Probing the basis of antibody reactivity with a panel of constrained peptide libraries displayed by filamentous phage

Active specific immunotherapy of malignant melanoma and peptide mimics of the human high-molecular-weight melanoma-associated antigen

The realization that tumor cells utilize multiple mechanisms to escape from immune recognition and destruction has stimulated interest in developing and applying immunotherapeutic strategies which target both humoral and cellular immunity to malignant cells. As a result, the tumor-associated antigens (TAA) used as targets have to be expressed on the cell surface membrane of malignant cells. Furthermore, since most of the TAA used for active specific immunotherapy are self-antigens, a challenge facing tumor immunologists is to develop strategies which are effective in breaking tolerance to self-antigens. This chapter describes one strategy which relies on the use of peptide mimics of the human high-molecular-weight melanoma-associated antigen (HMW-MAA) as immunogens to implement active specific immunotherapy in patients with malignant melanoma. These mimics, which are isolated from phage display peptide libraries by panning with anti-HMW-MAA monoclonal antibodies, are expected to induce both humoral and cellular anti-HMW-MAA immunity.

Profiling the immune response in patients with breast cancer by phage-displayed cDNA libraries

Display on the surface of filamentous phages has been shown to be well suited for the enrichment of serum antibody-binding ligands. Here, we have taken the advantage of this technology to analyze the humoral immune response in patients with cancer. The cDNA repertoires from breast cancer cell lines T47D and MCF-7 were fused to the 3'-end of the filamentous phage M13 gene VI in all three reading frames. When the libraries were biopanned on rabbit polyclonal IgG against the human Bcl-x(L) protein, positive clones were selected, thus confirming the utility of the libraries. Using serum antibodies from patients with breast cancer, we specifically selected IgG-binding phage-encoded cDNA products. Sequence analysis of the selected clones identified important antigens including p53, centromere-F, int-2, pentraxin I, integrin beta5, cathepsin L2 and S3 ribosomal protein. The selected phage-displayed cDNA products were recognized by a significant number of breast cancer sera as compared to sera from normal individuals. Although the human pentraxin I mRNA was reported to be exclusively localized in the nervous system, we found it also expressed by breast cancer cell lines. Four out of 30 patients with breast cancer (13 %) showed reactivity with the recombinant pentraxin expressed in Escherichia coli, while no reactivity was found in normal sera. The obtained results demonstrate that phage display could be a valuable method for the identification of antigens recognized by the humoral immune system in patients with cancer.

Delineation of a linear epitope by multiple peptide synthesis and phage display

Two different approaches, the phage display technique and the Spot peptide synthesis on cellulose membranes, were used to identify sequences recognized by Fab 57P, specific for tobacco mosaic virus protein (TMVP), and define the preferred chemical composition of a functional epitope. Kinetic measurements of the interaction between peptide variants and the antibody fragment were used to further refine the molecular basis of binding activity. Our results show that the functional epitope of Fab 57P requires precise physico-chemical properties at a limited number of positions, and that residues flanking these key residues can influence binding affinity. The phage display and Spot synthesis methods allowed the straightforward localization of the binding region and the identification of residues that are essential for recognition. However, these methods yielded slightly different views of accessory factors that are able to influence antibody binding. The influence on binding activity of these factors can only be assessed through quantitative affinity measurements.

Evidence that a protein-protein interaction 'hot spot' on heterotrimeric G protein betagamma subunits is used for recognition of a subclass of effectors

To understand the requirements for binding to G protein betagamma subunits, phage-displayed random peptide libraries were screened using immobilized biotinylated betagamma as the target. Selected peptides were grouped into four different families based on their sequence characteristics. One group (group I) had a clear conserved motif that has significant homology to peptides derived from phospholipase C beta (PLC beta) and to a short motif in phosducin that binds to G protein beta subunits. The other groups had weaker sequence homologies or no homology to the group I sequences. A synthetic peptide from the strongest consensus group blocked activation of PLC by G protein betagamma subunits. The peptide did not block betagamma-mediated inhibition of voltage-gated calcium channels and had little effect on betagamma-mediated inhibition of Gs-stimulated type I adenylate cyclase. Competition experiments indicated that peptides from all four families bound to a single site on betagamma. These peptides may bind to a protein-protein interaction 'hot spot' on the surface of betagamma subunits that is used by a subclass of effectors.

Construction of high-complexity combinatorial phage display peptide libraries

Random peptide libraries displayed on the surface of filamentous bacteriophage are widely used as tools for the discovery of ligands for biologically relevant macromolecules, including antibodies, enzymes, and cell surface receptors. Phage display results in linkage of an affinity-selectable function (the displayed peptide) to the DNA encoding that function, allowing selection of individual binding clones by iterative cycles of in vitro panning and in vivo amplification. Critical to the success of a panning experiment is the complexity of the library: the greater the diversity of clones within the library, the more likely the library contains sequences that will bind a given target with useful affinity. A method for construction of high-complexity (> or = 10(9) independent clones) random peptide libraries is presented. The key steps are highly efficient binary ligation under conditions where the vector is relatively dilute, with only a modest molar excess of insert, followed by efficient electrotransformation into Escherichia coli. Library design strategies and a protocol for rapid sequence characterization are also presented.

Identifying diagnostic peptides for lyme disease through epitope discovery

Serum antibodies from patients with Lyme disease (LD) were used to affinity select peptide epitopes from 12 large random peptide libraries in phage display format. The selected peptides were surveyed for reactivity with a panel of positive sera (from LD patients) and negative sera (from subjects without LD), thus identifying 17 peptides with a diagnostically useful binding pattern: reactivity with at least three positive sera and no reactivity with any of the negative sera. The peptides define eight sequence motifs, none of which can be matched convincingly with segments of proteins from Borrelia burgdorferi, the LD pathogen; evidently, then, they are "mimotopes," mimicking natural pathogen epitopes without matching contiguous amino acids of pathogen proteins. Peptides like these could be the basis of a new diagnostic enzyme-linked immunosorbent assay for LD, with sufficient specificity and sensitivity to replace expensive immunoblotting tests that are currently required for definitive serological diagnosis. Moreover, the method used to discover these peptides did not require any knowledge of the pathogen and involved generic procedures that are applicable to almost any infectious disease, including emerging diseases for which no pathogen has yet been identified.

Structural relatedness of distinct determinants recognized by monoclonal antibody TP25.99 on beta 2-microglobulin-associated and beta 2-microglobulin-free HLA class I heavy chains

The association of HLA class I heavy chains with beta2-microglobulin (beta2m) changes their antigenic profile. As a result, Abs react with either beta2m-free or beta2m-associated HLA class I heavy chains. An exception to this rule is the mAb TP25.99, which reacts with both beta2m-associated and beta2m-free HLA class I heavy chains. The reactivity with beta2m-associated HLA class I heavy chains is mediated by a conformational determinant expressed on all HLA-A, -B, and -C Ags. This determinant has been mapped to amino acid residues 194-198 in the alpha3 domain. The reactivity with beta2m-free HLA class I heavy chains is mediated by a linear determinant expressed on all HLA-B Ags except the HLA-B73 allospecificity and on <50% of HLA-A allospecificities. The latter determinant has been mapped to amino acid residues 239-242, 245, and 246 in the alpha3 domain. The conformational and the linear determinants share several structural features, but have no homology in their amino acid sequence. mAb TP25.99 represents the first example of a mAb recognizing two distinct and spatially distant determinants on a protein. The structural homology of a linear and a conformational determinant on an antigenic entity provides a molecular mechanism for the sharing of specificity by B and TCRs.

Homodimeric peptides displayed by the major coat protein of filamentous phage

Peptide libraries displayed by filamentous bacteriophage have proven a powerful tool for the discovery of novel peptide agonists, antagonists and epitope mimics. Most phage-displayed peptides are fused to the N terminus of either the minor coat protein, pIII, or the major coat protein, pVIII. We report here that peptides containing cysteine residues, displayed as N-terminal fusions to pVIII, can form disulfide-bridged homodimers on the phage coat. Phage clones were randomly selected from libraries containing one or two fixed Cys residues, and surveyed for the presence of peptide-pVIII homodimers by SDS-PAGE analysis that involved pretreatment of the phage with reducing or thiol-modifying agents. For all phage whose recombinant peptide contained a single Cys residue, a significant fraction of the peptide-pVIII molecules were displayed as dimers on the phage coat. The dimeric form was in greater abundance than the monomer in almost all cases in which both forms could be reliably observed. Occasionally, peptides containing two Cys residues also formed dimers. These results indicate that, for a given pVIII-displayed peptide bearing a single Cys residue, a significant fraction of the peptide (>40 %) will dimerize regardless of its sequence; however, sequence constraints probably determine whether all of the peptide will dimerize. Similarly, only occasionally do peptides bearing two Cys residues form intermolecular disulfide bridges instead of intramolecular ones; this indicates that sequence constraints may also determine dimerization versus cyclization. Sucrose-gradient analysis of membranes from cells expressing pVIII fused to a peptide containing a single Cys residue showed that dimeric pVIII is present in the cell prior to its assembly onto phage. A model of the peptide-pVIII homodimer is discussed in light of existing models of the structure and assembly of the phage coat. The unique secondary structures created by the covalent association of peptides on the phage surface suggest a role for homo- and heterodimeric peptide libraries as novel sources of bioactive peptides.

Characterization of murine coronavirus neutralization epitopes with phage-displayed peptides

Phage-displayed peptide libraries were used to map immunologically relevant epitopes on the surface (S) glycoprotein of a neurotropic murine coronavirus (MHV-A59). Three in vitro virus-neutralizing and in vivo protective mAbs against either continuous or discontinuous epitopes on the S glycoprotein were used to screen 12 different peptide libraries expressed on the pVIII major coat protein of the fd filamentous bacteriophage. Consensus sequences that matched short sequences within the S glycoprotein were identified. The sequence of a tight-binding, mAb-selected peptide suggested the location of a discontinuous epitope within the N-terminal S1 subunit. Several tightly binding phage were amplified and used directly as immunogens in BALB/c and C57BL/6 mice. Partial protection of C57BL/6 mice against a lethal acute virus infection was achieved with a phage preparation that displayed a linear epitope. Protection correlated with the presence of sufficient levels of specific antiviral antibodies recognizing the same immunodominant domain and 13-mer peptide, located within the C-terminal S2 subunit, as the selecting mAb. Thus, the direct use of phage-displayed peptides to evaluate protective antiviral immune responses complements their use to characterize antibody-binding epitopes. This is the first evaluation of protective immunization induced by mAb-selected phage-displayed peptides.

C1q-binding peptides share sequence similarity with C4 and induce complement activation

Two peptide motifs that bind to C1q have been identified from phage displayed libraries. A first panning cycle recovered phage that displayed a [N/S]PFxL motif. A synthetic peptide with that motif blocked those phage from binding to C1q. A second panning cycle was conducted with the [N/S]PFxL motif peptide present, leading to recovery of phage displaying a different motif, SHY. The two motifs are specific for C1q and are competed by DNA and the cognate synthetic peptide but not by immunoglobulins. Phage displayed peptide sequences containing the [N/S]PFxL have significant sequence similarity to a region of complement component C4, suggesting a possible site of interaction between C4, or one of its processed forms, and C1q. The SHY motif peptide induces C4 consumption in a hemolytic assay, suggesting that it activates C1 independent of immune complexes. This peptide may activate C1 by a mechanism similar to the beta-amyloid peptides found in Alzheimer's disease.

Identification of antigenic regions on VP2 of African horsesickness virus serotype 3 by using phage-displayed epitope libraries

VP2 is an outer capsid protein of African horsesickness virus (AHSV) and is recognized by serotype-discriminatory neutralizing antibodies. With the objective of locating its antigenic regions, a filamentous phage library was constructed that displayed peptides derived from the fragmentation of a cDNA copy of the gene encoding VP2. Peptides ranging in size from approximately 30 to 100 amino acids were fused with pIII, the attachment protein of the display vector, fUSE2. To ensure maximum diversity, the final library consisted of three sub-libraries. The first utilized enzymatically fragmented DNA encoding only the VP2 gene, the second included plasmid sequences, while the third included a PCR step designed to allow different peptide-encoding sequences to recombine before ligation into the vector. The resulting composite library was subjected to immunoaffinity selection with AHSV-specific polyclonal chicken IgY, polyclonal horse immunoglobulins and a monoclonal antibody (MAb) known to neutralize AHSV. Antigenic peptides were located by sequencing the DNA of phages bound by the antibodies. Most antigenic determinants capable of being mapped by this method were located in the N-terminal half of VP2. Important binding areas were mapped with high resolution by identifying the minimum overlapping areas of the selected peptides. The MAb was also used to screen a random 17-mer epitope library. Sequences that may be part of a discontinuous neutralization epitope were identified. The amino acid sequences of the antigenic regions on VP2 of serotype 3 were compared with corresponding regions on three other serotypes, revealing regions with the potential to discriminate AHSV serotypes serologically.

Affinity for the cognate monoclonal antibody of synthetic peptides derived from selection by phage display. Role of sequences flanking thebinding motif

Randomized peptide sequences displayed at the surface of filamentous phages are often used to select antibody ligands. The selected sequences are generally further used in the form of synthetic peptides; however, as such, their affinity for the selecting antibody is extremely variable and factors influencing this affinity have not been fully deciphered. We have used an f88.4 phage-displayed peptide library to identify ligands of mAb 11E12, an antibody reactive to human cardiac troponin I. A majority of the sequences thus selected showed a (T/A/I/L) EP(K/R/H) motif, homologous to the Y-TEPH motif identified by multiple peptide synthesis as the critical motif recognized by mAb 11E12 in the peptide epitope. A set of 15-mer synthetic peptides derived from the phage-selected sequences was used in BIACORE to characterize their interaction with mAb 11E12. Most peptides exhibited affinities in the 7-26 nM range. These affinities represented, however, only 1.9-7. 5% of the affinity of the 15-mer peptide epitope. In circular dichroism experiments, the peptide epitope showed a propensity to have some stabilized conformation, whereas a low-affinity peptide selected by phage-display did not. To try to decipher the molecular basis of this difference in affinity, new peptides were prepared by grafting the N- or the C-terminal sequence of the peptide epitope to the Y-TEPK motif of a low-affinity peptide selected by phage-display. These hybrid peptides showed marked increases both in affinity (as assessed using BIACORE) and in inhibitory potency (as assessed in competition ELISA), compared with the parent sequence. Thus, the sequences flanking the motif, although not containing critical residues, convey some determinants necessary for high affinity. The affinity of a given peptide strongly depends on its capacity to maintain the antigenically reactive structure it has on the phage, implying that it is impossible to predict whether high- or low-affinity peptides will be obtained from phage display.

Sequential determination of ligands binding to discrete components in heterogeneous mixtures by iterative panning and blocking (IPAB)

Biopanning has been used extensively in conjunction with purified components, but there are also examples in which mixtures of targets have been investigated. This study introduces a methodological innovation, termed iterative panning and blocking (IPAB), to extend the range of specific interactions that can be probed in mixtures. Here this procedure is used to probe a mixture of high molecular mass components of human cord blood with phage-peptide display libraries. The initial panning recovered phage that bore the consensus motif Gly-Pro-Arg-Pro, a known fibrinogen-binding motif. These phage bound specifically to purified fibrinogen. A series of peptides containing the Gly-Pro-Arg-Pro motif efficiently blocked the binding of phage having the same motif, presumably by binding to their common target. A second round of panning was performed against the same target mixture in the presence of this blocking peptide. Phage recovered from this second panning exhibited a motif (Ser-His-Tyr) that was subsequently shown to bind specifically to complement component C1q. A second peptide containing this motif specifically blocked the interaction of the phage with C1q. A third round of panning performed in the presence of both the fibrinogen- and the C1q- blocking peptides yielded phage with a new peptide motif (Asn-Pro-Phe) that also bound specifically to C1q, apparently at a new site. The three motifs isolated through this iterative process were distinct in that each was blocked only by its corresponding peptide. This IPAB strategy can be applied to many high diversity selection procedures that target complex mixtures.

Emergence of porcine reproductive and respiratory syndrome virus deletion mutants: correlation with the porcine antibody response to a hypervariable site in the ORF 3 structural glycoprotein

By using porcine immune sera to select a library of phage-displayed random peptides, we identified an antigenic sequence (RKASLSTS) in the C-terminus of the ORF 3 structural glycoprotein of European-type porcine reproductive and respiratory syndrome virus (PRRSV). Through the use of overlapping reading frames, the same PRRSV genetic locus codes for the ORF 3 "RKASLSTS" sequence, and a previously described ORF 4 epitope (Meulenberg, J. J. M., Van Nieuwstadt, A. P., Van Essen-Zandbergen, A., and Langeveld, J. P. M., 1997, J. Virol. 71, 6061-6067). Sequence analysis identified naturally occurring deletion mutants at this ORF 34 site. Phylogenetic analysis showed the presence of a highly accurate ORF 3 molecular clock, according to which deletion mutants and nondeleted viruses evolved at differing speeds. Furthermore, deletion mutants and nondeleted viruses evolved as separate lineages. These distinctions suggested that deletion mutants were a hitherto unrecognized subtype of European-type PRRSV. Currently, deletion mutants appear to be outcompeting nondeleted viruses in the field, highlighting the importance of the porcine antibody response against the minor structural glycoproteins of European-type PRRSV for viral evolution.

A rheumatoid factor specific mimotope identified by a peptide display library

We screened a 10 amino acid peptide display library in filamentous phage with B'20, a monoclonal high affinity IgM rheumatoid factor (RF) expressing the VkIIIa-dependent 4C9 idiotype. Using direct and indirect selection techniques, 12 B'20 reactive peptides were identified, 9 of which belonged to one of two motifs. Binding of B'20 to phage-bearing peptides was inhibited by both IgG and 4C9 antiidiotype. Synthetic peptides corresponding to the two motifs inhibited the Fc binding of a low avidity IgA B'20 construct. Purified IgM from 6/8 RF-positive RA patients and 8/11 monoclonal RFs with VkIII-encoded light chains bound to the phage, whereas none of the four monoclonal RFs with VkI or VkII encoded light chains bound. Phage binding appeared to be RF specific. Three 4C9 positive/RF negative cell lines from RA patients did not bind to phage nor did three B'20 mutants that had lost RF specificity, whereas two mutants that retained RF specificity also retained phage binding. We propose that there is a common epitope(s) recognized by VkIII encoded RFs that is mimicked by the structure of these peptides. Such mimotopes might be exploited to design novel agents that interfere with autoantibody binding.

Isolation of peptides that mimic epitopes on a malarial antigen from random peptide libraries displayed on phage

The ring-infected erythrocyte surface antigen (RESA) is a dense-granule protein of Plasmodium falciparum which binds to the cytoskeletal structure of the erythrocyte after parasite invasion. It is currently under trial as a vaccine candidate. In an effort to characterize further the antibody responses to this antigen, we have panned two independent libraries of random peptides expressed on the surface of filamentous phage with a monoclonal antibody (MAb 18/2) against RESA. One library consisted of a potentially constrained 17-mer peptide fused with the gpVIII phage coat protein, and the other displayed an unconstrained 15-mer as a fusion with the minor phage coat protein gpIII. Several rounds of biopanning resulted in enrichment from both libraries clones that interacted specifically with MAb 18/2 in protein-blotting and enzyme-linked immunosorbent assay experiments. Nucleotide sequencing of the random oligonucleotide insert revealed a common predominant motif: (S/T)AVDD. Several other clones had related but degenerate motifs. Thus, a monoclonal antibody against a malarial antigen can select common mimotopes from different random peptide libraries. We envisage many uses for this technology in malaria research.

Cell-Specific Peptide Binding by Human Neutrophils

Analysis of peptide binding to human neutrophils (PMN) using phage display techniques has revealed cell-specific motifs reactive with the PMN surface. Phage libraries displaying either linear 9-mer or cyclic 10-mer and 6-mer peptides were incubated with normal human neutrophils followed by elution of bound phage with low pH (pH 2.2) and non-ionic detergent. Three rounds of selection generated several related peptide sequences that bound with high avidity to PMN. Using the linear 9-mer library, PMN-binding phage expressed peptides with the motif (G/A)PNLTGRW. The binding of phage bearing this motif was highly specific since no binding was observed on lymphocytes, fibroblasts, epithelial, or endothelial cells. Functional assays revealed that phage bearing the sequence FGPNLTGRW induced a pertussis toxin-sensitive increase in PMN cytosolic calcium analogous to that observed with Gi coupled receptors. Other prominent motifs identified included phage bearing the consensus DLXTSK(M/L)X(V/I/L), where X represents a non-conserved position. Phage with this motif bound exclusively to a sub population of human PMN that comprised approximately 50% of the total and did not elicit a calcium response. The binding of such phage to PMN was prevented by co-incubation with competing peptides displaying identical or similar sequences (IC50 range from 0.6 μmol/L to 50 μmol/L for DLXTSK and GPNLTG, respectively). We speculate that these techniques will be useful in identifying functional cell-specific binding motifs and contribute to the development of new therapeutic and diagnostic strategies in human disease.

Cross-reactive epitope mimics in a fragmented-genome phage display library derived from the rickettsia, Cowdria ruminantium

BACKGROUND

Epitopes can be mapped by comparing immunoaffinity-selected peptides from fragmented-gene display libraries with the target gene. With larger libraries derived from unsequenced genomes, this is not possible. Spurious epitope mimics may be created by expressing DNA in a variety of meaningless reading frames and orientations.

OBJECTIVES

To determine empirically whether panning a large fragmented-genome phage display library with antibodies to MAP1, the major antigenic protein of the rickettsial parasite Cowdria ruminantium, would result in the selection of irrelevant, cross-reactive mimotopes.

STUDY DESIGN

A gene III phage library displaying peptides derived from C. ruminantium was constructed using cloned DNA from a bacteriophage lambda genomic library. After in vivo excision, plasmids were cleaved with PvuII followed by PCR. Genes with a PvuII site, including MAP1 were therefore not amplified. DNA was sonicated, partially digested with DNase and cloned into the display vector fUSE2. Affinity-purified MAP1 antibodies were used for panning. Peptides expressed by panned phages were tested for recognition in Western blot and ELISA. Oligonucleotides representing antigenic sequences were used to locate their encoding DNA sequences in the original lambda library. The phage display library was also panned with two monoclonal antibodies (Mabs) against bluetongue virus (BTV).

RESULTS

Five different peptide sequences were selected from the MAP1-deficient phage display library. None was identical to MAP1, but four peptides had regions that were similar, both to each other, and to the parasite protein. They produced strong signals in ELISA and Western blot. None could be located to any C. ruminantium open reading frame. Two BTV Mabs recognised a sequence similar to their authentic epitope.

CONCLUSION

Large genome-targeted phage display libraries may be sufficiently diverse to allow the selection of peptides that mimic actual antigenic determinants. This diversity may be exploited in the search for useful epitopes.

Cell-Specific Peptide Binding by Human Neutrophils

Analysis of peptide binding to human neutrophils (PMN) using phage display techniques has revealed cell-specific motifs reactive with the PMN surface. Phage libraries displaying either linear 9-mer or cyclic 10-mer and 6-mer peptides were incubated with normal human neutrophils followed by elution of bound phage with low pH (pH 2.2) and non-ionic detergent. Three rounds of selection generated several related peptide sequences that bound with high avidity to PMN. Using the linear 9-mer library, PMN-binding phage expressed peptides with the motif (G/A)PNLTGRW. The binding of phage bearing this motif was highly specific since no binding was observed on lymphocytes, fibroblasts, epithelial, or endothelial cells. Functional assays revealed that phage bearing the sequence FGPNLTGRW induced a pertussis toxin-sensitive increase in PMN cytosolic calcium analogous to that observed with Gi coupled receptors. Other prominent motifs identified included phage bearing the consensus DLXTSK(M/L)X(V/I/L), where X represents a non-conserved position. Phage with this motif bound exclusively to a sub population of human PMN that comprised approximately 50% of the total and did not elicit a calcium response. The binding of such phage to PMN was prevented by co-incubation with competing peptides displaying identical or similar sequences (IC50 range from 0.6 μmol/L to 50 μmol/L for DLXTSK and GPNLTG, respectively). We speculate that these techniques will be useful in identifying functional cell-specific binding motifs and contribute to the development of new therapeutic and diagnostic strategies in human disease.

Molecular mimicry of phage displayed peptides mimicking GD3 ganglioside

We have analyzed sequence homologies between nonimmunogenic phage displayed peptides mimicking GD3 ganglioside and human/mouse self-proteins. The GD3 ganglioside phagotopes showed homology to proteins involved in carbohydrate metabolism/transport. Besides this expected homology, molecular mimicry of critical regulatory proteins was found. These data contribute to our understanding of the structural relatedness of antigenic determinants defined by specific anti-GD3 monoclonal antibodies and, in addition, suggest that molecular mimicry might explain the nonimmunogenicity of these peptides otherwise characterized by specificity to the mAb counterpart. We conclude that construction of peptides harboring motifs absent or scarcely represented in endogenous self-proteins might be a useful approach in melanoma immunotherapy.

Patch engineering: a general approach for creating proteins that have new binding activities

Patch engineering is a technique for creating folded proteins that have new binding activities. Different protein scaffolds are used to present a patch of discontinuous residues on a folded-protein surface. By varying simultaneously the residues in these patches and displaying these mutant proteins on phage, one can select proteins that have new binding activities. Patch engineering is applicable to any protein fold. Novel proteins derived by this approach might replace antibodies in certain applications or provide lead molecules for the design of non-peptide analogues.

All-D peptides recognized by an anti-carbohydrate antibody identified from a positional scanning library

Monoclonal antibodies recognize antigens with high affinity and specificity, but the structural basis for molecular mimicry remains unclear. It is often assumed that cross-reactive antigens share some structural similarity that is specifically recognized by a monoclonal antibody. Recent studies using combinatorial libraries, which are composed of millions of sequences, have examined antibody cross-reactivity in a manner entirely different from traditional epitope mapping approaches. Here, peptide libraries were screened against an anti-carbohydrate monoclonal antibody for the identification of peptide mimics. Positional scanning libraries composed of all-l or all-d hexapeptides were screened for inhibition of monoclonal antibody HGAC 39.G3 binding to an antigen displaying N-acetyl-d-glucosamine (GlcNAc) residues on a polyrhamnose backbone. Inhibitory activity by mixtures from the all-d hexapeptide library was greater than the activity from the all-l libraries. The most active d-amino acid residues defined in each of the six positions of the library were selected to prepare 27 different individual hexapeptides. The sequence Ac-yryygl-NH2 was specifically recognized by mAb HGAC 39.G3 with a relative affinity of 300 nM when measured in a competitive binding assay. The contributions to overall specificity of the residues of the all-d peptide (Ac-yryygl-NH2) in binding to mAb HGAC 39.G3 were examined with a series of truncation, l and d-amino acid substitution, and retro analogs. Dimeric forms of the all-d peptide were recognized with tenfold to 100-fold greater affinities relative to the monomer. The all-d peptide was found to inhibit mAb HGAC 39.G3 binding to an anti-idiotype antibody with approximately 1000-fold greater affinity than GlcNAc. As demonstrated here, the study of immune recognition using combinatorial chemistry may offer new insights into the molecular basis of cross-reactivity.

The maltose-binding protein as a scaffold for monovalent display of peptides derived from phage libraries

Random peptide libraries are displayed on filamentous bacteriophage as fusions to either the minor coat protein, pIII, or the major coat protein, pVIII. We have devised a means of isolating the peptide displayed on a phage clone by transferring it to the N-terminus of the maltose-binding protein (MBP) of Escherichia coli encoded by malE. Transfer of a peptide sequence to monomeric MBP eliminates phage-encoded amino acids downstream of the insert peptide as well as avidity effects caused by multivalent display on phage. Peptide:MBP fusions are also easily affinity purified on amylose columns. The pMal-p2 vector was engineered to accept phage DNA encoding pIII- and pVIII-displayed peptides fused to their respective leader sequences. Both types of leader sequence were shown to target the peptide:MBP fusions to the periplasm of E. coli. A streamlined procedure for transferring peptides to MBP was applied to clones that had been isolated from a panel of pVIII-displayed peptide libraries by screening with an HIV-1-specific monoclonal antibody (Ab). By enzyme-linked immunosorbent assay, the Ab bound each of the peptide:MBP fusions and required the presence of a disulfide bridge within each peptide. Some of the peptide:MBP fusions were also analyzed using surface plasmon resonance. Thus, our study shows the value of malE fusion vectors in characterizing phage-displayed peptides.

The role of structure in antibody cross-reactivity between peptides and folded proteins

Peptides have the potential for targeting vaccines against pre-specified epitopes on folded proteins. When polyclonal antibodies against native proteins are used to screen peptide libraries, most of the peptides isolated align to linear epitopes on the proteins. The mechanism of cross-reactivity is unclear; both structural mimicry by the peptide and induced fit of the epitope may occur. The most effective peptide mimics of protein epitopes are likely to be those that best mimic both the chemistry and the structure of epitopes. Our goal in this work has been to establish a strategy for characterizing epitopes on a folded protein that are candidates for structural mimicry by peptides. We investigated the chemical and structural bases of peptide-protein cross-reactivity using phage-displayed peptide libraries in combination with computational structural analysis. Polyclonal antibodies against the well-characterized antigens, hen eggwhite lysozyme and worm myohemerythrin, were used to screen a panel of phage-displayed peptide libraries. Most of the selected peptide sequences aligned to linear epitopes on the corresponding protein; the critical binding sequence of each epitope was revealed from these alignments. The structures of the critical sequences as they occur in other non-homologous proteins were analyzed using the Sequery and Superpositional Structural Assignment computer programs. These allowed us to evaluate the extent of conformational preference inherent in each sequence independent of its protein context, and thus to predict the peptides most likely to have structural preferences that match their protein epitopes. Evidence for sequences having a clear structural bias emerged for several epitopes, and synthetic peptides representing three of these epitopes bound antibody with sub-micromolar affinities. The strong preference for a type II beta-turn predicted for one peptide was confirmed by NMR and circular dichroism analyses. Our strategy for identifying conformationally biased epitope sequences provides a new approach to the design of epitope-targeted, peptide-based vaccines.

Phage-displayed peptide libraries

Over the past year, significant advances have been achieved through the use of phage-displayed peptide libraries. A wide variety of bioactive molecules, including antibodies, receptors and enzymes, have selected high-affinity and/or highly-specific peptide ligands from a number of different types of peptide library. The demonstrated therapeutic potential of some of these peptides, as well as new insights into protein structure and function that peptide ligands have provided, highlight the progress made within this rapidly-expanding field.

Epitope mapping of a function-blocking beta 1 integrin antibody by phage display

Integrins are a major class of cell surface receptors involved in cell-cell and cell-matrix adhesion and communication. Ha2/11 is a function-blocking anti-rat beta 1 integrin hamster IgM that should be a useful reagent for understanding beta 1 integrin function. We demonstrate that Ha2/11 cross reacts with human, Xenopus, and Drosophila beta 1 integrins, and use phage display to map the epitope for Ha2/11 to residues within the sequence LRSGEPQTF which lies 18 amino acids proximal to the putative I domain in beta 1 integrins. Monoclonal antibody mapping experiments, mutational analyses, and direct binding assays have implicated integrin I domains in both cation and ligand binding. Our data therefore suggest that Ha2/11 blocks beta 1 integrin function by interfering with I domain-mediated ligand binding.

Induction of anti-carbohydrate antibodies by phage library-selected peptide mimics

One of the prerequisites for the development of polysaccharide subunit vaccines is the induction of an efficient immune response to carbohydrate antigens like lipopolysaccharide (LPS) or capsular polysaccharide antigens of pathogens. In an attempt to overcome the problems that arise from the T-independent immune response induced by such antigens, selecting peptide sequences that mimic protective carbohydrate epitopes has been proposed. In this study, we investigate a new selection strategy for immunogenic peptide mimics using the phage-displayed peptide library technology. Two monoclonal antibodies (mAb) of the A isotype (mIgA), mIgA C5 and mIgA I3, specific for the O-antigen (O-Ag) part of the human pathogen Shigella flexneri serotype 5a LPS and protective against homologous infection were used to screen two phage-displayed nonapeptide libraries in pVIII. Using mIgA C5, 13 different specific clones were selected, and 6 using mIgA I3; 5 of the latter also interacted in enzyme-linked immunosorbent assay with the first mAb. All of the 19 clones selected were separately used to immunize mice, but only 2 of them, p100c (mIgA I3-specific) and p115 (interacting with both mIgA) were able to induce anti-O-Ag antibodies. The immune response was specific for the O-Ag of the S. flexneri serotype 5a, and also selectively recognized the corresponding bacterial strain. The amino acid sequences of p100c and p115 immunogenic peptide mimics were YKPLGALTH (flanked by two Cys residues) and KVPPWARTA, respectively. These results are the first example of immunogenic mimicry of carbohydrates by phage-displayed peptides, and indicate a new strategy of selection of immunogens for the development of anti-polysaccharide vaccines.

Assaying phage-borne peptides by phage capture on fibrinogen or streptavidin

There is no simple and efficient method for assaying phage isolated from libraries without having to resort to PEG purification of the phage, or to the biotinylation or other labelling of the target molecule. We report here a method for producing 'bifunctional' phage that express two types of peptide; one peptide, fused to pVIII, will bind to immobilized fibrinogen, allowing capture of the phage out of culture supernatants; this allows the other peptide, fused to pIII or pVIII to be assayed by simple ELISA. This system has also been developed for the capture of phage bearing a streptavidin-binding peptide. The bifunctional phage are produced by bacterial cells bearing a plasmid that expresses pVIII fused either to the fibrinogen-binding peptide or to the streptavidin-binding one. Thus, when these cells are infected with a phage clone or pool to be assayed, phage will be produced whose 'capture-peptide' is produced from the plasmid and whose 'assay-peptide' is produced from the phage genome. We show here that, by this method, bifunctional phage can be produced that will bind to immobilized streptavidin or fibrinogen.

Antibody engineering

The development of recombinant techniques for the rapid cloning, expression, and characterization of cDNAs encoding antibody (Ab) subunits has revolutionized the field of antibody engineering. By fusion to heterologous protein domains, chain shuffling, and inclusion of self-assembly motifs, novel molecules such as bispecific Abs can now be generated which possess the subset of functional properties designed to fit the intended application. Rapid technological developments in phage display of peptides and proteins have led to a plethora of applications directed towards immunology and antibody engineering. Many of the problems associated with the therapeutic use of Abs are being addressed by the application of these new techniques.

Exploring the basis of peptide-carbohydrate crossreactivity: evidence for discrimination by peptides between closely related anti-carbohydrate antibodies

To investigate the molecular basis of antigenic mimicry by peptides, we studied a panel of closely related mAbs directed against the cell-wall polysaccharide of group A Streptococcus. These antibodies have restricted V-gene usage, indicating a shared mechanism of binding to a single epitope. Epitope mapping studies using synthetic fragments of the cell-wall polysaccharide supported this conclusion. All of the mAbs isolated crossreactive peptides from a panel of phage-displayed libraries, and competition studies indicated that many of the peptides bind at or near the carbohydrate binding site. Surprisingly, the peptides isolated by each mAb fell into distinct consensus-sequence groups that discriminated between the mAbs, and in general, the peptides bound only to the mAbs used for their isolation. Similar results were obtained with polyclonal antibodies directed against synthetic oligosaccharide fragments of the streptococcal cell-wall polysaccharide. Thus, the peptides appear to be specific for their isolating antibodies and are not recognized by the same mechanism as their carbohydrate counterparts.

Probing for cerebrospinal fluid antibody specificities by a panel of random peptide libraries

In a state of health the central nervous system is nearly devoid of macrophages, T and B cells. However, such cells are required to initiate the inflammation associated with multiple sclerosis (MS). An important issue in understanding the pathogenesis of such conditions is to fully define the T and B cell specificities. Using random peptide libraries, we have analysed the cerebrospinal fluid antibody specificities within one oligoclonal "band" from a patient with MS. The selected peptides from a 6-mer library revealed two peptide motifs S (S/T/Q) (R/S) (N/G) FP and PRn (G/P) FF. Interestingly, the first motif was also selected from a 15-mer library, while the second motif was selected from the 9-mer library. In addition, distinct but structurally related peptides to the mentioned motifs were also selected. Surprisingly, a SwissProt search with these motifs revealed a significant linear homology with collagen proteins, the 68 kDa neurofilament protein, versican and other proteins from viruses such as herpes simplex virus, human cytomegalovirus and human papillomavirus. Thus, the observation that antibodies present within one oligoclonal immunoglobulin band would recognize peptides, some that might be related by only the structure, raises the question of the involvement of multiple pathogens in MS.

Display of heterologous proteins on the surface of microorganisms: from the screening of combinatorial libraries to live recombinant vaccines

In recent years there has been considerable progress towards the development of expression systems for the display of heterologous polypeptides and, to a lesser extent, oligosaccharides on the surface of bacteria or yeast. The availability of protein display vectors has in turn provided the impetus for a range of exciting technologies. Polypeptide libraries can be displayed in bacteria and screened by cell sorting techniques, thus simplifying the isolation of proteins with high affinity for ligands. Expression of antigens on the surface of nonvirulent microorganisms is an attractive approach to the development of high-efficacy recombinant live vaccines. Finally, cells displaying protein receptors or antibodies are of use for analytical applications and bioseparations.