Directed Evolution of a Cyclized Peptoid–Peptide ... - ACS Publications

Mar 24, 2016 - Takashi Kawakami,* Koji Ogawa, Tomohisa Hatta, Naoki Goshima, and Tohru Natsume. Molecular Profiling Research Center for Drug ...
0 downloads 0 Views 4MB Size
Articles pubs.acs.org/acschemicalbiology

Directed Evolution of a Cyclized Peptoid−Peptide Chimera against a Cell-Free Expressed Protein and Proteomic Profiling of the Interacting Proteins to Create a Protein−Protein Interaction Inhibitor Takashi Kawakami,* Koji Ogawa, Tomohisa Hatta, Naoki Goshima, and Tohru Natsume Molecular Profiling Research Center for Drug Discovery (molprof), National Institute of Advanced Industrial Science and Technology (AIST), 2-4-7 Aomi, Koto-ku, Tokyo 135-0064, Japan S Supporting Information *

ABSTRACT: N-alkyl amino acids are useful building blocks for the in vitro display evolution of ribosomally synthesized peptides because they can increase the proteolytic stability and cell permeability of these peptides. However, the translation initiation substrate specificity of nonproteinogenic N-alkyl amino acids has not been investigated. In this study, we screened various N-alkyl amino acids and nonamino carboxylic acids for translation initiation with an Escherichia coli reconstituted cell-free translation system (PURE system) and identified those that efficiently initiated translation. Using seven of these efficiently initiating acids, we next performed in vitro display evolution of cyclized peptidomimetics against an arbitrarily chosen model human protein (β-catenin) cell-free expressed from its cloned cDNA (HUPEX) and identified a novel β-catenin-binding cyclized peptoid−peptide chimera. Furthermore, by a proteomic approach using direct nanoflow liquid chromatography−tandem mass spectrometry (DNLC−MS/MS), we successfully identified which protein−β-catenin interaction is inhibited by the chimera. The combination of in vitro display evolution of cyclized N-alkyl peptidomimetics and in vitro expression of human proteins would be a powerful approach for the high-speed discovery of diverse human protein-targeted cyclized N-alkyl peptidomimetics.

T

thioether-cyclized N-methylated peptides.30,49 Despite many reports on high permissiveness of acid substrates in the translation initiation and usefulness of ribosomal N-alkyl aa incorporation, initiation substrate specificity for nonproteinogenic N-alkyl aa has not been reported. In this study, we screened various ClAc N-alkyl aa and thiolreactive nonamino carboxylic acids for translation initiation using the PURE system and identified those that efficiently initiate translation. By using seven of these efficiently initiating acids, we next performed in vitro display evolution (selection) of cyclized peptidomimetics against an arbitrarily chosen model human protein (β-catenin) cell-free expressed from its cloned cDNA (HUPEX)50,51 and identified a novel β-catenin-binding cyclized peptoid−peptide chimera. Furthermore, by a proteomic approach using direct nanoflow liquid chromatography− tandem mass spectrometry (DNLC−MS/MS),52,53 we successfully identified which protein−β-catenin interaction is inhibited by the chimera.

he Escherichia coli (E. coli) ribosome initiates translation with formyl methionine (fMet) as a natural substrate. However, engineering initiator tRNAs,1,2 chemically modifying amino acids (aa) on initiator tRNAs,3,4 and developing novel tRNA-acylation methods5,6 have revealed high permissiveness of acid substrates in translation initiation. The substrates include formyl non-Met proteinogenic aa,7−10 formyl nonproteinogenic aa,11−15 nonformyl various N-acyl aa,16−21 a hydroxy acid,13 D-aa,22,23 backbone-elongated aa (such as βaa),22,24,25 and nonamino carboxylic acids.22,24 Translation initiation using non-fMet substrates has been used to produce N-terminal-labeled proteins4 or to probe acid substrate specificity in translation initiation.26 In addition to these applications, translation initiation using N-chloroacetyl (ClAc) aa has been used with the E. coli reconstituted cell-free translation system (PURE system)27,28 for in vitro selection of thioether-cyclized peptides.29,30 In addition to the initiation, various nonproteinogenic aa have been incorporated into ribosomally synthesized peptides in the translation elongation steps.29,31,32 In particular, N-alkyl aa33−48 are useful building blocks for short peptide selection because they can increase the proteolytic stability and cell permeability of peptides in vitro selected from highly diverse libraries. N-methyl aa incorporation in elongation has been combined with ClAc aa initiation for in vitro display selection of © XXXX American Chemical Society

Received: December 10, 2015 Accepted: March 8, 2016

A

DOI: 10.1021/acschembio.5b01014 ACS Chem. Biol. XXXX, XXX, XXX−XXX

Articles

ACS Chemical Biology

Figure 1. Translation initiation with diverse thiol-reactive N-alkyl amino acids and nonamino carboxylic acids in the Escherichia coli reconstituted cell-free translation system (PURE system). (a) Chemical structure of thiol-reactive N-alkyl amino acids and nonamino carboxylic acids used in this study. (b) Scheme of translation initiation with the thiol-reactive N-alkyl amino acids and nonamino carboxylic acids charged onto tRNACAUini (XtRNACAUini). Sequences of mRNA and peptides encoded in the mRNA are shown. The thiol-reactive N-alkyl amino acids and nonamino carboxylic acids (X) were reassigned to the vacant initiator AUG codon. (c) Tricine sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis of the expressed peptides labeled with [14C]-Asp for detection by autoradiography. The peptides were expressed in the presence of 100 μM of the designated acyl-tRNACAUini. The relative yield of each peptide based on its observed radioisotope count is shown in the plot.



RESULTS AND DISCUSSION Translation Initiation with N-Alkyl Amino Acids and Nonamino Carboxylic Acids in the PURE System. To obtain information regarding the structural requirements of efficient N-alkyl-aa initiators, we first chose the 16 ClAc N-alkyl aa shown in Figure 1a. It should be noted that all of the N-alkyl aa can be readily used as building blocks in chemical peptide synthesis because of the commercial availability of the

corresponding [(9-fluorenylmethyl) oxy] carbonyl (Fmoc) building blocks. Thus, a high amount of peptides containing these N-alkyl aa can be readily obtained, which is normally required after in vitro peptide selection. All 16 ClAc N-alkyl aa were chemically derivatized to activated esters for conversion to the corresponding tRNA-acylation substrates. The tRNAacylation conditions for the ClAc N-alkyl aa were optimized using these activated ester derivatives and a 22-mer microhelix B

DOI: 10.1021/acschembio.5b01014 ACS Chem. Biol. XXXX, XXX, XXX−XXX

Articles

ACS Chemical Biology

Figure 2. Scheme of mRNA display evolution of cyclized peptidomimetics against a human protein (β-catenin) cell-free expressed from its cloned cDNA (HUPEX). (a) The β-catenin protein is expressed in the cell-free translation system using human β-catenin cloned cDNA and site-specifically biotinylated via Avi-tag by E. coli biotin ligase (BirA). A DNA library containing randomized sequences, (NNK)8−15, is transcribed, and the library of the mRNAs is spontaneously modified with puromycin linkers in the release factor (RF)-free E. coli reconstituted cell-free transcription/translationcoupling system (PURE system). The puromycin-modified mRNA library is translated into a peptidomimetic library; the expressed peptidomimetics are spontaneously displayed on their encoding mRNAs through a puromycin-DNA linker and cyclized between the thiol-reactive initiator and the Cterminal cysteine residue in the PURE system. The mRNA-displayed peptidomimetic library is then reverse-transcribed and precleared, without purification, with β-catenin-free beads for the removal of streptavidin-bead-binding peptidomimetics. The precleared peptidomimetic library is then incubated with the biotinylated β-catenin immobilized on streptavidin beads. The beads are washed for the isolation of β-catenin-binding peptidomimetics. The recovered cDNAs encoding the β-catenin-binding peptidomimetics are amplified using polymerase chain reaction (PCR) and used for the next round of selection. (b) Progress during in vitro selection of cyclized peptidomimetics from random library (∼1013 in library size) against a cell-free expressed β-catenin. Recovery of cDNAs in each round of selection was determined by quantitative PCR. (c) Target-specificity of the cloned cyclized peptoid−peptide chimera. Binding of each protein to the cyclized peptoid−peptide chimera displayed on its encoding mRNA/ cDNA complex was evaluated. Binding efficiency was analyzed by quantification of the cDNA recovered after incubation with immobilized protein using quantitative PCR.

RNA45,47 that contained only an acceptor stem of tRNA. The ClAc N-alkyl-aa-microhelix RNAs were separated from unacylated microhelix RNAs on acid urea polyacrylamide gel electrophoresis (PAGE) based on the relative increase of their molecular weight. Quantification of their product yields showed that all of the ClAc N-alkyl aa could be charged onto a microhelix RNA with yields of over 30% under the optimized conditions (Supporting Information Table S1). An E. coli initiator tRNA (tRNACAUini) aminoacylated with the ClAc N-alkyl aa was next assayed by initiation of the ribosomal ClAc-N-alkyl-aa-(Lys)3-(Thr)2-Flag peptide synthe-

sis (Figure 1b). The peptide expression was performed using the PURE system that lacks Met. For comparison, the same DNA template was transcribed and translated into fMet-(Lys)3(Thr)2-Flag by inclusion of Met in the PURE system. The peptide yields were calculated based on the incorporation of [14C]-Asp into the ClAc-N-alkyl-aa-(Lys)3-(Thr)2-Flag peptide. The 60 min translation reaction was terminated by the addition of sodium dodecyl sulfate (SDS)-containing buffer, and the peptide products were separated from [14C]-Asp on tricine SDS-PAGE and detected by autoradiography. Figure 1c shows that translation could be initiated with all the tested ClAc NC

DOI: 10.1021/acschembio.5b01014 ACS Chem. Biol. XXXX, XXX, XXX−XXX

Articles

ACS Chemical Biology

amino carboxylic acids, we next performed in vitro display evolution (selection) of cyclized peptidomimetics against cellfree expressed β-catenin. To select cyclized peptidomimetics that bind to β-catenin, we designed the selection scheme shown in Figure 2a. Following the Gateway LR reaction, using an entry clone β-catenin cDNA (HUPEX) and a linearized destination vector DNA for the addition of His and Avi tags, we amplified a His-Avi-β-catenin DNA template by polymerase chain reaction (PCR) for transcription using the crude LR reaction mixture as a template. A His-Avi-β-catenin protein was expressed by transcription using the crude PCR product and translation with a wheat germ cell-free translation system. After in situ site-specific biotinylation of the His-Avi-β-catenin protein with E. coli biotin ligase (BirA), it was purified using the His tag (Supporting Information Figure S3a). High purity of the biotinylated β-catenin immobilized on streptavidin beads was confirmed by silver staining of an SDS-PAGE gel (Supporting Information Figure S3b). A DNA library was next designed to encode the peptide library consisting of 8−15 random amino acids between a thiolreactive initiator and a cysteine residue to be used for peptidomimetic cyclization. We prepared the cyclized peptidomimetic library by incorporating the corresponding mRNA library into a release factor (RF)-free PURE system containing a puromycin-DNA linker, together with 19 proteinogenic amino acids (without Met) and a thiol-reactive initiator charged on initiator tRNA. It should be noted that the mRNA is spontaneously modified with puromycin, and the expressed peptidomimetic is spontaneously displayed on the encoding mRNA in the RF-free PURE system58 (Figure 2a). After the reverse transcription step, the library of ∼1013 unique cyclized peptidomimetics was incubated, without prior purification, with biotinylated β-catenin immobilized on streptavidin beads. After the beads were washed to remove unbound cyclized peptidomimetics, the recovered cDNA was amplified using PCR and used as the input for the next round of selection. To monitor the β-catenin-binding ability of the library, we determined the ratio of recovered cDNA to input cDNA in each round of selection using quantitative PCR (Figure 2b). In the subsequent rounds of selection, the crude PCRamplified DNA library was directly added to the transcription/ translation-coupled PURE system containing a puromycinDNA linker, where an mRNA-displayed cyclized peptidomimetic library is produced from the corresponding DNA library in a one-pot approach (Figure 2a). β-Catenin-free beads were also used throughout the selection to remove undesired streptavidin-binding cyclized peptidomimetics. In the selection, the cDNA recovery rate was increased during rounds 9 and 10 for the ClAcBnGly library (Figure 2b), whereas the cDNA recovery rates were nearly constant for the other libraries (Supporting Information Figure S4). MALDI-TOF MS analysis of the peptide library expressed from the 10th ClAcBnGly DNA library in the PURE system containing RF indicated that the library was enriched to a few main sequences (Supporting Information Figure S5a). The cDNAs obtained from round 10 of the ClAcBnGly library were cloned, sequenced, and translated according to the universal genetic code table (Supporting Information Figure S5b). The sequence alignment of 19 clones revealed that the library was enriched mainly by three independent sequence families. Although the most abundant sequence in the most abundant family had an AUG codon, which is translated into Met in the universal genetic code table, the AUG codon is

alkyl aa. In particular, initiations with ClAc N-alkyl aa bearing aromatic side chains were more efficient (MePhe, MeTyr, MeTrp, Me Fcl, BnGly, and D-MePhe). This information is invaluable for the choice of appropriate ClAc N-alkyl aa and design thioethercyclized N-alkyl peptide libraries for in vitro display selection. Translation initiation using the designated ClAc N-alkyl aa and subsequent cyclization with an intramolecular cysteine residue was confirmed for all the tested ClAc N-alkyl aa by matrixassisted laser desorption/ionization time-of-flight mass spectrometric (MALDI-TOF MS) analysis (Supporting Information Figure S1). The ClAc N-alkyl aa are useful for constructing a cyclized peptide library containing a translation initiator that has specific functions such as light absorption for peptide quantification or affinity (reactivity) specific to its target (a so-called warhead) because such specific functions can be conferred by the sidechain of the ClAc N-alkyl aa. However, to increase the cell permeability of peptides, nonamino initiators such as ClAc (Figure 1a) would be preferable. Accordingly, to evaluate nonamino carboxylic acid substrates for their translation initiation ability, ClAc was charged onto tRNACAUini and tested for translation initiation by expressing the same model peptide as above (Figure 1b). Despite early reports of translation initiation with some nonamino carboxylic acids,22,24 the initiation efficiency of ClAc was extremely low (Figure 1c). Recently, a thiol-reactive halobenzyl group has been used by other groups for phage and mRNA display selection of peptides that are thioether-cyclized using cysteine residue(s).12,54,55 Because we found that translation initiation generally prefers aromatic N-alkyl aa irrespective of the position of the aromatic group against the carboxylic acid (Figure 1c), we next tested chlorobenzyl-containing carboxylic acids for translation initiation. Three of these (m-chloromethylbenzoic acid (mClPh), p-chloromethylbenzoic acid (pClPh), and m-chloromethylphenylacetic acid (mClBn), Figure 1a) were charged onto tRNACAUini and assayed by initiation of peptide synthesis (Figure 1b). We found that these chlorobenzyl-containing carboxylic acids could efficiently initiate translation (Figure 1c) and cyclize the expressed peptide with an intramolecular cysteine residue (Supporting Information Figure S1), supporting the idea that aromatic substrates are efficient translation initiators irrespective of the position of the aromatic group. To further develop a thiol-reactive translation initiator useful for in vitro display selection of cyclized peptidomimetics, we designed 3,5-bis-chloromethylbenzoic acid (Cl2Ph) for peptide bicyclization with two intramolecular cysteine residues, encouraged by the recent success of thioether-bicyclized peptide selection.56,57 We found that Cl2Ph charged onto tRNACAUini could initiate translation in a manner similar to the other aromatic nonamino carboxylic acid substrates (Figure 1c). We next added Cl2Ph-tRNACAUini to the PURE system containing DNA encoding a model peptide that contains two cysteine residues (Supporting Information Figure S2a). MALDI-TOF MS analysis showed that Cl2Ph was spontaneously bicyclized using two intramolecular cysteine residues via two redox-insensitive thioether bonds in the PURE system (Supporting Information Figure S2b and c). The peptide bicyclization method by Cl2Ph has a feature wherein the bicyclic structure is both spontaneously formed in the translation system and stable even under reducing conditions. mRNA Display Evolution of Cyclized Peptidomimetics against a Cell-Free Expressed Protein. Using seven of the efficiently initiating ClAc N-alkyl aa and thiol-reactive nonD

DOI: 10.1021/acschembio.5b01014 ACS Chem. Biol. XXXX, XXX, XXX−XXX

Articles

ACS Chemical Biology misread as Ile in the Met-lacking PURE system used for the selection. Thus, a Met to Ile mutant of the most abundant clone in the most abundant family (ClAc B n GlyRWLFLDSLATWFILPC), the most abundant clone in the second most abundant family (ClAcBnGly-RFARCWKILRAIFC), and the most abundant clone in the third most abundant family (ClAcBnGly-CFVIAYLLWC) were chosen and assayed (Supporting Information Figure S6). β-Catenin pulldown of the mRNA-displayed cyclized peptoid−peptide chimera showed that the cloned ClAcBnGly-RWLFLDSLATWFILPC had much higher β-catenin-binding ability. The higher binding ability probably occurred because the Met-to-Ile mutant is more efficiently expressed than its parent cyclized peptoid−peptide chimera. β-catenin-dependent binding of the Met-to-Ile mutant was also confirmed by comparing the pulldown with the β-catenin-free beads and different proteinimmobilized beads (Figure 2c). The clones identified from the ClAcMeTrp and mClPh library that were slightly recoveryincreased in the selection showed much weaker β-cateninbinding than the Met-to-Ile mutant (Supporting Information Figure S6). Thus, the Met-to-Ile mutant was named β-cateninbinding cyclized peptoid−peptide chimera 1 (BCP1) and further characterized. Proteomic Profiling Using DNLC-MS/MS for Characterization of the Cyclized Peptoid−Peptide Chimera. We chemically synthesized BCP1 fused with a C-terminal Flag peptide using Fmoc solid-phase peptide synthesis and the HPLC-purified BCP1-Flag was analyzed using MALDI-TOF MS (Supporting Information Figure S7). Pull-down of purified β-catenin from a wheat germ (WG) cell-free translation system with the BCP1-Flag immobilized on anti-Flag monoclonal antibody (Ab) beads showed that purified β-catenin could be immunoprecipitated with BCP1 in a BCP1-dependent manner (Figure 3a and b). Conversely, pull-down of β-catenin in human embryonic kidney (HEK) 293 cell lysate showed that βcatenin could not be efficiently immunoprecipitated using BCP1, although it was immunoprecipitated using the control anti-β-catenin Ab (Figure 3c). When BCP1-Flag-pull-down was performed using the HEK 293 lysate containing β-catenin overexpressed by transient transfection, immunoprecipitation of β-catenin was detected (Figure 3d). Accordingly, we hypothesized that the BCP1-binding site of β-catenin is masked by a certain β-catenin-interacting human protein present in HEK 293 cell lysate because β-catenin is known to interact with multiple other human proteins. Detection of β-catenin pulled down with BCP1 from crude β-catenin in the WG lysate also supports this hypothesis (Supporting Information Figure S8). To identify a human protein that masks the BCP1-binding site of β-catenin, we adopted a proteomic approach. The HEK 293 lysate containing overexpressed myc-tagged β-catenin was incubated with BCP1 beads or antimyc Ab beads at 4 °C to preserve the β-catenin/protein complex (Figure 3e). Following the endopeptidase digestion of isolated proteins, the peptide fragments were analyzed by direct nanoflow liquid chromatography−tandem mass spectrometry (DNLC−MS/MS).52,53 Whereas β-catenin was detected by immunoprecipitation using both BCP1 and antimyc Ab (Figure 4b), consistent with Western blotting analysis (Figure 3d), α-catenin, which is known to be a β-catenin-interacting protein, was detected only by antimyc Ab immunoprecipitation. Thus, the DNLC−MS/ MS analysis indicates that BCP1 binding to β-catenin was masked by α-catenin. When the β-catenin expression level in HEK 293 was increased by treatment using an inhibitor (LiCl)

Figure 3. Immunoprecipitation analysis of β-catenin using β-cateninbinding cyclized peptoid−peptide chimera (BCP1) obtained by mRNA display evolution. (a) Immunoprecipitation of β-catenin with the β-catenin-binding cyclized peptoid−peptide chimera (BCP1). BCP1 tagged with Flag-peptide was immobilized on protein G beads via anti-Flag antibody (Ab). (b) Immunoprecipitation of β-catenin expressed and purified from wheat germ (WG) lysate translation system. Immunoprecipitated β-catenin was analyzed by Western blotting analysis using anti-β-catenin antibody detection followed by anti-mouse secondary antibody detection. (c) Immunoprecipitation of endogenous β-catenin in HEK 293 cell lysate. Immunoprecipitation with anti-β-catenin antibody was used as a positive control. (d) Immunoprecipitation of β-catenin overexpressed in HEK 293 cell lysate.

of GSK3β that leads to β-catenin degradation instead of transient transfection, α-catenin was not detected in the DNLC−MS/MS analysis of the sample immunoprecipitated using BCP1 (Supporting Information Figure S9). This result is also consistent with the idea that the BCP1 binding to βcatenin is masked by α-catenin. Finally, to directly check whether β-catenin/α-catenin interaction is inhibited by BCP1, β-catenin beads were incubated with α-catenin and varying concentrations of BCP1 at 37 °C (Figure 4c) instead of 4 °C used in the above immunoprecipitation for preserving the β-catenin/α-catenin interaction. Western blotting analysis showed that the amount of α-catenin immunoprecipitated using β-catenin decreased with an increase in the amount of BCP1, whereas the amount of β-catenin used as loading control was constant (Figure 4d). This result clearly demonstrated that BCP1 is a β-catenin/αcatenin interaction inhibitor. E

DOI: 10.1021/acschembio.5b01014 ACS Chem. Biol. XXXX, XXX, XXX−XXX

Articles

ACS Chemical Biology

Figure 4. Proteomic profiling of interacting proteins using direct nanoflow liquid chromatography−tandem mass spectrometry (DNLC-MS/MS) for the identification of β-catenin/protein interaction inhibited by the β-catenin-binding cyclized peptoid−peptide chimera (BCP1). (a) Immunoprecipitation of myc-tagged β-catenin overexpressed in HEK 293 cell lysate with BCP1 or an antimyc antibody (Ab) and subsequent DNLC-MS/MS analysis of the digested immunoprecipitates. Human proteins identified by DNLC-MS/MS were compared between BCP1 and antimyc antibody. (b) The number of identified peptide fragments of β-catenin and α-catenin in the DNLC-MS/MS analysis. (c) Scheme of inhibition analysis of β-/α-catenin interaction by BCP1. β-catenin-immobilized beads were incubated with α-catenin and varying concentrations of BCP1. Pulled-down samples were analyzed by Western blotting. (d) Western blotting analysis of β-catenin and pulled-down α-catenin in the presence of varying concentrations of BCP1.



DISCUSSION In this study, by using continuous-type mRNA display selection30,45,47 of an extremely diverse library (∼1013) against a human protein cell-free-expressed from its cloned cDNA (HUPEX),50,51 we successfully identified a novel anti-β-catenin cyclized peptoid−peptide chimera. The salient feature of this ultrahigh-throughput screening system for discovering human protein-targeted cyclized N-alkyl peptidomimetic compounds is its simplicity. In the cell-free expression of a biotinylated target human protein from the corresponding cloned cDNA, only the serial addition of an aliquot of a previous reaction mixture to the subsequent one is required.50,51 It should be noted that the cyclized peptidomimetics selection using a cell-free expressed protein can be performed with the same level (amount) of the target protein as that using a live cell-overexpressed protein.30 In the cyclized N-alkyl peptidomimetic library preparation, the cDNA-displayed cyclized N-alkyl peptidomimetic library can be prepared from the corresponding DNA library in a one-pot approach.30,45,47 The use of these cell-free approaches would be suitable when combined with robotic system-based automation.59 By a proteomic approach using DNLC-MS/MS,52,53,60 we also successfully identified the resulting cyclized peptoid− peptide chimera (BCP1) as a β-catenin/α-catenin interaction inhibitor. A straightforward approach for the discovery of protein−protein interaction inhibitors by in vitro selection would be elution of target-binding active species with the existing competitive inhibitor(s). However, there is a possibility that highly tight binding inhibitors would not be eluted well and enriched even with an excess of existing competitive inhibitor because of extremely high affinity to the target

protein. Thus, combining binding-based selection with subsequent high-throughput compound characterization could be a complementary approach. Replacing the N-benzyl group of the cyclized peptoid− peptide chimera with an N-methyl group or hydrogen allowed for pulling-down of β-catenin at roughly the same level (Supporting Information Figure S10). This result suggests that the cyclized peptoid−peptide chimera binds to β-catenin at a point distant from the N-benzyl group. β-Catenin-binding constants of the cyclized peptoid−peptide chimera (KD = 600 nM), N-methyl analog (KD = 60 nM), and non-N-methyl analog (KD = 300 nM) determined using the Bio-Layer Interferometry method30 (ForteBio) also support this idea. A detailed β-catenin binding mechanism of the cyclized peptoid− peptide chimera will be determined by a structural biological approach. Although a structural biological approach is useful for the characterization of a compound/protein interaction site, a high amount of the compound and protein is generally required. Furthermore, when the compound allosterically inhibits protein−protein interaction, identifying the inhibited interaction may be difficult. On the other hand, a proteomic approach using DNLC-MS/MS requires only a small amount of cyclized N-alkyl peptidomimetic compound, which can be produced by a cell-free transcription/translation-coupled system, because of the high sensitivity of MS.52,53,60 Thus, identifying the protein−protein interaction inhibited by the cyclized N-alkyl peptidomimetic using DNLC-MS/MS can be achieved in a high-throughput manner. In conclusion, we screened diverse N-alkyl amino acids and nonamino carboxylic acids for translation initiation and found that acid substrates bearing aromatic side chains were F

DOI: 10.1021/acschembio.5b01014 ACS Chem. Biol. XXXX, XXX, XXX−XXX

Articles

ACS Chemical Biology

T7SDM2.F44, and 0.25 μM G5S-4an21.R41], and the cDNAs encoding the peptidomimetics bound to the β-catenin-immobilized beads were eluted at 95 °C for 5 min. The eluted cDNAs were quantified using SYBR Green-based quantitative PCR (StepOnePlus, Applied Biosystems) with primers T7SD8M2.F44 and CG5S4an21.R44. The eluted cDNAs were amplified for the next round of selection by PCR using Taq DNA polymerase (Genscript). From the second round, the mRNA-displayed peptidomimetic library was prepared using transcription and translation in the release factor-free transcription/translation coupling system (PURExpress ΔRF123 Kit, New England Biolabs) containing a puromycin linker and the crude PCR mixture, including the cDNA library, obtained from the previous round. The resultant mRNA-displayed peptidomimetic library was reverse-transcribed using ReverTraAce at 42 °C for 30 min. After quenching the reverse transcription using EDTA and neutralizing the solution using HEPES, the mRNA-displayed peptidomimetic library was incubated using β-catenin-free streptavidin-magnetic beads (Dynabeads, Life Technologies) at 25 °C for 5 min for negative selection. This negative selection was performed five times to remove peptidomimetics binding to the streptavidin-magnetic beads. The supernatant was then mixed with the β-cateninimmobilized beads at 25 °C for positive selection, and the beads were washed for around 1 min five times using HBS-T. The selected cDNAs on the beads were quantified using SYBR Green-based quantitative PCR using primers T7SD8M2.F44 and CG5S-4an21.R44. The cDNAs were then amplified by PCR using Taq DNA polymerase for the next round of selection. Immunoprecipitation of β-Catenin with a Cyclized Peptoid− Peptide Chimera (BCP1). Chemically synthesized BCP1-Flag was immobilized on Protein G-magnetic beads (Dynabeads, Life Technologies) via anti-Flag M2 mouse monoclonal antibody (Sigma-Aldrich). HEK 293 cell lysate was prepared by lysing the cells with lysis buffer containing 1% digitonin. The BCP1 beads were incubated with β-catenin-containing solution at 4 °C for 0.5 h and washed three times using cold HBS-T. For Western blotting analysis, immunoprecipitated samples were eluted using Flag peptide and analyzed by 10% SDS-PAGE. β-catenin was visualized using anti-βcatenin mouse antibody (BD Biosciences) and anti-mouse HRPconjugated secondary antibody followed by chemiluminescent detection. For DNLC-MS/MS analysis, immunoprecipitated samples were digested using Lys-C-endopeptidase and analyzed by DNLCMS/MS.52,53 Experiments were performed in quadruplicate.

particularly efficient initiators. We also isolated a cyclized peptoid−peptide chimera against a human β-catenin cell-free expressed from its cloned cDNA by in vitro display evolution. Furthermore, by a proteomic approach using DNLC-MS/MS, we successfully identified the anti-β-catenin cyclized peptoid− peptide chimera as a β-catenin/α-catenin interaction inhibitor. A combination of continuous-type mRNA display selection of cyclized N-alkyl peptidomimetics,30,45,47 in vitro expression of the human proteome,50,51 and robotic system-based automation59 would be a powerful approach for the high-speed discovery of diverse disease-associated human protein-targeted cyclized N-alkyl peptidomimetics.



METHODS

Ribosomal Synthesis of Peptides Initiated with ClAc N-Alkyl aa or a Nonamino Carboxylic Acid. Preparation of the DNA templates encoding the peptide and acyl-tRNACAUini is described in the Supporting Information. A translation reaction mixture containing a 0.04 μM DNA template, 0.5 mM Tyr, 0.5 mM Lys, 0.5 mM Thr, 50 μM [14C]-Asp, and 100 μM acyl-tRNACAUini was incubated in the PURE system (PURExpress, New England Biolabs) for 60 min at 37 °C. Products were analyzed using tricine SDS-PAGE and autoradiography (BAS2500, Fujifilm). For MALDI−TOF MS analysis, the reaction was performed using Asp and Cys instead of [14C]-Asp and Thr, respectively. The samples were desalted using C18-tip cartridges (C-TIP, Nikkyo Technos), eluted using 80% acetonitrile and 0.5% acetic acid, saturated with α-cyano-4-hydroxycinnamic acid (CHCA), and analyzed using MALDI-TOF MS (AXIMA-CFR, Shimadzu) operated in the linear positive mode. Selection of Cyclized Peptidomimetics against in Vitro Expressed β-Catenin. A human β-catenin protein was prepared from its cloned cDNA (HUPEX) as previously described50 using a robotic protein synthesizer (Protemist DT, CellFree Sciences). For the preparation of DNA templates encoding random peptides, SD8M2NNK8−15-CG5S oligo DNAs and CG5S-4an21.R44 oligo DNA (Supporting Information Table S2) were annealed and extended using KOD-plus-neo DNA polymerase (Toyobo). The resulting dsDNAs were amplified using primers T7SD8M2.F44 and CG5S-4an21.R44 with Taq DNA polymerase (Genscript). The amplified DNAs were purified by extraction with phenol/chloroform and ethanol precipitation. These random DNA templates were transcribed by runoff in vitro transcription using T7 RNA polymerase (Toyobo), and the random mRNAs were purified using isopropanol precipitation. An mRNA library was prepared by mixing the random mRNAs, which contained 8−15 repeated NNK sequences. A library of mRNAdisplayed peptidomimetics was prepared by translation for 25 min at 37 °C in the release factor-free translation system (PURExpress ΔRF123 Kit, New England Biolabs) on a 250-μL scale with 0.5 mM of 19 proteinogenic amino acids (Met), 2.5 μM of mRNA library, 2.5 μM of puromycin-linker (BEX; Supporting Information Table S2), and 100 μM of acyl-tRNACAUini. Ethylenediaminetetraacetic acid (EDTA) was then added to the solution of the mRNA-displayed peptidomimetic library to dissociate ribosomes. Reverse transcription of the mRNA-displayed peptidomimetic library was performed at 42 °C for 30 min prior to selection using RNase H-inactivated reverse transcriptase (ReverTraAce, Toyobo) and primer CG5S-4.R23 (Supporting Information Table S2). After quenching the reverse transcription using EDTA and neutralizing the solution using 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), site-specifically bioinylated β-catenin immobilized on streptavidin-magnetic beads (Dynabeads, Life Technologies) was mixed with a solution containing the peptidomimetic library at 25 °C for 10 min. After the supernatant was removed, the beads were washed for around 1 min three times using HEPESbuffered saline (HBS)-T (50 mM HEPES-K, pH 7.5, 300 mM NaCl, and 0.05% Tween 20). They were then diluted using a PCR premixture solution [10 mM Tris−HCl, pH 8.4, 50 mM KCl, 0.1% (v/ v) Triton X-100, 2 mM MgCl2, 0.25 mM each dNTP, 0.25 μM



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acschembio.5b01014. Supporting Figures S1−S10 and Tables S1 and S2 (PDF)



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was supported by a JSPS KAKENHI Grant-in-Aid for Young Scientists (A; 15H05372 to T.K.) and Grant-in-Aid for Scientific Research on Innovative Areas (26102715 to T.K.), JST A-STEP, the Sagawa Foundation for Promotion of Cancer Research, Osaka Cancer Research Foundation, Showa University Medical Foundation, the Waksman Foundation of Japan, and the Uehara Memorial Foundation and in part by JSPS KAKENHI Grant-in-Aid for Scientific Research (B; G

DOI: 10.1021/acschembio.5b01014 ACS Chem. Biol. XXXX, XXX, XXX−XXX

Articles

ACS Chemical Biology

acetyltransferase peptides affect translation and movement through the ribosome. J. Biol. Chem. 275, 1781−1786. (20) Forster, A. C., Cornish, V. W., and Blacklow, S. C. (2004) Pure translation display. Anal. Biochem. 333, 358−364. (21) Mamaev, S., Olejnik, J., Olejnik, E. K., and Rothschild, K. J. (2004) Cell-free N-terminal protein labeling using initiator suppressor tRNA. Anal. Biochem. 326, 25−32. (22) Heckler, T. G., Roesser, J. R., Xu, C., Chang, P. I., and Hecht, S. M. (1988) Ribosomal binding and dipeptide formation by misacylated tRNA(Phe),S. Biochemistry 27, 7254−7262. (23) Goto, Y., Murakami, H., and Suga, H. (2008) Initiating translation with D-amino acids. RNA 14, 1390−1398. (24) Muranaka, N., Miura, M., Taira, H., and Hohsaka, T. (2007) Incorporation of unnatural non-alpha-amino acids into the N terminus of proteins in a cell-free translation system. ChemBioChem 8, 1650− 1653. (25) Miura, M., Muranaka, N., Abe, R., and Hohsaka, T. (2010) Incorporation of fluorescent-labeled non-α-amino carboxylic acids into the N-terminus of proteins in response to amber initiation codon. Bull. Chem. Soc. Jpn. 83, 546−553. (26) Hecht, S. M. (1992) Probing the synthetic capabilities of a center of biochemical catalysis. Acc. Chem. Res. 25, 545−552. (27) Forster, A. C., Weissbach, H., and Blacklow, S. C. (2001) A simplified reconstitution of mRNA-directed peptide synthesis: activity of the epsilon enhancer and an unnatural amino acid. Anal. Biochem. 297, 60−70. (28) Shimizu, Y., Inoue, A., Tomari, Y., Suzuki, T., Yokogawa, T., Nishikawa, K., and Ueda, T. (2001) Cell-free translation reconstituted with purified components. Nat. Biotechnol. 19, 751−755. (29) Kawakami, T., and Murakami, H. (2012) Genetically encoded libraries of nonstandard peptides. J. Nucleic Acids 2012, 713510. (30) Kawakami, T., Ishizawa, T., Fujino, T., Reid, P. C., Suga, H., and Murakami, H. (2013) In vitro selection of multiple libraries created by genetic code reprogramming to discover macrocyclic peptides that antagonize VEGFR2 activity in living cells. ACS Chem. Biol. 8, 1205− 1214. (31) Uzawa, T., Tada, S., Wang, W., and Ito, Y. (2013) Expansion of the aptamer library from a ″natural soup″ to an ″unnatural soup″. Chem. Commun. 49, 1786−1795. (32) Wang, J., Kwiatkowski, M., and Forster, A. C. (2015) Kinetics of Ribosome-Catalyzed Polymerization Using Artificial Aminoacyl-tRNA Substrates Clarifies Inefficiencies and Improvements. ACS Chem. Biol. 10, 2187−2192. (33) Frankel, A., Millward, S. W., and Roberts, R. W. (2003) Encodamers: unnatural peptide oligomers encoded in RNA. Chem. Biol. 10, 1043−1050. (34) Merryman, C., and Green, R. (2004) Transformation of aminoacyl tRNAs for the in vitro selection of ″drug-like″ molecules. Chem. Biol. 11, 575−582. (35) Tan, Z., Forster, A. C., Blacklow, S. C., and Cornish, V. W. (2004) Amino acid backbone specificity of the Escherichia coli translation machinery. J. Am. Chem. Soc. 126, 12752−12753. (36) Millward, S. W., Fiacco, S., Austin, R. J., and Roberts, R. W. (2007) Design of cyclic peptides that bind protein surfaces with antibody-like affinity. ACS Chem. Biol. 2, 625−634. (37) Zhang, B., Tan, Z., Dickson, L. G., Nalam, M. N., Cornish, V. W., and Forster, A. C. (2007) Specificity of translation for N-alkyl amino acids. J. Am. Chem. Soc. 129, 11316−11317. (38) Kawakami, T., Murakami, H., and Suga, H. (2008) Messenger RNA-programmed incorporation of multiple N-methyl-amino acids into linear and cyclic peptides. Chem. Biol. 15, 32−42. (39) Kawakami, T., Murakami, H., and Suga, H. (2008) Ribosomal synthesis of polypeptoids and peptoid-peptide hybrids. J. Am. Chem. Soc. 130, 16861−16863. (40) Sando, S., Masu, H., Furutani, C., and Aoyama, Y. (2008) Enzymatic N-methylaminoacylation of tRNA using chemically misacylated AMP as a substrate. Org. Biomol. Chem. 6, 2666−2668.

15KT0071 to T.K.), Grant-in-Aid for Young Scientists (B; 25870146 to T.K.) and Grant-in-Aid for Challenging Exploratory Research (15K13751 to T.K.).



REFERENCES

(1) Chattapadhyay, R., Pelka, H., and Schulman, L. H. (1990) Initiation of in vivo protein synthesis with non-methionine amino acids. Biochemistry 29, 4263−4268. (2) Varshney, U., and RajBhandary, U. L. (1990) Initiation of protein synthesis from a termination codon. Proc. Natl. Acad. Sci. U. S. A. 87, 1586−1590. (3) Kudlicki, W., Odom, O. W., Kramer, G., and Hardesty, B. (1994) Chaperone-dependent folding and activation of ribosome-bound nascent rhodanese. Analysis by fluorescence. J. Mol. Biol. 244, 319− 331. (4) Olejnik, J., Gite, S., Mamaev, S., and Rothschild, K. J. (2005) Nterminal labeling of proteins using initiator tRNA. Methods 36, 252− 260. (5) Heckler, T. G., Chang, L. H., Zama, Y., Naka, T., Chorghade, M. S., and Hecht, S. M. (1984) T4 RNA ligase mediated preparation of novel ″chemically misacylated″ tRNAPheS. Biochemistry 23, 1468− 1473. (6) Kwiatkowski, M., Wang, J., and Forster, A. C. (2014) Facile synthesis of N-acyl-aminoacyl-pCpA for preparation of mischarged fully ribo tRNA. Bioconjugate Chem. 25, 2086−2091. (7) Pallanck, L., and Schulman, L. H. (1991) Anticodon-dependent aminoacylation of a noncognate tRNA with isoleucine, valine, and phenylalanine in vivo. Proc. Natl. Acad. Sci. U. S. A. 88, 3872−3876. (8) Li, S., Kumar, N. V., Varshney, U., and RajBhandary, U. L. (1996) Important role of the amino acid attached to tRNA in formylation and in initiation of protein synthesis in Escherichia coli. J. Biol. Chem. 271, 1022−1028. (9) Wu, X. Q., Iyengar, P., and RajBhandary, U. L. (1996) Ribosomeinitiator tRNA complex as an intermediate in translation initiation in Escherichia coli revealed by use of mutant initiator tRNAs and specialized ribosomes. EMBO J. 15, 4734−4739. (10) Goto, Y., Ohta, A., Sako, Y., Yamagishi, Y., Murakami, H., and Suga, H. (2008) Reprogramming the translation initiation for the synthesis of physiologically stable cyclic peptides. ACS Chem. Biol. 3, 120−129. (11) Josephson, K., Hartman, M. C. T., and Szostak, J. W. (2005) Ribosomal synthesis of unnatural peptides. J. Am. Chem. Soc. 127, 11727−11735. (12) Seebeck, F. P., and Szostak, J. W. (2006) Ribosomal synthesis of dehydroalanine-containing peptides. J. Am. Chem. Soc. 128, 7150− 7151. (13) Hartman, M. C., Josephson, K., Lin, C. W., and Szostak, J. W. (2007) An expanded set of amino acid analogs for the ribosomal translation of unnatural peptides. PLoS One 2, e972. (14) Matsubara, T., Iijima, K., Watanabe, T., Hohsaka, T., and Sato, T. (2013) Incorporation of glycosylated amino acid into protein by an in vitro translation system. Bioorg. Med. Chem. Lett. 23, 5634−5636. (15) Horiya, S., Bailey, J. K., Temme, J. S., Guillen Schlippe, Y. V., and Krauss, I. J. (2014) Directed evolution of multivalent glycopeptides tightly recognized by HIV antibody 2G12. J. Am. Chem. Soc. 136, 5407−5415. (16) Roesser, J. R., Chorghade, M. S., and Hecht, S. M. (1986) Ribosome-catalyzed formation of an abnormal peptide analogue. Biochemistry 25, 6361−6365. (17) Gite, S., Mamaev, S., Olejnik, J., and Rothschild, K. (2000) Ultrasensitive fluorescence-based detection of nascent proteins in gels. Anal. Biochem. 279, 218−225. (18) McIntosh, B., Ramachandiran, V., Kramer, G., and Hardesty, B. (2000) Initiation of protein synthesis with fluorophore-Met-tRNA(f) and the involvement of IF-2. Biochimie 82, 167−174. (19) Ramachandiran, V., Willms, C., Kramer, G., and Hardesty, B. (2000) Fluorophores at the N terminus of nascent chloramphenicol H

DOI: 10.1021/acschembio.5b01014 ACS Chem. Biol. XXXX, XXX, XXX−XXX

Articles

ACS Chemical Biology (41) Subtelny, A. O., Hartman, M. C., and Szostak, J. W. (2008) Ribosomal synthesis of N-methyl peptides. J. Am. Chem. Soc. 130, 6131−6136. (42) Kawakami, T., Ohta, A., Ohuchi, M., Ashigai, H., Murakami, H., and Suga, H. (2009) Diverse backbone-cyclized peptides via codon reprogramming. Nat. Chem. Biol. 5, 888−890. (43) Pavlov, M. Y., Watts, R. E., Tan, Z., Cornish, V. W., Ehrenberg, M., and Forster, A. C. (2009) Slow peptide bond formation by proline and other N-alkylamino acids in translation. Proc. Natl. Acad. Sci. U. S. A. 106, 50−54. (44) Subtelny, A. O., Hartman, M. C., and Szostak, J. W. (2012) Optimal codon choice can improve the efficiency and fidelity of Nmethyl amino acid incorporation into peptides by in-vitro translation. Angew. Chem., Int. Ed. 50, 3164−3167. (45) Kawakami, T., Ishizawa, T., and Murakami, H. (2013) Extensive reprogramming of the genetic code for genetically encoded synthesis of highly N-alkylated polycyclic peptidomimetics. J. Am. Chem. Soc. 135, 12297−12304. (46) Ieong, K. W., Pavlov, M. Y., Kwiatkowski, M., Ehrenberg, M., and Forster, A. C. (2014) A tRNA body with high affinity for EF-Tu hastens ribosomal incorporation of unnatural amino acids. RNA 20, 632−643. (47) Kawakami, T., Sasaki, T., Reid, P. C., and Murakami, H. (2014) Incorporation of electrically charged N-alkyl amino acids into ribosomally synthesized peptides via post-translational conversion. Chem. Sci. 5, 887−893. (48) Wang, J., Kwiatkowski, M., Pavlov, M. Y., Ehrenberg, M., and Forster, A. C. (2014) Peptide formation by N-methyl amino acids in translation is hastened by higher pH and tRNA(Pro). ACS Chem. Biol. 9, 1303−1311. (49) Yamagishi, Y., Shoji, I., Miyagawa, S., Kawakami, T., Katoh, T., Goto, Y., and Suga, H. (2011) Natural product-like macrocyclic Nmethyl-peptide inhibitors against a ubiquitin ligase uncovered from a ribosome-expressed de novo library. Chem. Biol. 18, 1562−1570. (50) Goshima, N., Kawamura, Y., Fukumoto, A., Miura, A., Honma, R., Satoh, R., Wakamatsu, A., Yamamoto, J., Kimura, K., Nishikawa, T., Andoh, T., Iida, Y., Ishikawa, K., Ito, E., Kagawa, N., Kaminaga, C., Kanehori, K., Kawakami, B., Kenmochi, K., Kimura, R., Kobayashi, M., Kuroita, T., Kuwayama, H., Maruyama, Y., Matsuo, K., Minami, K., Mitsubori, M., Mori, M., Morishita, R., Murase, A., Nishikawa, A., Nishikawa, S., Okamoto, T., Sakagami, N., Sakamoto, Y., Sasaki, Y., Seki, T., Sono, S., Sugiyama, A., Sumiya, T., Takayama, T., Takayama, Y., Takeda, H., Togashi, T., Yahata, K., Yamada, H., Yanagisawa, Y., Endo, Y., Imamoto, F., Kisu, Y., Tanaka, S., Isogai, T., Imai, J., Watanabe, S., and Nomura, N. (2008) Human protein factory for converting the transcriptome into an in vitro-expressed proteome. Nat. Methods 5, 1011−1017. (51) Maruyama, Y., Kawamura, Y., Nishikawa, T., Isogai, T., Nomura, N., and Goshima, N. (2012) HGPD: Human Gene and Protein Database, 2012 update. Nucleic Acids Res. 40, D924−929. (52) Natsume, T., Yamauchi, Y., Nakayama, H., Shinkawa, T., Yanagida, M., Takahashi, N., and Isobe, T. (2002) A direct nanoflow liquid chromatography-tandem mass spectrometry system for interaction proteomics. Anal. Chem. 74, 4725−4733. (53) Adachi, S., Homoto, M., Tanaka, R., Hioki, Y., Murakami, H., Suga, H., Matsumoto, M., Nakayama, K. I., Hatta, T., Iemura, S., and Natsume, T. (2014) ZFP36L1 and ZFP36L2 control LDLR mRNA stability via the ERK-RSK pathway. Nucleic Acids Res. 42, 10037− 10049. (54) Guillen Schlippe, Y. V., Hartman, M. C., Josephson, K., and Szostak, J. W. (2012) In vitro selection of highly modified cyclic peptides that act as tight binding inhibitors. J. Am. Chem. Soc. 134, 10469−10477. (55) White, E. R., Sun, L., Ma, Z., Beckta, J. M., Danzig, B. A., Hacker, D. E., Huie, M., Williams, D. C., Edwards, R. A., Valerie, K., Glover, J. N., and Hartman, M. C. (2015) Peptide library approach to uncover phosphomimetic inhibitors of the BRCA1 C-terminal domain. ACS Chem. Biol. 10, 1198−1208.

(56) Heinis, C., Rutherford, T., Freund, S., and Winter, G. (2009) Phage-encoded combinatorial chemical libraries based on bicyclic peptides. Nat. Chem. Biol. 5, 502−507. (57) Angelini, A., Cendron, L., Chen, S., Touati, J., Winter, G., Zanotti, G., and Heinis, C. (2012) Bicyclic peptide inhibitor reveals large contact interface with a protease target. ACS Chem. Biol. 7, 817− 821. (58) Kawakami, T., Ogawa, K., Goshima, N., and Natsume, T. (2015) DIVERSE System: De Novo Creation of Peptide Tags for Nonenzymatic Covalent Labeling by In Vitro Evolution for Protein Imaging Inside Living Cells. Chem. Biol. 22, 1671−1679. (59) Iemura, S., and Natsume, T. (2012) One-by-One Sample Preparation Method for Protein Network Analysis in Protein Interactions, (Cai, J., Ed.), pp 293−310, InTech. (60) Doi, T., Yoshida, M., Ohsawa, K., Shinya, K., Takagi, M., Uekusa, Y., Yamaguchi, T., Kato, K., Hirokawa, T., and Natsume, T. (2014) Total synthesis and characterization of thielocin B1 as a protein−protein interaction inhibitor of PAC3 homodimer. Chem. Sci. 5, 1860−1868.

I

DOI: 10.1021/acschembio.5b01014 ACS Chem. Biol. XXXX, XXX, XXX−XXX