Peptide ligands for a sugar-binding protein isolated from a random peptide library

Epitope and mimotope for an antibody to the Na, K-ATPase

The epitope of a monoclonal antibody specific for the alpha 2 isoform of the Na,K-ATPase was determined and its accessibility in native enzyme was examined. Protein fragmentation with N-chlorosuccinimide, formic acid, trypsin, and leucine aminopeptidase indicated binding near the Na,K-ATPase N-terminus but did not unambiguously delineate the extent of the epitope. The ability of the antibody to bind to denatured enzyme made it a good candidate for screening a random peptide library displayed on M13 phage, but the consensus sequence that emerged was not found in the Na,K-ATPase, Full-length cDNA for the Na,K-ATPase was randomly fragmented and cloned into beta-galactosidase to create a lambda gt11 expression library; screening with the antibody yielded a set of overlaps spanning 23 amino acids at the N-terminus. Chimeras of Na,K-ATPase alpha 1 and alpha 2 narrowed down the epitope to 14-19 amino acids. The antibody did not recognize fusion proteins constructed with shorter segments of this epitope. It did recognize a fusion protein containing the M13 library consensus sequence, however, indicating that this sequence, which is rich in proline and hydrophobic amino acids (FPPNFLFPPPP), was a mimotope. The natural epitope, unique to the Na,K-ATPase alpha 2 isoform, was GREYSPAATTAENG. Reconstitution of antibody binding in a foreign context such as M13 PIII protein or beta-galactosidase thus required a relatively large number of amino acids, indicating that antibody mapping approaches must allow for epitopes of significant size. The epitope was accessible in native enzyme and exposed on the cytoplasmic side, documenting the surface exposure of a stretch of amino acids at the N-terminus, where the Na,K-ATPase isoforms differ most.

Peptide mimicry of carbohydrate epitopes on human immunodeficiency virus

Cancer-related, mucin-type carbohydrate epitopes, principally mannose and sialo-syl residues, are expressed on the envelope protein gp 160 of the human immunodeficiency virus (HIV). Anticarbohydrate antibodies directed toward these and other carbohydrate epitopes are known to neutralize HIV-1 infection by cell-free virus. Carbohydrates, however, being T cell-independent antigens, typically elicit diminished immune responses. To overcome this potential draw back, we have examined the ability of peptides that mimic such epitopes to elicit immune responses that cross-react with carbohydrate structures. We report that mouse polyclonal antisera generated against peptides that mimic mucin-related carbohydrate epitopes have anti-HIV-1 activity. Generation of antibodies was not lr-gene restricted, as at least two different strains of mice. Balb/c (H-2d) and C57Bl/6 (H-2b), responded equally to the peptides. The antipeptide sera displayed neutralizing activity against HIV-I/MN and HIV-I/3B viral strains. This neutralization was as good as human anti-HIV sera. These results indicate that peptide mimics of carbohydrates provide a novel strategy for the further development of reagents that elicit immune responses to carbohydrate epitopes associated with many infectious organisms and tumor cells.

Identification of a novel type 1 diabetes-specific epitope by screening phage libraries with sera from pre-diabetic patients

We used random peptide libraries displayed on phage to search for ligands to insulin dependent diabetes mellitus-related antibodies and were able to identify several candidate disease-related peptides. One of them, clone 92, showed a significant difference in the frequency of reactivity with the sera of patients and normal controls. Human immunoglobulins immunopurified on phage 92 specifically stained the islets on human pancreatic sections. When injected into rabbits, the selected peptide elicited antibodies that also stained human and rat pancreatic sections, with a pattern similar to that observed with immunoglobulins purified from the sera of patients. No reactivity was observed in other tissues. Our results indicate that the peptide identified in this work mimics a novel, diabetes-related self-antigen.

Mimotopes of polyreactive anti-DNA antibodies identified using phage-display peptide libraries

Three monoclonal IgG2a anti-DNA polyreactive autoantibodies, derived from lupus-prone mice (NZB x NZW)F1, were studied by surface plasmon resonance (BIAcore) analysis using three different synthetic double-stranded (ds) oligonucleotides of 25, 30, and 25 base pairs (bp). These monoclonal antibodies (mAb) exhibited dissociation rate constants (k(off)), ranging from 0.0001 (mAb F14.6 and F4.1) to 0.01/s (mAb J20.8) and k(on) ranging from 2 x 10(5) to 2 x 10(6) /M/s. The screening of a constrained random peptide library displayed on M13 bacteriophages on these mAb allowed the determination of the specific consensus motifs (mimotopes) for mAb F14.6 and J20.8, but not for mAb F4.1. No cross-reaction was observed between F14.6- and J20.8-specific peptides (and vice versa). Binding of all phages selected on F14.6 was inhibited with 700 ng/ml soluble DNA. The binding of one group of peptides selected on J20.8 was inhibited by 400 ng/ml soluble DNA, of a second group by 2500 ng/ml, while binding of a third group could not be inhibited. The determined consensus sequences do not match with known sequences. Peptides specific for F14.6 share negative charges and aromatic rings that may mimic a DNA backbone, while peptides selected with J20.8 do not bear any negative charge, implying a different kind of molecular recognition, for example hydrogen or salt bonds. The peptides selected on J20.8 also bind serum antibodies from human patients with systemic lupus erythematosus. In addition, BALB/c mice immunized with some of the selected phages exhibit high serum titers of IgG3 anti-dsDNA antibodies, further supporting the hypothesis that peptide epitopes may mimic an oligonucleotide structure.

Peptide epitopes recognized by a human anti-cryptococcal glucuronoxylomannan antibody

Cryptococcus neoformans causes meningitis in 6 to 8% of individuals with AIDS. Recently, immunotherapeutic modalities including antibody therapy have been proposed for the treatment of cryptococcal meningitis in AIDS patients. This is a rational approach because existing antifungal agents fail to eradicate the infection in the setting of profound immunosuppression. Both murine and human antibodies elicited by the investigational cryptococcal capsular polysaccharide vaccine glucuronoxylomannan-tetanus toxoid (GXM-TT) have been shown to be biologically functional in different model systems. The human immunoglobulin M (lambda) GXM monoclonal antibody (MAb) 2E9 expresses idiotypes that are also found in naturally occurring anti-GXM antibodies and opsonic GXM-TT sera. However, the specificity of human anti-GXM antibodies and their possible role in protection against cryptococcosis are not known. In an effort to discover epitopes that are recognized by human anti-GXM antibodies, we screened a random decapeptide phage display library with the human anti-GXM MAb 2E9. An enzyme-linked immunosorbent assay (ELISA)-based screening method led to the selection of phages with peptide inserts that bound 2E9 and inhibited 2E9-GXM binding. Analysis of the amino acid sequences of these phages revealed an increased frequency of combinations of QTGLD residues. Inhibition ELISAs demonstrated that phages with QTG/TL/D motifs inhibited 2E9-GXM binding better than phages with different motifs. A peptide synthesized from one of the inhibitory phages, peptide 13 (GMDGT QLDRW), inhibited GXM binding to solid-phase 2E9 and 2E9 binding to solid-phase GXM. Peptide 13 also inhibited the GXM binding of GXM-TT immune sera and naturally occurring serum antibodies from human immunodeficiency virus (HIV)-negative, but not HIV-positive, individuals. Taken together, our data indicate that the peptide epitopes selected by 2E9 mimic GXM epitopes and that peptide 13 may be a mimotope of a GXM epitope that is recognized by human anti-GXM antibodies.

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.

Peptide inhibition of glomerular deposition of an anti-DNA antibody

Antibodies to double-stranded DNA are pathognomonic of systemic lupus erythematosus and deposit in the kidneys of lupus patients to cause glomerulonephritis. Recent data suggest that a significant proportion of anti-DNA antibodies may cross-react with renal antigens and be sequestered in the kidney by virtue of this cross-reactivity. If this is true, antigenic competition for pathogenic antibodies might prevent their deposition in kidneys and the ensuing tissue damage. To generate surrogate antigens that could be used for this purpose, we have used peptide display phage libraries to identify peptides that react with R4A, a pathogenic mouse monoclonal anti-DNA antibody that deposits in glomeruli. We have demonstrated that the peptides bind in or near the double-stranded DNA binding site. Furthermore, the peptides are bound preferentially by the R4A antibody as compared with two closely related antibodies derived from it, one of which deposits in renal tubules and one of which displays no renal pathogenicity. Administration of one of these peptides in a soluble form protects mice from renal deposition of the R4A anti-DNA antibody in vivo. This represents a new therapeutic approach in systemic lupus erythematosus that focuses on protecting target organs from antibody mediated injury.

Topological analysis of the functional mimicry between a peptide and a carbohydrate moiety

The shared surface topology of two chemically dissimilar but functionally equivalent molecular structures has been analyzed. A carbohydrate moiety (alpha-D-mannopyranoside) and a peptide molecule (DVFYPYPYASGS) bind to concanavalin A at a common binding site. The cross-reactivity of the polyclonal antibodies (pAbs) was used for understanding the topological relationship between these two independent ligands. The anti-alpha-D-mannopyranoside pAbs recognized various peptide ligands of concanavalin A, and the anti-DVFYPYPYASGS pAbs recognized the carbohydrate ligands, providing direct evidence of molecular mimicry. On the basis of differential binding of various rationally designed peptide analogs to the anti-alpha-D-mannopyranoside pAbs, it was possible to identify different peptide residues critical for the mimicry. The comparison of circular dichroism profiles of the designed analogs suggests that the carbohydrate mimicking conformation of the peptide ligand incorporates a polyproline type II structural fold. The concanavalin A binding activity of these analogs was found to have a direct correlation with the topological relationship between peptide and carbohydrate ligands.

Animal lectins

Protein and lipid glycosylation is no longer considered as a topic whose appeal is restricted to a limited number of analytical experts perseveringly pursuing the comprehensive cataloguing of structural variants. It is in fact arousing curiosity in various areas of basic and applied bioscience. Well founded by the conspicuous coding potential of the sugar part of cellular glycoconjugates which surpasses the storage capacity of oligonucleotide- or oligopeptide-based code systems, recognition of distinct oligosaccharide ligands by endogenous receptors, i.e. lectins and sugar-binding enzymes or antibodies, is increasingly being discovered to play salient roles in animal physiology. Having inevitably started with a descriptive stage, research on animal lectins has now undubitably reached maturity. Besides listing the current categories for lectin classification and providing presentations of the individual families and their presently delineated physiological significance, this review places special emphasis on tracing common structural and functional themes which appear to reverberate in nominally separated lectin and animal categories as well as lines of research which may come to fruition for medical sciences.

Peptide mimicry of adenocarcinoma-associated carbohydrate antigens

Carbohydrate antigens have been identified as significant antigens in many human tumors either by analyzing antibodies in patients' sera or by using monoclonal antibodies of either mouse or human origin. Three carbohydrate epitopes present on cancer-associated mucins [sialyl-Lewis A (SLA), sialyl-Lewis X (SLX), and sialyl-Tn (STn)] may have functional significance in metastasis. Subsequently, these antigens are considered as targets for active specific immunotherapy. Carbohydrates, as T-cell-independent antigens, often elicit diminished immune responses. To overcome this drawback, carbohydrates are typically coupled to protein carriers to elicit immunoglobulin G (IgG) responses as opposed to low-affinity IgM responses, which often times accompanies carbohydrate-based immunizations. In addition, some complex carbohydrates are difficult to synthesize. This latter aspect is further magnified if one considers that clustering of epitopes on neoglycoproteins must be emulated in the synthesis process, leading to multiple presentation or tandem repeats of the synthetic carbohydrate immunogen. Here, we examine the hypothesis that peptides that mimic carbohydrates might be developed to induce immune responses that target and mediate the killing of tumor cells, particularly breast cancer cells in an adjuvant-type setting. We have found that carbohydrate-mimicking peptides retain carbohydrate-like conformations, inducing anti-carbohydrate immune responses against breast tumor cells and mediating their killing by a complement-dependent mechanism.

Bacteriophage display and discovery of peptide leads for drug development

Phage display makes large-peptide diversity libraries readily attainable for identifying novel peptide ligands for receptors and other protein or non-protein targets. This technology kindles enthusiasm for the idea that large and protein-protein interaction surfaces (epitopes) can be distilled down to small pharmacophores. These may be accessible to organic scaffolding, yielding new orally active drugs that might otherwise have taken greater time and effort to be discovered through chemical-library screening. This review, though not comprehensive with respect to the explosive volume of phage display work over the last few years, focuses on recent developments in phage-displayed peptide technology.

Natural human anti-Gal alpha(1,3)Gal antibodies react with human mucin peptides

We have recently demonstrated that both antibodies to Gal alpha(1,3)Gal, and the Gal alpha(1,3)Gal binding lectin (IB4), bind a synthetic peptide (DAHWESWL), there being a similar recognition of carbohydrate and peptide structures. We now report that the anti-Gal alpha(1,3)Gal antibodies and IB4 lectin also react with peptides encoded by mucin genes (MUC 1, 3, 4)-sequences known to be rich in serine, threonine and proline. This activity was demonstrated (1) by the ability of mucin derived peptides to block the reaction of anti-Gal alpha(1,3)Gal antibodies and IB4 lectin with a Gal alpha(1,3)Gal+ pig endothelial cell line; the reactions were specific and did not occur with a random peptide containing the same sequences or with other mucin peptides; (2) by the fact that anti-mucin1 antibodies could react with the Gal alpha(1,3)Gal expressed after transfection of COS cells (Gal alpha(1,3)Gal-,Muc1-) with cDNA encoding the pig alpha, 3galactosyltransferase; and (3) that the IB4 lectin and anti-Gal alpha(1,3)Gal antibodies could react with mucin 1 found on the surface of human breast cancer cells. Thus natural occurring anti-Gal alpha(1,3)Gal antibodies found in all human serum can react with self (Muc1) peptides expressed in large amounts on the surface of tumour cells but not on normal cells. The findings are of interest and serve to explain the previously reported findings that human cells can, at times, express Gal alpha(1,3)Gal; such expression is an artefact, the reaction is due to the phenomenon described herein, i.e. that anti-Gal alpha(1,3)Gal antibodies react with mucin peptides.

Evidence for glycosylation of the juvenile-hormone-binding protein from Galleria mellonella hemolymph

The juvenile-hormone-binding protein (JHBP) from Galleria mellonella hemolymph, which is a member of the high-affinity/low-molecular-mass group of JHBP proteins, was found to be glycosylated. Glycosylation was confirmed by the following evidence. Carbohydrate gas-liquid chromatography analysis of the purified JHBP preparations showed the presence of a low amount of sugars (Man and GlcNAc were the major components). The JHBP electrophoretic band blotted onto nitrocellulose was stained with GlycoTrack (a reagent kit used for the detection of protein glycosylation) and showed strong binding of concanavalin A (ConA). JHBP was fractionated on a ConA-Sepharose 4B column into ConA-bound (strongly stained with ConA) and ConA-unbound (hardly stained with ConA) portions. Both fractions showed juvenile-hormone-binding activity and were glycosylated, as revealed by staining both of them with GlycoTrack. Electrospray-ionization mass spectrometry of JHBP suggested the presence of a small amount of presumably nonglycosylated protein (24988 Da) and five glycoforms, two of which (containing Man2GlcNAc, or Man2Fuc1GlcNAc2 chain) were not bound or were weakly bound to ConA, and three (with Man3GlcNAc2, Man5Fuc1GlcNAc2, or Man5GlcNAc2, chain) were present in the fraction strongly bound to ConA. In conclusion, the monosugar composition, GlycoTrack staining, ConA-binding properties and molecular mass analyses of JHBP supplied convincing evidence for its glycosylation and some information on the character of the oligosaccharide chains.

Random phage mimotopes recognized by monoclonal antibodies against the pyruvate dehydrogenase complex-E2 (PDC-E2)

Dihydrolipoamide acetyltransferase, the E2 component of the pyruvate dehydrogenase complex (PDC-E2), is the autoantigen most commonly recognized by autoantibodies in primary biliary cirrhosis (PBC). We identified a peptide mimotope(s) of PDC-E2 by screening a phage-epitope library expressing random dodecapeptides in the pIII coat protein of fd phage using C355.1, a murine monoclonal antibody (mAb) that recognizes a conformation-dependent epitope in the inner lipoyl domain of PDC-E2 and uniquely stains the apical region of bile duct epithelium (BDE) only in patients with PBC. Eight different sequences were identified in 36 phage clones. WMSYPDRTLRTS was present in 29 clones; WESYPFRVGTSL, APKTYVSVSGMV, LTYVSLQGRQGH, LDYVPLKHRHRH, AALWGVKVRHVS, KVLNRIMAGVRH and GNVALVSSRVNA were singly represented. Three common amino acid motifs (W-SYP, TYVS, and VRH) were shared among all peptide sequences. Competitive inhibition of the immunohistochemical staining of PBC BDE was performed by incubating the peptides WMSYPDRTLRTS, WESYPDRTLRTS, APKTYVSVSGMV, and AALWGVKVRHVS with either C355.1 or a second PDC-E2-specific mAb, C150.1. Both mAbs were originally generated to PDC-E2 but map to distinct regions of PDC-E2. Two of the peptides, although selected by reaction with C355.1, strongly inhibited the staining of BDE by C150.1, whereas the peptide APKTYVSVSGMV consistently inhibited the staining of C355.1 on biliary duct epithelium more strongly than the typical mitochondrial staining of hepatocytes. Rabbit sera raised against the peptide WMSYPDRTLRTS stained BDE of livers and isolated bile duct epithelial cells of PBC patients more intensively than controls. The rabbit sera stained all size ducts in normals, but only small/medium-sized ductules in PBC livers. These studies provide evidence that the antigen present in BDE is a molecular mimic of PDC-E2, and not PDC-E2 itself.

Analysis of novel streptavidin-binding peptides, identified using a phage display library, shows that amino acids external to a perfectly conserved consensus sequence and to the presented peptides contribute to binding

Streptavidin-binding peptides containing the consensus amino acid sequence motif EPDW were identified using a phage display library. Phage presenting peptides containing these sequences bound streptavidin in a biotin-sensitive fashion and could be eluted with biotin. The previously identified 'streptag' peptide sequence (AWRHPQGG) competed with phage presenting the EPDW consensus sequence for streptavidin binding. Furthermore, the EPDW sequence has two amino acids in common with yet another previously identified streptavidin-binding sequence, GDWVFI, which has similar biochemical properties. Binding inhibition studies revealed that residues flanking EPDW, as well as residues of the modified phage pIII product to which displayed peptides are fused, contributed to streptavidin binding. The derivation of small molecules based on the structure of peptides selected using display methods is a potentially important application of phage display technology. The relevance of the observations made here for that application are discussed. Finally, a group of 'nuisance' peptides of the consensus sequence WHWWXW, whose binding specificity has not been fully elucidated, but which have been isolated in a number of biopanning experiments, including those that do not utilize streptavidin, are also described.

Genetic selection of peptide aptamers that recognize and inhibit cyclin-dependent kinase 2

A network of interacting proteins controls the activity of cyclin-dependent kinase 2 (Cdk2) (refs 1,2) and governs the entry of higher eukaryotic cells into S phase. Analysis of this and other genetic regulatory networks would be facilitated by intracellular reagents that recognize specific targets and inhibit specific network connections. We report here the expression of a combinatorial library of constrained 20-residue peptides displayed by the active-site loop of Escherichia coli thioredoxin, and the use of a two-hybrid system to select those that bind human Cdk2. These peptide aptamers were designed to mimic the recognition function of the complementarity-determining regions of immunoglobulins. The aptamers recognized different epitopes on the Cdk2 surface with equilibrium dissociation constant in the nanomolar range; those tested inhibited Cdk2 activity. Our results show that peptide aptamers bear some analogies with monoclonal antibodies, with the advantages that they are isolated together with their coding genes, that their small size should allow their structures to be solved, and that they are designated to function inside cells.

Identification of D-peptide ligands through mirror-image phage display

Genetically encoded libraries of peptides and oligonucleotides are well suited for the identification of ligands for many macromolecules. A major drawback of these techniques is that the resultant ligands are subject to degradation by naturally occurring enzymes. Here, a method is described that uses a biologically encoded library for the identification of D-peptide ligands, which should be resistant to proteolytic degradation. In this approach, a protein is synthesized in the D-amino acid configuration and used to select peptides from a phage display library expressing random L-amino acid peptides. For reasons of symmetry, the mirror images of these phage-displayed peptides interact with the target protein of the natural handedness. The value of this approach was demonstrated by the identification of a cyclic D-peptide that interacts with the Src homology 3 domain of c- SRC. Nuclear magnetic resonance studies indicate that the binding site for this D-peptide partially overlaps the site for the physiological ligands of this domain.

Identification and characterization of a galactosyl peptide mimetic. Implications for use in removing xenoreactive anti-A Gal antibodies

Gal alpha 1,3 Gal is thought to be the major antigenic epitope present on pig tissues to which XNAs bind. Removal of antibodies directed against that structure may be critical to the success of pig to human xeno-transplantation. As a first step toward the development of ligands capable of removing XNAs, we have used a phage-displayed peptide library to identify a six-amino-acid peptide that binds to the lectin GS-1-B4 (which binds the carbohydrate Gal alpha 1,3 Gal). This peptide blocks the binding of GS-1-B4 to pig aortic endothelial cells. The carbohydrate Gal alpha 1,3 Gal competes with the binding of GS-1-B4 to the peptide, suggesting that they may bind the same site. Using a RBC agglutination assay, we show that this peptide inhibits the agglutination of pig RBCs by heat-inactivated human serum at concentrations similar to that of Gal alpha 1,3 Gal.

Solid phase organic synthesis (SPOS): a novel route to diketopiperazines and diketomorpholines

The solid phase synthesis of libraries containing a 1,3,4,6-tetrasubstituted-2,5-diketo-1,4-piperazine scaffold (DKP) or a 3,4,6-trisubstituted-2,5-diketo-1,4-morpholine scaffold (DKM) from alpha-bromocarboxylic acids and amines is described. Using a design strategy which we refer to as divergent library design, both templates were prepared from a common intermediate. The general utility of this synthetic route in creating novel, non-peptidyl chemical libraries is discussed.

Peptides which bind to E-selectin and block neutrophil adhesion

E-selectin is an inducible cell adhesion molecule which mediates rolling of neutrophils on the endothelium, an early event in the development of an inflammatory response. Inhibition of selectin-mediated rolling is a possible means for controlling inflammation-induced diseases, and several classes of compounds have been tested for this use. We describe here the use of recombinant peptide library screening for identification and optimization of novel ligands which bind to E-selectin. Several of these peptides bind with Kd values in the low nanomolar range and block E-selectin-mediated adhesion of neutrophils in static and flow-cell assays. Administration of the peptide to mice undergoing an acute inflammatory response reduced the extent of neutrophil transmigration to the site of inflammation, demonstrating the utility of this compound as a potential therapeutic. The identification of a peptide ligand for E-selectin suggests that the complete natural ligand for this adhesion molecule may include protein as well as carbohydrate moieties.

Peptide mimicry of the meningococcal group C capsular polysaccharide

Sequence analysis of the variable regions of the heavy and light chains of the anti-idiotypic antibody 6F9, which mimics the meningococcal group C capsular polysaccharide (MCP), was performed. The immunogenic site on 6F9 responsible for inducing an anti-MCP antibody response was determined by means of sequence and computer model analysis of these data. Complementarity-determining region 3 (CDR3) was found to be unique in that the sequence tract YRY was exposed on the surface. A synthetic peptide spanning the CDR3 domain was synthesized and complexed to proteosomes (meningococcal group B outer membrane protein). Immunizations of BALB/c mice with the peptide-proteosome complex resulted in a significant anti-MCP antibody response. Immunized mice were protected against infection with a lethal dose of Neisseria meningitidis serogroup C.

Expression of thioredoxin random peptide libraries on the Escherichia coli cell surface as functional fusions to flagellin: a system designed for exploring protein-protein interactions

We have developed a system for probing protein/protein interactions which makes use of the bacterial flagellum to display random peptide libraries on the surface of E. coli. In developing the system the entire coding sequence of E. coli thioredoxin (trxA) was inserted into a dispensable region of the gene for flagellin (fliC), the major structural component of the E. coli flagellum. The resulting fusion protein (FLITRX) was efficiently exported and assembled into partially functional flagella on the bacterial cell surface. A diverse library of random dodecapeptides were displayed in FLITRX on the exterior of E. coli as conformationally constrained insertions into the thioredoxin active-site loop, a location known to be a highly permissive site for the insertion of exogenous peptide sequences into native thioredoxin. To demonstrate that members of this library could be bound and selected via specific protein/protein interactions to a target protein, a method was devised to enable efficient isolation of those bacteria displaying peptides with affinity to immobilized antibodies. We have unambiguously mapped three different antibody epitopes using this method. Peptides selected as FLITRX active-site fusions retain their binding specificity when made as native thioredoxin active-site loop fusions. This will facilitate future structural characterizations and broaden the general utility of the system for exploring other classes of protein-protein interactions.

Identification of galectin-3 as a factor in pre-mRNA splicing

Galectin-3 (M(r) approximately 35,000) is a galactose/lactose-specific lectin found in association with ribonucleoprotein complexes in many animal cells. Cell-free-splicing assays have been carried out to study the requirement for galectin-3 in RNA processing by HeLa cell nuclear extracts by using 32P-labeled MINX as the pre-mRNA substrate. Addition of saccharides that bind galectin-3 with high affinity inhibited product formation in the splicing assay, while addition of carbohydrates that do not bind to the lectin did not inhibit product formation. Nuclear extracts depleted of galectin-3 by affinity adsorption on a lactose-agarose column were deficient in splicing activity. Extracts subjected to parallel adsorption on control cellobiose-agarose retained splicing activity. The activity of the galectin-3-depleted extract could be reconstituted by the addition of purified recombinant galectin-3, whereas the addition of other lectins, either with a similar saccharide binding specificity (soybean agglutinin) or with a different specificity (wheat germ agglutinin), did not restore splicing activity. The formation of splicing complexes was also sensitive to galectin-3 depletion and reconstitution. Together, these results define a requirement for galectin-3 in pre-mRNA splicing and identify it as a splicing factor.

Binding epitope of somatostatin defined by phage-displayed peptide libraries

We have developed a versatile phagemid system to display peptides on the surface of M13 bacteriophage at a copy number which approaches monovalency. In this system, a phagemid encodes a peptide fused to the amino-terminus of the second domain (dII) of the minor coat protein pIII under control of the inducible lac promoter. The fusion protein is displayed in combination with several copies of wild-type pIII on the surface of phage. Two diverse random octapeptide libraries, one linear and one which contained flanking cysteines capable of forming disulfide bridges, were were generated using an in vitro mutagenesis approach and affinity selected on an anti-somatostatin mAb. Peptides with high affinity for the mAb were enriched only from the cyclic library and the tetrapeptide, FWKT, was identified by consensus as the binding epitope. The selected peptides exhibited not only the primary amino acid sequence but also shared structural features with somatostatin. One peptide, CRFWKTWC, also exhibited nanomolar affinities for the five known somatostatin receptor subtypes. This system can easily be adapted to display individual peptides or a wide range of custom peptide libraries.

Chemical shift differences between free and Fab-bound peptide correlate with a two-stage selection of peptide sequences from a random phage display library to delineate critical and non-critical residues for antibody recognition

Epitope libraries provide a method to identify peptide ligands for antibodies, receptors or other binding proteins. As such, they provide a powerful tool to rapidly identify lead ligands in the drug discovery process. In an attempt to correlate structural information with the results from peptide screening, we have used NMR spectroscopy of peptide/antibody complexes to demonstrate that core residues identified through a two-stage selection process undergo a larger structural change upon binding antibody than do positions in the peptide amenable to a variety of side chains. The model system used was the M2 monoclonal antibody/Flag octapeptide epitope system. We have analyzed two peptides: Ac-Asp-Tyr-Lys-Leu-Gly-Asp-Asp-Leu-NH2 (peptide 1), which contains several non-core positions randomized, and Ac-Asp-Tyr-Lys-Asp-Asp-Asp-Asp-Leu-NH2 (peptide 2), which closely corresponds to the original Flag sequence. Enrichment of the peptides with 15N facilitated the investigation by permitting spectral editing of the peptide resonances in the presence of antibody. For peptide 1 the absolute shifts for the free vs. Fab-bound peptide were found to be largest for the amide groups of Asp-1 and Asp-6, in agreement with classification of these residues as critical by the phage display library selection process. For peptide 2 the largest absolute shifts were observed for Asp-1 and Asp-4, with the other aspartic acid residues also showing significant but smaller changes.

Structural analysis of two crystal forms of lentil lectin at 1.8 A resolution

The structures of two crystal forms of lentil lectin are determined and refined at high resolution. Orthorhombic lentil lectin is refined at 1.80 A resolution to an R-factor of 0.184 and monoclinic lentil lectin at 1.75 A resolution to an R-factor of 0.175. These two structures are compared to each other and to the other available legume lectin structures. The monosaccharide binding pocket of each lectin monomer contains a tightly bound phosphate ion. This phosphate makes hydrogen bonding contacts with Asp-81 beta, Gly-99 beta, and Asn-125 beta, three residues that are highly conserved in most of the known legume lectin sequences and essential for monosaccharide recognition in all legume lectin crystal structures described thus far. A detailed analysis of the composition and properties of the hydrophobic contact network and hydrophobic nuclei in lentil lectin is presented. Contact map calculations reveal that dense clusters of nonpolar as well as polar side chains play a major role in secondary structure packing. This is illustrated by a large cluster of 24 mainly hydrophobic amino acids that is responsible for the majority of packing interactions between the two beta-sheets. Another series of four smaller and less hydrophobic clusters is found to mediate the packing of a number of loop structures upon the front sheet. A very dense, but not very conserved cluster is found to stabilize the transition metal binding site. The highly conserved and invariant nonpolar residues are distributed asymmetrically over the protein.

Antigenic and immunogenic mimicry of the HER2/neu oncoprotein by phage-displayed peptides

To recover peptides that antigenically and immunogenically mimic the p185HER2 oncoprotein, we selected the phage-peptide libraries pVIII-9aa and pVIII-9aa. Cys using murine monoclonal antibodies (mAb) MGr2 and MGr6, directed against two distinct epitopes of the p185HER2 extracellular domain. Phage-displayed peptides containing consensus amino acid motifs were recovered and shown to compete specifically for mAb binding on tumor cells that overexpress p185HER2. The deduced amino acid sequence of the peptides suggests that both epitopes defined by the mAb on p185HER2 are discontinuous and that hydrophobic interactions are involved in binding with the mAb. A phage clone displaying the GPLDSLFAQ peptide elicited a specific immune response against the p185HER2 in BALB/c mice, demonstrating that this phage-displayed peptide represents an immunological equivalent of the MGr2 epitope on p185HER2 and might be used as a substitute for this oncoprotein in in vitro and in vivo immunological studies.

The mixotope: a combinatorial peptide library as a T cell and B cell immunogen

We report a new approach in peptide vaccine strategy based on combinatorial synthesis. A library of 7.5 x 10(5) related peptides, termed mixotope, was derived from the sequence of the third hypervariable domain (V3 loop) of the human immunodeficiency virus (HIV) envelope protein. This preparation induced a strong immune response in all syngeneic and outbred rodents tested. The response directed against the mixotope included antibodies, CD4+ T helper cells (TH1 and TH2) and CD8+ T cells. In rodents immunized with the mixotope, the T cell response directed against individual V3 peptide sequences (BRU, MN, RF, SF2, and ELI) as measured by T cell proliferation and interleukin (IL)-2 production, was found to be major histocompatibility complex haplotype-dependent. However, additional experiments performed in mice indicated that selectivity was less restrictive when using IL-3 secretion to explore T cell activation. This combinatorial antigen could be considered as a series of agretopic motifs framing a multiplicity of closely related epitopes for T cell recognition and able to elicit a T cell and B cell repertoire. This new construct may therefore provide a basis for the design of future vaccine strategies.

Recent advances in drug design methods: where will they lead?

Drug design methods have made significant new advances over the last ten years, mainly in the areas of molecular modelling. In more recent times important developments in theory have led to a different type of modelling becoming possible, the so-called de novo or automated design algorithms. In this new method the programs perform much of the chemist's thinking, in finding appropriately sized chemical groups to fit into a target site. However this is a combinatoric problem which has no general analytical solution; it is ripe for optimization. Other advances, such as combinatorial chemical synthesis and screening, will dramatically influence the search for new lead structures for target sites, which at present are poorly understood. Already these methods are being applied to peptide libraries. Peptides do not make good drug compounds because of their poor bioavailability; further, their flexibility reduces their affinity. In some cases peptide backbones can be removed and replaced with rigid non-peptide scaffolds.

A general strategy to identify mimotopes of pathological antigens using only random peptide libraries and human sera

A strategy to identify disease-specific epitopes from phage-displayed random peptide libraries using human sera is described. Peptides on phage (phagotopes) that react with antibodies present in patient sera are purified from > 10(7) different sequences by affinity selection and immunological screening of plaques. Disease-specific phagotopes can be identified out of this pool through an 'antigen independent' procedure which avails itself only of patient and normal human sera. Using this strategy, we have selected antigenic mimics (mimotopes) of two different epitopes from the human hepatitis B virus envelope protein (HBsAg). We could show that a humoral response to these mimotopes is widespread in the immunized population, suggesting that the strategy identifies phagotopes that have a potential role as diagnostic reagents. Immunization of mice with the selected phagotopes elicited a strong specific response against the HBsAg. These results open new inroads into disease-related epitope discovery and provide the potential for vaccine development without a requirement for the use of, or even information about, the aetiological agent or its antigens.