A High Performance Platform Based on cDNA Display for Efficient

Research Article pubs.acs.org/acscombsci

A High Performance Platform Based on cDNA Display for Efficient Synthesis of Protein Fusions and Accelerated Directed Evolution Mohammed Naimuddin*,†,‡ and Tai Kubo*,†,§ †

Biomedical Research Institute, National Institute of Advanced Industrial Science and Technology, Central 6, 1-1-1 Higashi, Tsukuba, Ibaraki 305-8566, Japan ‡ Janusys Corporation, #508, Saitama Industrial Technology Center, Skip City, 3-12-18 Kami-Aoki, Kawaguchi, Saitama 333-0844, Japan § Molecular Profiling Research Center for Drug Discovery, National Institute of Advanced Industrial Science and Technology (AIST), 2-4-7 Aomi, Koto-ku, Tokyo 135-0064, Japan S Supporting Information *

ABSTRACT: We describe a high performance platform based on cDNA display technology by developing a new modified puromycin linker-oligonucleotide. The linker consists of four major characteristics: a “ligation site” for hybridization and ligation of mRNA by T4 RNA ligase, a “puromycin arm” for covalent linkage of the protein, a “polyadenosine site” for a longer puromycin arm and purification of protein fusions (optional) using oligo-dT matrices, and a “reverse transcription site” for the formation of stable cDNA protein fusions whose cDNA is covalently linked to its encoded protein. The linker was synthesized by a novel branching strategy and provided >8-fold higher yield than previous linkers. This linker enables rapid and highly efficient ligation of mRNA (>90%) and synthesis of protein fusions (∼50−95%) in various cell-free expression systems. Overall, this new cDNA display method provides 10−200 fold higher end-usage fusions than previous methods and benefits higher diversity libraries crucial for directed protein/peptide evolution. With the increased efficiency, this system was able to reduce the time for one selection cycle to 95 not required

1:1−1:1.5 T4 RNA ligase 10 min >95 not required

1:1−1:4 T4 RNA ligase 60 min >90 required

1:200 T4 RNA ligase 2400 min 80% required

1:1 DNA ligase 60 min 20% required

≤65% ≤95% ∼60%

≤30% ≤30% 1%e

≤20% ND 0.2%

≤10% ≤70% ∼10%

≤40% ND ∼6.5%

a

Mochizuki et al.21 and Ueno et al.31 bYamaguchi et al.13 and Naimuddin et al.20 cMiyamoto-Sato et al.;26 dLiu et al.12 and Takahashi et al.14 eThis value represents the yield before purification via the C-terminal tags. The final fusion yield was not reported.

Figure 2. Formation of protein fusions in different cell-free lysates and purification. (a) Three linkers were prepared to test the efficiency for the formation of protein fusions in two lysates. These linkers differ in the repeats of adenosine residues. Linker-N contains 18 A, linker-N-10A contains 10 A, and linker-N-0A contains zero A residues. (b) Protein fusions in rabbit reticulocyte lysate. Proteins that form covalent fusion with puromycin are detected as a shifted band (indicated as mRNA-linker-protein) compared to mRNA-linker band of lower MW. Efficiency of fusion formation using the three linker-oligonucleotides is shown in the graph “RRL”. (c) Protein fusions in wheat germ lysate. Translation was performed for 10 min followed by incubation with high salts for 30 min to 1 h (see Experimental Section for details) using the four templates (PDO, BDA, 3F, and 6R14); 0 denotes the fraction sampled after 10 min of translation; 1, fraction incubated for 1 h without high salts; 2, fraction incubated for 30 min in high salts, and 3, for 1 h. (d) Performance of the fusion formation of other linker-oligonucleotides using the four templates. The left panel is for linker-N10A and the right panel for linker-N-0A. Fusion formation was performed for 1 h, and % fusion formation with the three linkers is shown in the graph indicating “WGE”. (e) One-step purification of His-tagged protein fusions by IMAC for separation from the untranslated mRNA-linker. F.T., flowthrough; E1, eluate 1; and E2, eluate 2. All fractions (F.T.−E2) were reverse transcribed and digested with RNase H to remove mRNA. PDO correspond to mRNA-protein-linker. cDNA-protein-linker bands migrate faster than mRNA-protein-linker. The bands have been adjusted from the same gel to show the purification of protein fusions. Error bars were calculated from three independent experiments.

use of ribonuclease that might create problems for controlling contamination for mRNA, although it is used after reverse transcription. cDNA Display. Translation and maturation into protein fusions is another important area that needs attention as this directly determines the library size and can influence downstream purification processes. In various cell-free lysates, the template mRNA-(linker-N) conjugates of PDO, BDA, 3F, and 6R14 were able to generate protein fusions very efficiently. Fusion formation was achieved at 50−65% in rabbit reticulocyte lysate (RRL) (Figure 2b), and 85−95% fusion formation was observed in wheat germ lysate (WGE) in all four

templates tested (Figure 2c). The presence of high salts (MgCl2 and KCl) for fusion maturation was an essential requirement for both lysates tested. In WGE, ∼20−30% fusion formation was observed in the absence of salts for an incubation time of 1 h (lane ‘1’, Figure 2c). However, an increase in fusion yields was evident with the inclusion of high salts and incubation for 30 min to 1 h (lanes ‘2’ and ‘3’). In general, even 30 min for maturation in high salts was sufficient to achieve 50−70% fusions. The intactness of mRNA-linker and mRNA-proteinlinker was calculated by the additive values of the individual band intensities and compared with the mRNA-linker of lane ‘0’. A benchmark of 0.9 (absolute value of 1 for lane ‘0’, i.e., D

DOI: 10.1021/acscombsci.5b00139 ACS Comb. Sci. XXXX, XXX, XXX−XXX

Research Article

ACS Combinatorial Science

Figure 3. Schematic for the formation of protein fusions with different linkers. (A) Translation (without linkers). Proteins are synthesized by ribosomes by reading out the mRNA framework. Ribosomes of eukaryotic origin have an average height of 25−30 nm. (b) Protein fusion with purolinker/SBP/L-N-0A. The short linkers (∼10−20 nts; 7 nm long) have been shown with respect to the length of ribosomes (assuming 15−20 nm of length from mRNA). Spacer is shown in orange, and puromycin (red) is at the top. The puromycin forms a covalent bond with the protein at the Psite of the ribosome and releases the protein. (c) Protein fusion with linker-N. Formation of fusion with the longer puromycin arm of linker-N (∼40 nts; 12−14 nm) consisting of oligo-dA (shown in green). (d) Depending on the position of ribosomes in solution after protein synthesis, linker-N with the oligo-dA “stem” and flexible spacers may provide the required flexibility to form a structure that may increase the probability of the formation of more fusions.

apparent difference observed between RRL and WGE.31 However, in an earlier report, higher efficiency of fusion formation (∼70%) was reported for WGE.26 Thus, the design of linker is important in that it can play a critical role in the efficiency of protein fusion formation plausibly based on the exploitation of ribosome concentration that is responsible for protein synthesis. Next, we investigated the effect of the length of the puromycin arm by modulating the number of adenosine residues. We constructed two more linkers: one that contained no adenosine residues (L-N-0A) and another that contained ten adenosine residues (L-N-10A) (Figure 2a). The formation of protein fusions in RRL was found to be ∼20% (L-N-0A) and ∼30% (L-N-10A) (Figure 2b). In WGE, fusion formation was ∼30% (L-N-0A) and ∼60% (L-N-10A) (Figure 2d). These experiments indicate that the number of repeats of adenosine plays a role in the formation of protein fusions and directly correlates with the difference in the efficiencies of the production of protein fusions. Furthermore, given the constant concentration of ribosomes in WGE, the three different linkers tested here have generated varying amounts of protein fusions. These results emphasize that the design of linker is important to probe the efficiencies of production of protein fusions. It may be noted that L-N-0A, which is similar to puro-linker,13 and SBP,21,31 as these linkers are devoid of oligo-dA, provided consistent efficiency of protein fusion formation (20−30%) in the two lysates.31 By considering the physical parameter of the linkers (with/without oligo-dA) and the ribosome, the difference in the efficiencies of the linkers may be partially explained. Eukaryotic ribosomes have an average height of 25−

mRNA-linker) was set, which indicates 90% of the mRNAlinkers and mRNA-protein-linkers were not degraded. Most lanes were in line with the benchmark value. Because of the presence of a fluorescein moiety for detection, some artifacts may be observed as FITC is influenced by several factors, such as exposure to light, ionic strength of buffers, and so forth. Other fluorescent moieties, such as Alexa and others that are not easily influenced by the environment, including buffers, may be used in future. We reasoned that the design of a longer puromycin linker arm (i.e., inclusion of repeats of adenosine residues) may have contributed to the higher yields achieved here.25 Several papers have been published that report the synthesis of ∼5−40% protein fusions in RRL12−14,25 and ∼70% in WGE using a single template.26 Our results show templateindependent higher fusion yields with various templates (Figure 2; Table 2). The difference in the fusion formation with two different lysates can be partially explained based on the number of ribosomes that are more than one order higher in WGE (2.5 × 1015 per ml or 4 μM) compared with RRL (1.2 × 1014 per ml or 0.2 μM).8,28−30 The ribosome and mRNA-linker ratio is 1:1−2:1 (0.1−0.2 μM ribosome and 0.1 μM mRNA-linker) in the case of RRL. The ratio is approximately 40:1 in the case of WGE. The near equimolar ratio in RRL may not be sufficient for higher fusion yields given the “single-turnover” in display technologies due to the absence of a stop codon.10−14 On the other hand, the high ratio in WGE may provide sufficient ribosomes for higher fusion yields obtained here. It is noteworthy that linker-N was also efficient in providing the distinction between the efficiencies of the two translation systems (RRL and WGE). In a recent report, there was no E


Research Article


Figure 4. cDNA display with linker-N and its demonstration for higher throughput of in vitro selection. (a) Schematic of the steps essential for performing cDNA display using wheat germ lysate. DNAs generated by PCR are transcribed to mRNAs, which are purified and quantified. mRNAs are then rapidly ligated to linker-N, translated, and matured to display proteins in high salt conditions. Purification is achieved by Ni-NTA magnetic beads, and cDNAs are synthesized on beads, eluted, and subjected to selection against the targets that are immobilized on streptavidin (SA) matrix. The selected/binding molecules are eluted and amplified for analysis. The schematic for selection using PDO/BDA mixture screened for IgG is shown. The time required for each step is indicated in green boxes, and the whole process can be completed in 6−8 h. (b) Selection and analysis by PCR. Four model proteins were prepared, and three parallel selections were performed according to the steps outlined in (a). Mixtures of a nonbinder and a binder (ratio of 20:1) were translated, processed, and subjected to selection against various targets (1−3). (1), IgG that binds to BDA; (2), AChBP that binds to 3F, and (3), IL6-R whose ligand is 6R14. Selection efficiency was monitored by the change in the ratio of “before” and “after” selection fractions and expressed as the “number of fold enrichment” shown in (c). Error bars were calculated from three independent experiments.

30 nm (29.4 nm for wheatgerm) and an average width of 27− 31 nm.32,33 Linker-N, which is approximately 40 nucleotides, has a length of 12−14 nm, and puro-linker/SBP/linker-N-0A is approximately 7 nm long (∼10−20 nts).31 Linker-N with a longer puromycin arm is geometrically better positioned to form protein fusions compared with other short linkers by approaching the tunnel of ribosome (assuming that the length of the ribosome is approximately 15−20 nm from mRNA out of the height of 29.4 nm for wheatgerm), as shown schematically in Figure 3a−c. On the basis of these observations, we present a model where the “stem” portion (polyadenosine) provides the required length and presumably stability to the flexible spacers (assuming oligo-dA has a rigid DNA structure) and increases the efficiency of the formation of

protein fusions (Figure 3d); however, the stability provided by oligo-dA requires further experimental investigation. For selection experiments, synthesis of cDNA has been suggested to be beneficial to reduce mRNA degradation and formation of secondary structures of mRNA, especially when libraries are used for selection.12−14,20,27 For performing reverse transcription, the protein fusions require purification for separation from the lysate and also to remove the stalled ribosomes (due to the absence of a stop codon). Furthermore, protein fusions also need to be purified utilizing the C-terminal tags to isolate full-length protein by eliminating proteins generated by premature termination codons in libraries. In general, two-step purification has been used in mRNA and cDNA displays. The incorporation of multiple purification steps F


Research Article


Figure 5. Selection of the complex mixture doped with Flag-tag. (a) Constructs adopted for the preparation of the complex mixture. The Flag-tag construct contained the residues of Flag-tag (DYKDDDDK) necessary for its binding to the anti-Flag M2 antibody flanked by the components required for cDNA display with linker-N. The 8-mer library contained random residues flanked by components for cDNA display. (b) Selection cycle. A complex mixture containing 8-mer library (nonbinder) was doped with 0.1% Flag-tag (binder) and subjected to selection cycle procedures of transcription, ligation to linker-N, translation in WGE to synthesize mRNA fusion proteins, purification utilizing the His-tag and reverse transcription to synthesize cDNA fusion proteins, selection on anti-Flag M2 antibody matrix, and PCR amplification. The amplified product undergoes the same process for 2 additional cycles and is then subjected to cloning and sequencing of the clones. The same complex mixture was also subjected to selection using procedures with “puro-linker”.13 (c) Summary of the selection result. The sequences of the clones of the complex mixture were obtained at R0 (prior to selection) and R3 (after 3 selection cycles) for both linkers. The % population indicates the fraction of Flag-tag sequences in the total number of sequences analyzed (given in parentheses).

often results in reduced yields in the range of 0.2−6.5%.12−14 Because of the formation of higher protein fusions, we have used gentle one-step IMAC purification via His-tag and performed reverse transcription. An estimated 60−80% of purified cDNA protein fusions were found to be isolated from lysates in less than 30 min (Figure 2e). Thus, taking into consideration all the steps starting from fusion formation to end-usage purified cDNA fusions, our method provides approximately ∼10−200 fold higher fusions than previously reported methods (Table 2).12,13 Higher yields of protein fusions will directly influence the diversity of proteins and increase the landscape of available proteins that may benefit in

vitro selection and evolution of proteins/peptides.34 Furthermore, the reduction of steps, minimizing purifications, and the utilization of magnetic matrices makes this system potentially amenable and friendly to high-throughput platforms. In Vitro Selection. The cDNA display proteins synthesized with linker-N were evaluated for performance in affinity selection experiments using various known targets according to the process outlined in Figure 4a. The whole process of one cycle of selection can be completed in an accelerated manner requiring 6−8 h. Two ligands that do not contain disulfide linkages (PDO and BDA)22,23 and two ligands that require disulfide linkages for the folded structure to maintain their G


Research Article


Figure 6. Analysis of the selection of the 3F library for VEGF. (a) The three-finger scaffold contains three fingers protruding from the globular head, and this topology is provided by the four disulfide linkages (shown in green). The randomization of the tips of loops facilitates preparation of library (shown by fluorescent green loops). The genetic construct for the preparation of the library is shown below. (b) Analysis of the R7 library by binding assays. The 3F library (cDNA fusions) after seven rounds of selection (R7) was subjected to a binding assay to evaluate its potential. The three fractions were allowed to bind to (1) streptavidin beads (SA; matrix negative control), (2) VEGF beads, and (4) Fc beads (protein negative control). The fourth fraction was the reverse transcribed R7 library (without fusion) to evaluate binding via the nucleic acid portion (3). The eluted fractions were PCR amplified and analyzed by gel electrophoresis. The left panel represents a schematic of the assay, and the right panel contains the gel electrophoresis profiles. (c) The three fractions, (1), (2), and (4), were subjected to the binding assay and detected using ELISA (see Experimental Section). (d) Summary of the sequences and affinities of clones. The sequences were divided into three groups based on the frequency of occurrence (given in parentheses). The affinities were measured by ELISA. Error bars were calculated from three independent experiments.

function (3F and 6R14, derived from a 3F library)20 were chosen for selection. PDO binds to DNA (OBD) and was chosen as a nonbinder to assess the selection efficiency. BDA binds to IgG,22 3F binds to AChBP, and 6R14 has affinity for IL-6R.20 The efficiency of selection was expressed as number of fold enrichment and found to be 30−40 for the three ligand/ receptor pairs after a single selection cycle (Figure 4b and c). These values are comparable to or better than those in previous reports.10,13 Thus, protein ligands (both disulfide-deficient and -rich) synthesized by the cDNA display method with linker-N can bind to diverse targets, such as macroprotein (IgG), soluble channel receptor (AChBP), and soluble signal transducing receptor (IL-6R). Selection from Complex Mixtures. An important aspect and final goal of display technologies is to select desired candidates (e.g., for binding) from undesirable (nonbinding) candidates contained in a library of molecules. We evaluated the performance of our method using (a) a doped mixture of known ligand (binder) and library molecules (nonbinder) for the natural target and (b) totally random libraries for a given target. We chose Flag-tag (known ligand; binder) for the purpose of preparing doped mixtures with a library of 8-mer peptide library (nonbinder) mixed at a ratio of 1:1000 (Flag-tag: library; or 0.1% Flag-tag) (Figure 5a). Please note that any nonbinding

peptide of comparable size can be used for this experiment. For comparison, the whole process was also performed with “purolinker”.13 Selection was carried out against immobilized antiFlag M2 antibody (natural target for Flag-tag) for three rounds simultaneously with both methods (Figure 5b). After selection, the pools were cloned and sequenced. The flag-tag sequences were 80% (16 of 20 clones) after 3 rounds of selection that utilized linker-N up from 0.1% compared with 20% that were observed with “puro-linker” (Figure 5c; see Supporting Table 1). It may be noted that similar results (17% sequences) were obtained with “puro-linker” in the previous report.13 These results demonstrate that, with linker-N, improved enrichment (4-fold) of Flag-tag sequences for its natural target, anti-Flag M2 antibody, was obtained relative to that of “puro-linker”. We speculate that, in the case of linker-N, the increased formation of protein fusions with a reduced “leftover” template mRNAlinker has facilitated improved selection. The leftover mRNAlinker is significantly higher in the case of “puro-linker” (∼80%) compared with the protein fusions (∼20%)13,20,21 and is difficult to purify completely even after utilizing protein purification via his-tag. The untranslated mRNA-linker converted cDNA-linker can potentially reduce the S/N ratio during PCR amplification by increasing the “noise” (i.e., unwanted/nonbinder DNA species) and interfere with the outcome of the selection process. H


Research Article

ACS Combinatorial Science Table 3. Comparison of Performance of Puromycin Linkers for cDNA Display

a

linker type

library/ml

library used

selection rounds

major group

highest affinity (Kd)

linker-Na puro-linkerb

6 × 1012 4 × 1010

3 × 1012 1.2 × 1011

7 (7 days) 10 (30 days)

60% 70%

2.1 ± 0.5 48 ± 8

This paper. bNaimuddin et al.20

process. Another interesting result was observed in the highest affinity molecule generated in each method, which is higher by approximately 1 order of magnitude obtained with linker-N (2.1 vs 48 nM); affinities varied by approximately 2 orders of magnitude of the major population generated in each method (2.1 vs 152 nM).20 This is consistent with the theory that selection outcome is dependent on the diversity of the initial library; however, this needs more investigation.28,34 These results demonstrate the advantages that can be useful for cDNA display by employing linker-N to obtain superior candidates at far lesser cost and labor in an accelerated manner. VEGF binding molecules have been isolated from antibody and cysteine constrained libraries by using phage display with affinities in the range of 1.8−470 nM.35,36 The candidates generated in this study are of comparable affinities to the reported molecules (Figure 6d). In the recent past, there have been notable developments in cDNA display technology in terms of time required for the execution of each round of cDNA display/selection, improvements in ligation reactions, restriction digestion, and increased purified fusions for higher library size.31,37−41 The progress in the faster execution of ligation reactions from 1 h to 10 min to 30 s is highly notable; however, the use of ribonucleases and endonucleases should be avoided to refrain from any sort of contamination that may negatively affect the whole selection process.21,31,39 These methods require 2−3 days/round. In a previous report, the purification steps have been modified to increase the efficiency of the system by reducing the time of processing to 2 days/round.38 Trap display smartly encounters the process to tremendously reduce the execution time to 2.5 h/round;40,41 however, when using libraries, inclusion of purification steps for obtaining full-length cDNA fusions is important to avoid selection of prematurely terminated fusions that can be a potential source of noise. Alternatively, custom designed libraries that do not contain stop codons should be used to counter this problem; however, it is noteworthy that custom designed libraries are expensive to prepare. Furthermore, the untranslated mRNA-linker present in the unpurified mixtures (∼70−80%) will participate in the selection process by the formation of aptamers and thus interfere with the outcome of selection.11,12,20 Recently, cDNA display was also improved by the introduction of purification protocols that increased the fusions by approximately 10-fold that can yield libraries by a higher order.37 However, this method requires an execution time of 2−3 days/round.

Next, we explored the efficiency of linker-N with a totally random library to obtain affinity molecules for the cytokine VEGF. For comparison with an earlier report,20 we chose to use the 3F library (Figure 6a). We prepared a 3F library containing 3 × 1012 molecules and performed seven rounds of selection against immobilized VEGF. Stringency of selection conditions was applied in terms of reducing concentration of VEGF (1 μM to 1 nM) and increasing the number of washes. After selection, the library was analyzed by binding assays and sequencing of clones. In the binding assay analyzed by PCR, the R7 library protein fusions showed significant binding to VEGF (more than 2%; Figure 6b (2)) compared to a negative control that did not contain VEGF (less than 0.2%; Figure 6b (1)). There is a possibility of enrichment of the library via the nucleic acid aptamers, i.e., mRNA/cDNA hybrids. We prepared an R7 mRNA/cDNA hybrid by reverse transcription of the mRNAlinker and tested for binding potential. There was no binding observed for VEGF via the mRNA/cDNA hybrid (less than 0.2%; Figure 6b (3)). The R7 protein fusions also did not show affinity for Fc (less than 0.2%; Figure 6b (4)), indicating target (VEGF)-specific enrichment of the 3F library. In the binding assay analyzed by ELISA, the R7 protein fusions showed 5-fold higher signal for immobilized VEGF compared with the negative control that did not contain VEGF and Fc as a nonspecific protein (Figure 6c). This result also indicates that the library was enriched specific to VEGF. Analysis of the sequences of clones revealed that the candidates were of three major groups: G-1 at 60% and G-2 and G-3 at 15% each. The affinities of these candidates were measured by competitive ELISA method20,24 and found to be in the range of 2.1−35 nM (Figure 6d). The potential advantages of the cDNA display with linker-N over “puro-linker” are summarized in Table 3. The major advantage lies in the preparation of the library where 6 × 1012 molecules can be obtained per milliliter of lysate used with the 3F library compared with 4 × 1010 molecules20 owing to the higher protein fusion yields. This library was prepared by NNS codon (NA, T, C or G; SC or G) coding for random amino acids that also contained stop codons. We prepared 3 × 1012 per 0.5 mL in this study compared to 1.2 × 1011 per 3 mL used in the earlier paper.20 The most interesting difference was evident in the number of rounds of selection required that led to similar sequence convergence (60−70%), which is seven rounds in the case of linker-N compared with ten rounds in the case of “puro-linker” (linker-N is approximately 25−30% more efficient). Furthermore, it may be noted that by using same target (anti-Flag M2 antibody), similar higher efficiencies (4fold) were established by this system over the previous cDNA display method (Figure 5c). This may be attributable to the reduction in “noise” (unwanted and unconverted cDNA/ mRNA hybrids) in the case of linker-N owing to the formation of higher protein fusions. The time required for the whole selection process with linker-N is 7 days, whereas it takes around 30 days with other linkers, and this appears as a huge advantage in light of the need to accelerate the discovery

■

CONCLUSIONS We have developed a high performance and robust platform based on cDNA display reinforced by a new modified linker-N that enables rapid and highly efficient preparation of templates and protein fusions for utilization in directed evolution. The whole cycle can be accomplished in 6−8 h. We expect it to be beneficial for iterative cycles for ligand discovery and evolution. Ten rounds of selection can now be estimated in 2 weeks, which can be conservatively supposed to outpace even phage display-based methods (2−3 weeks) in addition to the cell-free I


Research Article


for 1 h at 37 °C to generate capped transcripts, purified by RNeasy purification kit (Qiagen), and quantified by absorbance at 260 nm. Prior to ligation, mRNAs were annealed to linker-N (ratio 1:1; 50 pmoles) by heating at 94 °C and gradient-cooling to 4 °C. Ligation was performed by the addition of 3 U T4 Kinase (NEB, USA) and 20 U of T4 RNA ligase (Takara, Japan) at 25 °C for 10, 20, and 40 min. Ligation was confirmed by digestion of the mRNA-linker conjugates with RNase H (Ambion, Austin, TX, USA) that cleaves RNA hybrids. Ligation efficiency was checked by denaturing polyacrylamide gel electrophoresis (6% gel and 8 M urea) using FITC and Sybr Gold (Molecular Probes, USA) staining on a fluoroimager (Bio-Rad, Hercules, CA, USA). Ligation efficiency was calculated by the equation % ligation = (A/A + B)×100, where A is the band intensity of the ligated product and B is the intensity of the remaining mRNA. cDNA Display. The mRNA-linker conjugates (100 nM) were translated in a Rabbit Reticulocyte Lysate IVT Kit (Ambion, Austin, TX, USA) (hereby abbreviated as RRL) and Wheat Germ Extract (Promega, Madison, WI, USA) (hereby abbreviated as WGE). In the case of RRL, translation was performed for 10 min at 30 °C, and protein fusions (i.e., covalent linking of protein to the puromycin moiety) were matured in the presence of 65 mM MgCl2 and 750 mM KCl at 37 °C for 2 h.13,20 In WGE, translation was performed for 10 min at 25 °C, and proteins fusions were matured by the addition of 65 mM MgCl2 and 750 mM KCl at 25 °C for 1 h. Sampling was done for electrophoretic analysis at 10 min (i.e., before addition of high salts), 30 min, and 1 h (after addition of high salts). For monitoring the effect of high salts, proteins fusions were also matured in the absence of high salts for 1 h. Before reverse transcription (RT), the protein fusions were purified by immobilization on Ni-NTA magnetic beads (Qiagen) for 10 min at 25 °C. The beads were washed with 1× PBS buffer, and RT was performed by the addition of 20 U of M-MLV reverse transcriptase (Takara, Japan) at 42 °C for 10 min and placed on ice. After several washings, fusions were eluted twice with 250 mM imidazole for 5 min at 25 °C. For electrophoresis confirmation, the samples were incubated with RNaseH (Ambion, Austin, TX, USA) at 37 °C for 30 min to remove the mRNA. Analyses of all the samples were performed by 12% SDS-PAGE containing 8 M urea and detected by FITC on a fluoroimager. Efficiency of fusion formation was calculated by the equation % fusion formation = (A/A + B)×100, where A is the band intensity of the protein fusion (mRNA-linkerprotein) and B is the intensity of the remaining mRNA-linker. Protein Labeling and Immobilization. Immunoglobulin G (IgG) was purchased from Sigma, interleukin-6 receptor (IL6R) and VEGF were from PeproTech (London, UK), acetylcholine binding protein (AChBP) was cloned from Aplysia kurodai [manuscript in preparation], and the recombinant protein was expressed in E. coli and purified. These proteins were biotinylated using EZ-Link Sulfo-NHS-SS-biotin (Pierce, Rockford, USA) at a molar ratio of 1:10. Unreacted biotin was inactivated with 1 M Tris-HCl (pH 8.0) and removed by dialysis for 16 h at 4 °C. Biotinylation of proteins was confirmed by SDS-PAGE and Western blotting using streptavidin-HRP (GE Healthcare). Quantitative immobilization was checked by titration of biotinylated proteins against a fixed amount of streptavidin-coated magnetic beads (Takara, Japan) for 30 min at 25 °C. The supernatant was stored, and the immobilized proteins were eluted with 100 mM

advantages. We consider this as a good starting point to achieve versatility if combined with high-throughput automated systems for rapid and exponential discovery of ligands and lead molecules. For laboratory-scale throughput, as demonstrated here, multiple samples can be processed in parallel using microtubes/microplates far less laboriously and cost-effectively. Potential applications can be in proteome exploration for mapping multiple biomolecular interactions and in drug discovery, an area that always requires better throughput technologies for accelerated discovery and optimization of biomolecules. Thus, we expect this cDNA display method to provide higher diversity libraries for searching novel ligands at reduced cost, time, and labor and to combine the intrinsic throughput with high-throughput systems to accelerate the discovery process. The recent developments, including the optimized linkers and methods,31,37−41 and higher-throughput systems42,43 are expected to provide necessary support and significantly impact discovery research.

■

EXPERIMENTAL SECTION Synthesis of Puromycin Linker. The puromycin linkeroligonucleotide (hereby designated as linker-N) installed with multifunctional groups, 5′-CCCCCCCGCCGCCCCCCG(5Me-dC)A18(Spec18) (Spec18) (Spec18)(F-dT) (Spec18)CC(Puro)-3′, was custom synthesized (BEX, Tokyo, Japan). The symbol 5-Me-dC denotes 5′-dimethoxytrityl-N4-(O-levulinyl-6oxyhexyl)-5-methyl-2′-deoxycytidine; Spec18, C18 spacer phosphoramidite; F-dT, fluorescein-dT; and Puro, puromycin CPG. Linker-N was synthesized by the DNA synthesizer ABI394 (Applied Biosystems, Japan), and the 5′-end of the linker was protected by acetylation. After synthesis, the levulinyl protecting group in the 5-Me-dC residue was removed by 0.5 M hydrazine hydrate in 1:1 pyridine/acetic acid, and the column was further washed by pyridine/acetic acid (1:1) and then by acetonitrile. The primer sequence 5′-CCTG-3′ was branched from the activated brancher 5-Me-dC in the linkeroligonucleotide. The product, i.e., primer-branched linkeroligonucleotide, was cleaved off from the column by K2CO3 in methanol followed by deprotection of the acetyl group. Additional deprotection was performed by reaction in 25% ammonium hydroxide for 1.5 h at 65 °C. These conditions were similar or milder than commonly used.44,45 The linkeroligonucleotide was purified by reversed-phase HPLC and confirmed by TOF-MS and gel electrophoresis. All phosphoramidite reagents used for modifications were from Glen Research (Sterling, VA, USA). Oligonucleotides, Vectors, and Template DNA Constructs. The B-domain of Protein A (BDA)22 was amplified from the pEZZ 18 protein A gene fusion vector (GE Healthcare) and Pou-domain of Oct-1 (PDO), according to the previous report.23 Three-finger (3F) was cloned from the South American coral snake Micrurus coralinus,20 and 6R14 was obtained from the pBADThio/TOPO vector IL-6R10-14.20 Custom oligonucleotides containing SP6 promoter, cap site, Xenopus β-globin untranslated sequence (UTR), and translation initiation codon (ATG) were added at the 5′ end and spacer (G3S)2, 6xHis, G3S, and Y-tag sequences were added at the 3′ end by PCR. Amplification was carried out by denaturation (94 °C) for 20 s, annealing (60 °C) for 15 s, and extension (72 °C) for 30 s. Transcription and Ligation of Linker-N. Purified DNAs were transcribed by SP6 RNA polymerase in the RiboMAX Large Scale Production System (Promega, Madison, WI, USA) J


Research Article

ACS Combinatorial Science dithiothreitol (DTT; Wako, Japan) for 10 min at 25 °C. The supernatant and eluted fractions were analyzed by SDS-PAGE to confirm the ratio of streptavidin-coated magnetic beads and proteins required for complete immobilization. In Vitro Selection. The mRNA-(linker-N) conjugates of a nonbinder and a binder were mixed at a ratio of 20:1, translated, and processed to cDNA display proteins. In the case of disulfide-rich ligands, such as 3F and 6R14, translation and maturation was carried out in the presence of protein disulfide isomerase (PDI) (1:1 molar ratio to display protein assuming that more than 70% fusion is formed with respect to input template), 10 mM oxidized glutathione (GSSG), and 1 mM reduced glutathione (GSH).24 DTT was excluded from all of the buffers. The His-tag eluted samples (50 μL) were mixed with 150 μL of selection buffer (SB; 1× PBS, 100 mM NaCl, and 0.1% Tween-20) and incubated with respective immobilized targets at 25 °C for 30 min. The beads were washed 10 times with 200 μL of the same buffer and eluted with 100 mM DTT and desalted (Microspin columns, GE Healthcare). The “before selection” and the eluted (“after selection”) fractions were amplified by PCR using 0.2 μM of the primers 5′ATTTAGGTGACACTATAGAATACAAGCTTGCT-3′ and 5′-TTTCCCCGCCGCCCCCCGTCCTGCTTCCGCCGTGATGAT-3′ for 25 cycles (denaturation (94 °C), 20 s; annealing (60 °C), 15 s; and elongation (72 °C), 30 s). Quantitative analysis was performed by denaturing gel electrophoresis (4.5% gel and 8 M urea), Sybr Gold staining, and fluoroimaging. Selection Using a Mixture of Flag-Tag and 8-Mer Peptide Library. An 8-mer peptide library was prepared by constructing the oligonucleotide fragment containing the random region flanked by complementary “overlap sequence” on the 3′ end and His-tag and Y-tag sequences on the 5′ end. The random region was coded by NNS codon (NA, T, C or G; SC or G). The other oligonucleotide fragment contained SP6 promotor and cap site at the 5′ end followed by 5′-UTR, Kozak, and ATG. The two fragments were joined by overlap PCR for 15 cycles of denaturation (94 °C) for 20 s, annealing (60 °C) for 15 s, and extension (72 °C) for 30 s. The Flag-tag construct was prepared in a similar manner except that the random region was replaced by Flag-tag sequence. Sequences were confirmed by cloning and sequencing. For selection experiments, 10 pmoles of 8-mer library mRNA-linker was mixed with 10 fmoles of Flag-tag mRNAlinker. Translation and maturation was performed in Wheat germ extract (WGE) and processed to cDNA protein fusions. For “puro-linker” the earlier protocol was followed.13 The cDNA protein fusions were mixed with 200 nM of anti-Flag M2 antibody (Sigma, USA) in 200 μL of SB and incubated for 30 min at 25 °C. The beads were washed 5 times (Round 1), 8 times (Round 2), and 10 times (Round 3) with an equal volume of SB and eluted with 100 mM DTT followed by desalting. The eluted products were PCR amplified and proceeded to the next round of selection for a total of 3 cycles. After Round 3, the PCR products were cloned (TA cloning, Invitrogen) and sequenced by random picking of the clones. Selection of 3F Library against VEGF. A 3F library was prepared according to the previous report.20 The library mRNA-(linker-N) (200 nM) was translated in 0.5 mL of lysate and processed to cDNA fusions. Estimation of cDNA fusions was performed by SDS-PAGE with a known amount of mRNAlinker as a concentration standard. An estimated 3 × 1012 molecules were obtained as cDNA fusions.

The cDNA fusions were incubated with immobilized VEGF for 1 h at 25 °C in 200 μL of SB followed by several washings with SB and WB (1× PBS, 100 mM NaCl, and 0.5% Tween20). The bound molecules were eluted with 100 mM DTT. The eluted fractions were desalted and PCR amplified (denaturation (94 °C) for 20 s, annealing (60 °C) for 15 s, and extension (72 °C) for 30 s). For applying selection pressure, the concentration of VEGF was varied from 1 μM to 1 nM, and the number of washes was progressively increased with WB. After Round 7 (R7), PCR products were cloned and sequenced. Binding Assays. R7 mRNA-(linker-N) (200 nM) was translated and processed to cDNA fusions. These fusions were divided into three parts and incubated with streptavidin beads, VEGF beads (200 nM), and Fc beads (200 nM) for 30 min at 25 °C in 200 μL of SB. In the fourth reaction, mRNA-(linkerN) was reverse transcribed, and an equivalent amount was incubated with VEGF beads. The beads were washed 10 times with the same buffer and eluted with 100 mM DTT. The eluted fractions were desalted and PCR amplified with the following conditions: denaturation (94 °C) for 20 s, annealing (60 °C) for 15 s, and extension (72 °C) for 30 s. The R7 fusion proteins were also analyzed by the ELISA method.24 R7 cDNA fusions were divided into 3 parts and incubated with SA beads, 200 nM VEGF beads, and 200 nM Fc beads in PBS-BSA (0.01% BSA) at 25 °C for 1 h. The mixtures were washed with PBS-T (0.1% Tween) and subsequently incubated with anti-His-antibody (R & D Systems) in PBS-T for 30 min at 25 °C. The mixtures were then washed several times with PBS-T followed by addition of the substrate, 3,3′,5,5′-tetramethylbenzidine (TMB; Sigma). After color development, the reaction was stopped by the addition of 0.5 M H2SO4; the samples were centrifuged, and absorbance was measured at 450 nm. Measurement of Dissociation Constant. Dissociation constants were measured according to the previously reported method.24 A constant amount of 3F display protein (∼1 nM) was incubated with varying amounts of VEGF (0.1 nM to 1 μM) for 1 h. The mixture was applied to a constant amount of immobilized VEGF (∼100 nM) and incubated further for 30 min. After several washings, anti-His-antibody (R & D Systems) was added at 1/2000 dilution and incubated for 1 h. The substrate TMB was added after several washings of the incubation mixture. After proper color development, the reaction was stopped by the addition of 0.5 M H2SO4, and absorbance was measured at 450 nm.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acscombsci.5b00139. Sequences of Round 0 (initial library) and Round 3 selection for anti-Flag M2 antibody (PDF)

■

AUTHOR INFORMATION

Corresponding Authors

*Tel: +81 48 2621247. Fax: +81 48 2621248. E-mail: [email protected]. *E-mail: [email protected]. Notes

The authors declare no competing financial interest. K


Research Article


■

(20) Naimuddin, M.; Kobayashi, S.; Tsutsui, C.; Machida, M.; Nemoto, N.; Sakai, T.; Kubo, T. Directed evolution of a three-finger neurotoxin by using cDNA display yields antagonists as well as agonists of interleukin-6 receptor signaling. Mol. Brain 2011, 4, 2. (21) Mochizuki, Y.; Biyani, M.; Tsuji-ueno, S.; Suzuki, M.; Nishigaki, K.; Husimi, Y.; Nemoto, N. One-pot preparation of mRNA/cDNA display by a novel and versatile puromycin-linker DNA. ACS Comb. Sci. 2011, 13, 478−485. (22) Gouda, H.; Torigoe, H.; Saito, A.; Sato, M.; Arata, Y.; Shimada, I. Three-dimensional solution structure of the B domain of staphylococcal protein A: comparisons of the solution and crystal structures. Biochemistry 1992, 31, 9665−9672. (23) Dekker, N.; Cox, M.; Boelens, R.; Verrijzer, C. P.; van der Vliet, P. C.; Kaptein, R. Solution structure of the Pou-specific domain of Oct-1. Nature 1993, 362, 852−855. (24) Naimuddin, M.; Kubo, T. Display of disulfide-rich proteins by complementary DNA display and disulfide shuffling assisted by protein disulfide isomerase. Anal. Biochem. 2011, 419, 33−39. (25) Kurz, M.; Gu, K.; Lohse, P. A. Psoralen photo-crosslinked mRNA-puromycin conjugates: a novel template for the rapid and facile preparation of mRNA-protein fusions. Nucleic Acids Res. 2000, 28, e83. (26) Miyamoto-Sato, E.; Takashima, H.; Fuse, S.; Sue, K.; Ishizaka, M.; Tateyama, S.; Horisawa, K.; Sawasaki, T.; Endo, Y.; Yanagawa, H. Highly stable and efficient mRNA templates for mRNA-protein fusions and C-terminally labeled proteins. Nucleic Acids Res. 2003, 31, e78. (27) Tabuchi, I.; Soramoto, S.; Nemoto, N.; Husimi, Y. An in vitro DNA virus for in vitro protein evolution. FEBS Lett. 2001, 508, 309− 312. (28) He, M.; Khan, F. Ribosome display: next-generation display technologies for production of antibodies in vitro. Expert Rev. Proteomics 2005, 2, 421−430. (29) Madin, K.; Sawasaki, T.; Ogasawara, T.; Endo, Y. A highly efficient and robust cell-free synthesis system prepared from wheat embryos: Plants apparently contain a suicide system directed at ribosomes. Proc. Natl. Acad. Sci. U. S. A. 2000, 97, 559−564. (30) Kopeina, G. S.; Afonina, Z. A.; Gromova, K. V.; Shirokov, V. A.; Vasiliev, V. D.; Spirin, A. S. Step-wise formation of eukaryotic doublerow polyribosomes and circular translation of polysomal mRNA. Nucleic Acids Res. 2008, 36, 2476−2488. (31) Ueno, S.; Kimura, S.; Ichiki, T.; Nemoto, N. Improvement of a puromycin-linker to extend the selection target varieties in cDNA display method. J. Biotechnol. 2012, 162, 299−302. (32) Verschoor, A.; Srivastava, S.; Grassucci, R.; Frank, J. Native 3D structure of eukaryotic 80s ribosome: morphological homology with E. coli 70S ribosome. J. Cell Biol. 1996, 133, 495−505. (33) Verschoor, A.; Warner, J. R.; Srivastava, S.; Grassucci, R.; Frank, J. Three-dimensional structure of the yeast ribosome. Nucleic Acids Res. 1998, 26, 655−661. (34) Gold, L. mRNA display: diversity matters during in vitro selection. Proc. Natl. Acad. Sci. U. S. A. 2001, 98, 4825−4826. (35) Presta, L. G.; Chen, H.; O’Connor, S. J.; Chisholm, V.; Meng, G.; Krummen, l.; Winkler, M.; Ferrara, N. Humanization of an antivascular endothelial growth factor monoclonal antibody for the therapy of solid tumors and other disorders. Cancer Res. 1997, 57, 4593−4599. (36) Kenrick, S. A.; Daugherty, P. S. Bacterial display enables efficient and quantitative peptide affinity maturation. Protein Eng., Des. Sel. 2010, 23, 9−17. (37) Naimuddin, M.; Ohtsuka, I.; Kitamura, K.; Kudou, M.; Kimura, S. Role of messenger RNA-ribosome complex in cDNA display. Anal. Biochem. 2013, 438, 97−103. (38) Barendt, P. A.; Ng, D. T.; McQuade, C. N.; Sarkar, C. A. Streamlined protocol for mRNA display. ACS Comb. Sci. 2013, 15, 77−81. (39) Mochizuki, Y.; Suzuki, T.; Fujimoto, K.; Nemoto, N. A versatile puromycin-linker using cnvk for high-throughput in vitro selection by cDNA display. J. Biotechnol. 2015, 212, 174−180.

ACKNOWLEDGMENTS We thank S. Kobayashi for technical support. We gratefully acknowledge Dr. Shozeb Haider, Queen’s University, UK, for critical reading of the manuscript. This work was supported by New Energy and Industrial Technology Development Organization (NEDO) (04A02542a), Japan to T.K. and in part by Janusys Corporation to M.N.

■

REFERENCES

(1) Bradbury, A. R. M.; Sidhu, S.; Dubel, S.; McCafferty, J. Beyond natural antibodies: the power of in vitro display technologies. Nat. Biotechnol. 2011, 29, 245−254. (2) Lohse, P. A.; Wright, M. C. In vitro protein display in drug discovery. Curr. Opin Drug Discov Devel. 2001, 4, 198−204. (3) Rothe, A.; Hosse, R. J.; Power, B. E. In vitro display technologies reveal novel biopharmaceutics. FASEB J. 2006, 20, 1599−1610. (4) Smith, G. P.; Petrenko, V. A. Phage Display. Chem. Rev. 1997, 97, 391−410. (5) Rowley, M. J.; O’Connor, K.; Wijeyewickrema, L. Phage display for epitope determination: a paradigm for identifying receptor-ligand interactions. Biotechnol. Annu. Rev. 2004, 10, 151−188. (6) Brissette, R.; Goldstein, N. I. The use of phage display peptide libraries for basic and translational research. Methods Mol. Biol. 2007, 383, 203−213. (7) Rivinoja, A.; Laakkonen, P. Identification of homing peptides using the in vivo phage display technology. Methods Mol. Biol. 2011, 683, 401−415. (8) Roberts, R. W. Totally in vitro protein selection using mRNAprotein fusions and ribosome display. Curr. Opin. Chem. Biol. 1999, 3, 268−273. (9) Nemoto, N.; Miyamoto-Sato, E.; Husimi, Y.; Yanagawa, H. In vitro virus: Bonding of mRNA bearing puromycin at the 3′-terminal end to the C-terminal end of its encoded protein on the ribosome in vitro. FEBS Lett. 1997, 414, 405−408. (10) Roberts, R. W.; Szostak, J. W. RNA-peptide fusions for the in vitro selection of peptides and proteins. Proc. Natl. Acad. Sci. U. S. A. 1997, 94, 12297−12302. (11) Lipovsek, D.; Plückthun, A. In vitro protein evolution by ribosome display and mRNA display. J. Immunol. Methods 2004, 290, 51−67. (12) Liu, R.; Barrick, J. E.; Szostak, J. W.; Roberts, R. W. Optimized synthesis of RNA-protein fusions for in vitro protein selection. Methods Enzymol. 2000, 318, 268−293. (13) Yamaguchi, J.; Naimuddin, M.; Biyani, M.; Sasaki, T.; Machida, M.; Kubo, T.; Funatsu, T.; Husimi, Y.; Nemoto, N. cDNA display: a novel screening method for functional disulfide-rich peptides by solidphase synthesis and stabilization of mRNA-protein fusions. Nucleic Acids Res. 2009, 37, e108. (14) Takahashi, T. T.; Roberts, R. W. In vitro selection of protein and peptide libraries using mRNA display. Methods Mol. Biol. 2009, 535, 293−314. (15) Cox, J. C.; Hayhurst, A.; Hesselberth, J.; Bayer, T. S.; Georgiou, G.; Ellington, A. D. Automated selection of aptamers against protein targets translated in vitro: from gene to aptamer. Nucleic Acids Res. 2002, 30, e108. (16) Crameri, R.; Kodzius, R. The powerful combination of phage surface display of cDNA libraries and high throughput screening. Comb. Chem. High Throughput Screening 2001, 4, 145−155. (17) Turunen, L.; Takkinen, K.; Soderlund, H.; Pulli, T. Automated panning and screening procedure on microplates for antibody generation from phage display libraries. J. Biomol. Screening 2009, 14, 282−293. (18) Tabata, N.; Sakuma, Y.; Honda, Y.; Doi, N.; Takashima, H.; Miyamoto-Sato, E.; Yanagawa, H. Rapid antibody selection by mRNA display on a microfluidic chip. Nucleic Acids Res. 2009, 37, e64. (19) Nishigaki, K.; Taguchi, K.; Kinoshita, Y.; Aita, T.; Husimi, Y. Yligation: an efficient method for ligating single-stranded DNAs and RNAs with T4 RNA ligase. Mol. Diversity 1998, 4, 187−190. L


Research Article

ACS Combinatorial Science (40) Ito, K.; Passioura, T.; Suga, H. Technologies for the synthesis of mRNA-encoding libraries and discovery of bioactive natural productinspired non-traditional macrocyclic peptides. Molecules 2013, 18, 3502−3528. (41) Ishizawa, T.; Kawakami, T.; Reid, P. C.; Murakami, H. TRAP display” A high-speed selection method for the generation of functional polypeptides. J. Am. Chem. Soc. 2013, 135, 5433−5440. (42) Kinoshita, Y.; Tayama, T.; Kitamura, K.; Salimullah, M.; Uchida, H.; Suzuki, M.; Husimi, Y.; Nishigaki, K. Novel concept microarray enabling PCR and multistep reactions through pipette-free apertureto-aperture parallel transfer. BMC Biotechnol. 2010, 10, 71. (43) Olson, C. A.; Adams, J. D.; Takahashi, T. T.; Qi, H.; Howell, S. M.; Wu, T. T.; Roberts, R. W.; Sun, R.; Soh, H. T. Rapid mRNAdisplay selection of an IL-6 inhibitor using continuous-flow magnetic separation. Angew. Chem., Int. Ed. 2011, 50, 8295−8298. (44) The Glen Report. Deprotection supplement (March 2011). Glen Research, Sterling, VA, USA. (45) Horn, T.; Chang, C.-A.; Urdea, M. S. Chemical synthesis and characterization of branched oligodeoxyribonucleotides (bDNA) for use as signal amplifiers in nucleic acid quantification assays. Nucleic Acids Res. 1997, 25, 4842−4849.

M


A High Performance Platform Based on cDNA Display for Efficient

Recommend Documents