Nascent MSKIK peptide cancels ribosomal stalling by arrest peptides in Escherichia coli

The insertion of the DNA sequence encoding SKIK peptide adjacent to the M start codon of a difficult-to-express protein enhances protein production in Escherichia coli. In this report, we reveal that the increased production of the SKIK-tagged protein is not due to codon usage of the SKIK sequence. Furthermore, we found that insertion of SKIK or MSKIK just before the SecM arrest peptide (FSTPVWISQAQGIRAGP), which causes ribosomal stalling on mRNA, greatly increased the production of the protein containing the SecM arrest peptide in the E. coli–reconstituted cell-free protein synthesis system (PURE system). A similar translation enhancement phenomenon by MSKIK was observed for the CmlA leader peptide, a ribosome arrest peptide, whose arrest is induced by chloramphenicol. These results strongly suggest that the nascent MSKIK peptide prevents or releases ribosomal stalling immediately following its generation during the translation process, resulting in an increase of protein production.

Despite protein production by microorganisms being an essential technology in life science research and industry, researchers still often encounter low productivity of target proteins in microbial hosts and mammalian cells such as Chinese hamster ovary cells (1)(2)(3). A general method to enhance the production of any protein in any organism without laborious trial and error will surely revitalize the bioeconomy that includes manufacturing bio-based products and bioenergy.
We reported previously that the insertion of codons 5 0 -TCT AAA ATA AAA-3 0 or synonymous codons 5 0 -TCG AAG ATC AAG-3 0 that encode four amino acids, Ser-Lys-Ile-Lys (SKIK), immediately after the Met start codon markedly increased the production level of proteins that were difficult to express in Escherichia coli (4). Although we have shown that SKIK tagging may promote translation but not transcription in E. coli in vivo expression experiments, the detailed mechanism remains unclear. Several recent studies showed that an optimal codon composition increases mRNA stability, resulting in improved protein production in E. coli, Saccharomyces cerevisiae, and humans (5)(6)(7). Based on these observations, a possible hypothesis is that the codon composition encoding SKIK may influence protein production.
Newly generated polypeptides produced during translation are called nascent chains. These chains have important roles in various cellular processes including their maturation and quality control of proteins and mRNA and can be involved in the control and regulation of cell-wide biological phenomena (8,9). Among them, nascent chains that cause ribosomal stalling are called ribosome arrest peptides (RAPs). RAPs contain specific amino acid sequences that interact with internal components of ribosomes in the nascent state. Various RAPs with different arresting mechanisms are distributed widely across species (10)(11)(12)(13)(14).
SecM, a bacterial secretion monitor protein composed of 170 amino acid residues, is a well-known RAP in E. coli (15). Ribosomal stalling by SecM is caused by the exit tunnel of the ribosome 50S subunit interacting with the nascent chain FXXXXWIXXXXGIRAGP (X is any amino acid) present in the C-terminal region of SecM (16,17). This RAP functions as an in vivo or in vitro translation arrester in E. coli even when inserted into another protein of interest (18). Other types of RAPs such as the CmlA leader peptide (KNAD) induced by chloramphenicol, TnaC (WFNIDNKIVDHRP*, where star represents a stop codon) induced by Trp, and the artificially designed WPPP sequence (FQKYGIWPPP) are also profiled in E. coli (15).
In this study, we demonstrate how SKIK-encoding codon usage affects protein production in E. coli in vivo and in vitro expression systems to verify whether the codons of SKIK or the amino acid sequence are important for translation enhancement. We then hypothesize that the SKIK peptide, the nascent chain, may act as a translation enhancer, preventing translational slowdown caused possibly by the interaction between the ribosome and nascent polypeptides. We also evaluate the effect of combinations and orders of SKIK or MSKIK with various RAPs. Finally, we reveal that nascent MSKIK peptide cancels ribosomal stalling that RAPs may cause.

Results
SKIK-encoding codons do not affect protein production in E. coli in vivo and in vitro expression systems In our previous study, two patterns of SKIK-encoding codons (5 0 -TCT AAA ATA AAA-3 0 and 5 0 -TCG AAG ATC AAG-3 0 ) were shown to affect protein synthesis to a similar extent (4). In this study, we evaluated all synonymous codons except the last lysine codon (AAA) and quantified the yield of His-tagged proteins by dot blot analysis using anti-His tag horseradish peroxidase (HRP) and ImageJ software.
In IPTG-induced E. coli samples, the production of each SKIK-tagged rabbit single chain variable fragment (scFv) was approximately 20 to 28 times higher than that of no-SKIKtagged scFv, and the 36 variants showed almost the same expression level regardless of the different codon usage (Figs. 1 and S1). ANOVA showed no significant differences in protein production levels among these 36 samples. Thus, the results showed no correlation between the codon patterns encoding the SKIK peptide tag and the amount of the expressed protein.
The same 36 variants of SKIK-tagged scFv were expressed by autoinduction for 72 h in TB medium at 30 C. Protein productivity was analyzed by SDS-PAGE, followed by Coomassie Brilliant Blue (CBB) staining, yielding clear bands of all SKIK-tagged scFvs, whereas no target band was observed without the SKIK tag (Fig. S2). There was no significant difference in the amount of protein produced among the 36 variants. These results indicated that adding the SKIK tag increases the production of target proteins uniformly, and this increase is independent of codon usage or induction methods. Moreover, the in vitro E. coli-reconstituted cell-free protein synthesis system yielded good expression at similar levels for the 36 variants, whereas the nontagged protein was not detected (Fig. S3). On the basis of the above results obtained for in vivo expression using different induction methods and in vitro expression, we concluded that the SKIK peptide tag increases the amount of protein produced both in vivo and in vitro in the E. coli expression system and is not affected by codon usage in either case.
The minimum free energy of the thermodynamic ensemble from the entire transcripts (mRNAs from transcript start position GGG to T7 terminator) calculated by RNAfold was −311.1 ± 12.6 kcal/mol for all SKIK-tagged clones, whereas the value was −273.0 kcal/mol for the nontagged clone. Specific nucleotides from the 5 0 UTR (−19) TTAA-GAAGGAGATATACAT to ACCCAGACTCCAGCCTCC in r1scFv were calculated to be −7.3 ± 1.0 kcal/mol in SKIKtagged variants and −6.5 kcal/mol in the no-tag variant (Fig. S4).
Increased production of the SKIK-tagged protein is caused by improving translation and not transcription To evaluate whether the addition of the SKIK peptide tag affects transcription or translation, in vitro transcription and translation were performed separately using rabbit fragment of antibody binding (Fab) genes L-and H-chains (GenBank accession numbers BAX56586 and BAX56587: protein molecular mass is approximately 30 kDa each) (19), with and without SKIK tags. The amounts of mRNA transcribed from each template were similar when equal amounts of PCR products were used as templates, regardless of the presence or absence of the SKIK tag ( Fig. 2A). In contrast, adjusting the template mRNA amounts to be equivalent and performing the translation reaction in the PURE system showed an increase in protein production of both the H-and L-chains when the SKIK tag was present (Fig. 2B). Thus, adding the SKIK tag for in vitro protein synthesis does not alter transcription but improves translation significantly. The entire gel and membrane images are shown in Fig. S5.   Figure 1. Protein production in Escherichia coli of an antibody fragment tagged with SKIK composed of various codons. In vivo protein (single chain antibody fragment from rabbit fused with a C-terminal His-tag) expression in E. coli was performed by IPTG induction. The graph shows mean±S. D. values (bars) with scatter plots of replicated experiments. The dot-blotted proteins in cell lysates were visualized and quantified with the HRP-conjugated antibody against the His-tag to react 3,3 0 ,5,5 0 -tetramethylbenzidine (TMB). Images were analyzed with ImageJ software. Sample numbers correspond to the SKIKencoding codon patterns listed in Table 1. HRP, horseradish peroxidase.
Peptide tags with only three amino acid residues SKX, KKX, and AKX can improve protein production The above results indicate that the amino acid sequence of the SKIK tag is important for promoting translation. Therefore, the effect of other amino acids at the second residue position and adjacent to the starting M was examined because this position was reported to be important for protein production (5,20). Here, the production of a rabbit scFv in E. coli . Surprisingly, all of the N-terminal-tagged scFvs were more productive when compared with the scFvs without the N-terminal tag (Fig. S6).

MSKIK peptide tag counteracts ribosome arrest caused by SecM
The obtained results suggest that the amino acid sequence, but not the mRNA sequence, of the SKIK peptide tag contributes to translation enhancement. We then evaluated how combining a typical SecM arrest peptide (SecM AP, FSTPVWISQAQGIRAGP) and the SKIK peptide affected translation in an effort to understand the mechanism of the SKIK tag on translation. The amino acid sequences of the ten constructs are shown in Figure 3A. The PURE system was used for in vitro transcription and translation reactions, and the produced proteins were analyzed by Western blotting with the anti-His tag antibody conjugated with HRP. The amount of proteins containing SecM AP was below the detection limit ( Fig. 3B, lanes one and 5), whereas the proteins without SecM AP were detected (Fig. 3B, lanes three and 4). This observation implies translation arrest by SecM AP occurred in the PURE system, resulting in poor protein production. Comparing lanes 3 and 4 which are lacking SecM AP, protein production was higher in lane 4 with SKIK.
Surprisingly, N-terminal SKIK placed immediately before the SecM AP increased the amount of synthesized protein significantly (lane 6). This effect was not as noticeable when the distance between the SKIK tag and SecM AP was 112 amino acid residues (lanes 1 and 2). Interestingly, the insertion of the sequence MSKIK or SKIK adjacent to SecM AP in the middle region of the protein (lanes 7 and 8) gave higher production than the constructs with no SKIK or N-terminal SKIK (lanes 1 and 2). This result implies that the presence of this tag not only at the N-terminus but also in the middle of ORFs, just before the AP, may have a translation-promoting effect on entire genes. In addition, we investigated MSKIK and SKIK insertions immediately after SecM AP. These two samples showed no bands possibly because the ribosome stalls at the SecM AP site. This data strongly supports our hypothesis that generation of nascent SKIK peptide before AP is important for promoting translation.
These results revealed that both MSKIK and SKIK tag placed just before the SecM AP releases or reduces ribosome arrest and promotes translation of the full-length gene. In addition, the physical distance from the AP are important factors that define the strength of the translation enhancement.

Discussion
We showed that a short nascent chain, MSKIK peptide, increases the production of difficult-to-express proteins by likely preventing or releasing ribosomal stalling caused by RAPs. Various studies have demonstrated that the codon usage of a gene affects the amount of protein produced (21). In particular, codon bias encoding the N-terminus affects mRNA expression levels but has been reported to be largely independent of mRNA translation or mRNA stability (22). Bo€ el et al. reported that the initial 16 codons more likely influence mRNA folding, rather than the entire codon content, by directly modulating translation efficiency and mRNA stability (5).
In the present study, we evaluated protein production in all synonymous codon patterns in SKIK except for the last K and showed that all codon patterns yielded the same protein production enhancement effect as observed for the original SKIK tag. In general, the presence of rare codons reduces protein expression (23). Interestingly, SKIK containing rare codons such as ATA (Ser) and TCG (Leu) (24) gave protein levels that were consistently increased. The energy and structure of the RNA were examined for each codon pattern, with no distinct differences found (Fig. S4). This finding is also consistent with the observation that there was no significant difference at the protein level among the 36 variants.
In addition, SKX, KKX, and AKX immediately after the starting M increased protein production (Fig. S6). This observation supports the hypothesis that this increased protein production is codon-independent. Because the second K was fixed and the third amino acid was evaluated by arbitrarily selecting representative amino acids with different properties, this limited evaluation is incomplete, but at least SK, AK, and KK, after the M may be important for enhancing protein production.
Separately conducted in vitro transcription and translation experiments revealed that the SKIK tag did not affect the amount of transcription product but only increased translation (Figs. 2 and S5). In our previous E. coli in vivo experiment, the mRNA level quantified with real-time PCR and the amount of mRNA per cell decreased when using the SKIK tag, whereas the amount of protein increased by 30-fold (4). Taken together, SKIK tagging significantly increases the efficiency of  E. coli in vivo and in vitro translation. However, because the present study only compares the amount of products, further detailed analysis is required to determine whether there is a change in the transcription/translation rate due to SKIK tagging.
Combining SecM AP and SKIKs revealed unexpectedly that placing MSKIK or SKIK immediately before the AP counteracts the effect of the arrest and increases protein production (Fig. 3). This observation suggests that the nascent chain, MSKIK or SKIK, at the N-terminus and inside the ORF cancels the arrest caused by SecM immediately after this nascent chain is generated. The insertion of MSKIK and SKIK immediately after the SecM AP did not increase the protein production and no bands were detected (Fig. 3, lane 9 and 10), suggesting that generating nascent SKIK peptide before SecM AP in translation process is important to cancel the arrest. In our previous studies, N-terminal SKIK peptide tagging showed a protein production-enhancing effect on difficult-to-express proteins (4), but it remains unknown whether these proteins contain RAPs or not. In this study, inserting SKIK peptide to the SecM AP-free construct also increased its production compared to the no SKIK (Fig. 3,  lane 3 and 4), suggesting that SKIK may simply have the effect of increasing the efficiency of translation itself independent of the presence of the APs. The kinetics of translation elongation are critical for driving gene expression and maintaining proteostasis, including nascent chain folding and binding to different ribosome-associated machinery (25). The ribosome arrest mechanism of SecM has been analyzed extensively (26)(27)(28), and the nascent SecM chain is known to interact with the negatively charged surface of the outer ribosome to stabilize the arrest (29). The arrest by SecM AP is thought to be susceptible to a pulling force exerted by the nascent chain outside the ribosome (30). Simulations of the stalling by SecM AP showed that the structure of the nascent peptide tens of angstrom distal from the active catalytic center of the ribosome and its interaction with the exit tunnel are important for ribosome arrest (31). On the basis of the proposed mechanism of SecM AP and our results combining SecM AP and MSKIK on translation efficiency, we postulate that the nascent chain MSKIK, which presents just before SecM AP within the ribosomal exit tunnel, alters the native interaction between SecM AP and the ribosome to prevent or disrupt the arrest. Understanding how the nascent MSKIK peptide causes structural changes to SecM AP or acts directly on the movement of the ribosome is of great interest. The distance between SKIK and AP appears to be an important factor because the translation enhancing effect was reduced when the distance between the SKIK and the AP was increased. Chloramphenicol physically prevents the formation of the transition state of peptide-bond formation during translation. Chloramphenicol stalls the ribosome at a specific mRNA location, and the CmlA leader ORF is a chloramphenicoldependent AP (32). Surprisingly, N-terminal SKIK tagging increased protein production even in the presence of chloramphenicol (Fig. 4). Inhibition of protein synthesis by chloramphenicol is most effective when A, S, and T are present in the nascent peptide inside the ribosome, whereas the inhibitory effect of the drug is counteracted by the presence of G at the C-terminus of the nascent chain or the presence of G in the incoming aminoacyl-tRNA (33). However, the details of these mechanisms remain unresolved. Based on these reports and the results herein, it is plausible that the nascent chain MSKIK promotes peptidyl bond formation that occurs directly after MSKIK is created. We hypothesize that the nascent chain prevents the subsequent arrest because the MSKIK peptide is generated before AP in the translation process. As for WPPP, no arrest cancellation was observed by SKIK tagging. Because of its pyrrolyl ring structure of proline, two or more repeat of prolines (PPX motif) is known to stall the ribosome by delaying peptide bond formation (34). The stalling can be released by translation elongation factor EF-P, which interacts with the P-site tRNA in the ribosome and facilitates peptide bond formation between proline residues (35,36). These reports suggest that external factors such as EF-P may be important to cancel the ribosomal stalling caused by PPX motif. On contrast, Elgamal et al. described that amino acids encoded immediately upstream of PPX motif strongly influence on an EF-P-alleviated ribosomal stalling (37). Taken together such a report with the fact that the actual sequence of WPPP is FQKWGIWPPP and the results of Figure 3, which suggest the importance of the distance between SKIK and AP in the case of SecM AP, it will be necessary to investigate the effect of distance between PPX motif and SKIK. Leininger et al. recently demonstrated using multiscale modeling simulations that consecutive positively charged residues slow down translation by generating large forces that move the P-site amino acid away from the A-site amino acid (38). Using such a simulation approach, we postulate that explaining the translation-enhancing mechanism may be through mechanical forces generated between MSKIK and RAPs during translation. In addition, because many RAPs other than SecM or CmlA leader have been found (39) and a recent study showed a force-sensitive arrest by MifM in Bacillus subtilis was canceled by N terminally-adjacent dynamic nascent chain (40), we expect that there are still many translation-promoting peptides remain to be undiscovered.
In conclusion, we have discovered that the nascent chain MSKIK peptide cancels the ribosomal stalling caused by RAPs, such as SecM and CmlA leader. Although our data cannot reveal whether MSKIK prevents or releases the arrest, the new insights obtained herein provide important clues to further understand the translation process by the ribosome in terms of regulation of translation by nascent chains.

Construction of 36 SKIK codon variants
The originally used codons encoding SKIK, that is, TCT AAA ATA AAA, were converted into 35 different patterns consisting of Ser (TCT, TCC, TCA, TCG, AGT, or AGC), Lys (AAA or AAG), Ile (ATT, ATC, or ATA), and Lys (AAA). The final Lys codon was fixed as AAA (Table 1). To convert the codons, we performed PCR using each primer listed in Table S1 and the plasmid pET22b-r1scFv-His tag harboring the rabbit scFv antibody gene as the template, which was challenging to express in E. coli (4,41). The PCR program was as follows: 94 C 3 min; (94 C, 30 s; 55 C, 30 s; 72 C, 5 min) × 25 cycles; and 72 C, 3 min, using KOD plus (Toyobo). The DpnI-(Takara Bio Inc.) treated PCR products were directly incorporated into E. coli XL10-Gold competent cells (Agilent Technology). The colonies formed on LB agar were checked by colony-directed PCR (94 C, 3 min; [94 C, 30 s; 55 C, 15 s; 72 C, 30 s] × 25 cycles; 72 C, 3 min; 12 C, ∞) using Tks Gflex DNA polymerase (Takara Bio Inc.) with primers F1 and R1, both of which anneal outside the T7 promoter and terminator regions (Table S1). Unless otherwise stated, 100 mg/L ampicillin was included in all cultures. The sequences of whole DNA fragments were confirmed by Sanger sequencing with the same primers. Plasmids from the selected clones were prepared using the FastGene Plasmid Mini Kit (NIPPON Genetics Co., Ltd), according to the manufacturer's manual. Table 1 The codon patterns encoding the SKIK peptide tag Clone no. S K I K   1  TCT  AAA  ATT  AAA  2  TCT  AAA  ATC  AAA  3  TCT  AAA  ATA  AAA  4  TCT  AAG  ATT  AAA  5  TCT  AAG  ATC  AAA  6  TCT  AAG  ATA  AAA  7  TCC  AAA  ATT  AAA  8  TCC  AAA  ATC  AAA  9  TCC  AAA  ATA  AAA  10  TCC  AAG  ATT  AAA  11  TCC  AAG  ATC  AAA  12  TCC  AAG  ATA  AAA  13  TCA  AAA  ATT  AAA  14  TCA  AAA  ATC  AAA  15  TCA  AAA  ATA  AAA  16  TCA  AAG  ATT  AAA  17  TCA  AAG  ATC  AAA  18  TCA  AAG  ATA  AAA  19  TCG  AAA  ATT  AAA  20  TCG  AAA  ATC  AAA  21  TCG  AAA  ATA  AAA  22  TCG  AAG  ATT  AAA  23  TCG  AAG  ATC  AAA  24  TCG  AAG  ATA  AAA  25  AGT  AAA  ATT  AAA  26  AGT  AAA  ATC  AAA  27  AGT  AAA  ATA  AAA  28  AGT  AAG  ATT  AAA  29  AGT  AAG  ATC  AAA  30  AGT  AAG  ATA  AAA  31  AGC  AAA  ATT  AAA  32  AGC  AAA  ATC  AAA  33  AGC  AAA  ATA  AAA  34  AGC  AAG  ATT  AAA  35  AGC  AAG  ATC  AAA  36 AGC AAG ATA AAA Expression in E. coli by IPTG induction E. coli SHuffle T7 Express (New England Biolabs), which is used to express disulfide bond-containing proteins such as antibodies (42,43), was used as a host in this study. E. coli SHuffle T7 Express cells were transformed with each of the 36 plasmids that contained different codons for SKIK (Table 1). The transformants obtained on LB agar plates were inoculated into 4 ml of liquid LB medium and cultured at 37 C for 20 h as precultures. A portion of these cultures (40 μl) was transferred to 4 ml of fresh LB medium and cultivated with shaking at 37 C for 4 h until the A 600 reached 0.4 to 0.5. Four microliters of 0.1 M IPTG were then added. Protein expression was performed at 30 C for 3 h. All cultures were grown with a shaking rate of 110 rpm. The cells were harvested by centrifugation (13,000g, 2 min, 4 C), and the cell density was adjusted to be the same in each sample by suspending with an appropriate volume of PBS (137 mM NaCl, 8

Expression in E. coli by autoinduction
The preculture cells (40 μl) described above were transferred into 4 ml of terrific broth liquid medium (20 g tryptone, 5 g yeast extract, 5 g NaCl, 6 g Na 2 HPO 4 , and 3 g KH 2 PO 4 per 1 L, pH 7.2) containing 2 μl 10% glucose and 100 μl 8% lactose and incubated at 30 C for 72 h with shaking at 110 rpm (44). Cell samples were harvested as described above and suspended in PBS to give equivalent A 600 values. Five microliters of the bacterial suspensions was subjected to SDS-PAGE and stained with CBB.

In vitro transcription
DNA fragments containing the T7 promoter, rabbit antibody genes (HcG_22 and Lc_22 in Table S2), and T7 terminator were amplified with primer pair F1 and R1 by Tks Gflex DNA polymerase using a plasmid based on pET22b as a template. PCR products were purified by using a silica membrane column (EconoSpin, Ajinomoto BioPharma) and eluted with Tris-EDTA buffer (10 mM Tris-HCl, 1 mM EDTA-2Na, pH 8.0). In vitro transcription and purification of mRNA were carried out according to the cell-free protein synthesis kit instructions (NU Protein). In detail, the following were mixed and incubated at 37 C for 3 h: 2.5 μl 10× transcription buffer, 2.5 μl 25 mM dNTPs, 1.25 μl 0.1 M DTT, 15 μl DEPC water, 1 μl T7 RNA polymerase, and 2.5 μl DNA fragment (40-60 ng) as the template. After the reaction, 10 μl 4 M ammonium acetate and 100 μl chilled 99.5% ethanol were added sequentially. Precipitants after centrifugation (13,000g, 20 min, 4 C) were dried under atmospheric conditions and dissolved with Tris-EDTA buffer.
Protein expression in the E. coli in vitro expression system The above-transcribed products (mRNA) or purified PCR products containing the T7 promoter and terminator were used as templates for the E. coli-reconstituted cell-free protein synthesis system (PURE system) (PUREfrex 2.0 and DS supplement, Gene Frontier Inc.). The reaction was performed at the 10 or 20 μl scale and 37 C for 90 min, as recommended by the manufacturer. Fluorescent Lys (fLys)-tRNA (Promega) was also included at 1/10th the reaction volume if required. For analysis, 5 μl of the reaction solution, 2 μl PBS, 1 μl 100 mM DTT, and 2 μl 5 × sample buffer were mixed and boiled for 3 min, subjected to SDS-PAGE, and analyzed using the Typhoon FLA 9000 (Cytiva) laser imaging scanner to detect proteins incorporating fLys. A reaction without the template was also performed as a negative control. ). The amino acid residues corresponding to X were selected to cover different properties. Insertion of tags was carried out with the same method as for constructing 36 variants using each primer pair named as three letters amino acid -F and -R, as listed in Table S1. In vivo expression following IPTG induction and whole cell SDS-PAGE analysis were performed using the same procedure described above.

Construction of SecM-containing plasmids
The SecM gene (Gene ID: 944831) was amplified from the E. coli K12 strain (NBRC3301) obtained from National Biological Resource Center. The genome was extracted by boiling the cell pellet collected from a one-night 1 ml culture in LB medium with 0.2 ml 25 mM NaOH solution for 10 min and then neutralized by adding 32 μl 1 M Tris-HCl (pH 7.0). This genome was used as the template for the amplification of SecM with KOD plus DNA polymerase. The primers 5 0 -GCCGAA CCAAACGCGCCCGCAAAAG-3 0 and 5 0 -GGTGAGGCG TTGAGGGCCAGCAC-3 0 were designed for cloning the fulllength SecM gene except for the signal sequence (Table S1). A total of eight plasmids encoding polypeptides presented in Figure 3A were constructed by rearranging the order of the genes and connecting them using the HiFi assembly technique (New England Biolabs). The ATG start codon containing the NdeI site of pET22b was retained in all constructs. The DNA sequences constructed here are summarized in Table S2.

Evaluating the combination of SKIK-CmlA leader and SKIK-WPPP
We constructed plasmids with the basic structure of AP-sfGFP-His tag with and without N-terminal SKIK to investigate the effect of SKIK peptides on non-SecM RAPs. The E. coli CmlA leader (KNAD), whose arrest is induced by chloramphenicol (33), and the artificially designed AP, WPPP Nascent MSKIK peptide cancels ribosomal stalling (FQKYGIWPPP), were selected (45). The sfGFP gene was synthesized by GenScript (46) and cloned into pET22b. In general, each plasmid was constructed by inverse PCR with primers containing the corresponding peptide-coding nucleotides and pET22b-sfGFP as the template and subsequently selfcyclized by HiFi assembly followed by transformation into E. coli. In the case of constructing pET22b-SecM AP-sfGFP-His tag and pET22b-SKIK-SecM AP-sfGFP-His tag, the SecM(-AP) ORFs in the constructs 5 and 6 shown in Figure 3A were replaced to sfGFP by assembling with linearized vectors. After purifying the plasmid, PCR-amplified products with primer F1 and R1 were subjected to cell-free protein synthesis, as described above. For CmlA leader, the final concentration of chloramphenicol was adjusted to 0 to 10 μg/ml by adding 1/ 10th volume of 0 to 100 μg/ml chloramphenicol in ethanol into the reaction mixture.

Protein analysis by CBB staining, dot blot, and Western blotting
After SDS-PAGE, proteins were analyzed by CBB staining and fLys detection for in vivo and in vitro expression, respectively. For scFv expressed by IPTG induction, dot blot analysis was also performed to quantify the expressed proteins in several series. In detail, the collected bacterial cells were suspended in 1 ml PBS, centrifuged (13,000g, 2 min, 4 C), and resuspended in 3 ml 6 M guanidine hydrochloride (GuHCl) buffer (200 mM NaCl, 100 mM Tris-HCl, 6 M GuHCl, 10 mM DTT, pH 8.0). The cells were disrupted by sonication; then 1 μl of this sample was dropwise added to a nitrocellulose membrane and allowed to dry. Purified scFv of known concentrations were also spotted for calibration. Next, the sample was soaked in blocking buffer (1 w/v% Block Ace, KAC Ltd) for 1 h. After washing with PBS-T (10 min x 3 times), the membrane was reacted with a solution of anti-His HRP-conjugate (Medical & Biological Laboratories Co, Ltd, 5,000× dilution with PBS) for 1 h at room temperature with gentle shaking. Then, after washing with PBS-T (as described above), 1-Step Ultra TMB Blotting Solution (Thermo Fisher Scientific K.K.) was added to visualize the His-tagged proteins. All reactions by dot blotting were performed in six series, and the results were quantified using ImageJ software (47) (https://imagej.nih.gov/ij/index.html). The standard His-tagged scFv used here was purified with Ni-sepharose resin in advance (0.61 mg/ml; calculated by the Bradford method). For Western blotting, 1 μl of the reaction solution after cell-free protein synthesis, 2.5 μl sterile water, 0.5 μl 0.1 M DTT, and 1 μl 5 × sample buffer were mixed, boiled for 3 min, and subjected to SDS-PAGE analysis. The proteins were transferred onto a nitrocellulose membrane by a Transblot SD cell system (Bio-Rad) with standard buffer (25 mM Tris, 192 mM Glycine, 20 % MeOH, pH 8.3) at 10 V for 30 min. Subsequent operations were the same as above, but anti-HA tag-HRP (FUJIFILM Wako Pure Chemical, 10,000× dilution with PBS) was used for detecting HA-tagfused proteins.

Fluorescence analysis
The GFP fluorescence intensity was measured as follows. Reaction solutions in the cell-free protein synthesis system were each diluted 10-fold with water following the reaction and 40 μl per well dispensed into Flat Bottom Microfluor Plates Black (Thermo Fisher Scientific), and the fluorescence intensity was measured (excitation 485 nm (bandwidth 9 nm)/ emission 535 nm (bandwidth 20 nm)) using a microplate reader (Tecan Infinite 200 pro (Tecan)).

Data availability
All data are contained within the manuscript. Information of DNA sequences constructed in this study is available in Table S2. All unique/stable materials generated in this study are available from the lead contact Teruyo Ojima-Kato (teruyo@agr.nagoya-u.ac.jp) with a completed materials transfer agreement.
Supporting information-This article contains supporting information.