ARTIFICIAL LECTIN MIMICS AND USES THEREOF

BACKGROUND OF THE INVENTION

This application claims the priority of U.S. Provisional Patent Application Serial No. 61/116,548, filed November 20, 2008, the entire disclosure of which is specifically incorporated herein by reference.

1. Field of the Invention

The present invention relates generally to the field of biology. More particularly, it relates to a high-throughput screening platform for identification of carbohydrate binding peptide sequences.

2. Description of the Related Art

Complex carbohydrates are among the most important biomolecules; however, the studies of their molecular interactions have lagged behind studies of other important macromolecules, such as nucleic acids and proteins (Turnbull and Field, 2007). Glycans, a form of complex carbohydrate, form complex and often branched structures in biological systems. Historically they have been notoriously difficult to synthesize and analyze (Murrell et ah, 2004). An outgrowth of interest in glycan function has been termed functional glycomics, which focuses on the nature of glycan interaction with glycan binding proteins, such as lectins and anti-carbohydrate antibodies (Paulson et ah, 2006). Such glycan-protein interactions are important in protein functionality, pathogenic infection, cell adhesion, cell signaling, and a multitude of other biological pathways. Therefore, these are considered important targets of biomedical research. Increasing awareness of the importance of glycosylation to biological systems has led to recognition of the need to develop better tools for the analysis of protein-carbohydrate interactions (Campbell and Yarema, 2005).

In contrast to template-driven nucleic acid and protein sequences which aid function assignments, the need for a more empirical, high throughput analysis of potential carbohydrate patterns has resulted in a variety of approaches (Gemeiner et ah, 2008). Two key newcomers in the area of functional glycomics are glycan (Mellet and Fernandez, 2002) and lectin microarrays (Pilobello and Mahal, 2007), in which glycans or lectins are immobilized on glass slides for investigating the specificity of glycan-binding proteins or glycoconjugates, respectively.

In lectin microarrays, carbohydrate-binding proteins, such as lectins and anti- carbohydrate antibodies, are immobilized on a solid support in a high spatial density. Interrogation of these arrays with fluorescently-labeled samples creates a pattern of binding (glycoprofile) that depends on the carbohydrate structures, providing a method for the rapid characterization of glycans on glycoproteins (Pilobello et al., 2005), bacteria (Hsu and Mahal, 2006), or mammalian cells (Pilobello et al., 2007; Zheng et al., 2005).

Recently, a lectin-based method for rapid analysis of glycosylation patterns of glycoproteins was developed. The method consists of a chip having a set of 20-30 lectins arrayed on a nitrocellulose membrane-coated glass slide. Each lectin is arrayed at several concentrations and in replicates on each slide. The concentration ranges are tailored for each of the lectins and calibrated to provide a linear response within the same range, regardless of the affinity of the lectin. A sample of intact glycoprotein or other glycoconjugate is applied to the array, and its binding pattern is detected by fluorescent label, which is placed either on the glycoprotein itself or on a probe directed toward either the protein or carbohydrate moieties of the glycoprotein. Common labeled probes include polyclonal antibodies, lectins or streptavidin for biotinylated samples. This technology provides two levels of data: a fingerprint and an interpretation table derived from deconvolution of the fingerprint using proprietary software based on a database of lectin specificities.

The microarray format allows rapid parallel analysis of multiple carbohydrate- protein interactions with a minimal amount of sample. However, lectin microarrays are intrinsically handicapped by limited availability (Pilobello and Mahal, 2007), and the limited and often unexpected specificities of natural lectins (Manimala et al., 2005 Alvarez et al., 2005). Only about 60 lectins are commercially available with the ability to recognize only a fraction of glycans present on either mammalian or especially on microbial cells (Pilobello and Mahal, 2007). The common problems inherent to protein arrays, such as linking chemistry, orientation-dependent binding activity, and storability, are also important factors arguing in favor of alternative approaches. Proteins are not the only molecules that can bind biomolecules. Aptamers, short nucleotides that bind specific proteins, have also been explored (Svarovsky and Joshi, 2008). Use of well-defined binding affinity agents facilitates the uniform and reproducible arraying chemistries and allows higher flexibility in screening the arrays. Synthetic peptides are known as highly versatile molecules for a variety of biological applications. Unlike proteins, which unfold readily and subsequently lose their biological activities, peptides are functionally stable and capable of retaining their activities under most reaction conditions, making them the preferred molecules for facile and robust screening assays, especially in microarray-based formats.

Lipopolysaccharides (LPS) have been implicated in the systemic inflammatory response and septic shock claiming more than 200,000 lives each year in the US alone (David, 2001). Hence, there is a great deal of interest in developing therapeutic agents that can efficiently bind LPS (Wood et ah, 2004). Finally, although the therapeutic strategy directed against viral glycans was a success, a similar approach to antibacterial therapies has not been systematically explored mainly due to difficulties finding molecules that can selectively bind the bacterial glycans (Balzarini, 2007).

SUMMARY OF THE INVENTION

In some aspects, the present invention concerns methods for identifying carbohydrate-binding peptides and kits containing the identified peptides. In other aspects, the invention provides methods for a template-driven assembly of carbohydrate binding peptides into small fully synthetic lectin mimics.

In one aspect, the invention provides a method for identifying a glycan- binding peptide comprising providing a microarray having a plurality of peptides and probing the microarray with a lipoglycan fluorescently labeled with a quantum dot to detect glycan-binding peptides or peptide-like molecules. In another aspect, the invention provides a probe useful to identify glycan-binding peptides comprising a plurality of lipoglycans bound to a quantum dot. In some aspects, the lipoglycan is directly fluorescently labeled with the quantum dot, wherein the lipoglycan is attached to the surface of the quantum dots. In other aspects, the lipoglycan is indirectly labeled.

The glycan-binding peptide may be any peptide or peptide-like molecule. For convenience, the term "peptide" is used broadly herein to mean peptides, polypeptides, proteins and fragments of proteins, peptoids, peptidomimetics, and peptide-like molecules that contain non-naturally occurring amino acids, peptoids and the like.

The lipoglycan may be any lipid-glycan glycoconjugate. In some embodiments, the lipoglycan is a lipopolysaccharide (LPS). In certain embodiments, the LPS is smooth type. In some embodiments, the LPS may be a bacterial LPS. The bacterial LPS may be derived from any appropriate bacteria known to those having skill in the art. Bacteria that can be used as an LPS source include, but is not limited to Salmonella typhi, S. enteritidis, S. typhimirium, S. newport; Shigella dysenteriae, S. flexneri, S. boydii, S. sonnei; Neisseria gonorrhoeae, N. meningitides; Klebsiella pneumonia; and/or Legionella philia. In some embodiments, the bacteria is E. coli or P. aeruginosa. In particular embodiments, the E. coli is ECoiii or EC ₀ 55. In other embodiments, the P. aeruginosa bacteria is of serotype 10 (PAlO).

A quantum dot (QDot) is a semiconductor whose conducting characteristics are closely related to the size and shape of the individual crystal and is useful as a luminescent labeling agent. In some embodiments, the quantum dot is hydrophobic. In particular embodiments, the hydrophobic quantum dot is directly conjugated to a lipoglycan. The conjugated Qdot and lipoglycan may also be known as a glycoprobe. The glycoprobe may be any glycoprobe known to those of skill in the art. In some embodiments, the glycoprobe is Tn antigen, lactose, TF antigen, 3'-sialyllactose, or 6'-sialyllactose. In some embodiments, the control is a QDot with no glycan.

The microarray may contain a plurality of peptides immobilized on a solid support. In some embodiments, the microarray comprises 1,000 to 20,000 peptides. In some embodiments, the microarray comprises 5,000 to 15,000 peptides. In particular embodiments, the microarray comprises 10,000 peptides to 100,000 peptides. The peptides may be known or unknown and may be any appropriate length as known to those having skill in the art. In some embodiments, the plurality of peptides are 20mers. In particular embodiments, the peptides comprise a C-terminal sequence of -GSC.

The probing and detecting may be performed by any method known to those of skill in the art. In particular embodiments, the probing comprises contacting the microarray with the labeled lipoglycan and scanning the microarray to detect fluorescence. In some embodiments, the detection of carbohydrate -peptide interactions can be direct, indirect, optical, electrical, magnetic, chemical, enzymatic, or by any other means available in the art. In particular embodiments, the detecting is optical. In some embodiments, the method may further comprise additional probing. The additional probing may be performed by any appropriate means known to a person having skill in the art. In some embodiments, the method further comprises probing the microarray with a lipoglycan fluorescently labeled with fluorescein isothiocyanate (FITC) to detect glycan-binding peptides.

In some embodiments, the method further comprises validating the peptides. In some embodiments, the peptides may be validated by hemagglutination, flow cytometry, surface plasma resonance, haptenic inhibition, or fluorescence assisted cell sorting.

In yet another aspect, the invention provides a kit containing a plurality of antagonists of a bacteria identified by the methods of the invention. The bacteria may be any appropriate bacteria known to those having skill in the art. Bacteria that can be targeted include, but is not limited to Salmonella typhi, S. enteritidis, S. typhimirium, S. newport; Shigella dysenteriae, S. flexneri, S. boydii, S. sonnei; Neisseria gonorrhoeae, N. meningitides; Klebsiella pneumonia; and/or Legionella philia. In some embodiments, the bacteria is E. coli. In particular embodiments, the E. coli is ECo ₁₁₁ . In other embodiments, the E. coli is EC ₀ 55. In some embodiments, the bacteria is P. aeruginosa. The antagonist of the bacteria may be any type antagonist known to those having skill in the art. In certain embodiments, the antagonists are antagonists of E. coli and comprise a peptide having SEQ ID NOs: 1- 24 (QF1-QF16 from Table 2 and ECl-8 from Table 3). In other embodiments, the plurality of antagonists are antagonists against P. aeruginosa and comprise a peptide having SEQ ID NOs: 25-33 (PAl -9 from Table 3).

In still another aspect, the invention provides a method of differentiating between a first and a second bacterial serotypes comprising providing a first microarray having a plurality of peptides, probing the microarray with a first LPS derived from a first bacterial serotype directly fluorescently labeled with a hydrophobic quantum dot to detect first bacterial serotype-binding peptides, providing a second microarray having a plurality of peptides, probing the microarray with a second LPS derived from a second bacterial serotype directly fluorescently labeled with a hydrophobic quantum dot to detect second bacterial serotype-binding peptides, and comparing the binding peptides of the first microarray and the second microarray to differentiate between the first and second bacterial serotypes. The first and second bacterial serotype may be any bacterial serotype, such as E. coli 026:B6 or H157:H7 and the like. In some embodiments, both bacterial serotypes are from E. coli. In particular embodiments, one bacterial serotype is ECoiii and the other bacterial serotype is EC055.

In yet another aspect, the invention provides a method for identifying a carbohydrate-binding peptide comprising providing a microarray having a plurality of peptides, probing the microarray with a control and eliminating peptides that show binding, probing the microarray with a labeled quantum dots glyco-nanoparticle (glyco-QD) to detect binding peptides, blocking the microarray to create a blocked microarray, probing the blocked microarray with a labeled glyco-QDot to detect non- binding peptides, and comparing the identified binding peptides and the identified non-binding peptides to identify a carbohydrate-binding peptide. The microarray may be blocked by any method known to those of skill in the art. In particular embodiments, the microarray is blocked with a non-luminescent gold glyco- nanoparticle or an excess of free glycan.

In some embodiments, the method further comprises validating the peptides. The peptides may be validated by any appropriate method known to those having skill in the art. In some embodiments, the peptides may be validated by hemagglutination, flow cytometry, surface plasmon resonance, haptenic inhibition, or fluorescence assisted cell sorting.

Any embodiment discussed with respect to one aspect of the invention applies to other aspects of the invention as well.

The embodiments in the Example section are understood to be embodiments of the invention that are applicable to all aspects of the invention.

The use of the term "or" in the claims is used to mean "and/or" unless explicitly indicated to refer to alternatives only or the alternatives are mutually exclusive, although the disclosure supports a definition that refers to only alternatives and "and/or."

Throughout this application, the term "about" is used to indicate that a value includes the standard deviation of error for the device or method being employed to determine the value.

Following long-standing patent law, the words "a" and "an," when used in conjunction with the word "comprising" in the claims or specification, denotes one or more, unless specifically noted. Other objects, features and advantages of the present invention will become apparent from the following detailed description. It should be understood, however, that the detailed description and the specific examples, while indicating specific embodiments of the invention, are given by way of illustration only, since various changes and modifications within the spirit and scope of the invention will become apparent to those skilled in the art from this detailed description.

BRIEF DESCRIPTION OF THE FIGURES

The following drawings form part of the present specification and are included to further demonstrate certain aspects of the present invention. The invention may be better understood by reference to one or more of these drawings in combination with the detailed description of specific embodiments presented herein.

FIG. 1 General application of the random sequence peptide microarray to the discovery of glycan-binding peptides (artificial lectins).

FIGS. 2A-B (FIG. 2A) organic dye labeled lipopolysaccharides (LPS); (FIG. 2B) Qdot labeled LPS. The m designates number of O-antigen repeating units, n - number of LPS molecules attached to the Qdot.

FIGS. 3A-D Screening process (FIG. 3A) Screen with glyco-QDots (green) that will bind specifically, non-specifically or not bind at all. (FIG. 3B) Screen with a control QDot without glycan. Only non-specific and some specific peptides are detected. (FIG. 3C) Statistically subtract non-specific binders, leaving mostly specific peptides. (FIG. 3D) Block with non- luminescent glyco-gold nanoparticles (glyco- GNPs) (orange) or with excess of free glycan and then screen with glyco-QDot. Most specific peptides will not light up due to blocking. Statistically superimpose C and D to select positive hits.

FIG. 4 Strategy for detection of cancer based on the well-established evidence that cancer cells display aberrantly glycosylated glycans.

FIG. 5 Application of artificial lectin microarrays in glycopro filing of proteins. Sequential enzymatic removal of terminal glycans from human transferrin will lead to sequential exposure and detection of Sia, Gal, GIcNAc, and Man. The device can distinguish between different glycovariants of a protein.

FIGS. 6 Scaffolds A (FIG. 6A) and Scaffold B (FIG. 6B) selected as templates for the Synlecs described. FIGS. 7A-B (FIG. 7A) Synthesis of Synlecs using scaffold A. (FIG. 7B) Synthesis of Synlecs using scaffold B.

FIG. 8 Structures of glycans conjugated to QDots and gold nanoparticles and screened on CIM10,000 random sequence peptide microarray.

FIG. 9 Structures of lactose-, 3'-sialyllactose-, 6'-sialyllactose.

FIGS. 10A-B Venn diagrams showing selected artificial lectin libraries. (FIG. 10A) No overlapping is observed with structural isomers. (FIG. 10B) Some overlapping is observed with stereoisomers but specific artificial lectin binders to each isomer were selected.

FIGS. 11A-D FIG. HA is an illustration of high resolution differential surface plasmon resonance (HRD-SPR). FIG. HB illustrates the sensing and reference areas of HRD-SPR. FIGS. HC-D depict representative HRD-SPR data comparing peptide 3'SL-Pl with 2mM 3'sialyllactose (3'SL) and 2mM 6'sialyllactose (6'SL) injections with nonbinding peptide as control.

FIGS. 12A-B Flow cytometry analysis of peptides binding to Chinese hamster cells (CHO-Kl cells) known to express 3 'SL and, to a lesser degree, 6 'SL, imaged through biotin-streptavidin labels, representative data. (FIG. 12A) Streptavidin label control (blue), Maackia amurensis (MAA) lectin to 3 'SL (black), 3'SL-Pl (green), and 3'SL-P2 (brown). More peptide than lectin is observed binding with the 3 'SL peptides. (FIG. 12B) Streptavidin label control (blue), Sambucus Nigra (SNA) lectin to 6 'SL (black), 6'SL-Pl (green), and 6'SL-P2 (brown). The 6 'sialic acid lectin (SNA) does not bind these cells, but the peptides have binding to these cells.

FIG. 13 Comparison of raw signal intensity histograms for EC ₀ 55. The first three correspond to experiments with AlexaFluor488 label, and the last two with QDot label: (A) before LPS hybridization, (B) unblocked, and (C) blocked, all three with detection at 488 nm. (D) before LPS hybridization and (E) after hybridization with detection at 605nm. Blocked in experiment C is done with native unlabeled EC ₀₅₅ . X-axis represents number of peptides; Y-axis represents raw fluorescent intensity in logarithmic (log ₂ ) scale. Coloring in all panels is based on the signal intensity of the respective peptides in the center panel.

FIG. 14 FITC labeled ECoiii LPS vs. Qdot-labeled EC _O n 1 LPS correlation (R=O.824). Annotated black dots indicate selected LPS-binding peptides . Both axis show normalized signal in a logarithmic (log ₂ ) scale. Blue lines delimit the 2-fold change.

FIG. 15 Growth inhibition activities of peptides QF1-16 tested against

E. coli DHlOB. The error bars are based on standard deviation of triplicate measurements. Negl is a negative control peptide.

FIG. 16 Flow cytometry of AlexaFluor488-labeled streptavidin, Negl control peptide, QF5, QF8, and QF5 and QF8 after one hour pre -incubation with 100- fold excess of ECoiii. The y-axes show the count of cells, and the x-axes show the AlexaFluor488.

FIG. 17 HRD SPR responses of peptides Negl, QF5, QF8 to ECoiii.

FIG. 18 E. coli O111 :B4 vs. E. coli 055 :B5 correlation for triplicate experiments of each (R=O.907). The black dot corresponds to the ECoiii specific peptide FPKDQW (abbreviated sequence, shown in the insert, with ECo ₁₁₁ on the left and ECo ₅₅ on the right). Both axis show normalized signal in a logarithmic (log ₂ ) scale. Blue lines delimit the 2-fold change. Insert shows close up of the peptide microarray binding patterns of (A) ECoiii and (B) EC ₀ 55. The first 6 (of 20) amino acids are shown.

FIG. 19 Chemical structures of the repeating units of LPS used.

FIG. 20 LPS binding patterns ('glycosignatures') on CIM10,000 of (A)

P. aeruginosa 10 LPS and (B) E. coli 0111 :B4 LPS. Abbreviated sequences in yellow indicate peptides unique to P. aeruginosa; sequences in green are unique to E. coli; sequences in white are common binders. Only first 6 (of 20) amino acids are shown.

FIG. 21 ECoiii vs. PA ₁₀ correlation (R=0.630). Annotated blue dots correspond to peptides in Table 1 ; annotated black dots correspond to the peptides in Table 2. Both axes show normalized fluorescence signal at 605nm on a logarithmic scale. Blue lines delimit the 2-fold change.

FIG. 22 Scheme of the full process of synthesis of synthetic lectin mimics (Synlecs).

DESCRIPTION OF THE ILLUSTRATIVE EMBODIMENTS

In some aspects, the present invention provides a random peptide sequence microarray approach to rapidly and transparently identify carbohydrate-binding peptides or artificial lectin mimics of natural lectins (FIG. 1). The method allows carbohydrate-binding peptides with desired glycan binding properties to be selected from large combinatorial libraries printed on the surface of random sequence peptide microarrays. This method allows selection of binding peptides for known or unknown carbohydrates by the detection of the interaction between potential binding peptides and a glycoprobe to be examined.

This method allows selection of highly specific non-overlapping sets of peptides having micromolar affinity for their corresponding glycotargets. The selected peptides may be validated by any known method, such as hemagglutination assays, haptenic inhibition, surface plasmon resonance (SPR), and fluorescence assisted cell sorting (FACS). The binding peptides identified are highly useful. For example, the peptides can be used in bead format as bioanalytical tools for extraction and purification of glycoconjugates and cell sorting.

One particular embodiment provides for screening of 10,000 random 20-mer peptide sequences printed on a glass slide with the luminescent saccharidic portion of bacterial lipopolysaccharides (LPS) glycoprobes for potential lectinomimetic activity. The diversity of glycan structures unique to bacteria would allow the widest dynamic range of molecules to be tested.

A set of specific lectinomimetic antagonists of the LPS molecules have been discovered that can be used as a new class of diagnostic, anti-endotoxic, and antimicrobial peptide leads in both on- and off-array formats. For example, these binding peptides can be used for (1) rapid glycoprofϊling of proteins, (2) detection of pathogens, (3) diagnosing various disease states, (4) cell sorting, (5) affinity purification and (6) other applications where carbohydrate protein interactions are involved or implied. The methods of use are simple enough to be utilized outside of the analytical laboratory, for example, in a manufacturing, field or point-of-care settings. Kits for carrying out the foregoing applications are also described.

Such peptide sequence microarrays are capable of differentiating between gram-negative bacterial strains based on differences in their LPS structures. Furthermore, the parallel analysis platform that is inherent to the microarray format allows rapid and direct identification of multiple LPS interactions at the same time. In contrast to screening by phage-displayed or other solution-based combinatorial peptide libraries, the microarray format allows systematic analysis and statistical deconvolution of post-selection data. One of the most attractive features of the peptide sequence microarray approach is the ability to select carbohydrate binding agents without prior knowledge of the exact structure of the glycotargets. For example, a relation between virulent and non-virulent serotypes of a genus can be examined without actual knowledge of the changes in the carbohydrate composition of the bacterial cell membranes. Membrane glyco components can be extracted from unknown pathogens and analyzed against known pathogens to compare their "glycosignatures." The transparency of the peptide sequence microarray approach also allows selection and exclusion of peptides active towards membranes associated with beneficial bacteria or human cells. This information can be translated into powerful small peptides and peptide constructs that can detect or interfere with host-bacterial interactions, thus advancing the systematic discovery and validation of new pharmaceutical targets as well as the development of novel diagnostic and affinity purification devices. This platform may enable direct discovery of protease-resistant peptidomimetics, as peptides can be readily synthesized with non-natural functionalities, e.g., D-amino acids, cyclic structures, and unnatural side chains (peptoids) to facilitate the transition of discovered leads into the clinic.

I. Selection of Binding Peptides A. Glycoprobes

The binding peptide selection process involves the use of glycoprobes for microarray interrogation. Since most glycans are small molecules, simple fluorescent labeling is not applicable due to large relative size of fluorescent dyes, which interfere with glycan binding. Low binding affinity of simple glycans and difficulties with purification of labeled from unlabeled glycans are also serious deterrents for direct fluorescent labeling of glycans.

1. Glyco-nanoparticles

Glyco-nanoparticles are useful as non-interfering scaffolds for polyvalent presentation of glycans. While the glyco-nanoparticles have not been used in an array format before, they have shown great promise as biomimetics of natural glycoconjugates in other studies (Svarovsky and Barchi 2007, de Ia Fuente and Penades 2006). Two types of glycoprobes may be used in the binding peptide selection process: (1) non-luminescent gold glyco-nanoparticles (glyco-GNPs); (2) luminescent quantum dots glyco-nanoparticles (glyco-QDots). Glyco-GNPs are composed of gold nano-core with an approximate diameter of 2-5 nanometers to which glycans are attached with a linker using gold-sulfur covalent bond. The gold nanoparticles are optically silent and are used as blocking glycoprobes (see below) but can also be used as detection probes. The glyco-QDots may be composed of luminescent semiconductor nano-core with an approximate diameter of 2-5 nanometers to which glycans are attached with a linker using strong coordination of sulfur on a linker with the surface atoms of the nano-core. The semiconductor nano-core can be composed of various elements including but not limited to CdTe, CdSe, CdS, HgTe, etc. and can be optically active in a variety of wavelengths from near-ultraviolet to optical to far-infrared.

In other embodiments, free sugars, other monovalent and polyvalent carbohydrate conjugates, such as polyacrylamide (PAA)-conjugated glycans, bovine serum albumin (BSA)-conjugated glycans, other polyvalent constructs can be used for interrogation and/or blocking peptide microarrays. The detection of carbohydrate- peptide interactions can be direct, indirect, optical, electrical, magnetic, chemical, enzymatic, or by any other means available in the art.

2. Lipoglycan-Qdot complexes

In some aspects, lipoglycan {i.e., a molecule having at least one lipid group and at least one carbohydrate group) screening technology is also provided. This method allows quick identification of lipopolysaccharide (LPS) binding peptides that are specific to the highly variable saccharidic branch of LPS. LPS is a complex, negatively-charged lipoglycan composed of three distinct regions: (1) a fatty acid region called Lipid A that has very low variability; (2) a conserved glycosidic "core" consisting of approximately 10 monosaccharides; and (3) a highly variable region called O-antigen consisting of repetitive subunits of one to eight monosaccharides repeated up to 100 times (Caroff and Karibian, 2003). The O-antigen region defines the strain, serotype, and the virulence of the bacteria.

Existing LPS labeling strategies rely on chemical modification of LPS molecules with organic dyes (FIG. 2A). This method requires complex manipulation and purification steps, is not site-specific, and depends on availability of reactive groups in LPS molecule (Triantafϊlou et ah, 2000). When such groups are unavailable, an extra functionality is introduced into the saccharidic branch of LPS that may affect its physical properties and biomolecular recognition events thus making it not ideally suitable for the purposes of this study.

In some aspects, the lipoglycan is directly fluorescently labeled, wherein the lipoglycan is attached to the surface of the quantum dots. Directly labeling the LPS for probing the peptide microarray avoids the need for a secondary detection method, thus avoiding the extra steps in screening which complicate the interpretation and de- convolution of the data. This consideration is especially important for the random peptide microarrays because each peptide on the array de facto is not specific and serves as a putative ligand for any target of choice. Not only may the secondary probe bind to the array, but it also may compete with the primary probe. In the case of carbohydrates, whose binding affinities are typically weak, the latter can present a serious problem.

For these reasons, a conjugation protocol for the attachment of LPS to the surface of hydrophobic quantum dots was developed that takes advantage of the amphipathic (micelle-forming) nature of LPS molecule and does not introduce any new chemical modalities into the structure. The protocol is universally applicable to other lipoglycans and yields robust water soluble luminescent constructs where the lipid core is occupied by the luminescent quantum dots and the saccharidic portion is exposed outwards for unobstructed binding. Further, QDot labeled LPS takes advantage of the amphipathic (micelle-forming) nature of LPS molecule and does not introduce any new chemical modalities into the structure.

Nanometer-sized crystals of semiconductors known as quantum dots (Qdots) that have recently emerged as useful luminescent labeling agents may be used (Resch- Genger et al., 2008). Coating of hydrophobic quantum dots with phospholipids (Dubertret et al., 2002) and synthetic amphiphilic polymers has been previously described (Anderson and Chan, 2008). Both methods rely on phase transfer of hydrophobic Qdots from organic solvent to an aqueous solution of an amphiphile. Using a similar approach, smooth type LPS from E. coli and P. aeruginosa were conjugated to hydrophobic Qdots (FIG. 2B) (see Example 2). In this case, Lipid A, which is responsible for the self-aggregation, also confers the ability of LPS to bind to hydrophobic surfaces of the Qdots. Since the lipid functionality is attached directly to the label, LPS-Qdot constructs are especially useful for studying the variable saccharidic moiety of LPS. In other aspects, the lipoglycan is indirectly labeled. B. Microarrays

The peptide microarrays may be constructed by spotting a plurality of random peptide sequences in duplicates on a microscope glass slide. The random peptide library may be produced by conventional solid-phase synthesis (Alta Biosciences, Birmingham, UK) based on computer-generated random sequences of a specified number of amino acids, excluding cysteine, for all but the 3 amino acids found at the C-terminal. A C-terminal -GSC sequence can be incorporated into each peptide to facilitate coupling to the array surface.

C. Microarray Interrogation

1. Interrogation with Glyconanoparticles

As shown in FIG. 3, in the first probing, glyco-QDots bind both specifically and non-specifically to the peptide features present on the array (FIG. 3A). To eliminate non-specific peptide binders, the second probing is conducted with QDots alone (no glycan) (FIG. 3B). In this step, most non-specific and some specific peptide binders are selected and subtracted from the ones selected by the first probing, leaving mostly specific binders (FIG. 3C). In the blocking probing, non-luminescent glyco- GNPs or free glycans are used to block the specific sites, followed by screening of the blocked slide with glyco-QDots as in the first probing. Since the specific sites are already being occupied by non-luminescent glyco-GNPs, the luminescent glyco- QDots are unable to bind to them (FIG. 3D). Only peptides that do not light up in the blocking probing but were selected as positive hits in the first and second probings are selected. The blocking experiment increases the probability that peptides selected in the first two experiments are valid.

2. Interrogation with Lipoglycan-Qdot complexes

Alternatively, the arrays may be probed directly with the quantum dot (QDot) labeled LPS in order to specifically single out carbohydrate interactions. Probing the peptide microarray with these LPS-Qdot constructs allows identification of O-antigen binding peptides. For example, screening of LPS from O111 :B4 and 055 :B5 serotypes of E. coli revealed only minor differences, while LPS from E. coli 0111 :B4 and P. aeruginosa 10 produced unique sets of peptides (see Example 2). Similar to the existing antimicrobial peptides of indolicidin and histatin families, E. coli specific peptides were prominently enriched in aromatic and cationic amino acids. Most of these peptides inhibited growth of E. coli bacteria. In contrast, peptide selections against P. aeruginosa were largely composed of hydrogen-bond forming and aliphatic amino acids in accord with dramatic differences in composition of O-antigenic glycans between E. coli and P. aeruginosa.

II. Comprehensive Glycomic Kits

One embodiment of this invention provides diagnostic kits comprising the binding peptides identified by the screening method. As glycan libraries expand in scope and availability, the screening method described here can be extended to include all available biologically active glycans present in a particular organism for the construction of glycome-wide peptide microarrays. Such microarrays may quickly provide a glimpse of the functional roles of glycosylation at a systemic level in nearly real time.

In some aspects, the present invention provides binding peptides which are useful for recognizing differences in carbohydrates in biological systems. The peptides may be placed in arrayed format on a solid support in a suitable order in predetermined positions. The region on which the binding peptide is immobilized may be in the form of a well or a spot. Individual peptides can be combined on a spot or linked to a spot to give more specificity. The substrate may be composed of any material having any shape, such as chip, plate and bead, insofar as peptides can be immobilized thereon. The substrate may be made of any suitable material and may contain synthetic resin (including plastics), metal (including platinum, silver, copper, gold, silicon etc.), mica or a mixture thereof. The positions where individual binding peptides are to be immobilized on a chip can be suitably changed so as to optimize the diagnosis of objective cells or diseases.

The ability to efficiently profile the variation in glycosylation in complex biological samples would be useful for a variety of purposes, such as characterization of disease-associated alterations, pathogen detection, detection of protein glyco forms, protein glycoprofiling, identification of new diagnostic biomarkers, and studying the factors that regulate glycan structures.

Any of the compositions or peptides described herein may be comprised in a kit. In some embodiments, carbohydrate binding peptides identified by the current method are included in a kit. In certain aspects the kit is portable.

The components of the kits may be packaged either in an aqueous, powdered or lyophilized form. The container means of the kits will generally include at least one vial, test tube, flask, bottle, syringe or other container means, into which a component may be placed, and preferably, suitably aliquoted. Where there is more than one component in the kit, the kit also will generally contain a second, third or other additional container into which the additional components may be separately placed. However, various combinations of components may be comprised in a canister. Such containers may include injection or blow molded plastic containers into which the desired vials are retained.

When the components of the kit are provided in one and/or more liquid solutions, the liquid solution is an aqueous solution, with a sterile aqueous solution being particularly preferred, but not required. However, the components of the kit may be provided as dried powder(s). When reagents and/or components are provided as a dry powder, the powder may be reconstituted by the addition of a suitable solvent or administered in a powdered form. It is envisioned that the solvent may also be provided in another container means.

A kit may also include instructions for employing the kit components as well the use of any other reagent not included in the kit. Instructions may include variations that can be implemented. It is contemplated that such reagents are embodiments of kits of the invention. Such kits, however, are not limited to the particular items identified above. Examples of such kits include, but are not limited to, those disclosed below.

1. Cancer Diagnostic Kits

New approaches for improved disease diagnostic are continually being explored. For example, altered glycosylation is a universal feature of cancer cells that stems from major disturbances in glycosylation machinery and multiple publications strongly suggest the use of glycan biomarkers for cancer diagnostics. Glycan biomarkers, however, are currently being discovered in a painstakingly slow process using mostly mass spectroscopy as an analytical tool. Since glycosylation changes may occur rapidly and unpredictably, this situation provides an important research incentive to address the need for high-throughput, large-scale technologies for rapid screening of glycan biomarkers. Future research on glycosylation and cancer will benefit from the emergence of new-large scale analytical tools, such as the one shown in FIG. 4 and FIG. 5. 2. Pathogen Detection Kits

Many pathogens express highly characteristic glycosignatures. In most cases there are no readily available lectins or antibodies that can detect these glycan structures. These glycans may be used for the selection of binding peptides that can be used in the construction of pathogen-specific binding peptide chips. These chips will provide pathogen-specific fingerprints directly from the human sera. Most important, due to unique features of bacterial glycomes, this task can be accomplished with or without prior knowledge of carbohydrate identity. Third world countries may especially benefit from this inexpensive technology that does not require high capital investments in instrumentation and expertise.

3. Glycoprotein Profiling Kits

Detection of carbohydrate chain changes in glycoproteins associated with various diseases allows clinical diagnosis and early detection differentially. Binding peptide chips may also be used for quality control of glycoproteins designed for use in developing biopharmaceuticals. As with glycoprotein profiling, the technology of cell surface glycoprofϊling using binding peptides to detect carbohydrate chain changes can be applied to clinical diagnostic examinations. Binding peptide chips may also be used for cellular quality control and evaluation of equivalence and the development of cultured cells in regenerative medicine. The binding peptides identified by the method of the present invention can be used to provide a tool for detection of antibody glycoforms. There can also be provided a diagnostic reagent by which glycoforms of various serum proteins, including antibodies, can be easily and rapidly diagnosed in vitro, and there can also be provided a standard design tool for guaranteeing qualities of cells necessary for a practical stage of regenerative medicine or cell therapy. The lectin library can be applied for example to early finding of diseases, to accurate understanding of morbid states or to therapeutic agent and prophylactic agent in cases where there occur glycosylation of immunoglobulin in rheumatism or autoimmune diseases, glycosylation of specific hormones or proteins upon change from hepatitis to hepatoma, or detection of unique or characteristic glycosignatures in blood doping agents in sports. Finally, reading out of degradation products of carbohydrate chains can be used to sequence the resulting components. III. Applications

A. Glycobiomarkers

Changes in glycosylation have been shown to be associated with various physiological and pathological processes, including cellular differentiation, cancer and autoimmune diseases. The use of the random sequence microarrays for glycobiomarker discovery combines the convenience of the printed array approach with the 'shotgun' selection strategies used in biocombinatorial methods where detailed knowledge of the distinct glycotargets is not required, thus allowing identification of novel glycomarkers in a high-throughput fashion.

B. Biosensors

The use of immobilized peptides to probe for the presence, structure and function of glycotargets is promising for the development of glycan-based biosensors. Most recent efforts have been focused on identification and detection of protein cancer biomarkers overshadowing the importance of glycobiomarkers that could serve as a critical confirmation for the presence or absence of the disease. The FDA- approved biomarkers for ovarian and breast cancer detection, CAl 25 and MUCl, are heavily glycosylated with a glycosylation pattern which is known to be altered at the onset of the disease. The reliable detection of these alterations with specific glycan- binding peptides can significantly aid the development of early cancer diagnostic tests, an important goal which still remains a significant challenge in clinical oncology.

C. Therapeutics

Because of their small size, peptides have favorable pharmacokinetic properties in biomedical applications, such as clearance, improved target-to- background ratio and fairly unimpeded accessibility to target organs/sites. Such small synthetic products could become potent tools in carbohydrate-mediated drug delivery, imaging, or anti-adhesion/proliferation approaches. Chemical synthesis that has virtually no limitations regarding the incorporation of non-natural building blocks, organic residues, branched or cyclized structures may help to circumvent the inherently low affinity of carbohydrate-binding peptides.

D. Drug Discovery

Although united by their defining feature of carbohydrate -binding activity, lectins have been shown to bind small molecules that are predominantly hydrophobic in nature. This feature opens up a possibility of screening selected glycan-binding peptides for binding to combinatorial libraries of drug-like compounds for glycomimetic activity.

IV. Synthetic Lectins (Synlecs)

Currently there is a great deal of interest in synthetic biology, a discipline involved in the design and fabrication of biological components and systems that do not exist in the natural world. Fully synthetic antibodies (Synbodies) and enzymes are being developed as attractive candidates designed to eventually outperform their natural counterparts. Lectins represent another important class of biomolecules that are crucial for proper function of any multicellular organism. The fundamental weakness of using lectins in biomedical applications remains their large size and limited specificity towards their carbohydrate ligands. Despite their widespread distribution in nature, there are only about 60 lectins that are commercially available because they are expensive to isolate, scale-up and purify. An ideal solution to these problems would be to substitute naturally occurring lectins with artificial mimics that are small, highly specific, and readily available.

After identifying the binding peptides, the next step is the selection of an appropriate template-driven assembly to be used as a support to obtain small fully synthetic lectin mimics. A large number of different template models have been used before to mimic the folding of native proteins and, depending on their topology, they can induce and stabilize secondary and/or tertiary structures of the peptides attached to them or give well-defined orientation to the attached peptides by limiting their degree of freedom. Such templates displaying functional groups/ligands in defined spatial orientations are ideal candidates for the synthesis of Synlecs. Several aspects have to be considered in the election of the most appropriate template:

(1) A cyclic structure will facilitate the placement of the peptides in a more cooperative binding disposition

(2) A more rigid molecule will mimic better the slides of the microarray where the peptides were selected

(3) The orthogonality of the protecting groups will give more flexibility for the use of a higher number of different peptides on one single template, and will facilitate the synthesis of several Synlecs on one scaffold (4) The presence of a functionalized position that can be used for the attachment of labels to the Synlec, or the attachment of the whole Synlec to a microarray slide will help during the validation process

(5) The simplicity of the synthesis will help to achieve the goal

FIG. 6 and FIG. 7A and 7B show two scaffolds that combine most of the characteristics above described. Both are cyclic structures, the first one constituted by four amino acids, and the second one by ten. Both syntheses start on solid phase, but while the synthesis of the scaffold shown in FIG. 6A can be completely carried out on a single solid support, the cyclization of the scaffold shown in FIG. 6B occurs in solution.

A template-driven assembly of carbohydrate binding peptides into small fully synthetic lectin mimics (SynLecs) is disclosed (see Example 6). These peptides are placed on a topological template that will work as a scaffold to orientate them and cluster them together.

V. Examples

The following examples are included to demonstrate preferred embodiments of the invention. It should be appreciated by those of skill in the art that the techniques disclosed in the examples which follow represent techniques discovered by the inventor to function well in the practice of the invention, and thus can be considered to constitute preferred modes for its practice. However, those of skill in the art should, in light of the present disclosure, appreciate that many changes can be made in the specific embodiments which are disclosed and still obtain a like or similar result without departing from the spirit and scope of the invention.

EXAMPLE 1 MICROARRAY SELECTION OF GLYCAN BINDING PEPTIDES

The screening method is exemplified by selection of artificial lectin library for the following glycans: monosaccharide Tn antigen, disaccharides lactose and TF antigen, and trisaccharides 3'-sialyllactose and 6'-sialyllactose (FIGS. 8 and 9). These structures were chosen not only because of structural considerations but also based on their important biological functions. For example, both Tn and TF antigens are well-known tumor-prognostic biomarkers. Sialyllactoses, on the other hand, have multiple functions, most important of which is their role in influenza virus adhesion and proliferation.

The peptide microarrays were constructed by spotting 10,000 random 20-mer sequences in duplicates on a maleimide-functionalized microscope glass slide using a robotic pin spotter. The random peptide library was produced by conventional solid- phase synthesis (Alta Biosciences, Birmingham, UK) based on computer-generated random sequences of 19 amino acids, excluding cysteine, for the first 17 amino acids. A C-terminal -GSC sequence was incorporated into each peptide to facilitate coupling to the array surface. These arrays were probed directly with fluorescently labeled LPS.

Microarray interrogation. Peptides of interest were identified by probing a microarray consisting of 10,000 random sequence spotted peptides (CIM10,000). Lactose-, 3'-sialyllactose-, 6'-sialyllactose-encapsulated quantum dots were used as multivalent luminescent glycoprobes to probe an addressable random sequence peptide microarray consisting of 10,000 twenty-mer features printed on polylysine glass slides. The peptides on the CIM10,000 peptide microarray were 20-mers pre- synthesized and spotted in duplicate at a known concentration. The C-terminal sequence of each peptide was -GSC. The sequence of the other 17 amino acids was generated by a random peptide sequence generator using 19 amino acids, excluding cysteine. The thiol group of the C-terminal cysteine was used to attach the peptides to maleimide-activated low-noise poly-L-lysine modified glass slides. Glycan functionalized quantum dots (glyco-QDots) and gold nanoparticles (glyco-GNPs) were prepwered and used to interrogate the microarrays (FIG. 3). The microarrays were probed in three ways (FIG. 3):

QDot Control (FIG. 3B): Each microarray slide was first probed with control QDots derivatized with a linker alone (linker-QD), without any glycans attached. Peptide microarray slides are probed with the control 2-AA labeled QDot. Peptides that show binding here were eliminated when analyzing data from the two following experiments.

QDot -Glvcans (FIG. 3A): Each microarray was probed with glycan labeled QDots (glyco-QDots) fluorescent at 605nm. Slides were probed with QDot labeled glycans (Tn, lactose, TF, 3'-sialyllactose, or 6'- sialyllactose). Glycan Blocking (FIG. 3D): Slides were first blocked with 20OnM unlabeled glycan (Tn, lactose, TF, 3'-sialyllactose, or 6'-sialyllactose). In other words, the microarray was blocked with the sugars attached to non- fluorescent gold nanoparticles (glyco-GNPs). The microarrays were subsequently probed again with QDot labeled glycan fluorescent at 605nm (Tn, lactose, TF, 3'-sialyllactose, or 6'-sialyllactose). Peptides whose binding to QDot glycan was abolished (or significantly reduced) by glycan blocking were considered strong hits. Each of these experiments were performed in triplicate and statistical analysis of the data was done using GeneSpring 7 to identify peptides that bind to QDot glycans only.

Peptide microarray data showed distinct non-overlapping populations of peptides binding to structural isomers 3'- and 6'-silalylactose, and the related lactose (FIG. 10A). The peptide populations for very closely related stereoisomers β-Tn antigen and α-Tn antigen showed non-overlapping and overlapping sets of peptides (FIG. 10B). The validation of selected peptides by various assays (hemagglutination, flow cytometry, SPR) showed that these peptides function similar to the lectins (FIG. 11-12). Representative HRD-SPR data of peptide 3'SL-Pl with 2mM 3'sialyllactose or 2mM 6'sialyllactose injections with nonbinding peptide as control are shown in FIG. HC-D.

EXAMPLE 2 HAPTENIC INHIBITION OF HEMAGGLUTINATION

Hemagglutination of red blood cells (rbc's) has been used to study the ability of flu viruses to cause rbc's to form a matte instead of a button through the multivalent binding of the virus particles to sialic acid residues expressed on the cell surfaces. The titer column indicates the amount of peptide required to inhibit button formation. For example, 3'SL-P2 has the most potent ability to form the matte, while 6'SL-P2 is the least. By preincubating the peptides with sugars, the ability of the sugar to prevent the matte formation indicated that the peptide is binding to the sugars. The peptide 3'SL-Pl required low peptide concentrations to form the matte and high sugar to peptide molar ratios to prevent matte formation, suggesting that this peptide binds sugars on the cell surface preferentially to those in solution. Peptide 6'SL-P2 required a high peptide concentration to form the matte, but relatively low sugar concentration on a molar ratio basis inhibited the matte formation, which suggests that this peptide binds sugars on the surface better than those in solution. All peptides required high concentrations of sialic acid and lactose to inhibit matte formation, but not as high concentrations of 3'sialyllactose and 6'sialyllactose, suggesting that the peptides bind the complete sugar preferentially to the components of the sugar.

Table 1: Haptenic inhibition of hemagglutination by pre-incubating selected peptides with relevant sugars

Peptide Titer 3'Sialyllactose 6'Sialyllactose Sialic Acid Lactose μM μM μM μM μM

3'SL-Pl 49 30 30 390 390

3'SL-P2 9.2 30 45 580 580

6'SL-Pl 39 30 30 580 580

6'SL-P2 90 67 67 1350 900

EXAMPLE 3 COMPARISON OF QDOT- AND ORGANIC DYE-LABELED LPS

The peptide microarrays were constructed by spotting 10,000 random 20-mer sequences in duplicates on a maleimide-functionalized microscope glass slide using a robotic pin spotter. The random peptide library was produced by conventional solid- phase synthesis (Alta Biosciences, Birmingham, UK) based on computer-generated random sequences of 19 amino acids, excluding cysteine, for the first 17 amino acids. A C-terminal -GSC sequence was incorporated into each peptide to facilitate coupling to the array surface. These arrays were probed directly with conventional organic dye-labeled and the QDot labeled LPS in order to specifically single out carbohydrate interactions.

In order to demonstrate that the peptides on a microarray do bind LPS and the interaction is not label- or lipid-induced, several experiments were conducted. First, identical concentrations of LPS from E. coli O111:B4 (ECo ₁₁₁ ) labeled with FITC (FITC-ECoiii) (Invitrogen, Inc.) were used to probe the microarray slides. Each experiment was done in triplicate with randomized slides to test reproducibility and to avoid systematic printing errors. The slides were scanned prior to hybridization with the LPS, so the autofluorescent peptides that equal or exceed the post-hybridization results could be eliminated from the final data. The "unblocked" sample was used as is, while the "blocked" sample was spiked with 100-fold excess of unlabeled ECo ₁₁₁ (Sigma, Inc.) during the hybridization step.

FIG. 13 shows a comparison of raw signal intensity histograms for EC ₀ 55 corresponding to the scanned images before treatment of the slide with BSA prior to hybridization at the correct wavelength detection for the QDot labeled-ECoss (FIG. 13D) and the AlexaFluor488-labeled EC ₀₅₅ (FIG.13A), and the unblocked slides with the AlexaFluor488-labeled EC ₀ 55 (FIG.13B) versus the blocked slides for AF488- ECO ₅₅ (FIG. 13C) and QD-EC ₀ 55 (FIG. 13E) after hybridization with detection at 605nm. First, it is evident that pre-hybridization intensity (FIGS. 13A and D) of the slides is lower than post-hybridization intensity (FIGS. 13B and E) for both AF488- EC055 and QD-EC055 probes. Second, in the AF488-ECo55 unblocked hybridization experiment (FIG. 13B), the histogram is strongly skewed towards high intensity values. Blocking with unlabeled LPS completely eliminates the skew and returns the histogram to its symmetrical bell-shaped curve (FIG. 13C) indicative of the nonspecific interactions. Furthermore, the histogram colors in FIG. 13 are arranged by the level of peptides' fluorescent intensity (red is high, green is low). It is noticeable that most of the top binders in the central histogram become low binders when excess of unlabeled LPS is used to inhibit the specific peptide-LPS interactions in the histogram to the right. This simple test effectively eliminated any unspecific dye- induced interactions. The same test was applied to all hybridization experiments described thereon. In all cases, the histogram shift to the high intensity values along with the opposite shift upon blocking was interpreted as a positive test for specific binding.

In order to test the applicability of the Qdot-LPS probes to detect specifically the carbohydrate binding events, an experiment was conducted to compare the results with those obtained with the conventionally labeled LPS. Separate experiments compared FITC-labeled and Qdots-labeled LPS from E. coli Ol 11 :B4 (EC0111). FIG. 13 shows reasonable correlation (R=O.824) between the two experiments. Although some unique hits were present both in Qdot-LPS and FITC-LPS binding peptides, they may be attributed to differences in the probe construction and photophysical properties of the labels. The LPS is presented on the multivalent Qdots in a defined orientation since the hydrophobic Qdots only exist in aqueous solution when they are enclosed in the hydrophilic environment created by the lipid portion of the LPS molecules. This orientation exposes the saccharidic branch of the LPS, similar to their orientation in the micellar (or cellular) state of LPS. In contrast, the monovalent FITC-LPS probe would detect both saccharidic and lipid-induced interactions. As it has been found that aggregated micellar FITC-LPS has strongly diminished fluorescence, while the disaggregation of single FITC-LPS molecules from the micelles leads to enhancement in fluorescence, it is likely that oriented micellar LPS molecules would not be observable, and the most significant signal detected would come from the single LPS molecules, which include non-saccharidic lipid components. Due to the well-known cluster glycoside effect (Lundquist and Toone, 2002), the interaction of multiple sugars is also stronger than a single LPS molecule. These observations highlight the utility of Qdot-LPS for the studying of the variable saccharidic components.

EXAMPLE 4 BACTERIAL GLYCOPROFILING

Selection of peptides binding saccharidic branch of LPS. Since the Qdot labels were novel for LPS, it was asserted that only peptides that bind both FITC-LPS and Qdot-LPS with high affinity were the most reliable saccharidic LPS-binding peptides. Such peptides were identified by statistical analysis using image-processed data. Lipopolysaccharides (LPS) were labeled with quantum dots and with FITC organic dye to detect bacteria. The microarray was probed and only high intensity binders with a minimal standard deviation (σ<0.2) that were present in both Qdot-LPS and FITC-LPS experiments were selected (FIG. 14). Autofluorescent peptides were filtered out and each hit was independently confirmed by careful visual inspection of the slides. The data revealed 16 peptides QFl -QF 16 that bind with high affinity to E.coli O111 :B4 LPS (Table 2). Most of these peptides contain noticeably abundant cationic arginine (R), lysine (K) and histidine (H) along with clusters of aromatic hydrophobic tryptophan (W) and phenylalanine (F) and/or tyrosine (Y). Table 2. ECoiii LPS binding peptides arranged by pi - isoelectric point; NR - number of negative residues; PR- number of positive residues; AI - aliphatic index; APP - antibacterial peptides prediction score (Lata et al., 2007).

As many of the existing LPS-binding peptides are also antimicrobial (Mancek et Ia., 2002), it was hypothesized that if the selections were valid then at least some of the peptides should share sequence similarity with the existing antimicrobial peptides (AMPs). To test this hypothesis, the selected sequences were compared against several AMP databases. In particular, it was found that the above amino acids were also abundant in indolicidin-like AMPs (Chan et al., 2006). Moreover, using Antimicrobial Peptide Database (Wang and Wang, 2004), it was found that some of these peptides (QF1-8) shared 30 to 40% similarity to human histatins-2, -6, or -9, which are histidine-rich AMPs found in oral cavities. Finally, an algorithm that predicts antibacterial sequences based on similarity was applied to the existing 486 AMPs (Lata et al., 2007). The higher the antibacterial peptides prediction (APP) score, the more probable the antibacterial activity, while negative scores suggest no antibacterial activity. The APP scores shown in Table 2 suggested that 11 out of 16 selected peptides had potential antibacterial properties. Because most known AMPs are dominated by amphipathic cationic sequences, only peptides QFl-I l were predicted to have antibacterial properties. These observations are further supported by two independent investigations. In one study, peptides were selected to bind components of the bacterial membrane devoid of polysaccharides. This selection led to peptides containing only cationic arginine and lysine but no aromatic residues (Xie et al, 2006). In a second study, peptides that bind LPS from Salmonella enteriditis were identified by screening phage displayed peptide libraries against bead-immobilized LPS (Kim et al, 2006). All of these peptides were found to be enriched in aromatic hydrophobic residues such as tryptophan and phenylalanine. These peptides were also capable of discriminating between various bacterial species, which strongly supports their ability to target the distinctly variable O-antigenic domains.

Antimicrobial properties of LPS binding peptides. It was hypothesized that the LPS-binding peptides may possess antibacterial properties. The ability of these peptides to inhibit E. coli growth was assayed and the LPS-binding peptides were compared to 150 LPS non-binding peptides. FIG. 15 shows that nearly 70% of the LPS-binding peptides demonstrated some growth inhibition activity against E. coli DHlOB, while none of the 150 non-binding peptides inhibited the growth by more than 280%. Interestingly, the peptides QF 12, 13, 14, and 16 demonstrated enhancement of bacterial growth (FIG. 15). In agreement with the APP scores (Table 2), peptides QFl-10 displayed antibacterial activity. An evident outlier QF 15 was also identified that departs from the conventional cationic amphipathic motifs associated with AMPs, highlighting the utility of this method for discovering unconventional antibiotics. Further testing through kinetic growth curves showed that these peptides are bacteriostatic, not bactericidal. This agrees with recent work demonstrating that the biophysical properties required to kill bacteria differ from those to bind LPS (Rosenfeld et al, 2008). In addition to affinity for LPS, bactericidal activity requires the abilities to traverse the LPS layer and to disaggregate LPS micelles. The concentration dependent studies (data not shown) demonstrated that even at lOμM concentration the peptides QF7, QF8, and QFlO retained their ability to inhibit 30-50% of E. coli growth.

The strategy of screening a specific set of antimicrobial targets against an array of chemically diverse peptide leads is fundamentally different from current antimicrobial screening protocols (Hancock and Sahl, 2006), which utilize peptides derived from innate immune defenses. The strategy gives precise knowledge of the target and the mechanism of action, creating an advantage in further lead optimization. The lectin- like activity of the LPS-binding peptides may be exploited in combination therapies where the LPS-binding peptides are used to depolarize the bacterial membrane, allowing conventional antibiotics to penetrate the dense hydrophilic LPS barrier of gram-negative bacteria.

Flow cytometry studies of the LPS binding peptides. Flow cytometry studies were conducted to quantify the in vivo binding abilities of the selected peptides to E. coli and the ability of pre-incubation with LPS to block the binding. The LPS non-binding peptide Negl was used as a negative control. The peptides were biotinylated and their specificities were compared through quantifying the cell surface staining of E. coli DHlOB cells with AlexaFluor488-labeled streptavidin. The peptides QFl through QFlO bound the cells almost completely in the Ml region, distinguishing the binding of the peptides from the streptavidin only cell binding. The results for streptavidin, the control peptide Negl, QF5 and QF8 are summarized in FIG. 16. Both QF5 and QF8 bound to the DHlOB cells and their cell surface binding was nearly eliminated after pre-incubation with E. coli O111 :B4 LPS. This result strongly suggests that the observed binding demonstrates the interaction of the peptides with the cell surface LPS. The peptide QF2 was also pre-incubated with E. coli O111 :B4 LPS and its binding to the cells was reduced by one third, suggesting that QF2 is binding LPS but may also be binding to other cell surface components.

Surface Plasmon Resonance. High resolution differential surface plasmon resonance was used to compare the binding of QF5, QF8 and Negl to LPS from E. coli O111 :B4 (FIG. 17). This technique has sufficient sensitivity to detect direct binding of free glycans to lectins immobilized on a sensor chip, allowing the evaluation of sugar-lectin dissociation constants as small as in the nM range. Both QF5 and QF8 were strong antimicrobial candidates, while Negl, a peptide showing no binding to LPS on the peptide microarrays, was used as a negative control. The responses of these peptides were compared using normalization based on the immobilization density and the molecular weight of the respective peptides. Both QF5 and QF8 peptides had similar abilities to bind LPS, while Negl had negligible binding.

Differentiation of E. coli serotypes. Gram-negative bacteria are classified by serological types (serotypes) based on the composition of the LPS O-antigen domains. Thus, the O-antigen, responsible for much of the immunospecifϊcity of the bacterial cells, essentially serves as a 'fingerprint' for a bacterium (Caroff and Karibian, 2003). To test whether different serotypes of a bacterium can be distinguished using the peptide microarray, Qdot-labeled LPS derived from two different serotypes of E. coli (O111 :B4 (ECoin) and O55:B5 (EC ₀₅₅ )) were screened. FIG. 18 shows the ID- scatter plot associated with these experiments. Overall, an excellent correlation (R=O.907) was observed between the two serotypes, indicating that there are only marginal differences detectable by the microarray. The high correlation coefficient is in agreement with the compositional similarity of the LPS molecules derived from the two serotypes (FIG. 19). The O-antigen repeating units of ECoiii (Gupta et al, 1995) and ECo ₅₅ (Samuel et al, 2004) are composed of five neutral monosaccharides that include glucose (GIc), galactose (Gal), N-acetyl-galactosamine (GaINAc), N- acetylglucosamine (GIcNAc), and colitose (Col, 3,6-dideoxy-L-galactose). Although both structures differ in branching and sequence, the overall sugar content remains quite similar. Despite the negligible statistical differences, a close visual inspection of the slides reveals several distinct hits that are unique to ECo ₁₁₁ (insert to FIG. 20) and to ECo ₅₅ . In all the cases, for a hit to be statistically significant it must reproduce in all replicate slides with a standard deviation of less than 0.2.

Differentiation of E. coli and P. aeruginosa. While the two E. coli serotypes have subtle compositional differences, more prominent differences are apparent when the LPS structures of P. aeruginosa 10 (PA ₁ O) and ECoiii are compared (FIG. 21). The repeating unit of PA ₁ O consists of three unusual sugars: 2-O-acetyl-L-rhamnose (RhaAc), 2-N-acetyl-L-galacturonic acid (GaINA), and 2-N-acetyl-2,6-dideoxy-D- glucosamine (QuiN). One of these sugars (GaINA) contains carboxylic acid group that can carry negative charge and form strong hydrogen bonds. Screening of the PA ₁ O LPS labeled with Qdots and statistical correlation of the results with ECoiii revealed a number of distinct hits for ECo ₁₁₁ and PA _1O . Even a superficial visual inspection of the slides immediately shows binding pattern differences between the two experiments (FIG. 20). Table 3. Peptides specific to EC LPS vs. PAi ₀ LPS. Column headers indicate the pi - isoelectric point; NR - negative residues; PR - positive residues; AI - aliphatic index.

FIG. 21 shows statistical correlation between QdOt-PA _1O and Qdot-ECom experiments. The correlation coefficient is far lower (R=O.630) than in the case of ECoiii vs. EC ₀₅₅ (R=0.907) (c/ FIG. 18). The peptides ECl-8 (Table 3) that specifically bind ECo ₁₁₁ but not PA _1O were identified by minimizing the error (standard deviation σ<0.2) while maximizing the ratio of normalized ECo ₁₁₁ to PA _1O signals. These comparisons independently validate the first selection of ECoiii binding peptides QF1-16 (Table 1) which are annotated in blue in Fig. 21. A similar selection strategy seeking peptides that specifically bind PA ₁₀ but not EC _{011 I} yielded peptides PA1-8 (Table 3). Most of the peptides unique to EC ₀₁₁₁ are enriched in aromatic tryptophan and cationic arginine, lysine, and histidine, while the peptides specific to PA ₁ O tend to contain aliphatic amino acids, anionic aspartic and glutamic acids, and especially hydrogen bond forming glycine, proline, serine and threonine (see FIG. S3). These differences are reflected in the consistent differences in pi values and aliphatic indices (AI) of the selected peptides (Table 3). This can be explained by the prominent compositional differences between ECo ₁₁₁ ^an d PA _1O LPS (FIG. 19).

In contrast to the neutral ECoiii and EC ₀ 55 repeating units, the repeating unit of PA ₁ O contains 33% of negatively charged galacturonic acid GaINA (FIG. 19), which can form strong hydrogen bonds with the aspartic and glutamic acids as well as with hydrophilic glycine, proline, serine and threonine that are prominently over- represented in the selected PAio-specifϊc peptides. Using the microarray data obtained during the selection of PAio-specific peptides, all the non-specific top peptide binders with σ<0.2 for PA _1O were examined. These top binders had a wide range of pi values and no evident preference for any particular amino acids, except cationic H/K/R, in contrast to the specific binders shown in Table 2.

To test the contribution of electrostatic interactions to the microarray binding of the LPS to the peptides, the z-potential (ζ) of ECo ₁₁₁ ^an d PA _1O LPS were measured. Z-potential is the overall charge a particle acquires in a specific medium, and is a measure of the potential at the slipping plane, which is the layer just past the bulk solution layer of ions surrounding the particle. Under the conditions identical to those used in the microarray probing experiments, the ECo ₁₁₁ LPS had a charge of ζ= - 6.7±1.4mV, while the PA _1O LPS was also negative and of significantly greater magnitude at ζ= -25.7±3.1mV consistent with the presence of negatively-charged galacturonic acid. Since the ECo ₁ 11 LPS has only hydroxyls in the structure and thus lacks the ability to form strong hydrogen bonds in aqueous solutions, its interactions are dominated by CH-π interactions and by electrostatic attraction, driving the selection towards hydrophobic aromatic and cationic amino acids. On the other hand, the galacturonic acid in the repeating unit OfPA ₁ O LPS has a strong propensity to form hydrogen bonds, which evidently overpowers the electrostatic forces, thus driving the selection towards hydrogen-bond forming amino acids. Thus, the specific interactions of peptides with LPS on the microarray are not driven by electrostatic forces alone, but involve far more specific molecular interactions such as hydrogen bonds and hydrophobic forces. This makes the peptide microarray a suitable tool for studying carbohydrate interactions. EXAMPLE 5 MATERIALS AND METHODS

Smooth type LPS from E. coli serotype O111 :B4 was from Fluka (cat# 62325), serotype 055 :B5 was from Sigma (cat# 62326), P. aeruginosa 10 was from Sigma (cat# L8643). FITC-labeled LPS from E. coli O111 :B4 was from Sigma (cat# F3665), AlexaFluor-488 labeled LPS from E. coli 055 :B5 was from Invitrogen (cat# L23351).

Unless noted otherwise, all chemicals were purchased from Sigma-Aldrich, Inc (Milwaukee, WI) and used without further purification. Deionized water was obtained from a Millipore ultrapure water filtration unit. PEPscreen ^® peptides were synthesized by Sigma-Genosys, Inc with 100% quality control and used as received for initial screens. Lead peptides were re-synthesized in-house using Fmoc chemistry and purified to 95% by HPLC. Organic QDots were purchased from Invitrogen (Carlsbad, CA, cat# Q2170 IMP). Sephacryl HiPrep 16/60 (S-200 HR) was from GE Healthcare. Transmission electron microscopy (TEM) was done on a Philips CM12S microscope. Scanning electron microscopy was performed on a Leica-Cambridge 360FE microscope. Confocal microscopy was performed on a Carl-Zeiss LSM501 confocal microscope. In-solution nanosizing and zeta potential was measured on a Zetasizer Nano-ZS instrument (Malvern Instruments, UK). Spectrophotometry measurements were carried out on a NanoDrop ^® ND- 1000 instrument.

Labeling LPS with Qdots. The supplied solution of organic Qdots (QDot ^® 605 ITK™, cat# Q2170 IMP, Invitrogen, Inc) in decane (lμM) was evaporated to dryness on a SpeedVac ^® at room temperature (rt) and re-dissolved in equal amount of chloroform. A lOOμL aliquot of the chloroform solution was diluted to 500μL with chloroform and mixed with lOOμL of lOmg/mL aqueous solution of corresponding LPS (E. coli O111 :B4, E. coli O55:B5, and P. aeruginosa 10). Methanol was added dropwise with occasional vortexing until complete mixing of both phases was achieved (ca. 400 μL of MeOH). The mixture was then evaporated to dryness on a SpeedVac ^® and the solid residue was suspended in lOOμL of ddH ₂ O. A saturated solution of tetramethylammonium hydroxide pentahydrate (Me ₄ NOHxSH ₂ O) was added until the mixture was at pHl l-12 (ca. 25 μL). The latter basification step is critical as it allows the transfer of the Qdots into the aqueous phase; no transfer occurs in non-basified solutions. The colored solution was then passed through two consecutive Zeba columns (2mL, Pierce) to remove salts and excess of free LPS. The LPS-coated Qdots were further purified by size-exclusion chromatography using Sephacryl HiPrep 16/60 (S-200 HR) column (5 Ox lcm). The LPS-Qdot constructs eluted in a narrow color band and were stored in the dark at 4 ⁰ C. Under these conditions, the LPS-Qdots are stable for at least one month without any visible signs of deterioration. In a control experiment, the above procedure was repeated without LPS. No solubilization of Qdots was observed without LPS as determined by measuring absorbance of Qdots in the supernatant.

Peptide Microarray Design and Construction. The peptide microarray consists of 10,000 20-residue peptides of random sequence, with a C-terminal linker of Gly-Ser-Cys-COOH. Each peptide was synthesized by Alta Biosciences Ltd (Birmingham, UK) based on amino acid sequences provided by in-house custom software (Hunter, Preston and Uemura, Yusuke, CIM, The Biodesign Institute). Nineteen amino acids (cysteine was excluded) were selected at random for each of the first seventeen positions with -GSC as the carboxy-terminal linker. The synthesis scale was lmg total at 95% purity with 2% of the peptides tested at random by mass spectrometry as quality control. Dry peptides were brought up in 100% N ₅ N'- dimethylformamide until dissolved, then diluted 1:1 with purified water at pH 5.5 to a master concentration of 2mg/mL. The original 96-deep-well plates were robotically transferred to 384-well spotting plates, and the peptides were diluted to a final spotting concentration of lmg/mL concentration in phosphate buffered saline at pH 7.2. High-quality pre-cleaned Gold Seal glass microscope slides were obtained from Fisher (Fair Lawn, NJ, cat# 3010). Each slide was treated with amino-silane, activated with sulfo-SMCC (Pierce Biotechnology, Rockford, IL, cat# 22622) to create a maleimide-activated surface, and quality checked for coating efficiency. During spotting, a Telechem Nanoprint 60 using 48 Telechem series SMP2 style titanium pins was employed. Each pin spots approximately 500 pL of lmg/mL peptide per spot, as estimated based on pin trajectory, surface dwell time, and the amount of liquid each pin holds. The spotting environment is 25°C at 55% humidity. The maleimide-activated surface reacts with the sulfhydryl group on the peptide's terminal cysteine. Each peptide is spotted twice per array. The arrays are spotted in an orange-crate packing pattern to maximize spot density. Six fiducials are applied asymmetrically using AlexaFluor-647, -555 and -488 labeled peptide. The fiducials are used to align each subarray during image processing. The printed slides are stored under an argon atmosphere at 4°C until used. Quality control includes imaging the arrays by laser scanner (Perkin-Elmer ProScanArray HT, Perkin Elmer, Wellesley, MA) at 647nm to image the spot morphology. If the batch passes this test, further testing of randomly selected slides with known proteins and antibodies allows QC of precision spot intensity. Array batches that fail to meet an array-to-array variability of 30% CV are discarded.

Microarray Probing. Each microarray probing was performed in triplicate. The slides were placed in a humidity chamber and blocked for 1 hr at rt with 650 μL of 3% BSA solution and ImM mPEG ₄ -thiol (Zheng et ah, 2004) in IxPBS with 0.05% Tween-20 (TBS-T). The slides were washed with IxTBS-T (3x30 swishes in a Coplin jar) and ddH ₂ O (3x30 swishes in a Coplin jar). The slides were then dried by centrifugation at 1500 rpm for 3 min, with the barcode label at the bottom to avoid spreading the label glue onto the slide surface. The slides were then scanned at the appropriate wavelength to note any peptide autofluorescence. An AbGene frame was then attached to the surface of each slide to confine the 260μL solution of labeled LPS in IxPBS (0.154 mg/mL for FITC-LPS and 0.630 mg/mL for Qdot-LPS) was added to the printed area. A plastic coverslip was used to spread the solution on the surface of the slide and seal the frame while avoiding bubbles. The slides were incubated for 1 hr in the dark at rt in a humidity chamber. The coverslips and AbGene frames then were removed and the slides were washed by dipping two times in ddH ₂ O, then incubating the slides for 5 min in ddH ₂ O, and then dipping 2 more times in ddH ₂ O, changing the solution each time. Finally the slides were dried by centrifugation at 1500rpm for 3 min at RT and scanned.

Microarray Scanning and Image Analysis. Microarrays were scanned using a Perkin Elmer ProScanArray HT Microarray Scanner using the 488 and 543 nm excitation lasers at 100% power and 70% photomultiplier tube gain. The detection was done at 605nm for Qdot probes and at 543nm for the FITC probes. All scanned images were analyzed using GenePix Pro 6.0 software (Axon Instruments, Union City, CA). Upon careful visual inspection, bad spots were eliminated by flagging them "absent". Median spot intensities were used in further analyses. Statistical analysis comparison of microarray data was done with GeneSpring 7.2 (Agilent, Inc, Palo Alto, CA) by importing image-processed data from GenePix Pro 6.0 (Molecular Devices, Inc.). Median signal intensities were used in the calculations. For statistical comparisons, each slide was normalized to 50 ^th percentile. Measurements of less than 0.01 were set to 0.01. No per "gene" normalization was included. Autofluorescent peptides were identified by scanning the slides prior to hybridization with LPS and the peptides whose fluorescent intensity was comparable to the post-hybridization intensity were eliminated from the final selections.

Flow cytometry. Peptides were conjugated to biotin by incubation with heterobifunctional maleimide-PEO ₂ -biotin linker (Pierce Biotechnology, Inc., cat# 21901) in IxPBS at pH 7.2 overnight at rt. Excess of biotin was removed by overnight dialysis on IkD cutoff membrane (Spectrum Laboratories, Inc). DHlOB E. coli bacterial cells (MAX Efficiency ^® DHl 0B™ Competent Cells, Cat# 18297-010, Invitrogen Inc) were cultured under routine conditions, pelleted by centrifugation and washed 3x with IxPBS to remove traces of media. The harvested cells were resuspended in blocking buffer (IxPBS, 0.05% FBS). Experiments with LPS preincubation involved mixing lOul of 400μM biotinylated peptide solution with 20μl of a 200μM LPS solution for one hour at rt. Then, 10 xlO ⁶ cells were mixed with lOμl of 400μM biotinylated peptides, biotinylated peptides pre-incubated with LPS, or biotinylated WGA lectin (EY Laboratories, Inc.) (10 μL of lmg/mL solution) and incubated for lhr on ice. The cells were then washed three times with ImL of blocking buffer to remove unbound peptides or lectins. For detection of bound peptides and lectins, lOOμL of 4μg/ml streptavidin-AlexaFluor488 (Invitrogen, Inc., cat# S-11223) was added to the cells and incubated for 1 hr. Cells were washed three times in blocking buffer, re-suspended in 300 μL of blocking buffer and analyzed for cell surface staining using the FACS Caliber machine (BD Biosciences, Inc.). Cells stained with streptavidin-AlexaFluor488 only were used a control.

Surface Plasmon Resonance. Bare gold SPR sensor chips (Biosensing Instruments, Tempe AZ) were functionalized by adding ImM 8-amino-l-octanethiol in ddH ₂ O(AOT, Dojindo Molecular Technologies, Inc, cat# A424) to the ethanol pre- washed gold surface and incubating for 2 hrs at rt in a humidity chamber. The surface was then washed with ddH ₂ O and dried using ultrapure Argon gas. A ImM solution of sulfo-SMCC linker (bio-WORLD, Dublin OH) in IxPBS was added to the gold surface, incubated for 30 min at rt in a humidity chamber, then washed and dried as above. The SPR instrument sensitivity was calibrated using the response from 1% ethanol in water on a bare gold sensor chip as a standard. The AOT/sulfo-SMCC modified chip was mounted on the instrument, and the peptide was immobilized in the sample channel by injecting a lOOμM solution of peptide in TBS-T. A solution of 0.01% sodium dodecylsulfate (SDS) in TBS-T was injected to dissociate any peptide aggregates. The SPR response after the SDS wash was used to calculate the immobilization density, where lRU=lpg/mm ² . To abrogate possible non-specific interactions, a ImM solution of methoxytetraethyleneglycol thiol (InPEG ₄ -SH), prepared as described by Zheng et al (2004), was used to block unreacted maleimide groups on the sample channel and to act as a non-binding control on the reference channel. A lmg/mL solution of LPS from E. coli serotype O111 :B4 was injected at 20μL/min flow rate in the TBS-T analyte solution. Regeneration was accomplished using 0.05% SDS followed by an injection of 1OmM glycine (pH 2.5).

Antimicrobial assays. DHlOB E. coli cells (MAX Efficiency ^® DHl 0B™ Competent Cells, Cat# 18297-010, Invitrogen Inc) were grown overnight at 37 ⁰ C at 270 rpm rotation in TB media with 0.1% Streptomycin to a cell density of 2,00OxIO ⁶ CFLVmL. An aliquot of 1 μL of these cultured cells were then mixed with 1 mL of fresh media containing individual peptide at concentrations of 25, 50 and lOOμM and allowed to grow overnight. The McFarland turbidity scale for E. coli. (Koch, 1994; Smibert and Keig, 1994) was used to quantify the overnight growth of the cells by comparing the optical density of the cells at 600 nm to the turbidity equivalent of 1% BaCl ₂ A % H ₂ SO ₄ in the micro plate reader (Spectra MAX 190, Molecular Devices, Inc). In control experiments, the above procedure was repeated with no peptide in the culture and with non-binding peptide (Negl, sequence EFSNPTAQVFPDFWMSDGSC (SEQ ID NO:34)) as a negative control.

EXAMPLE 6 SYNTHETIC LECTINS

Two cyclic regioselectively addressable functionalized templates (RAFTs) were chosen as scaffolds to place the previously selected peptides (see Example 4) in an upward 'crown- like' fashion. In this format, the attachment of three carbohydrate binding peptides of differing specificity (Pl, P2, P3) to a particular sugar not only amplifies the binding affinity of the Synlecs, but also their specificity due to complementary binding mechanisms. This strategy allows efficient de novo design of the novel template assembled synthetic lectins. FIG. 22.

Scaffold selection. After identifying the binding peptides, the next step was the selection of an appropriate template-driven assembly to be used as a support to obtain small fully synthetic lectin mimics. The template used as a support for the carbohydrate binding peptides was a cyclic beta-tetrapeptide. The methylene group between the amino and carboxy function in each beta-amino acid residue orientates the NH and CO groups to opposite faces of the beta-peptide ring that lets the axial and interior positions of the ring unobstructed. The use of an orthogonally protected cyclotetrapeptide gives total flexibility in placing up to three different peptides onto the scaffold in a "crown-like" fashion. In this format, the attachment of three carbohydrate binding peptides of differing specificity to a particular sugar not only amplifies the binding affinity of the SynLecs, but also their specificity due to complementarity of binding mechanisms. FIG. 6 shows two scaffolds that combine most of the desired characteristics. Both are cyclic structures, the first one constituted by four amino acids, and the second one by ten. Both syntheses start on solid phase, but while the synthesis of the scaffold shown in FIG. 6A can be completely carried out on a single solid support, the cyclization of the scaffold shown in FIG. 6B occurs in solution.

The cyclic β-tetrapeptide a presents a very rigid structure, and the extra methylene group between the amino and carboxy function in each β-amino acid residue allows all the peptides to be placed on the same face of the scaffold. The presence of three different protective groups allows the introduction of up to three different peptides in the Synlec. One of the most challenging steps in the synthesis of a cyclic peptide is the cyclization. Here, the cyclization is carried out on the solid phase and when the scaffold is completely synthesized, it is cleaved from the resin. This makes the synthesis very attractive in terms of simplicity.

The cyclic decapeptide b contains two linked β-turns. Consequently, the lysine side chains used in the attachment of the peptide branches are orientated to the same face of the peptide ring. However, in this case the template was more flexible and allowed the peptides to move more freely. With the use of orthogonal protecting groups, up to four different peptides can be introduced. In addition, there is an extra position on the opposite face of the structure that can be used either to attach a label to the Synlec or to place the Synlec in an array in order to validate the carbohydrate- Synlec binding. Synthesis of the scaffolds is described in FIGS. 7A-B. 1. Cyclic tetrapeptide linker synlec

The synthesis of a cyclic tetrapeptide having three orthogonally protected conjugation sites for attachment of peptide or other affinity elements is demonstrated. The structure shown in FIG. 7A is synthesized from three modified amino acids, and a fourth one that is commercially available, as shown. The three amino acids are first synthesized and the resin modified. The synthesis of the tetrapeptide is then carried out and peptides or other affinity elements are added. Thus, the tetrapeptide serves as a linker for construction of a synbody. a. Synthesis of the modified amino acids

1-Methyl-l-phenylethyl 3-aminopropanoate (3): Over a suspension of NaH (50 mg, 2.1 mmol) in diethyl ether (2 mL), a solution of 2-phenyl-2-propanol (2.5 g, 18.36 mmol) in 2 mL of diethyl ether was added dropwise. The mixture was stirred at room temperature for 20 min and then cooled at 0 ⁰ C. Trichloroacetonitrile (1.9 mL) was slowly added (for 15 min) and the mixture was allowed to reach room temperature. After 1 hour of stirring, the mixture was concentrated to dryness and the resultant oil was dissolved in pentane (2 mL) and the solution was filtered. The filtrate was evaporated to dryness, to get a very dark oil, that is used immediately in the next reaction. The freshly prepared 1-methyl-l-l -phenyl ethyl trichloroacetimidate (2.7 g, 6.424 mmol) was added over a solution of Fmoc-β-alanine, 1, (1 g, 3.212 mmol) in DCM (8 mL). After overnight stirring, the precipitated trichloroacetamide was removed by filtration, and the filtrate mixture was evaporated to dryness and purified by flash chromatography CH ₂ Cl ₂ /Me0H (0% to 1%) to yield 1.158 g (84%) of compound 2 as a colorless oil.

In a flask, 2 (1.158 g, 2.698 mmol) was solved in DCM (4 mL), and diethylamine (12 mL) was added. Inmediately, the mixture becomes clear. The mixture was stirred for 2 hours. After adding 20 mL of toluene, the mixture was concentrated to dryness and the separation carried out by flash chromatography, using 10% of CH ₂ CVMeOH and 2% OfEt ₃ N to yield 526 mg (94%) of 3 as a colorless oil.

7V ² -(allyloxycarbonyl)-7V ³ -(9-fluorenylmethoxycarbonyl)-2,3- diaminopropanoic acid (7): Over a solution of 2 g of asparagine (4, 15.138 mmol) in 3.78 mL of 4M NaOH solution cooled in an ice-bath, 1.615 mL of allyl chloroformate (15.138 mmol) and 3.78 mL of 4M NaOH solution in portions were added. The reaction was kept alkaline and stirred for 15 minutes at room temperature. The mixture was extracted with ether and acidified with concentrated HCl, so the product was crystallized, filtrated, and lyophilized to afford 5 (2.816 g, 86%) as a white solid.

[Bis(trifluoroacetoxy)iodo]benzene (8.402 g, 19.539 mmol) was added to a mixture of 5 (2.816 g, 13.026 mmol) and aqueous DMF (140 mL, 1 :1, v/v). The mixture was stirred for 15 min, and DIEA (4.54 mL, 26.052 mmol) was added. After 8 hours the reaction, only half of the reaction went. So, the same quantities of [Bis(trifluoroacetoxy)iodo]benzene and DIEA were added, and the reaction was stirred overnight. The next day, the solution was concentrated to dryness, the residue solved in 100 mL of water and the organic side products were removed by repeated washings with diethyl ether (4 x 100 mL). The water phase was evaporated to dryness to yield product 6 that was used in the next reaction without further purification.

The oil previously obtained (6) was redissolved in water (20 mL), and DIEA (2.24 mL, 13.026 mmol) and FmocOSu (4.393 g, 13.026 mmol) in acetonitrile (15 mL) were added, and the reaction was allowed to stir for 1.5 h. The mixture was acidified (to pH 2.0) by addition of HCl, and the product was extracted in DCM (5 x 40 mL). The organic phases were combined, dried with Na ₂ SO ₄ , and evaporated to dryness. The crude product mixture was purified by flash chromatography (10% MeOH in DCM). Hexane was added to the combined product fractions, and the precipitate formed was filtered and washed with hexane, and dried to yield 7 as a white solid.

2-azido-3-[(9-fluorenylmethyloxycarbonyl)amino] -propanoic acid (10): A solution of NaN ₃ (9.841 g, 151.38 mmol) in 25 mL of H ₂ O was cooled in an ice bath and treated with 50 mL of CH ₂ Cl ₂ . The biphasic mixture was stirred vigorously and treated with Tf ₂ O (8.542 g, 282.14 mmol) for over a period of 30 min. The reaction mixture was stirred at ice bath temperature for 2 h. After quenching with aqueous NaHCO3, the layers were separated, and the aqueous layer was extracted twice with CH ₂ Cl ₂ (2 x 50 mL). The organic layers were combined to afford 100 mL Of TfN ₃ solution that was washed once with Na ₂ CO ₃ and used in the next reaction without further purification.

To a solution of L-asparagine 4 (2 g, 15.138 mmol) in 50 mL of H ₂ O and 100 mL of MeOH were added: K ₂ CO ₃ (3.138 g, 22.707 mmol), CuSO ₄ (38 mg, 0.151 mmol), and the solution of TfN ₃ in CH ₂ Cl ₂ previously prepared. The reaction was stirred at room temperature overnight. Then, solid NaHCO ₃ (10 g) was added carefully, and the organic solvents evaporated. Concentrated HCl was added to the aqueous solution to obtain pH=6, and 100 rnL of 0.25 M PBS was added. Then, ethyl acetate (3 x 150 rnL) was used to do extractions. Next, more concentrated HCl was used to reach pH=2 and new extractions were carried out with ethyl acetate (5 x 150 mL) and the extract concentrated to dryness to afford 8 as a yellow oil, that was used in the next reaction without further purification.

[Bis(trifluoroacetoxy)iodo]benzene (19.529 g, 45.414 mmol) was added to a mixture of the crude 8 (15.138 mmol) and aqueous DMF (120 mL mL, 1 :1, v/v). The mixture was stirred for 15 min, and DIEA (10.546 mL, 60.552 mmol) was added. The reaction continue overnight. The next day, the solution was concentrated to dryness, the residue dissolved in 100 mL of water and the organic products were removed by repeated washings with diethyl ether (3 x 100 mL). The water phase was evaporated to dryness to yield product 9 as a pale oil, that was used in the next reaction without further purification.

The oil previously obtained (9) was redissolved in water (20 mL), and DIEA (2.6 mL, 15.138 mmol) and FmocOSu (5.106 g, 15.138 mmol) in acetonitrile (15 mL) were added, and the reaction was allowed to stir for 1.5 h. The mixture was acidified (to pH 2.0) by addition of HCl, and the product was extracted in DCM (5 x 40 mL). The organic phases were combined, dried with Na ₂ SO ₄ , and evaporated to dryness. The crude product mixture was purified by flash chromatography (10% MeOH in DCM). Hexane was added to the combined product fractions, and the precipitate formed was filtered and washed with hexane, and dried to yield 10 as a white solid. b. Derivatization of the resin

A mixture of Boc- and Fmoc- [beta] -alanine (2.0 eq of both, 4.0 equiv of TBTU, 8 equiv of DIEA in DMG, Ih at 25 ⁰ C) was coupled to aminomethyl polystyrene resin (1.0 g, 0.5 mmol/g). 50% TFA in DCM was used to remove the Boc groups, and the exposed amino groups were capped with acetanhydride treatment. Thus, the loading of the resin was reduced to 0.16 mmol/g. A treatment of 20% piperidine in DMF was used to remove the Fmoc groups, and 4-(4-formyl-3,5- dimethoxyphenoxy)butyric acid was attached by HATU-promoted coupling to obtain the derivatized resin. c. Synthesis of the scaffold on the resin

Previously derivatized resin (1.0 g, a loading of 0.16 mmol/g) was treated for 1 h at room temperature with a mixture of 1 -methyl- 1-phenylethyl 3-aminopropanoate (3, 160 mg, 4 equiv) and NaCNBH3 (48 mg, 4 equiv) in DMF, containing 1% (v/v) AcOH (16 mL). The resin was washed with DMF, DCM, and MeOH and dried on a filter.

The secondary amine was acylated with Aloc-Dpr(Fmoc)-OH 7 (5.0 equiv), using 5 equiv of PyAOP and 10 equiv of DIEA in DMF-DCM, 1 :9, v/v for 2 h at 25 ⁰ C. The Fmoc group was removed by treatment of piperidine-DMF, 1 :4, v/v, for 20 min at 25 ⁰ C. Couplings of 2-azido-3-[(9- fluorenylmethyloxycarbonyl)amino]propanoic acid (10) and Fmoc-Dpr-(Mtt)-OH (11) were carried out in each case, by treatment with 5 equiv of the aminoacid, 5 equiv of HATU and 10 equiv of collidine in DMF for 1 h at 25 ⁰ C to afford product 12. The removal of Mtt and PhiPr protections was carried out by treatment with a solution of TFA in DCM (1 :99, v/v, for 6 min at 25 ⁰ C), followed by immediate neutralization by washings with a mixture of Py in DCM (1 :5, v/v).

Cyclization of the peptide (13) was then performed using PyAOP as an activator (5 equiv of PyAOP, 5 equiv of DIEA in DMF for 2 h at 25 ⁰ C). After each coupling (including the cyclization step), potentially remaining free amino groups were capped by an acetic anhydride treatment.

Then, the resin was treated with TFA in DCM (1 :1, v/v, 30 min at 25 ⁰ C) to release the final product 14. d. Sequential addition of peptides to the scaffold

The three amino acid residues can be sequentially deprotected, reacted with sulfosuccinimidyl-4-(N-maleimidomethyl)cyclohexane- 1 -carboxylate (Sulfo-SMCC) or other heterobifunctional linker, and the corresponding peptide added. Thus, this scaffold allows incorporation of up to three same or different peptides as shown in FIG. 7A.2. Peptides are chosen based on screening of target on a random peptide microarray as described in preceding examples.

2. Cyclic decapeptide linker synlec.

The synthesis of a cyclic decapeptide scaffold from commercial Fmoc amino acids by solid phase synthesis, using Trt-Lys(Fmoc)OH as the N-terminal amino acid, and SASRIN resin as shown in FIG. 7B.1 is demonstrated. The cyclization of the decapeptide is carried out in high dilution. This decapeptide structure provides orthogonally protected conjugation sites enabling attachment of up to four distinct peptides or other affinity elements, and serves as a linker for the synlec. a. Synthesis of the decapeptide

H ₂ NLyS(FmOC)PrOGIyLyS(PNz)LyS(BoC)PrOGIy-LyS(AlOc)AIaOH (19)

Assembly of the protected peptide was carried out manually. Fmoc-Ala- SASRIN (0.5 g, 0.75 equiv/g) was washed and swollen with CH2C12 (2 x 10 mL x 15 min) and DMF (2 x 50 mL x 15 min). Coupling reactions were performed using, relative to the resin loading, 4 equiv of N-α-Fmoc-protected amino acid activated in situ with 4 equiv of PyBOP and 8 equiv of DIEA in 8 mL of DMF for 30 min. The completeness of each coupling was confirmed by Kaiser tests. N-α-Fmoc protecting groups were removed by treatment with piperidine:DMF 1 :4 (10 mL x 4 x 10 min), the completeness of each deprotection being verified by the UV absorption of the piperidine washings at 299 nm.

Peptide resin was treated repeatedly with TFAiCH ₂ Cl ₂ 1 :99 until the resin beads became dark purple (10 x 10 mL x 3 min). Each washing solution was neutralized with pyridine:MeOH 1 :4 (5 mL). The combined washings were concentrated under reduced pressure, and white solid was obtained by precipitation from EtO Ac/petroleum ether. This solid was dissolved in EtOAc, and pyridinium salts were extracted with water. The organic layer was dried over Na ₂ SO ₄ , filtered, and concentrated to dryness. Precipitation from CH ₂ Cl ₂ /Et ₂ 0 afford white solid which was further desalted by solid-phase extraction and lyophilized to afford the linear peptide. This material was used in the next step without further purification.

Peptide resin was treated repeatedly with TFA/CH2C12 1 :99 until the resin beads became dark purple (IO x 10 mL x 3 min). Each washing solution was neutralized with pyridine: MeOH 1 :4 (5 mL). The combined washings were concentrated under reduced pressure, and white solid was obtained by precipitation from EtO Ac/petroleum ether. This solid was dissolved in EtOAc, and pyridinium salts were extracted with water. The organic layer was dried over Na2SO4, filtered, and concentrated to dryness. Precipitation from CH2C12/Et20 afford white solid which was further desalted by solid-phase extraction and lyophilized to afford the linear peptide. This material was used in the next step without further purification. b. Cyclization in solution (20)

The above linear peptide was dissolved in DMF (100 mL), and the pH was adjusted to 8-9 by addition of DIEA. HATU (1.1 equiv) was added, and the solution was stirred at room temperature for 3 h. Solvent was removed in a vacuum; the residue was dissolved in TFAiCH ₂ Cl ₂ 1:1 (15 mL) and allowed to stand for 45 min at room temperature. The solution was then concentrated under reduced pressure and the residue was triturated with Et ₂ O and filtered to yield crude scaffold shown in FIG. 7B.1. The scaffold can be functionalized in order to attach it to different surfaces, or to add a dye that will help in the studies. c. Addition of linker

The scaffold can be functionalized in order to attach it to different surfaces, or to add a dye that will help in the studies. Thus, the linker in can be engineered to have a thiol (SH) group at a terminal position. The functionalization takes place at the free NH2 group as shown in compound 21 in FIG. 7B.2. As an example, this amino group can be acylated using tert-butylthio protected thioglycolic acid or an acetylated thiol. At this point, the scaffold is ready for sequential addition of peptides of interest. d. Sequential addition of peptides to the scaffold

The four lysine residues can be orthogonally (without affecting each other) deprotected, reacted with sulfosuccinimidyl-4-(N-maleimidomethyl)cyclohexane- 1 - carboxylate (Sulfo-SMCC) or other similar heterobifunctional linker, and the corresponding NH2 -protected peptide added. Thus, this scaffold allows incorporation of up to four different peptides as shown in FIG. 7B.2.

3. Forming Multimeric Synlecs

The linker in Scheme 4 can be engineered to have a thiol (SH) group at a terminal position. This thiol can be oxidized to yield a dimer of the structure 24 as shown in FIG. 7B.3. The thiol can also be used to attach structure 24 to various other scaffolds and surfaces.

All of the compositions and/or methods disclosed and claimed herein can be made and executed without undue experimentation in light of the present disclosure. While the compositions and methods of this invention have been described in terms of some embodiments, it will be apparent to those of skill in the art that variations may be applied to the compositions and methods and in the steps or in the sequence of steps of the method described herein without departing from the concept, spirit and scope of the invention. More specifically, it will be apparent that certain agents which are both chemically and physiologically related may be substituted for the agents described herein while the same or similar results would be achieved. All such similar substitutes and modifications apparent to those skilled in the art are deemed to be within the spirit, scope and concept of the invention as defined by the appended claims.

REFERENCES

The following references, to the extent that they provide exemplary procedural or other details supplementary to those set forth herein, are specifically incorporated herein by reference.

Alvarez et al, Gfycobiology, 15:1207, 2005.

Anderson and Chan. Acs Nano, 2(7): 1341-1352, 2008.

Balzarini, Nature Rev. Microbiol, 5(8):583-597. 2007.

Campbell and Yarema, Genome Biol, 6, 2005..

Caroff and Karibian, Carbohydrate Res., 33&(23):2431-2447, 2003.

Caroff and Karibian. Carbohydrate Res., 338(23):2431-2447, 2003.

Chan et al, Biochimica et Biophysica Acta-Biomembranes, 1758(9): 1184-1202, 2006.

David, J. Molec. Recog., 14(6):370-387, 2001. de Haas et al., J. Immunol. Methods, 242(l-2):79-89, 2000. de Ia Fuente and Penades, Biochimica et Biophysica Acta, 1760:636-651, 2006.

Ding et al, Cell. Molec. Life Set, 65(7-8): 1202-1219, 2008.

Drickamer, Structure, 5(4) :465-468, 1997.

Dubertret et al, Science, 298(5599): 1759-1762, 2002.

Gemeiner et al, Biotechnol Advaces, 2008 (epub ahead of print).

Gupta et al, Infect. Immunity, 63(8):2805-2810, 1995.

Hancock and Sahl, Nature Biotechnol, 24(12): 1551-1557, 2006.

Hsu and Mahal, Nature Protocols, l(2):543-549, 2006.

Kim et al, Biotechnol. Lett., 28(2):79-84, 2006.

Koch, In: Methods for General and Molecular Bacteriology, Gerhardt (Eds.), Washington

DC, Amer. Soci. Microbiol, 248-277, 1994. Lata et al, BMC Bioinformatics, 8:263, 2007. Lundquist and Toone, Chem. Rev., 102(2):555-578, 2002. Mancek et al, Biochem. Biophysical Res. Commun., 292(4):880-885, 2002. Manimala et al, Chembiochem., 6(12):2229-2241, 2005. Mellet and Fernandez, Chembiochem., 3:819, 2002. Murrell et al, Chembiochem., 5(10):1334-1347, 2004. Murrell et al, Chembiochem., 5:1334-1347, 2004. Nahtman et al, J. Immunol. Methods, 328(1-2): 1-13, 2007. Paulson et al, Nat. Chem. Biol, 2:238-248, 2006. Paulson et al, Nature Chemical Biology, 2(5):238-248, 2006.

Pilobello and Mahal, Curr. Opin. Chem. Biol, 300-305, 2007. Pilobello et al. Proc. Natl. Acad. ScL USA, 104(28):l 1534-11539, 2007.

Pilobello et al, Chembiochem., 6(6):985-989, 2005.

Resch-Genger et ah, Nature Methodsm 5(9) :763-775, 2008.

Rodi et al, J. Molec. Biol, 322(5): 1039-1052, 2002.

Rosenfeld et al, Biochemistry, 47(24):6468-6478, 2008.

Rosenfeld et al, J. Biochem. Biophys. Methods, 70:415-426, 2007.

Royce et al , Dna Microarrays, Part B: Databases and Statistics, 411 :282-311 , 2006.

Samuel et al, J. Bacteriol, 186(19):6536-6543, 2004.

Smibert and Keig, In: Methods for General and Molecular Bacteriology, Gerhardt (Eds.),

Washington DC, Amer. Soci. Microbiol, 607-654, 1994. Sujatha and Balaji, Proteins-Structure Func. Bioinform., 55(l):44-65, 2004. Svarovsky and Barchi, In: De novo synthesis of biofunctional carbohydrate -encapsulated quantum dots, 375-392, ACS Symposium Series, Amer. Chem. Soc, Washington

DC, 2007.

Svarovsky and Joshi, Biocomb. SeI Carboh. Binding Agents Therap. Signif, 20-28, 2008. Triantafϊlou et al, Cytometry, 41(4):316-320, 2000. Turnbull and Field. Nature Chemical Biology, 3(2):74-77, 2007. Wang and Wang. Nucleic Acids Res., 32:D590-D592, 2004. Wood et al, Combin. Chem. High Throughput Screen., 7(3):239-249, 2004. Xie et al, J. Peptide ScL, 12(10):643-652, 2006. Zheng et al, J. Amer. Chem. Soc, 127(28):9982-9983, 2005. Zheng et al, Langmuir, 20(10):4226-4235, 2004.