal fit to rearranged human Ig variable region genes, we have made a comprehensive comparison of both protein and nucleotide sequences. The resulting set of 15 primers was able to amplify a wide range of rearranged antibody variable region genes. Restriction sites included in the primers facilitate cloning of the PCR products into various expression vectors. Sequence analyses of PCR-amplified cDNA derived from a polyclonal B cell population showed that maximal enrichment is obtained for highly represented variable Ig gene subgroups. Rarely occurring V kappa 4 and V lambda 5 subgroups were not detected. Rearranged Ig variable region genes from each of 19 human B cell lines were also amplified. Comparisons to germline sequences allowed the allocation of rearranged genes to the original Ig genes. This primer set should be very useful for generating large repertoires of rearranged V genes and for amplifying genes of individual B cell clones.

Bifunctional and multimeric complexes of streptavidin fused to single chain antibodies (scFv)

Multivalent and multispecific antibodies with defined stoichiometry could provide valuable tools for biological and medical research and for the diagnosis and therapy of cancer. We have therefore fused single chain antibodies (scFv) with core-streptavidin. This chimeric protein, expressed by the vector pSTE-215 (plasmid for streptavidin-tagged expression), can form tetrameric complexes, binds antigen and contains the biotin binding site which may be used for further complex formation. An additional cysteine was inserted near the carboxy terminus to facilitate the construction of covalently linked bifunctional molecules. The scFv fusion protein could be purified by affinity chromatography using biotin analogues. We have also shown that the scFv fusion protein could be used for direct detection of its antigen in ELISA and Western blots when stained with biotinylated horseradish peroxidase.

Recombinant single-chain Fv fragments carrying C-terminal cysteine residues: production of bivalent and biotinylated miniantibodies

A murine antibody single-chain Fv (scFv) fragment carrying five C-terminal histidine residues preceded by a cysteine residue and a marker peptide was expressed in Escherichia coli. Its variable heavy (VH) and light (VL) domains are derived from the mouse monoclonal antibody mAb215, which is specific for the largest subunit of RNA polymerase II of Drosophila melanogaster. ScFv' monomers, covalently linked (scFv')2 and non-covalent dimers, as well as aggregated antibody fragments, were isolated from an E. coli cell paste by immobilized metal affinity chromatography in 6 M urea followed by a renaturation procedure that does not use any sulfhydryl agents. In a final step, the components were separated by size exclusion chromatography. All the recombinant antibody fractions demonstrated high antigen-binding activity and specificity as shown by ELISA and Western blot analysis. Affinity measurements carried out by competitive immunoassays showed that covalently linked (scFv')2 have binding constants quite close to those of the parental monoclonal antibodies and four-fold higher than scFv' monomers. ScFv derivatives, specifically biotinylated through the free sulfhydryl group, recognize the corresponding antigen in ELISA and Western blot analysis, thus demonstrating the possibility of using chemically modified scFv antibodies for immunodetection.

Isolation and characterization of a cDNA clone encoding the pokeweed antiviral protein II from Phytolacca americana and its expression in E. coli

Three distinct ribosome-inactivating proteins (RIPs) were isolated from pokeweed (Phytolacca americana). We identified and sequenced for the first time a complete cDNA encoding the pokeweed antiviral protein II (PAP II), which is expressed in the late summer leaves of pokeweed. The cDNA of PAP II consists of 1,187 nucleotides and encodes a mature protein of 285 amino acids. Its predicted amino acid sequence is only 33% similar to PAP and PAP-S. The NH2 terminal extrapeptide (25 amino acid residues) was similar but not identical to that of PAP's extrapeptide. The cDNA of PAP II was expressed in E. coli. The growth of the transformants was strongly inhibited after induction of the gene. Furthermore, PAP II, which was produced in E. coli, inhibited protein synthesis in a rabbit reticulocyte translation system. Thus, recombinant PAP II would appear to be as functional as native PAP in inhibiting protein synthesis in both prokaryotes and eukaryotes.

Surface display of antibodies

To screen antibody libraries that contain many millions of different clones, a selection system is required with an efficiency comparable to that of the immune system. This can be achieved by displaying antibodies on the surface of microorganisms containing the antibody's gene, analogous to the expression of the IgM antigen receptor on the surface of unactivated B-lymphocytes. Specific clones can then be selected using immobilized antigens. The minor coat protein of filamentous phages, pIII, which initiates the infection of E.coli by binding to their F-pili, and the major coat protein, pVIII, have been used as carriers for displaying antibodies on the phage surface. Recombinant antibodies have also been targeted to the cell surface of bacteria by fusing them with outer membrane components derived from lipoproteins, OmpA and an IgA protease. However, only the pIII system has been routinely used for screening antibody libraries. Here we describe the various antibody surface display systems and the screening of antibody libraries generated from the gene repertoire of lymphocytes and by gene synthesis. Finally, we have made a short comparison of the bacterial production of Fabs versus single chain antibodies (scFv).

Use of living columns to select specific phage antibodies

Here we demonstrate that it is possible to confront two recombinant microorganisms in order to select one using the other. We have shown that an epitope derived from p21ras expressed within the outer membrane protein, LamB, can be recognized both by the monoclonal antibody Y13-259, as well as the single chain Fv fragment derived from it. This specificity, which is maintained when the Y13-259 single chain Fv is expressed as a fusion protein with the phage fd gene 3 protein, has allowed us to use the living column of LamB-ras to purify Y13-259 phage from a background of non-binding phage, even at dilutions as high as 10 phage in 10(10) irrelevant phage.

Efficient method for constructing comprehensive murine Fab antibody libraries displayed on phage

We have developed efficient methodologies for construction and expression of comprehensive phage display libraries of murine Fab antibody fragments in E. coli cells. Our methods optimize several critical steps of the polymerase chain reaction (PCR) amplification of transcripts of the re-arranged immunoglobulin genes and of their subsequent assembly and expression: Firstly, we have designed exhaustive sets of PCR primers of low degeneracy for the amplification of transcripts of the Fab region of the heavy and light-chain genes. These primers proved effective in amplification of Fab gene fragments from a large panel of hybridoma cell lines of different specificity and family sub-type. Secondly, we have developed a 'jumping PCR' technique that effectively assembled and recombined the amplified heavy and light-chain gene fragments into a bi-cistronic operon. Thirdly, we have constructed expression vectors for insertion of the combinatorial Fab gene-cassette in fusion with a truncated version of the phage surface protein, gIIIp. The heavy chain and the light chain-gIII fusion are transcribed as a polycistronic mRNA from the lacZ promoter and efficient transcriptional control is provided by wildtype lacI present on the vector. The utility of the system was demonstrated by isolating several antigen-binding clones from hybridomas and libraries made from immunized mice.

A family of vectors for surface display and production of antibodies

Expression vectors for surface display and production of single-chain (Fv) antibodies (scAb) have been constructed based on the phagemid pSEX, which expresses DNA encoding a scAb fused to the gene III product of filamentous phage [Breitling et al., Gene 104 (1991) 147-153]. A smaller version of this phagemid, pSEX20, was made by removing an unnecessary cat. To produce a vector for the surface display of other proteins and peptides, the scAb of pSEX20 was substituted by a polycloning site (MCS) to give pSEX40. For the presentation of Ab on the surface of Escherichia coli, phagemid pAP10 was derived from pSEX20 by substituting gene III with a gene encoding the peptidoglycan-associated lipoprotein (PAL). Vectors for producing scAb that can be purified by antibody and metal affinity chromatography were constructed by substituting gene III in the vector pSEX20 with DNA encoding a peptide with a C-terminal epitope recognised by a monoclonal antibody (phagemid pOPE40) or with five C-terminal histidines (pOPE 90).

Production of human antibodies using bacteriophage

The immune system produces antibodies by a process of antigen-driven selection. An in vitro process of antigen-driven selection, based on the display of antibody fragments on filamentous bacteriophage, has recently been developed. This enables human antibody fragments of high affinity and specificity to be produced without immunization.

Antigen recognition and targeted delivery by the single-chain Fv

The single-chain Fv (sFv) has proven attractive for immuno-targeting, both alone and as a targeting element within sFv fusion proteins. This chapter summarizes the features of sFv proteins that have sparked this interest, starting with the conservation of Fv architecture that makes general sFv design practical. The length and composition of linkers used to bridge V domains are discussed based on the sFv literature; special emphasis is given to the (Gly4Ser)3 15-residue linker that has proven of broad utility for constructing Fv regions of antibodies and other members of the immunoglobulin superfamily. The refolding properties of sFv proteins are summarized and examples given from our laboratory. Spontaneous refolding from the fully reduced and denatured state, typified by 26-10 sFv, is contrasted with disulfide-restricted refolding, exemplified by MOPC 315 and R11D10 sFv proteins, which recover antigen binding only if their disulfides have been oxidized prior to removal of denaturant. The medical value of sFv proteins hinges on their reliability in antigen recognition and rapidity in targeted delivery. Detailed analysis of specificity and affinity of antigen binding by the 26-10 antidigoxin sFv has demonstrated very high fidelity to the binding properties of the parent 26-10 sFv. These results gave confidence to the pursuit of more complex biomedical applications of these proteins, which is indicated by our work with the R11D10 sFv for the imaging of myocardial infarctions. Diagnostic imaging and therapeutic immunotargeting by sFv present significant opportunities, particularly as a result of their pharmacokinetic properties. Intravenously administered sFv offers much faster clearance than conventional Fab fragments or intact immunoglobulin with minimal background binding.

Medical applications of single-chain antibodies

A single-chain antibody or single-chain Fv (sFv) incorporates the complete antibody binding site in a single polypeptide chain of minimal size, with an approximate molecular weight of 26,000. In antibodies, the antigen combining site is part of the Fv region, which is composed of the VH and VL variable domains on separate heavy and light chains. Efforts over nearly two decades have indicated that Fv fragments can only rarely be prepared from IgG and IgA antibodies by proteolytic dissection. Beginning in 1988, single-chain analogues of Fv fragments and their fusion proteins have been reliably generated by antibody engineering methods. The first step involves obtaining the genes encoding VH and VL domains with desired binding properties; these V genes may be isolated from a specific hybridoma cell line, selected from a combinatorial V-gene library, or made by V gene synthesis. The single-chain Fv is formed by connecting the component V genes with an oligonucleotide that encodes an appropriately designed linker peptide, such as (Gly4-Ser)3. The linker bridges the C-terminus of the first V region and N-terminus of the second, ordered as either VH-linker-VL or VL-linker-VH. In principle, the sFv binding site can faithfully replicate both the affinity and specificity of its parent antibody combining site, as demonstrated in our model studies with the 26-10 anti-digoxin sFv. Furthermore, the sFv remains stable at low concentrations that promote VH and VL dissociation from the Fv heterodimer, resulting in loss of Fv binding. Intravenously administered sFv proteins exhibit accelerated biodistribution and exceptionally fast clearance compared to IgG or Fab. These pharmacokinetic properties allow rapid imaging by sFv, which therefore may be labeled with a short-lived isotope such as Tc-99m. Expression of a single gene product from fused sFv and effector genes facilitates immunotargeting of the effector protein, as shown for single-chain Fv toxin fusion proteins.

Protease inhibitor display M13 phage: selection of high-affinity neutrophil elastase inhibitors

We report display of the complete protease inhibitor (Kunitz) domain, BPTI, on the surface of bacteriophage M13 as a fusion to the gene III product. Phage that display BPTI bind specifically to anti-BPTI antibodies, trypsin and anhydrotrypsin. A point mutation of BPTI [Lys15-->Leu(K15L)] alters the binding specificity of fusion phage such that a human neutrophil elastase-binding phenotype is conferred while a trypsin-binding phenotype is eliminated. Phage were eluted from an immobilized protease with step gradients of decreasing pH. Phage that display Kunitz domains having higher affinity for the immobilized protease exhibit characteristic pH elution phenotypes, indicating that bound display phage can be selectively recovered from an affinity matrix. Utilization of this technology should enable the selection of remodeled protease inhibitors exhibiting novel binding specificities.

Antibody expression in bacteriophage systems: The future of monoclonal antibodies?

Bacteriophage systems have been utilized to express and isolate antibodies. This promising technology has been evolving rapidly and has the potential to revolutionize the way in which monoclonal antibodies are generated. This review focuses on the many recent advances that have been made in obtaining monoclonal antibodies from bacteriophage systems.

By-passing immunisation. Human antibodies from synthetic repertoires of germline VH gene segments rearranged in vitro

By display of antibody repertoires on the surface of a filamentous bacteriophage and selection of the phage by binding to antigen, we can mimic immune selection. Recently, by tapping the repertoire of rearranged V-genes from the peripheral blood lymphocytes of unimmunised donors, we succeeded in making human antibody fragments with different specificities, including both haptens and proteins, from the same library of phage. Now we have built a repertoire of human VH genes from 49 human germline VH gene segments rearranged in vitro to create a synthetic third complementarity determining region (CDR) of five or eight residues. The rearranged VH genes were cloned with a human V lambda 3 light chain as single chain Fv fragments for phage display, and the library of phage panned by binding to each of two haptens, 2-phenyl-5-oxazolone (phOx) or 3-iodo-4-hydroxy-5-nitrophenyl-acetate (NIP) coupled to bovine serum albumin (BSA). Many different antibody fragments were isolated which bound specifically to hapten, some with affinities in the micromolar range. The in vitro "immune response" to the hapten NIP was dominated by the 9-1 segment (VH3 family), and that to phOx by the VH26 segment (VH3 family) with an invariant aromatic residue (Tyr, Phe, Trp) at residue 97 of CDR3. However, the isolation of phage against protein antigens proved more elusive, with a single phage binding to human tumour necrosis factor, and none to bovine serum albumin, turkey egg-white lysozyme or human thyroglobulin. Nevertheless, the work shows that human antibody fragments with specific binding activities can be made entirely in vitro.

Selection of phage antibodies by binding affinity. Mimicking affinity maturation

We describe a process, based on display of antibodies on the surface of filamentous bacteriophage, for selecting antibodies either by their affinity for antigen or by their kinetics of dissociation (off-rate) from antigen. For affinity selection, phage are mixed with small amounts of soluble biotinylated antigen (less than 1 microgram) such that the antigen is in excess over phage but with the concentration of antigen lower than the dissociation constant (Kd) of the antibody. Those phage bound to antigen are then selected using streptavidin-coated paramagnetic beads. The process can distinguish between antibodies with closely related affinities. For off-rate selection, antibodies are preloaded with biotinylated antigen and diluted into excess unlabelled antigen for variable times prior to capture on streptavidin-coated paramagnetic beads. To mimic the affinity maturation process of the immune system, we introduced random mutations into the antibody genes in vitro using an error-prone polymerase, and used affinity selection to isolate mutants with improved affinity. Starting with a small library (40,000 clones) of mutants (average 1.7 base changes per VH gene) of the mouse antibody B1.8, and using several rounds of affinity selection, we isolated a mutant with a fourfold improved affinity to the hapten 4-hydroxy-5-iodo-3-nitrophenacetyl-(NIP)-caproic acid (mutant Kd = 9.4(+/- 0.3) nM compared with B1.8 Kd = 41.9(+/- 1.6) nm). The relative increase in affinity of the mutant is comparable to the increase seen in the anti-4-hydroxy-3-nitrophenylacetyl/NIP-caproic acid murine secondary immune response.

Recombinant human monoclonal antibodies. Basic principles of the immune system transferred to E. coli

To produce human monoclonal antibodies in bacteria, a gene repertoire of IgM variable regions was isolated from human peripheral B lymphocytes by the polymerase chain reaction. Alternatively, synthetic antibody genes with random hypervariable regions are being generated that may provide libraries of even higher complexity. For the selection of specific monoclonal antibodies from these libraries, we have developed two E. coli vector systems that facilitate the surface display of an antibody physically linked to its own gene. The phagemid pSEX encodes a fusion protein of an antigen binding domain (Fv-antibody) with the docking protein (pIII) of filamentous phages. Specific antibody genes can therefore be enriched by antigen affinity chromatography. The plasmid pAP1 encodes a fusion protein of an Fv-antibody with a bacterial cell-wall protein. Bacteria carrying this plasmid express functional Fv-antibodies tightly bound to their surface. This should enable the selection of single cells with a fluorescence-assisted cell sorter (FACS) using labeled antigen or by adsorption to immobilized antigen. These vectors permit three major principles of the antibody response to be mimicked in E. coli: 1. Generation of a highly complex antibody repertoire; 2. Clonal selection procedures for library screening; and 3. The possibility of increasing a given affinity by repeated rounds of mutation and selection.

Rapid evolution of peptide and protein binding properties in vitro

A significant bottleneck in protein engineering arises from the problem of identifying particular molecules with new functions from a potentially enormous range of peptide or protein variants. Two areas of emerging technology, phage display and multiple peptide synthesis, provide new means of screening huge libraries in vitro for novel binding properties. This review is also published in Current Opinion in Structural Biology 1992, 2:597-604.

Regulated secretion and purification of recombinant antibodies in E. coli

A plasmid for optimized protein expression of recombinant Fv antibodies (pOPE) in E. coli was used to express the variable domains of the murine monoclonal antibody HD39 specific for the human B-cell surface antigen CD22. The production of Fv antibodies by pOPE can be regulated over a wide range by varying the IPTG concentration. Antibodies that can discriminate between secreted and nonsecreted Fv antibody fragments were used to show that secretion is the limiting step for the production of functional Fv antibodies. IPTG concentrations above 20 microM increased the total antibody production, but did not yield larger amounts of secreted Fv antibodies. The addition of five histidines to the C terminus facilitates an easy single-step enrichment procedure based on immobilized metal affinity chromatography.