key: cord-007755-o2r8ktie authors: Kokoszka, Malgorzata E.; Kay, Brian K. title: Mapping Protein–Protein Interactions with Phage-Displayed Combinatorial Peptide Libraries and Alanine Scanning date: 2014-10-20 journal: Peptide Libraries DOI: 10.1007/978-1-4939-2020-4_12 sha: doc_id: 7755 cord_uid: o2r8ktie One avenue for inferring the function of a protein is to learn what proteins it may bind to in the cell. Among the various methodologies, one way for doing so is to affinity select peptide ligands from a phage-displayed combinatorial peptide library and then to examine if the proteins that carry such peptide sequences interact with the target protein in the cell. With the protocols described in this chapter, a laboratory with skills in microbiology, molecular biology, and protein biochemistry can readily identify peptides in the library that bind selectively, and with micromolar affinity, to a given target protein on the time scale of 2 months. To illustrate this approach, we use a library of bacteriophage M13 particles, which display 12-mer combinatorial peptides, to affinity select different peptide ligands for two different targets, the SH3 domain of the human Lyn protein tyrosine kinase and a segment of the yeast serine/threonine protein kinase Cbk1. The binding properties of the selected peptide ligands are then dissected by sequence alignment, Kunkel mutagenesis, and alanine scanning. Finally, the peptide ligands can be used to predict cellular interacting proteins and serve as the starting point for drug discovery. Very often in research projects, there is interest in mapping the protein-protein interactions of a protein of interest as a way of understanding its function in the cell or virus. While a variety of techniques exist for this purpose, such as yeast two-hybrid screening, mass spectrometry, and protein complementation assays, another approach is the use of phage-displayed combinatorial peptide libraries. In such an approach, one takes a purifi ed, recombinant protein and isolates peptide ligands to it through affi nity selection (Fig. 1 ) . Interestingly, the phage-displayed peptides bind at "hot spots" for molecular interaction and very often share the same primary structure as short regions within cellular interacting proteins. Several examples where this approach has proven useful include such targets as protein interacting domains [ 1 , 2 ] . To demonstrate the utility of this approach, we describe its application to two targets, human protein tyrosine kinase, Lyn, and the yeast serine/threonine-protein kinase Cbk1. Lyn was fi rst discovered as a viral oncogene [ 3 ] and later appreciated to be a proto-oncogene in humans. It is a non-receptor protein tyrosine Fig. 1 A general workfl ow diagram for isolating and characterizing the peptide ligands to protein domains or fragments using phage display methods. In principle, coding regions of any protein domain of interest can be converted into recombinant DNA and expressed as a fusion protein to a partner such as glutathione S-transferase (GST). With a soluble protein in hand, one can perform affi nity selections with phage-displayed combinatorial peptide libraries to identify its peptide ligands. The binding properties of those ligands can be then further characterized through affi nity and specifi city measurements. To assess the importance of each residue in the peptide ligand, a mutagenic analysis known as alanine scanning can be performed on the recombinant DNA. With that knowledge, potential interacting partners of the protein of interest can be predicted [ 14 ] . Finally, improving affi nity and specifi city of selected ligands through directed evolution may lead to the development of antagonists, which could be used as inhibitors of specifi c cellular interaction for proof-of-principle experiments of drug development kinase, which is a member of the Src family of proteins. It has a modular architecture: a Src homology 3 (SH3) domain, a Src homology 2 (SH2) domain, several linker regions, and a catalytic domain that phosphorylate tyrosines in proteins [ 4 ] . The SH3 domain plays a role in mediating protein-protein interactions and has been described to bind proline-rich motifs in proteins. The Cbk1 belongs to a large family of NDR/LATS protein kinases, which is conserved across eukaryotes and includes members such as myotonic dystrophy kinase [ 5 ] . Cbk1 plays a role in controlling cell separation after cytokinesis, cell integrity, and polarized growth in Saccharomyces cerevisiae [ 6 ] . To date, only a few substrates of Cbk1 have been reported. One of them, Ace2, a transcription factor that is activated by Cbk1 via three phosphorylation sites [ 7 ] , is responsible for transcriptional control of enzymes required for septum degradation after cytokinesis [ 6 ] . The other, Ssd1, is an RNAbinding protein that suppresses translation of certain mRNAs and which loses activity when phosphorylated by Cbk1 [ 8 ] . We chose these two proteins as targets in parallel affi nity selection experiments to illustrate that the same phage-displayed combinatorial peptide library can yield very different peptide ligands to different targets. The Lyn SH3 domain has previously been used in affi nity selection experiments, and so it served as a positive control. The Cbk1 protein had not been used previously in affi nity selection of combinatorial peptide libraries. We also show that the Kunkel mutagenesis in combination with alanine scanning and ELISA are very simple and expedient methods for evaluating the contribution of certain amino acids in the peptide ligands for binding to their targets. Prepare using deionized water (diH 2 O). Autoclave or fi lter sterilize and store at room temperature, unless indicated otherwise. Single-stranded DNA samples were isolated with the QIAprep Spin M13 purifi cation kit (Qiagen). 6. Covalently closed, circular double-stranded M13 DNA, synthesized via Kunkel mutagenesis [ 9 , 10 ] , was purifi ed with QIAquick PCR Purifi cation Kit (Qiagen). 7. Concentrations of DNA samples were determined using NanoDrop ND-1000 UV spectrophotometer (Thermo Fisher Scientifi c, Inc.) and measuring the optical absorbance at 260 nm. 8. The following steps: phosphorylation of oligonucleotides, annealing to the template, and synthesis of covalently closed circular double-stranded DNA (cccDNA) were performed using DNA Engine Dyad ® thermal cycler (Bio-Rad). In described methods, affi nity selections with phage-displayed combinatorial peptide libraries (Figs. 2 and 3 ) were employed to identify peptide ligands that bind to human Lyn SH3 domain and yeast Cbk1 protein kinase. In general, once peptide ligands have been identifi ed for a protein target, an interacting motif can be deduced via sequence alignment (Logo, Fig. 4 ) of the primary structures of the displayed binding peptides. However, to assess the functionality of this motif, amino acid replacements of the critical residues are generally necessary. While this is possible through the chemical synthesis of variant peptides, it is more expedient to use mutagenesis techniques, such as alanine scanning (Fig. 5 , [ 11 ] ). For the purpose of this selection protocol, the target domain was overexpressed in Escherichia coli in the form of a glutathione S-transferase (GST) fusion protein. Recombinant protein can be prepared via commercially available pGEX vectors (GE Healthcare). Figure 2 presents a schematic diagram of the selection protocol utilizing glutathione-conjugated magnetic beads. To maintain optimum sterility and eliminate carryover, only fi ltered tips should be used throughout the selection procedure. 1. Add 20 μL of glutathione magnetic bead slurry (5 μL of settled beads) into 2 mL Eppendorf Tube, and wash twice with 600 μL of ice-cold wash buffer #1 (PBS) on magnetic separator. Remove the supernatant. 4. Remove the supernatant. Resuspend in 200 μL of blocking buffer #1. Add 50 μg of GST protein to remove potential GST-binding phage clones and 20 library equivalents (e.g., if the library size is 10 10 clones and the titer 10 13 pfu/mL use 20 μL). Bring the volume to 600 μL with ice-cold wash buffer #1. Incubate at 4 °C for 1 h on the rotating shaker. 5. Remove the supernatant. Wash three times with 800 μL of icecold wash buffer #2 (PBST). Remove supernatant, add 800 μL of ice-cold wash buffer #1, and transfer into a new 2 mL Eppendorf Tube to eliminate any plastic-bound phage. 6. Repeat the washing step two more times with wash buffer #1. Remove the supernatant. 7. Elute the phage with 200 μL of elution buffer. Incubate for 10 min at room temperature, and gently mix the content every couple minutes. 8. Place the tube on magnetic separator, and transfer the eluted phage supernatant into a new 1.5 mL Eppendorf Tube containing 12 μL of neutralization buffer, and mix well. Fig. 3c ) revealed a known Y/FxFP docking motif to Cbk1 [ 17 ] . Also, it should be noted that sequence 3 (Fig. 3c ) contains the FKFP motif, which is present in Ssd1, a known Cbk1 substrate [ 8 , 17 ] . Our fi ndings clearly indicate that potential binding partners of certain targets can be deduced via selections with phage-displayed peptide libraries. As only three sequences were used for alignment, any additional amino acid preference analysis could be biased. ( c ) Further characterization of the isolated Y/FxFP motif showed preference to positively charged residues, R and K, at the "x" position. Other allowed residues at that position included V, M, T, Q, when at least one positively charged residue was observed outside the motif, or C, when a second cysteine was present following the motif. All sequences incorporated in the LOGO (32 isolates, data not shown) were isolated by screening phage-displayed focused library. The library was synthesized via Kunkel mutagenesis (as described in Subheading 3. 13. If blue-white screening can be performed, add 45 μL of 100 mM IPTG and 80 μL of 2 % X-Gal. 14. Immediately before plating, add 4 mL of 2 × YT top agar (0.8 %), gently swirl, and pour over prewarmed 2 × YT agar plates. Incubate overnight at 37 °C. 7. Add anti-M13 HRP-conjugated antibody diluted 1:5,000 in PBST (same per well volume as the target protein), and incubate 30 min to 1 h at room temperature. 8. Wash the plates with 200 μL of PBST three more times. 9. Add chromogenic substrate solution (Subheading 2.1 , item 21 ) (same per well volume as the target protein), and incubate for 10-30 min. 10. Quantify the results by measuring the optical absorbance at 405 nm, using a microtiter plate spectrophotometer. All binding phage clones isolated via ELISA are amplifi ed and their binding regions subsequently identifi ed by DNA sequencing ( see Note 7 ). If the selection generates enough unique isolates, the binding motif of the target can be predicted by sequence alignment known as LOGO plot (Fig. 4 ) . With that knowledge, one can attempt to better defi ne the known interactions and to predict new substrates of the target. To further assess the importance of each residue in the motif, mutagenic analysis known as alanine scanning can be performed (Fig. 5 , [ 11 ] ). In this method, each residue is replaced, one by one, by alanine (or by glycine if originally occupied by alanine). One way to generate all the variants is to incorporate them into M13 phage genome via Kunkel mutagenesis (described in Subheading 3.3.1 ) [ 10 , 12 , 13 ] . Once the phage-displayed mutant pool is generated, the effect of the substitutions on their binding to the target is evaluated by phage ELISA (described in Subheading 3.2 ). For the purpose of this protocol, a modifi ed M13 phage vector containing an amber stop codon has to be fi rst obtained [ 9 ] . In the vector constructed by Scholle et al., an amber stop codon, TAG, has been placed at the N-terminus of the coding region of gene III, following the signal sequence of protein III (pIII). That modifi cation eliminates the need for generation of uracil-containing single-stranded DNA (U-ssDNA) template, using E. coli strain CJ236 ( dut − ung − ). A mutagenic oligonucleotide, containing both complementary and exogenous DNA fragments, is then annealed to the ssDNA template, with the exogenous fragment looping over the TAG-containing vector region. Once the double-stranded DNA (dsDNA) is synthesized from the ssDNA template, it is electroporated into non-suppressor E. coli strain (e.g., SS320). If the TAG triplet-containing region has not been replaced by the exogenous fragment, the phage will not be propagated as the minor coat protein, pIII, cannot be translated. The wild-type phage can be propagated via the suppressor E. coli strain such as TG1 cells. The amber stop codon is then translated into glutamine. The use of TAG-containing template facilitates nearly 100 % recombination effi ciency [ 9 ] . 2. If low quantities of target are available, a small amount can be used to coat the beads (e.g., 10 μg of protein and 10 μL of slurry, respectively). Also, decreasing the amount of used target can facilitate isolation of higher affi nity clones. 3. For convenience, target-bound beads can be prepared for the entire selection procedure and stored at 4 °C in the blocking buffer #1 (PBS containing 3 % BSA (w/v)) for up to a week. 4. As long as no contamination is introduced, phage particles can be stored in a culture medium at 4 °C for many years. 5. Since the phage titer, recovered after fi nal round of selection, depends on many controllable and uncontrollable factors such as stringency of the selections (i.e., number of washes, amount of target used) or type and structure of the target, from our experience, we recommend to cover a wide dilution range from at least 10 −4 -10 −8 . 6. If fi ltered tips are used for phage transfer, it is convenient to briefl y rinse and remove the tips with multichannel pipette. 7. In order to identify DNA sequences of potential "binders," the phage is fi rst amplifi ed by infecting 200 μL of TG1 cells (from a mid-log phase culture) with 10 μL of phage supernatant (or single plaque) and incubated overnight in a shaking incubator (240 rpm), in 5 mL fi nal volume. The pelleted cells are then used for isolation of dsDNA that consists of replicative form of phage DNA. Subsequently, the samples are analyzed via sequencing. Remaining phage supernatant can be stored at 4 °C ( see Note 4 ). 8 . In order to isolate high quantity of ssDNA template, large amount of phage particles has to be fi rst collected. Amplifi cation of phage by infecting cells with just a single plaque may result in low quantity of template. Thus, we suggest a two-step preparation process, where the PEG-precipitated concentrated virions are used to infect the fresh mid-log cells. 9. To remove any polyethylene glycol (PEG) residues from the surface of the tube prior to phage reconstitution, it is recommended to fi rst briefl y rinse the side of the tube with PBS, avoiding the phage pellet. Usually 1-3 mL of room temperature PBS is used to resuspend the pelleted virions. One of the major advantages of affi nity selection of phage-displayed combinatorial peptide libraries is the potential for rapid discovery of peptide ligands to a target protein. It is relatively straightforward to use homemade or commercially purchased libraries to isolate peptide ligands to a given target and then deduce a binding motif. A peptide with a consensus motif can then be used to study the biology of the target (i.e., predict cellular interacting partner, solve the three-dimensional structure of the peptide and target in a complex, and to inhibit the target in cells). However, when only one or a few peptide ligands are isolated to a target, it is diffi cult to know a priori what residues in the phage-displayed peptide ligand contribute to binding. While one can explore the binding parameters of the peptide sequence through chemical synthesis of peptides with systematic amino acid replacements across the primary structure, we present an alternative, faster, and easier approach involving Kunkel mutagenesis, alanine scanning, and ELISA. We fi nd that by coupling these three techniques, one can readily determine at fi rst pass many important aspects of the binding interaction. Exploring protein -protein interactions using peptide libraries displayed on phage Mapping intracellular protein networks The yesrelated cellular gene lyn encodes a possible tyrosine kinase similar to p56lck Src family kinases: regulation of their activities, levels and identifi cation of new pathways Cbk1p, a protein similar to the human myotonic dystrophy kinase, is essential for normal morphogenesis in Saccharomyces cerevisiae The Saccharomyces cerevisiae Mob2p-Cbk1p kinase complex promotes polarized growth and acts with the mitotic exit network to facilitate daughter cell-specifi c localization of Ace2p transcription factor The NDR/LATS family kinase Cbk1 directly controls transcriptional asymmetry Cbk1 regulation of the RNA-binding protein Ssd1 integrates cell fate with translational control Effi cient construction of a large collection of phage-displayed combinatorial peptide libraries Effi cient site-directed mutagenesis using uracilcontaining DNA Highresolution epitope mapping of hGH-receptor interactions by alanine-scanning mutagenesis Improvements to the Kunkel mutagenesis protocol for constructing primary and secondary phage-display libraries Rapid and effi cient sitespecifi c mutagenesis without phenotypic selection Convergent evolution with combinatorial peptides Can we infer peptide recognition specifi city mediated by SH3 domains? Distinct ligand preferences of Src homology 3 domains from Src, Yes, Abl, Cortactin, p53bp2, PLCy, Crk, and Grb2 Proteome-wide discovery of evolutionary conserved sequences in disordered regions This work was funded by the Chicago Biomedical Consortium, with support from the Searle Funds at the Chicago Community Trust. We would like to thank Mr. Kyle Schneider, Dr. Brian Yeh, and Dr. Eric Weiss (NU) for providing the GST-Cbk1 DNA constructs and purifi ed GST-Cbk1 fusion protein. We are grateful to Dr. Michael Kierny and Dr. Renhua Huang (UIC) for their helpful comments on this manuscript.