Synthesis and Screening of a Random Dimeric Peptide Library Using

Bioconjugate Chem. 2006, 17, 335−340

335

Synthesis and Screening of a Random Dimeric Peptide Library Using the One-Bead-One-Dimer Combinatorial Approach Saurabh Aggarwal,† James L. Harden,† and Samuel R. Denmeade†,§,* Department of Chemical and Biomolecular Engineering, The Johns Hopkins University Whiting School of Engineering, and The Sidney Kimmel Comprehensive Cancer Center, The Johns Hopkins University School of Medicine (SRD), Baltimore, Maryland 21231. Received September 1, 2005; Revised Manuscript Received October 24, 2005

Large combinatorial libraries of random peptides have been used for a variety of applications that include analysis of protein-protein interactions, epitope mapping, and drug targeting. The major obstacle in screening these libraries is the loss of specific but low affinity binding peptides during washing steps. Loss of these specific binders often results in isolation of peptides that bind nonspecifically to components used in the selection process. Previously, it has been demonstrated that dimerizing or multimerizing a peptide can remarkably improve its binding kinetics by 10- to 1000-fold due to an avidity effect. To take advantage of this observation, we constructed a random library of 12 amino acid dimeric peptides on polyethylene glycol acrylamide (PEGA) beads by modifying the ‘one-bead-one-compound’ approach. The chemical synthesis of 100 000 peptides as dimers can be problematic due to steric and aggregation effects and the presence of many peptide sequences that are difficult to synthesize. We have designed a method, described in detail here, to minimize the problems inherent in the synthesis of a dimeric library by modifying the existing ‘split and pool’ synthetic method. Using this approach the dimeric library was used to isolate a series of peptides that bound selectively to epithelial cancer cells. One peptide with the amino acid sequence QMARIPKRLARH bound as a dimer to prostate cancer cells spiked into the blood but did not bind to circulating hematopoeitic cells. The monomeric form of this peptide, however, did not bind well to the same LNCaP cell line. These data demonstrate that “hits” obtained from such a ‘one-bead-one-dimer’ library can be used directly for the final application or used as leads for construction of second generation libraries.

INTRODUCTION Peptides are versatile small molecules that are increasingly being used alone or as part of bioconjugates in the therapy of a variety of human diseases, including cancer, diabetes, and osteoporosis. For example, peptide analogues of the luteinizing hormone-releasing hormone (LHRH) are the mainstay of treatment for metastatic prostate cancer. Peptides derived from fragments of human proteins such as the Arg-Gly-Asp (RGD)based motifs from extracellular matrix (ECM)(1), endostatin from collagen, and angiostatin from plasminogen have been tested as anticancer therapies (2-4). Targeted peptides holds special promise for cancer, because many of the antibody-based targeted drugs have been shown to have difficulty permeating into tumors (5, 6). Because of their size, small peptides can be more effective in delivering a therapeutic dose deep into tumors. On this basis, peptide-cytotoxin conjugates are under development that selectively target proteases and receptors within sites of cancer (7-9). Peptides are also under study as targeted imaging agents (10). As an additional application, our laboratory is interested in using peptides to develop indwelling probes or ex vivo blood filtration systems that can continuously extract circulating cancer cells from blood over an appropriate diagnostic period (11). As an alternative approach to antibody-based cell binding, we are instead identifying small peptides that bind selectively to epithelial cells. Small peptides of 12 amino acids are easily synthesized and are not highly immunogenic. Compared to antibodies, these peptides are relatively inexpensive, stable, and easily coupled to a probe surface. * To whom all correspondence should be sent: Bunting Blaustein Cancer Research Building, 1650 Orleans St., Baltimore, MD 21231. Phone: 410.502.3941. Fax: 410.614.9397. E-mail: [email protected]. † The Johns Hopkins University Whiting School of Engineering. § The Johns Hopkins University School of Medicine.

Previously it has been demonstrated that dimerizing or multimerizing a peptide ligand can improve its binding kinetics 10-1000-fold due to an avidity effect (12-15). On this basis, using the one-bead one-peptide approach we synthesized a random dimeric peptide library using a C-terminal lysine residue as a branch point on poly(ethylene glycol) acrylamide (PEGA) beads (11). The production of a cell-binding surface coated with peptide ligand requires that the surface of the probe be relatively inert and not adsorb proteins and/or cells nonspecifically. Therefore, for this application, poly(ethylene glycol) (16) was selected as the screening surface because it has been demonstrated to be one of the most effective biomaterials for coating surfaces to prevent nonspecific adsorption (17-19). In this case, the synthesis of peptides on PEGA beads, which have a PEG outer surface, mimics potential catheter surface conditions from the outset in the selection of cell-binding peptides. In this previous study, no cell binding was observed on PEGA beads when peptides were synthesized as monomers (11). In contrast, screening of a dimeric peptide library containing approximately 100 000 peptides yielded a series of peptide ligands that bound selectively to human cancer cells in the blood, but did not bind to circulating leukocytes (11). The synthesis and screening of a novel dimeric library, to our knowledge, has not been described previously. The use of such a dimeric library in screening strategies enhances the isolation of cell and/or protein specific ligands for applications such as drug targeting, imaging, and biosensors. The goal of this current study is to provide a detailed description of the methods used to synthesize, screen, and sequence a large combinatorial dimeric peptide library with the intent that this would allow for more widespread application in the future of the ‘one-bead-one-compound’ combinatorial approach in the selection of peptide binding ligands.

10.1021/bc0502659 CCC: $33.50 © 2006 American Chemical Society Published on Web 02/08/2006

336 Bioconjugate Chem., Vol. 17, No. 2, 2006

EXPERIMENTAL PROCEDURES Materials. Poly(ethylene glycol) amide (PEGA) beads were from Polymer Labs (Amherst, MA). Fmoc-blocked amino acids, 1-hydroxybenzotriazole (HOBT), and 1,3-diisopropylcarbodiimide (DIC) were purchased from Advanced ChemTech, (Louisville, KY). N,N-Diisopropylethylamine (DIEPA), N-methyl-2-pyrrolidinone (NMP), dimethylformamide (DMF), and methylene chloride (DCM) were from Sigma. The human prostate cancer cell line LNCaP was obtained from ATCC (Manassas, VA). Cell Tracker Green and Orange were from Invitrogen (Rockville, MD). Unless otherwise specified, all other reagents were from Sigma (St. Louis, MO). Synthesis of AG-73 Peptide on PEGA Beads. Heparin binding peptide sequence from laminin AG73 (RKRLQVQLSIRT)(20) was synthesized on Fmoc-Gly-loaded PEGA beads using an automated peptide synthesizer (PS-3, Rainin, Tucson, AZ). Dimeric peptides were synthesized by precoupling FmocLys(Fmoc)-OH to Gly-PEGA followed by solid-phase synthesis of the remaining peptide. Preparation of PEGA Beads and Fmoc Amino Acids. PEGA beads with lower substitution level of 0.2 mmol/g were used to avoid any aggregation effect from higher loading. Five grams of PEGA beads [supplied beads are 10.7% w/w (i.e. 5 g ) 46.7 g of suspension)] were washed with DMF. For the first overnight coupling, 1.5 mmol of Fmoc-Lys(Fmoc)-OH (0.88 g), 1.5 mmol of HOBT (0.208 g), and 1.69 mmoles of DIC (261 µL), and 6 mL of 10% DEIPA were added to 10 mL of DMF and activated for 30 min. Activated complex was added to the beads and allowed to react overnight at room temperature. Completion of coupling was verified by Kaiser test with a small subset of beads (21). Repeated couplings with 2-fold molar excess of the above-mentioned reagents were performed. The second coupling was done for 2 h at room temperature. Completion of coupling was again verified by Kaiser test. Beads were washed 5-6 times with DCM and DMF and stored at 4 °C in DMF until later use. Fmoc was removed from preloaded Lys in the presence of 20% piperidine in NMP. Deprotection was repeated four times for 5 min each. After extensive washing, beads were split into 19 pools for coupling of all natural amino acids with the exception of cysteine. Pools were divided to have 50 µmol of total peptide in each reaction vessel. A combinatorial peptide synthesis platform from Advanced ChemTech was used for the split and pool couplings. This apparatus is designed specifically to generate split and mix peptide libraries. The library was always pooled in a polypropylene vessel. Concentrated stock solutions of all 19 Fmoc amino acids with HOBT were prepared and stored at -20 °C. Fmoc amino acids (0.5 M) with HOBT (0.49 M) stocks were prepared in DMF (except for Fmoc-phenylalanine which was prepared in NMP). Coupling of Fmoc Amino Acids To Generate 12-mer Library. Double coupling was done for all Fmoc amino acids. For the first coupling, 4× molar excess i.e., 200 µmol of Fmoc amino acid, and HOBT (400 µL of 0.5 M stock) and 50 µL of DIC were allowed to react for 45 min at room temp and then added to 19 reaction vessels. Beads were mixed for 1 min with the activated fmoc amino acids. DIEPA (0.5 mL of 10% solution) was added, and the reaction was allowed to continue for 2-4 h. After this coupling, the beads were washed extensively with NMP. Coupling was repeated for 30 min with 2× molar excess of the activated Fmoc amino acids. Beads were again washed with DMF. A final wash was done with methanol. A small number of beads from all 19 vessels were taken for the Kaiser test. Deprotection was done twice with 20% piperidine in NMP for 30 min each. Beads were washed extensively and later pooled into a beaker with methanol. Beads were stored at -20 °C till later use.

Aggarwal et al.

Before the removal of the final N-terminal Fmoc, the library was acetylated with a 1:3 dilution of acetic anhydride in 10% DIEPA in NMP. Acetylation was done for 30 min. Completion of reaction was confirmed by the Kaiser test. The final Fmoc was removed as mentioned above. After 8-9× volume wash with NMP and DCM the library was stored at -20 °C till later use. The library was washed with methanol and partially dried under an aspiration setup. Beads were suspended in 20 mL of Reagent K (TFA:H2O:thioanisole:ethanedithiol:phenol 85:5:5: 2.5:2.5). This Reagent K solution was aspirated after 10 min, and the beads were suspended in another 20 mL. Deprotection was done for 3 h with gentle shaking. Beads were washed 10 times with NMP and DCM and finally stored in methanol at -20 °C. Dimeric Library Screening for Cell Binding. An amount of 100 000 beads suspended in methanol were washed three times with water. Beads were washed and equilibrated in PBS and finally suspended in RPMI 1640 Media containing 10% fetal calf serum (Invitrogen, Rockville, MD). Beads were incubated with 10% serum containing media for 1 h at room temperature. LNCaP cells were removed from culture flasks with trypsin/EDTA, washed and re-suspended in RPMI 1640. Cells were incubated with Cell Tracker orange (10 µM for 30 min at 37 °C). To remove excess dye, cells were pelleted and resuspended in 10% serum containing media at a final density of 5 million cells/ml. The cell suspension was mixed with the beads and incubated at 37 °C with gentle shaking for 30-40 min. Beads were gently washed with Hanks Balanced Salt Solution (HBSS) and fixed for visualization with 80% Methanol in dH2O. After fixation, the bead pool was evaluated under a fluorescent microscope equipped with the appropriate filter set for either FITC or rhodamine. Isolation and Preparation of Beads for Sequencing. Single positive beads containing fluorescently tagged cells were identified using a fluorescent microscope under 2.5× objective magnification and removed manually with a 200 µL pipetter using pipet tips in which the last third of the tip had been removed with a razor blade. Selected beads were washed with water and stripped of cells with excess 8 M urea or 6 M guanidine HCl for 24-48 h with intense shaking. Urea or guanidine HCl was removed by 4-5 washes with water. A dissecting microscope was used to confirm that the beads were not lost during washing. Single beads suspended in water were sent for sequencing to University of Arizona Peptide Sequencing Core. Sequencing of Single Beads with Edman Degradation. Peptide sequencing was completed by Wallace Clark at the University of Arizona using an Applied Biosystem 477A Protein/Peptide sequencer (Edman chemistry) interfaced with a 120A HPLC (C-18 PTH column, reverse-phase chromatography) analyzer to determine phenylthiohydantoin (PTH) amino acids. One sample bead was analyzed per sequence by placing onto a glass fiber filter, which was then assembled within a quartz cylinder and sequenced.

RESULTS Dimerization Enhances Cell Binding. AG-73 (RKRLQVQLSIRT) is a heparin-binding peptide sequence derived from laminin that has previously been demonstrated to support cell attachment. The AG-73 peptide was synthesized on the PEGA beads as both a monomer and a dimer. Both types of beads were blocked with 10% serum containing media and then incubated with Cell Tracker Green labeled LNCaP cells and compared for the degree of cell binding. After 30 min incubation, excess cells were washed, and beads were fixed and visualized under a fluorescence microscope. No cells bound to

Synthesis of a Random Dimeric Peptide Library

Figure 1. LNCaP cell binding to monomeric and dimeric peptides on PEGA beads. (a) Binding of Cell Tracker Green loaded cells to monomeric AG-73 (10× objective); (b) binding to dimeric AG-73 (10× objective); (c) binding of Cell Tracker Orange labeled cells to monomeric QMARIPKRLARH (20× objective); (d) dimeric QMARIPKRLARH peptide (20× objective).

monomeric AG-73-containing beads, (Figure 1a). In contrast, LNCaP cells bound readily to the dimeric form of AG-73 (Figure 1b). These results support previous findings that dimeric peptides have enhanced cell binding capabilities due to an avidity effect. Substitution Level of PEGA Beads with Fmoc-Lys (Fmoc)OH Affects Synthesis. While these results with the AG-73 peptide demonstrated the superior binding characteristics of the dimeric compared to the monomeric peptide, the AG-73 peptide was initially selected solely for its ability to bind to a target antigen. Our goal in these studies was to identify peptides that can bind selectively to circulating epithelial cancer cells while not binding to highly abundant circulating leukocytes. Therefore, we opted to use the combinatorial-based peptide library approach to identify peptides meeting these two characteristics. As proof of concept, a strategy was developed to synthesize a library of dimeric peptides in which each “arm” of the peptide consisted of 12 positions of randomized to one of 19 amino acids (i.e. cysteine omitted) connected by a lysine linker. On this basis a library of ∼105 peptides out of a total of 1912 (i.e. 2.2 × 1015) possible members was generated and screened for selective epithelial cancer cell binding. The steps used to synthesize this dimeric library are outlined in Figure 2. The synthesis of dimeric library of 12-mer peptides can be limiting for a fraction of a library because of steric hindrance of bulky side functional groups and bulky protection agents. Standard substitution levels of 0.4-0.5 mmol/g used in synthesis of a test dimeric peptides (e.g. Wang resin) markedly decreased peptide yield, Figure 3a. To minimize such a steric effect, PEGA beads with a lower substitution level of 0.2 mmol/ ml were utilized. To produce a branched peptide, Fmoc-Lys-

Bioconjugate Chem., Vol. 17, No. 2, 2006 337

Fmoc-OH was coupled to the low substitution PEGA. Coupling was confirmed to be positive by a Kaiser test before proceeding with synthesis of the library using the two amino groups of deprotected lysine. The use of the lower substitution PEGA beads enhanced the yield of the test peptide and minimized incomplete coupling observed with the higher substitution PEGA, Figure 3b. Quality Control of Library Synthesis with the Kaiser Test. The large scale manual nature of this procedure makes the syntheses liable to error, and hence it proved critical to check all 19 reaction vessels for completion of coupling (21). Although the Kaiser test is a qualitative test and will not be able to detect rare and partial incomplete couplings, it was, in general useful to rapidly confirm overall coupling efficiency. Subsequent coupling of amino acids were done twice and were checked for completion by the Kaiser test. In some cases, the couplings were repeated till a negative Kaiser test was confirmed. Final Capping of Library before Removal of Terminal Fmoc. The sequencing of individual selected beads using Edman degradation depends on the completion of coupling after each round. In such a combinatorial synthesis of a large number of peptides there are some peptide sequences that contain difficult coupling steps that can lead to some percentage of partially synthesized peptide on an individual bead. These incomplete peptides can complicate interpretation of sequencing results and can make it difficult to determine the full sequence of the actual peptide, Figure 4a. Acetylating the amino terminal amine of any incomplete peptides using acetic anhydride can help to resolve this problem. For best results the capping should be done after each coupling. However, we determined that capping the terminal amine was sufficient and gave enough signal upon sequencing to identify the peptide sequence, Figure 4b. Capping was accomplished by acetylating the library with 1: 3 dilution of acetic anhydride in 10% DIEPA in NMP for 30 min prior to the removal of final N-terminal Fmoc of the full length peptide. Screening the Library for Binding to Cells. The dimeric library on low substitution PEGA beads was blocked and incubated with LNCaP prostate cancer cells preloaded with Cell Tracker Orange. Cell Tracker Orange was used instead of the Cell Tracker Green used to screen the AG-73 peptide because the Orange dye provided better signal-to-noise and markedly less autofluorescence. This made the few positive beads in a field of thousands of negative beads easier to rapidly visualize and select. An amount of 100 000 beads was screened and, of these, approximately 100 cell-coated beads (i.e. “hits”) were observed. The remaining 99.9% of the beads were totally negative for cell binding. The 100 positive beads were difficult to grade because each of them was highly coated with cells. Without a bead sorter these beads were picked manually based on the relative number of bound cells under a fluorescence microscope. On this basis, 10 beads were picked and subsequently sequenced by Edman degradation. The amount of peptide that can be obtained from a single PEGA bead with 0.2 mmol/mL substitution on average is ∼40 pmol. This limits the number of analytical techniques which can be used for determining the sequence.

Figure 2. Schematic representation of random peptide library synthesized by ‘split and mix’ method where each bead contains multiple copies of a unique 12 amino acid peptide. One of 19 defined amino acids (i.e. cystein omitted) was added during each step of synthesis followed by mixing and then splitting into 19 separate flasks prior to addition of next set of defined amino acids. The scheme illustrates the utility of capping incompletely synthesized peptides with acetic anhydride prior to the removal of the final Fmoc protecting group on the N-terminal amine.

338 Bioconjugate Chem., Vol. 17, No. 2, 2006

Aggarwal et al.

selected peptide sequences bound only malignant epithelial cancer cells but not leukocytes.

DISCUSSION

Figure 3. Reversed phase HPLC C18 chromatograms of dimeric peptide (QMARIPKRLARH)2KG synthesized on glycine Wang resin. (a) 0.4 mmol/g substitution level yields broad multiple peaks indicating mixture of peptides (b) 0.2 mmole/g substitution level yields narrow discrete peak consistent with high yield of single peptide.

Selected peptide sequences were synthesized again as either a monomer or a dimer on PEGA beads and rescreened for binding to different cell lines. LNCaP cells did not bind to resynthesized monomeric peptide beads to any significant degree, Figure 1c. In contrast, all of the dimeric versions of the peptides selected from the screen bound the LNCaP cells, Figure 1d. Additional analysis demonstrated that a subset of these

Peptides that bind selectively to a specific cell type or protein have been identified from large combinatorial libraries of peptides. Such libraries can be either biologically synthesized as phage display libraries (∼109 different peptides) (22) or they can be chemically synthesized as the ‘one-bead-one-peptide′ libraries (∼105-106 different peptides) (23). Phage display libraries are widely used and have been employed to successfully isolate peptide ligands that bind to purified protein targets such as erythropoietin (EPO)(12), thrombopoetin (TPO) (15), and fibroblast growth factor (FGF) (24). However, the selection of cell-binding peptides from phage display libraries screened on whole cells has been problematic (25, 26). A phage display library has the limitation of displaying only five copies of monomeric peptides on the pIII coat protein per individual phage. Initial screens of phage libraries typically yield peptides with low binding affinities and rapid off rates (on the order of few minutes). The fast dissociation of these low-affinity, but cell-type specific, binding peptides often results in isolation of more nonspecific than specific ones binding phage (26). To address the problem of low affinities, Zwick et al. (27) developed a method to make a phage display library of homodimeric peptides. However, as described by Zwick et al., the construction of this homodimeric phage library was complicated and would not be applicable for widespread use. Synthetic peptide libraries have a number of unique attributes when compared to phage libraries. Such libraries allow for the incorporation of D-amino acids as well as unnatural amino acids which is not possible using phage display. In contrast to phage libraries, thousands to millions of copies of a unique peptide sequence can be displayed on the surface of a synthetic bead.

Figure 4. Sample chromatograms for three positions of Edman sequencing of a dimeric peptide on 0.2 mmol/g PEGA beads following (a) synthesis without final capping and (b) synthesis with a final capping. Without capping, sequencing reveals multiple amino acid peaks for each position, whereas with capping, distinct single peaks are observed.

Synthesis of a Random Dimeric Peptide Library

Thus, bead-based libraries seem particularly suited for cell capture applications because the beads can provide multiple binding sites across the cell surface. However, since the first report describing this combinatorial method in 1991, there have been few reports on the use of such libraries for selection of cell binding ligands. The technical expertise and sophistication required in implementing the protocol for synthesis and screening such libraries, particularly those with cyclic peptides, appears to have limited the application of this screening method to only a few laboratories. In addition, like phage display, screening of bead-based libraries often yields low affinity peptides or nonspecific binders. Proteins for which peptides are not the natural ligands typically display low affinity interactions (28), and this leads to fast dissociation kinetics with koff of the order of few minutes (26). Many of these specific but low affinity binders are lost during multiple washing steps involved in the screening process. A number of studies have now demonstrated that dimerizing or multimerizing a peptide ligand can improve its binding kinetics due to an avidity effect. For example, Wrighton et al. had demonstrated that affinity for peptides binding to the erythropoietin receptor could be increased 100-fold when peptides were synthesized in a dimeric form (12). Previously a dimeric version of an RGD integrin-binding peptide was shown to have enhanced binding characteristics and nanomolar affinity. Dimeric RGD peptide has also been used by many groups for targeting drugs and for imaging applications (29). Previously, Terskikh et al. demonstrated that the affinity of a peptide isolated from phage display screening could be increased 200 000-fold by making a pentamer (i.e. “peptabody”) using the pentamerization domain of a cartilage oligomeric protein (13). In our own studies we have demonstrated that incubation of biotinylated peptides with avidin forms a tetrameric species with enhanced binding characteristics compared to monomeric peptide (11). To demonstrate the utility of screening a dimeric peptide library, a strategy was developed to synthesize a library of dimeric peptides in which each “arm” of the peptide consisted of 12 positions of randomized to one of 19 amino acids (i.e. cysteine omitted) connected by a lysine linker. On this basis a library of ∼105 peptides out of a total of 1912 (i.e. 2.2 × 1015) possible members was generated and screened for selective epithelial cancer cell binding. This library represents only a small fraction of potential peptide sequences. However, the use of the dimeric library can increase the chance of obtaining “hits” in the first round of screening. These “hits” can then later be optimized by synthesizing a second random library in which different positions in the original dimeric “hit” can be modified to evaluate the effect of modification on overall binding. This second round of synthesis and screening, based on the QMARIPKRLARH peptide, is currently underway in our laboratory The solid-phase “split and mix” method used to synthesize this dimeric library was similar in many ways to previously published methods for synthesizing monomeric libraries. However, to optimize both the synthesis and identification of individual dimeric peptides in the library, a few modifications were made to the standard methodologies. First, synthesis of a test dimeric peptide (QMARIPKRLARH)2KG on standard substitution resins (0.4-0.5 mmol/g) yielded a mixture of peptides on an individual bead. This incomplete coupling of amino acids was due, most likely, to intermolecular steric hindrance of bulky side functional groups and bulky protection agents. Synthesis of the test peptide on resin with a lower substitution level of 0.2 mmol/g minimized such steric effects and resulted in higher yields of the dimeric test peptide. Additionally, with the use of lower substitution resin in the production of the dimeric library, the majority (i.e. >90%) of

Bioconjugate Chem., Vol. 17, No. 2, 2006 339

the selected beads from the screen yielded a complete 12 amino acid peptide sequence. However, even with this lower substitution resin, some of the selected dimeric peptides were difficult to resynthesize in large quantities in high yield again due to sequence-dependent intramolecular steric hindrance. A second important modification to the standard screening methodology was the use of capping by acetylation of the amino terminus of incompletely synthesized peptides. This capping step is not always performed in the standard synthesis of monomeric libraries. In this case, however, the more complex nature of the dimeric peptide results in a higher percentage of incomplete couplings on an individual resin bead. Capping of incomplete peptides before removal of the final Fmoc group proved critical to the elucidation of the peptide sequence. Without capping, mixtures of amino acids were frequently observed after each round of Edman degradation. In this study we provide further evidence demonstrating enhanced binding characteristics of dimeric peptides compared to monomers. However, unlike other reports in which the monomeric peptide sequence was first selected from a random pool and then dimerized, in this study we have directly screened a random dimeric peptide library. The use of this dimeric library improves chances of selecting binding peptides that might otherwise not been identified from a monomeric library screen. In fact, in each case, the monomeric version of the dimeric peptide selected in this screen did not bind to cells to any significant degree. Studies are underway in our laboratory to identify the cognate protein receptor for the QMARIPKRLARH)2KG peptide. The fact that the dimeric peptide binds better than the monomeric version suggests that the cognate receptor itself may be a dimeric protein. In this regard, Wrighton et al. synthesized a dimeric version of a 20 amino acid peptide using a single lysine as a branch point and demonstrated that this dimeric peptide had 100-fold increased affinity for the erythropoietin receptor compared to the monomeric peptide. The authors of this study concluded that the proposed mechanism of increased potency of the dimer over the monomer was that the dimeric peptide was able to span the cleft and interact with both molecules of the EPOR extracellular domain. This conclusion was based on the structure of a cocrystal of the EPOR binding peptide and the extracellular domain of the EPOR in which a noncovalent peptide dimer is seen spanning the cleft between two molecules of the EPOR extracellular domain (30). On the basis of these results, therefore, we conclude that synthetic dimeric peptide libraries should be considered in the future for a variety of screening applications. We hope that the detailed methodologies described here can serve as a reference to aid other investigators interested in using this powerful combinatorial approach in peptide screening applications.

ACKNOWLEDGMENT The authors acknowledge Jacqueline Robinson for excellent assistance in the preparation of the manuscript. They also acknowledge Mr. Wallace Clark from the University of Arizona Peptide Sequencing Facility for expertise in sequencing peptides. This work was supported by funding from Alexander and Margaret Stewart Trust to S.R.D. and partial support from the NASA Human Exploration and Development of Space program NAG 9-1345 to J.L.H.

LITERATURE CITED (1) Aumailley, M., Gurrath, M., Muller, G., Calvete, J., Timpl, R., and Kessler, H. (1991) Arg-Gly-Asp constrained within cyclic pentapeptides. Strong and selective inhibitors of cell adhesion to vitronectin and laminin fragment P1. FEBS Lett. 291, 50-4.

340 Bioconjugate Chem., Vol. 17, No. 2, 2006 (2) O’Reilly, M. S., Boehm, T., Shing, Y., Fukai, N., Vasios, G., Lane, W. S., Flynn, E., Birkhead, J. R., Olsen, B. R., and Folkman, J. (1997) Endostatin: an endogenous inhibitor of angiogenesis and tumor growth. Cell 88, 277-85. (3) Folkman, J. (1996) Fighting cancer by attacking its blood supply. Sci. Am. 275, 150-4. (4) Mendoza, F. J., Espino, P. S., Cann, K. L., Bristow, N., McCrea, K., and Los, M. (2005) Antitumor chemotherapy utilizing peptidebased approachessapoptotic pathways, kinases, and proteasome as targets. Arch. Immunol. Ther Exp (Warsz) 53, 47-60. (5) Christiansen, J., and Rajasekaran, A. K. (2004) Biological impediments to monoclonal antibody-based cancer immunotherapy. Mol. Cancer Ther. 3, 1493-501. (6) Aina, O. H., Sroka, T. C., Chen, M. L., and Lam, K. S. (2002) Therapeutic cancer targeting peptides. Biopolymers 66, 184-99. (7) Denmeade, S. R., Nagy, A., Gao, J., Lilja, H., Schally, A. V., and Isaacs, J. T. (1998) Enzymatic activation of a doxorubicin-peptide prodrug by prostate-specific antigen. Cancer Res. 58, 2537-40. (8) Denmeade, S. R., Jakobsen, C. M., Janssen, S., Khan, S. R., Garrett, E. S., Lilja, H., Christensen, S. B., and Isaacs, J. T. (2003) Prostatespecific antigen-activated thapsigargin prodrug as targeted therapy for prostate cancer. J. Natl. Cancer Inst. 95, 990-1000. (9) Schally, A. V., and Nagy, A. (2003) New approaches to treatment of various cancers based on cytotoxic analogues of LHRH, somatostatin and bombesin. Life Sci. 72, 2305-20. (10) Chen, X., Conti, P. S., and Moats, R. A. (2004) In vivo nearinfrared fluorescence imaging of integrin alphavbeta3 in brain tumor xenografts. Cancer Res. 64, 8009-14. (11) Aggarwal, S., Janssen, S., Wadkins, R. M., Harden, J. L., and Denmeade, S. R. (2005) A combinatorial approach to the selective capture of circulating malignant epithelial cells by peptide ligands. Biomaterials 26, 6077-86. (12) Wrighton, N. C., Farrell, F. X., Chang, R., Kashyap, A. K., Barbone, F. P., Mulcahy, L. S., Johnson, D. L., Barrett, R. W., Jolliffe, L. K., and Dower, W. J. (1996) Small peptides as potent mimetics of the protein hormone erythropoietin. Science 273, 45864. (13) Terskikh, A. V., Le Doussal, J. M., Crameri, R., Fisch, I., Mach, J. P., and Kajava, A. V. (1997) “Peptabody”: a new type of high avidity binding protein. Proc. Natl. Acad. Sci. U.S.A. 94, 1663-8. (14) Janssen, M., Oyen, W. J., Massuger, L. F., Frielink, C., Dijkgraaf, I., Edwards, D. S., Radjopadhye, M., Corstens, F. H., and Boerman, O. C. (2002) Comparison of a monomeric and dimeric radiolabeled RGD-peptide for tumor targeting. Cancer Biother. Radiopharm. 17, 641-6. (15) Cwirla, S. E., Balasubramanian, P., Duffin, D. J., Wagstrom, C. R., Gates, C. M., Singer, S. C., Davis, A. M., Tansik, R. L., Mattheakis, L. C., Boytos, C. M., Schatz, P. J., Baccanari, D. P., Wrighton, N. C., Barrett, R. W., and Dower, W. J. (1997) Peptide agonist of the thrombopoietin receptor as potent as the natural cytokine. Science 276, 1696-9.

Aggarwal et al. (16) Pegram, M. D., Pietras, R., Bajamonde, A., Klein, P., and Fyfe, G. (2005) Targeted therapy: wave of the future. J. Clin. Oncol. 23, 1776-81. (17) Harris, J. M., and Chess, R. B. (2003) Effect of pegylation on pharmaceuticals. Nat. ReV. Drug DiscoVery 2, 214-21. (18) Ratner, B. D., and Bryant, S. J. (2004) Biomaterials: where we have been and where we are going. Annu. ReV. Biomed. Eng. 6, 41-75. (19) Bures, P., Huang, Y., Oral, E., and Peppas, N. A. (2001) Surface modifications and molecular imprinting of polymers in medical and pharmaceutical applications. J Controlled Release 72, 25-33. (20) Nomizu, M., Kim, W. H., Yamamura, K., Utani, A., Song, S. Y., Otaka, A., Roller, P. P., Kleinman, H. K., and Yamada, Y. (1995) Identification of cell binding sites in the laminin alpha 1 chain carboxyl-terminal globular domain by systematic screening of synthetic peptides. J. Biol. Chem. 270, 20583-90. (21) Kaiser, E., Colescott, R. L., Bossinger, C. D., and Cook, P. I. (1970) Color test for detection of free terminal amino groups in the solid-phase synthesis of peptides. Anal.Biochem. 34, 595-8. (22) Smith, G. P., and Petrenko, V. A. (1997) Phage Display. Chem ReV 97, 391-410. (23) Lam, K. S., Salmon, S. E., Hersh, E. M., Hruby, V. J., Kazmierski, W. M., and Knapp, R. J. (1991) A new type of synthetic peptide library for identifying ligand-binding activity. Nature 354, 82-4. (24) Ballinger, M. D., Shyamala, V., Forrest, L. D., Deuter-Reinhard, M., Doyle, L. V., Wang, J. X., Panganiban-Lustan, L., Stratton, J. R., Apell, G., Winter, J. A., Doyle, M. V., Rosenberg, S., and Kavanaugh, W. M. (1999) Semirational design of a potent, artificial agonist of fibroblast growth factor receptors. Nat. Biotechnol. 17, 1199-204. (25) Mori, T. (2004) Cancer-specific ligands identified from screening of peptide-display libraries. Curr. Pharm. Des. 10, 2335-43. (26) Giordano, R. J., Cardo-Vila, M., Lahdenranta, J., Pasqualini, R., and Arap, W. (2001) Biopanning and rapid analysis of selective interactive ligands. Nat. Med. 7, 1249-53. (27) Zwick, M. B., Shen, J., and Scott, J. K. (2000) Homodimeric peptides displayed by the major coat protein of filamentous phage. J. Mol. Biol. 300, 307-20. (28) Menendez, A., and Scott, J. K. (2005) The nature of targetunrelated peptides recovered in the screening of phage-displayed random peptide libraries with antibodies. Anal.Biochem 336, 14557. (29) Chen, X., Sievers, E., Hou, Y., Park, R., Tohme, M., Bart, R., Bremner, R., Bading, J. R., and Conti, P. S. (2005) Integrin alpha v beta 3-targeted imaging of lung cancer. Neoplasia 7, 271-9. (30) Livnah O, Stura EA, Johnson DL, Middleton SA, Mulcahy LS, Wrighton NC, Dower WJ, Jolliffe LK, Wilson IA. (1996) Functional mimicry of a protein hormone by a peptide agonist: the EPO receptor complex at 2.8 A. Science 273, 464-71. BC0502659