Highly stable and efficient mRNA templates for mRNA–protein fusions and C-terminally labeled proteins

Abstract

For high-throughput in vitro protein selection using genotype (mRNA)–phenotype (protein) fusion formation and C-terminal protein labeling as a post-selection analysis, it is important to improve the stability and efficiency of mRNA templates for both technologies. Here we describe an efficient single-strand ligation (90% of the input mRNAs) using a fluorescein-conjugated polyethylene glycol puromycin (Fluor-PEG Puro) spacer. This ligation provides a stable c-jun mRNA with a flexible Fluor-PEG Puro spacer for efficient fusion formation (70% of the input mRNA with the PEG spacer) in a cell-free wheat germ translation system. When using a 5′ untranslated region including SP6 promoter and Ω29 enhancer (a part of tobacco mosaic virus Ω), an A₈ sequence (eight consecutive adenylate residues) at the 3′ end is suitable for fusion formation, while an XA₈ sequence (XhoI and the A₈ sequence) is suitable for C-terminal protein labeling. Further, we report that Fluor-PEG N-t-butyloxycarbonylpuromycin [Puro(Boc)] spacer enhances the stability and efficiency of c-jun mRNA template for C-terminal protein labeling. These mRNA templates should be useful for puromycin-based technologies (fusion formation and C-terminal protein labeling) to facilitate high-throughput in vitro protein selection for not only evolutionary protein engineering, but also proteome exploration.

INTRODUCTION

In vitro protein selection experiments using ‘in vitro virus’ (IVV) (1), RNA–peptide fusion (2–4), mRNA display (5–7), ribosome display (8–10) and STABLE (11) have been developed for evolutionary protein engineering and are expected to be applicable for genome analyses (12,13). However, current in vitro protein selection techniques (1–7,12,13) require an improvement in the stability of mRNA–protein fusions to minimize mRNA degradation in cell-free translation systems (11). They also require an improvement in the efficiency of formation of mRNA–protein fusions to provide large libraries and highly efficient enrichment, as well as to simplify tedious processes such as the post-translational maturation step (4).

In the course of development of IVV (1), two useful puromycin-based techniques, named ‘puromycin technology’, have been established on the basis of the interesting phenomenon that puromycin can bond to a full-length protein at the C-terminus (14). One technique is IVV formation (1), in which an in vitro-translated full-length protein (phenotype) is attached to its encoding mRNA (genotype), and the other is C-terminal protein labeling (14–16), in which a puromycin derivative bearing a fluorescein moiety is used to label the C-terminal end of a full-length protein. Accordingly, high-throughput in vitro protein selection can be achieved by combining IVV and post-selection using C-terminally labeled proteins, especially for proteome exploration (post-selection analysis is necessary to confirm protein–protein interactions). In fact, we have already reported in vitro analyses of protein–protein interactions using C-terminally labeled proteins (15,16). Thus, in the present study, we attempted to obtain efficient mRNA templates that could improve puromycin technology (IVV formation and C-terminal labeling of proteins).

Since the flexibility of mRNA–puromycin conjugates is an important factor in mRNA–protein fusion formation (3,17), we chose simple enzymatic ligation using T4 RNA ligase. The previous ligation methods, such as splint ligation (4) and photo-linked ligation (17), require hybridization sequences that impair the flexibility at the 3′ end of mRNA, as opposed to our single-strand ligation. Here we demonstrate that, with c-jun mRNA template, single-strand ligation using the flexible fluorescein-conjugated polyethylene glycol puromycin (Fluor-PEG Puro) spacer provides an improvement of mRNA stability and IVV formation. By utilizing particular 5′ and 3′ end sequences of c-jun mRNA (Fig. 1A), we can obtain efficient mRNA templates for IVV formation (Fig. 1B) and C-terminal protein labeling (Fig. 1C, I). Further, we show that the Fluor-PEG N-t-butyloxycarbonylpuromycin [Puro(Boc)] spacer (Fig. 1C, II) enhances the stability and efficiency of c-jun mRNA for C-terminal protein labeling.

Figure 1.

The mRNA template and puromycin-based technologies of IVV formation and C-terminal labeling of proteins. (A) The mRNA template is comprised of 5′ UTR, T7 tag, ORF, Flag tag and 3′ tail. Five mRNA templates have different 5′ UTR sequences: SP6 + Ω29 [SP6 RNA polymerase transcription promoter (SP6) and translation enhancer Ω29 (CAACAACAACAACAAACAACAACAAAATG), which is a part of tobacco mosaic virus Ω sequence (23)], SP6 + Ω, T7 + Ω29 [T7 RNA polymerase promoter (T7) and Ω29], T7 + delta-TMV [T7 and delta-TMV sequence (2)] and T7 + K [T7 and Kozak (K) sequence (24)]. Four mRNA templates have different 3′ tail sequences: XA₈ (CTCGAGAAAAAAAA), X′A₈ (CATCACAAAAAAAA), X(CTCGAG) and A₈ (AAAAAAAA). The standard template has SP6 + Ω29 as a 5′ UTR and XA₈ as a 3′ tail. (B) IVV formation on the ribosome (1). Puromycin at the 3′-terminal end of a spacer (Fluor-PEG Puro) ligated to mRNA can enter the ribosomal A-site to covalently bind to the C-terminal end of its encoding protein. Puro represents puromycin. (C) C-terminal labeling of proteins on the ribosome (14,15). A puromycin derivative (Fluor-dCpuro; 16) can enter the ribosomal A-site to bind covalently to the C-terminal end of the protein when mRNA without a spacer (I) or mRNA with a Fluor-PEG Puro(Boc) spacer (II) is used as a template. Puro(Boc) represents N-t-butyloxycarbonylpuromycin, which is incapable of forming IVV.

MATERIALS AND METHODS

Synthesis of Fluor-PEG Puro and Puro(Boc) spacers

Protected deoxycytidine phosphoramidite (dC-amidite), thymidine(fluorescein) phosphoramidite [T(Fluor)-amidite] and chemical phosphorylation reagent II (CPR II) were purchased from Glen Research Corporation (VA). Polyethylene glycol (PEG, average mol. wt 2000) was purchased from NOF Corporation (Tokyo, Japan). N-Fluorenylmethoxycarbonyl puromycin and N-t-butyloxycarbonylpuromycin attached to controlled-pore glass supports [Puro(Fmoc)-CPG and Puro(Boc)-CPG] were synthesized according to the published procedure (18). Synthesis of (4,4′-dimethoxytrityl)-PEG-(2-cyanoethyl-N,N-diisopropyl)phosphoramidite (PEG-amidite) was done according to the published procedure (19). Fluor-PEG Puro spacer [p(dCp)₂-T(Fluor)p-PEGp-(dCp)₂-puromycin] was synthesized from Puro(Fmoc)-CPG, and Fluor-PEG Puro(Boc) spacer [p(dCp)₂-T(Fluor)p-PEGp-(dCp)₂-N-t-butyloxycarbonylpuromycin] was synthesized from Puro(Boc)-CPG, with dC-amidite, PEG-amidite, T(Fluor)-amidite and CPR II according to the standard phosphoramidite method (19). After deprotection with concentrated NH₄OH/EtOH (3:1) at 25°C for 14 h, the spacers were purified by reverse-phase HPLC on a YMC-Pack ODS-A SH-343-5 (YMC, Kyoto, Japan) with 0.1 M triethylammonium acetate (pH 7.0) as solvent A and acetonitrile as solvent B at a flow rate of 10 ml/min. A linear gradient of 10–60% solvent B over 30 min was used for elution.

Preparation of DNA

PCR amplification was conducted through 35 cycles (98°C, 20 s; 55°C, 60 s; 72°C, 180 s) with the primers listed in Table 1. DNA templates of c-jun from pUC-Jun containing a part of c-Jun (179–335) cloned from a mouse testis cDNA library (Takara) were amplified in two PCR steps. The first PCR was performed with primers 5′T7Jun and 3′JunFlagA, and then the second PCR was done using the first PCR product as a template with the 5′ primers 5′SP6(O-29)T7 having SP6 + Ω29, 5′T7g-SP6 having SP6 + Ω, and 5′T7(O-29)T7 having T7 + Ω29, and the 3′ primers 3′FlagA having XA₈, 3′Flag2 having X, 3′FlagA(C3) having X′A₈, 3′Flag1A having A₈, 3′FlagHisXA having XA₈, and 3′FlagHisA having A₈. DNA templates of c-fos from pCMV-FosCBPzz including a part of c-Fos (118–211) cloned from a mouse testis cDNA library (Takara) were amplified with the 5′ primer 5′SP6(O29/F.H)Fos and the 3′ primer 3′FosFlagA with XA₈ or 3′FosFlag1A with A₈. DNA templates of cyclin B1 (cB1) from pCMVzzCBPcB1 containing full-length cB1 were amplified with the 5′ primer 5′SP6(O29)cB1 and the 3′ primer 3′(646)FlagA with XA₈ or 3′(646)Flag1A with A₈.

Table 1. Primers used in this study.

Gene	Primer	Sequence
c-jun	5′T7Jun	CATGACCCATGGCTAGCATGACTGGTGGACAGCAAATGGGTGCGGCCGCGCCGGAGATGCCGGGAGGAC
	3′JunFlag	ACCCAAGCTTTTTTTTTCTCGAGCTTGTCGTCATCGTCCTTGTAGTCCCGCGGTCTAAACGTTTGCAACTGCTGCGT TAGCATGAGTTG
	5′SP6(O-29)T7	ATTTAGGTGACACTATAGAACAACAACAACAACAAACAACAACAAAATGGCTAGCATGACTGGTGGACCAAATG
	5′T7g-SP6	TAATACGACTCACTATAGGGAGACCACAACGGTTTCCCATTTAGGTGACACTATAGAATACACGGAATTGCG
	5′T7(O-29)T7	TAATACGACTCACTATAGGGCAACAACAACAACAAACAACAACAAAATGGCTAGCATGACTGGTGGACCAAATG
	3′FlagA	TTTTTTTTCTCGAGCTTGTCGTCATCGTCCTTGTAG
	3′Flag2	CTCGAGCTTGTCGTCATCGTCCTTGTAG
	3′FlagA(C3)	TTTTTTTTGTGATGCTTGTCGTCATCGTCCTTGTAGTCCC
	3′Flag1A	TTTTTTTTCTTGTCGTCATCGTCCTTGTAGTCCCG
	3′FlagHisXA	TTTTTTTTCTCGAGGTGATGGTGATGGTGATGCTTGTCGTCATCGTCCTTGTAGTCCC
	3′FlagHis	ATTTTTTTTGTGATGGTGATGGTGATGCTTGTCGTCATCGTCCTTGTAGTCCC
c-fos	5′SP6(O29/F.H)Fos	ATTTAGGTGACACTATAGAACAACAACAACAACAAACAACAACAAAATGCTTGGATCCGGCAGAGCGCAG
	3′FosFlagA	TTTTTTTTCTCGAGCTTGTCGTCATCGTCCTTGTAGTCCGGCTCGAGGCCAAGGTCATCGGGGAT
	3′FosFlag1A	TTTTTTTTCTTGTCGTCATCGTCCTTGTAGTCCTCGAGGCCAAGGTCATCGGGGAT
cyclin B1	5′SP6(O29)cB1	ATTTAGGTGACACTATAGAACAACAACAACAACAAACAACAACAAAATGGCGCTCCGAGTCACCAGGA
	3′(646)FlagA	TTTTTTTTCTCGAGCTTGTCGTCATCGTCCTTGTAGTCGTTTCCAGTGACTTCCCGACCCAGTAG
	3′ (646)Flag1A	TTTTTTTTTGTCGTCATCGTCCTTGTAGTCGTTTCCAGTGACTTCCCGACCCAGTAG

Ligation of mRNA to a Fluor-PEG Puro or Puro(Boc) spacer

The transcription was performed using the RiboMax™ Large Scale RNA Production System (Promega) with the PCR products as DNA templates, as described in Preparation of DNA. The transcription product of 200 nM mRNA was ligated to 40 µM Fluor-PEG Puro spacer or Puro(Boc) spacer with T4 RNA ligase (Takara) at 15°C in the presence of 120 µM free PEG (average mol. wt 2000) and then excess spacer was removed by using RNeasy Mini Kits (Qiagen). Ligation products were analyzed by 8 M urea/4% PAGE after staining with ethidium bromide. The fluorescence of mRNA with a Fluor-PEG spacer on the gel was easily and directly visualized with a fluorescence image analyzer (Molecular Imager FX, Bio-Rad).

IVV formation and C-terminal protein labeling

The standard IVV formation protocol was performed by using 200 nM mRNA ligated to a Fluor-PEG Puro spacer as a template in a cell-free wheat germ translation system (Wheat Germ Extract, Promega) at 26°C for 1 h. The translation product was analyzed by 8 M urea/10% SDS–PAGE. The fluorescence of the resulting IVV on the gel was easily and directly visualized with Molecular Imager FX (Bio-Rad). The Fluor-PEG Puro spacer offers the easy detection of IVV without using any radioisotopes. The C-terminal labeling of proteins was carried out using 200 nM mRNA as a template in the presence of 20 µM Fluor-dCpuro (16) at 26°C for 1 h. The yield of C-terminal labeled proteins was evaluated by scanning the fluorescence on a 15% SDS–polyacrylamide gel with a Molecular Imager FX (Bio-Rad). IVV formation and C-terminal labeling of proteins were also confirmed by western blotting. Western blots were probed with antibodies against T7-tag (Novagen) or Flag-tag (Sigma) using ECF Western blotting reagent packs (Amersham).

RESULTS AND DISCUSSION

Ligation efficiency with Fluor-PEG Puro spacer

Enzymatic ligation for the preparation of IVV requires a cumbersome gel purification of the mRNA template ligated to the spacer because of low ligation efficiency (1–3), thus slowing down template preparation. We took note of the fact that T4 RNA ligase prefers a C-rich sequence at the 5′ end as a donor and an A-rich sequence at the 3′ end as an acceptor (20). Ligation by using T4 RNA ligase was done with a Fluor-PEG Puro spacer with two consecutive deoxycytidylate residues as a donor. Two different mRNA templates with eight consecutive adenylate residues as a common acceptor were efficiently ligated to the spacer (Fig. 2A, Ib and IIb) compared with the templates without the acceptor (Fig. 2A, Ia and IIa). The ligation yield was finally increased from 80% (Fig. 2A, Ib, Fig. 2B, filled circles) to 90% (Fig. 2B, open circles; 80% within 1 h) of the input mRNA by the addition of free PEG (mol. wt 2000). This ligation can be applied not only to a Fluor-PEG Puro spacer for IVV formation (Fig. 1B), but also to a Fluor-PEG Puro(Boc) spacer for C-terminal labeling of proteins (Fig. 1C, II) in the same manner. This ligation method should allow efficient formation of both IVV and C-terminally labeled proteins without requiring a cumbersome gel purification of the mRNA template ligated to the spacer, thus speeding up template preparation.

Figure 2.

Efficiency of ligation and stability of mRNA with a Fluor-PEG Puro spacer. (A) The ligation of c-jun mRNA having different 3′ tail sequences described in Figure 1A, X (I) or X′ (II) without (a) or with an A₈ sequence (b) was analyzed with 8 M urea/4% PAGE followed by staining with ethidium bromide. Ligation efficiency is given as a percentage relative to input mRNA. (B) The time course of ligation of c-jun mRNA to a spacer described in (A, Ib) in the presence (open circles) or absence (closed circles) of free PEG (mol. wt 2000). (C) The remaining mRNA and IVV of c-jun mRNA were ligated to Fluor-PEG Puro spacer (I) containing 20% DNA, TEG spacer (II) (pdA₂₀dT(fluorescein)[C9]₃dAdCdCPuromycin) (3) containing 80% DNA, or DNA spacer (III) (pdA₂₁dT(fluorescein)dA₆dCdCPuromycin) (1,2) containing 100% DNA in a cell-free wheat germ translation system for 60 min in the presence (lane 2 of I–III) or absence (lane 3 of I–III) of 400 µM puromycin. The length of three spacers corresponds to approximately 30 nucleotide units. The input mRNA ligated to each spacer corresponds to 100% (lane 1 of I–III).

Stability of mRNA with Fluor-PEG Puro spacer

We examined the stability of mRNA with different flexible spacers (Fig. 2C). The Fluor-PEG Puro spacer most markedly enhanced the stability of c-jun mRNA template among these spacers (Fig. 2C, lane 2 in I–III) and the template remained almost entirely intact (>90%) within 60 min in a cell-free wheat germ translation system (Fig. 2C, lane 2 in I). The PEG spacer (Fig. 2C, I), which is the most flexible spacer containing the least percentage of DNA (20%), also gave the most efficient c-Jun IVV formation among these spacers (Fig. 2C, lane 3 in I–III). This result is consistent with previous findings that flexibility of the spacer is an important factor in mRNA–protein fusion formation (3,17). Thus, we obtained a stable and flexible c-jun mRNA template that could be easily handled in a cell-free translation system for IVV formation by means of ligation using the Fluor-PEG Puro spacer. Further, previous ligation methods, such as splint ligation (4), require sequences capable of hybridizing with a splint DNA, whereas single-strand ligation does not require hybridization sequences. Therefore, we can examine the effects of not only 5′-, but also 3′-terminal sequences on the efficiency of mRNA templates for puromycin technology (IVV and C-terminal protein labeling).

Optimized 5′- and 3′-terminal sequences of c-jun mRNA for puromycin technology

We have investigated the efficiency of IVV formation (Fig. 1B) and the formation efficiency of C-terminally labeled proteins (Fig. 1C, I) using five c-jun mRNA templates with different 5′ untranslated region (UTR) sequences (Fig. 3A, lanes 1–5) in a cell-free wheat germ translation system. The efficiency of IVV formation increased (Fig. 3A, I) in parallel with the formation efficiency of C-terminally labeled proteins (Fig. 3A, II). The combination of SP6 as a promoter and Ω29 as an enhancer gave the best c-jun mRNA template for both IVV formation (Fig. 3A, I, lane 1) and C-terminal labeling of proteins (Fig. 3A, II, lane 1). Also, SP6 promoter (Fig. 3A, lane 1) is more suitable than T7 promoter (Fig. 3A, lane 3), and full-length Ω enhancer (Fig. 3A, lane 2) is not always better than shorter enhancers such as Ω29 (Fig. 3A, lane 1) in a cell-free wheat germ translation system (T.Sawasaki and Y.Endo, unpublished data). In contrast, the efficiency of IVV formation (Fig. 3B, I) of four c-jun mRNA templates with different 3′ sequences (Fig. 3B, lanes 1–4) did not increase in parallel with increasing the formation efficiency of C-terminally labeled proteins (Fig. 3B, II). Thus, the results suggest that the formation efficiency of C-terminally labeled proteins does not necessarily correspond to the efficiency of IVV formation (Fig. 3A and B).

Figure 3.

Effect of 5′ UTR and 3′ tail sequences of mRNA on puromycin technology. (A) The efficiency of IVV formation (I) and C-terminally labeled protein formation (II) with five different 5′ UTR sequences (lanes 1–5). Individual sequences of 5′ UTR are as described in Figure 1A. Lane 1, SP6 + Ω29; lane 2, SP6 + Ω; lane 3, T7 + Ω29; lane 4, T7 + delta-TMV; lane 5, T7 + K. Efficiency is normalized to the efficiency of lane 1. Data represent the mean SD of two separate experiments. (B) The efficiency of IVV formation (I) and C-terminally labeled protein formation (II) with four different 3′ tail sequences (lanes 1–4). Individual sequences of 3′ tail are as described in Figure 1A. Lane 1, XA₈; lane 2, X′A₈; lane 3, X; lane 4, A₈. Efficiency is normalized to the efficiency of lane 1. Data represent the mean ± SD of two separate experiments.

3′ tail sequences of mRNA favoring efficient IVV formation

Two factors influence the efficiency of IVV formation, i.e. the input mRNA ligated to a spacer (Fig. 3A and B, I) and the total synthesized protein. The total synthesized protein corresponds to the sum of IVV and free proteins produced during IVV formation. We examined the effect of concentrations of c-jun mRNA with XA₈ and A₈ sequences (Fig. 3B, lanes 1 and 4) on the efficiency of IVV formation. The efficiency of IVV formation of the input c-jun mRNA with the XA₈ sequence ligated to the spacer was much the same as that with the A₈ sequence (Fig. 4A, XA₈ and A₈). This is consistent with the finding that the efficiency of IVV formation of the input mRNA ligated to the spacer was hardly affected by these 3′ tail sequences (Fig. 3B, I). In contrast, the efficiency of IVV formation of the total synthesized protein was affected (Fig. 4B, XA₈ and A₈), and the A₈ sequence is more suitable for IVV formation. Accordingly, we optimized IVV formation of the input mRNA with the spacer (Fig. 4C, I; 70%) and IVV formation of the total synthesized proteins (Fig. 4C, II; 90%) using c-jun mRNA with SP6 + Ω29 as a 5′ UTR and the A₈ sequence as a 3′ tail.

Figure 4.

Effect of the 3′ tail of mRNA on IVV formation. (A) The IVV formation was examined at various concentrations of the input c-jun mRNA. Lanes 1–6 correspond to 10, 50, 100, 200, 400 and 800 nM input c-jun mRNA, respectively. IVV formation efficiency represents a relative percentage with respect to the input mRNA ligated to a Fluor-PEG Puro spacer. (B) The formation of free protein was examined by C-terminal protein labeling at various concentrations of the input mRNA. Formation efficiency of free proteins represents a relative percentage with respect to resulting total synthesized proteins, corresponding to IVV and free proteins. Lanes indicate the concentrations of input mRNA as described in (A). (C) IVV formation using 50 nM c-jun mRNA ligated to a Fluor-PEG Puro spacer was carried out in a cell-free wheat germ translation system for 1 h. The efficiency was determined by detecting fluorescein (I) in the spacer and by western blotting for T7 tag (II). (D) The efficiency of IVV formation and the yield of free proteins examined using 200 nM c-fos (lane 1), c-jun (lane 2) and cB1 (lane 3) mRNAs having SP6 + Ω29 as a 5′ UTR and XA₈ or A₈ sequence as a 3′ tail ligated to Fluor-PEG Puro spacer, respectively.

Furthermore, we investigated the effect of the 3′ tail sequences on the IVV formation using not only c-Jun, but also c-Fos and cyclin B1 to examine whether the effect on the efficiency of IVV formation is general or not. The IVV formation of all c-fos, c-jun and cB1 mRNA templates having the SP6 + Ω29 sequence as a 5′ UTR and the XA₈ or A₈ sequence as a 3′ tail was efficient and in much the same range of 45–55% at a concentration of 200 nM mRNA (the concentration used in Fig. 4A, lane 4), though the yield of free proteins followed the order of XA₈ > A₈ sequence (Fig. 4D). This suggests that when using the SP6 + Ω29 sequence as a 5′ UTR, the effect of the A₈ sequence as a 3′ tail would apply to general mRNA templates in an IVV library for in vitro protein selection.

Efficient IVV formation using a Fluor-PEG Puro spacer

According to previous reports, the optimized efficiency of formation of fusions was 40% of the input mRNA ligated to a spacer and 50% of the total synthesized protein in a cell-free rabbit reticulocyte lysate translation system (3,4,17), and the rabbit reticulocyte lysate system was superior to the wheat germ system (3). However, we obtained the optimized efficient IVV formation in a cell-free wheat germ translation system (Fig. 4C). We have examined IVV formations of more than 10 genes based on both the total proteins and the input mRNAs with the A₈ sequence ligated to the Fluor-PEG Puro spacer, and the range of the efficiency was much the same as that in Figure 4D, A₈ (data not shown). In addition, this IVV formation procedure does not require any post-translational treatment, such as a maturation step (3,4,17). This indicates that the single-strand mRNA template ligated to the Fluor-PEG Puro spacer is able to form IVV rapidly and efficiently because of its superior flexibility. Here, the flexibility is provided by the flexible Fluor-PEG Puro spacer containing the least percentage of DNA (20%; Fig. 2C) and the single-strand mRNA template without hybridization sequences that impair the flexibility at the 3′ end of mRNA. We consider that highly efficient IVV formation thus obtained is a key factor in allowing the preparation of larger libraries [up to 10¹⁴/ml, compared with the previously reported 10¹³/ml (3); calculated from Fig. 4A] with few free proteins in a cell-free wheat germ translation system for in vitro protein selection experiments. Such highly efficient IVV formation should also allow high-throughput in vitro selection both without a maturation procedure (3,4,17) and without purification (3,4,17) of mRNA–protein fusions after translation.

3′ tail sequences of mRNA favoring efficient C-terminal protein labeling

We expected that the effect of XA₈ > A₈ sequence on the yield of free proteins (Fig. 4B, XA₈ and A₈) might not be irrelevant to the effect on the yield of proteins (Fig. 3B, lanes 1 and 4 in II). To establish whether the XA₈ sequence enhances the yield of C-terminally labeled proteins more effectively than does the A₈ sequence (Fig. 3B, lanes 1 and 4 in II) in general, we investigated the formation efficiency of C-terminally labeled proteins of c-Jun, c-Fos and cB1 with a Flag tag (Fig. 5A, Jun-Flag, Fos-Flag and CB1-Flag), and also c-Jun having a His tag (Fig. 5A, Jun-His). The formation efficiency of C-terminally labeled proteins of all mRNA templates of c-jun, c-fos and cB1 with an XA₈ sequence at the 3′ end was 3- to 5-fold higher than that in the case of an A₈ sequence, confirming the XA₈ > A₈ sequence effect on the formation efficiency of C-terminally labeled proteins (Fig. 5A). Figure 5A also indicates that the formation efficiency of C-terminally labeled proteins was independent of the upstream sequence of XA₈, such as a Flag tag or His tag. We can conclude that mRNA templates with the XA₈ sequence are clearly preferable to those with the A₈ sequence for the C-terminal labeling of proteins.

Figure 5.

Effect of the 3′ tail of mRNA on C-terminal protein labeling. (A) The formation of C-terminally labeled proteins was examined using 200 nM c-jun (Jun-Flag), c-fos (Fos-Flag), cB1 (cB1-Flag) and c-jun (Jun-His) mRNAs having a Flag tag or His tag followed by an XA₈ or A₈ sequence, respectively. The formation efficiency of C-terminally labeled proteins was normalized to the formation efficiency using Jun-Flag (XA₈). (B) The remaining mRNA template having an XA₈ (closed circles) or XA₈ sequence ligated to the Fluor-PEG Puro(Boc) spacer (open circles) was examined in a cell-free wheat germ translation system in the presence of 400 µM puromycin. (C) The yield of C-terminally labeled proteins was examined using an mRNA template having an A₈ (I), XA₈ (II) or XA₈ sequence ligated to the Fluor-PEG Puro(Boc) spacer (III). Translation was done in a cell-free wheat germ translation system at 26°C for 3 h. The yield of C-terminally labeled proteins was normalized to the yield of c-jun mRNA having an A₈ sequence (I).

Efficient C-terminal protein labeling using a Fluor-PEG Puro(Boc) spacer

The half-life of c-jun mRNA with an XA₈ sequence ligated to a Fluor-PEG Puro(Boc) spacer at the 3′ end (Fig. 5B, open circles) was 5 h, whereas that of mRNA with an XA₈ sequence (Fig. 5B, filled circles), but without a Fluor-PEG Puro(Boc) spacer at the 3′ end, was 1 h. This result is consistent with the finding that the use of the Fluor-PEG Puro spacer led to an improvement of mRNA stability (Fig. 2C). Since enhanced stability of mRNA affects the improvement of translation efficiency (21,22), a spacer like the Fluor-PEG Puro(Boc) spacer that is unable to form IVV may enhance the formation efficiency of C-terminally labeled proteins (Fig. 1C, II). As expected, the formation efficiency of C-terminally labeled proteins of c-jun mRNA with an XA₈ sequence ligated to the Fluor-PEG Puro(Boc) spacer (Fig. 5C, III) was twice that of the XA₈ sequence without the spacer (Fig. 5C, II) and eight times higher than that in the case of an A₈ sequence without the spacer (Fig. 5C, I). The result suggests that the Fluor-PEG Puro(Boc) spacer enhances the yield of C-terminally labeled proteins. Thus, we obtained an mRNA template with a Fluor-PEG Puro(Boc) spacer as a new tool for the efficient C-terminal labeling of proteins (Fig. 5C).

In conclusion, we obtained highly stable and efficient mRNA templates for puromycin technology by single-strand ligation with a flexible PEG spacer including the Fluor-PEG Puro spacer for IVV or the Fluor-PEG(Boc) Puro spacer for C-terminally labeled proteins. After the optimization of 5′- and 3′-terminal sequences using c-jun mRNA, the mRNA template offers the advantages of easy handling because of high stability and efficiency for IVV formation (70% of the input mRNA with the Fluor-PEG Puro spacer; 90% of total proteins), simplifying formerly tedious processes. To facilitate high-throughput in vitro protein selection for proteome exploration, as well as evolutionary protein engineering, the mRNA template with SP6 + Ω29 as a 5′ UTR and the A₈ sequence as a 3′ sequence should be utilized for IVV formation in a selection of functional proteins, and the mRNA template with SP6 + Ω29 as a 5′ UTR and XA₈ sequence as a 3′ sequence for C-terminal protein labeling in conjunction with protein microarrays and fluorescence cross-correlation spectroscopy (16) in a post-selection of functional proteins.