Selection of Short 5′-UTR of Chemically Synthesized mRNA to Improve Translation Efficiency

Abstract

The advent of mRNA medicine, initially implemented as a vaccine during the coronavirus disease 2019 (COVID-19) pandemic, has attracted interest in diverse therapeutic applications, including cancer vaccines and protein replacement therapies. Our group recently established a method for the complete chemical synthesis of mRNA. Although this approach has some advantages, chemically synthesized mRNA is limited to approximately 150 nucleotides in length and necessitates optimized designs for untranslated regions (UTRs) and coding sequences. To address this challenge, we investigated whether the non-reporter-based selection methods, including ribosome profiling and polysome profiling, which were often used for UTR optimization in long mRNA, could be adapted for short mRNA to identify highly translated UTR sequences. Using these methods, we collected mRNAs that interacted with ribosomes and analyzed their 5′-UTR sequences. We successfully identified a 9-nucleotide 5′-UTR that demonstrated approximately double the translation efficiency of the Kozak sequence, a widely used positive control. This work highlights the adaptability of ribosome-focused selection techniques for short, chemically synthesized mRNA and provides a foundation for effective sequence design. These findings advance the development of chemically synthesized mRNA as a viable alternative to in vitro-transcribed mRNA, paving the way for innovative therapeutic applications.

Introduction

mRNA was first implemented as a vaccine for the coronavirus disease 2019 (COVID-19) pandemic.^1,2) In the future, various applications such as cancer vaccines and protein replacement therapies are expected.³⁾ Generally, mRNA is prepared using an in vitro transcription (IVT) method involving enzymes. The production of mRNA via IVT requires multiple steps, starting with the design and synthesis of the DNA template, followed by the transcription reaction to synthesize and purify the mRNA.

In contrast, our group recently achieved the complete chemical synthesis of mRNA.⁴⁾ In this method, during RNA synthesis using a nucleic acid synthesizer, the 5′ end is monophosphorylated, and then, the cap structure is chemically introduced by reacting the cap reagent, 7-methylguanosine 5′-diphosphate imidazolide (Im-m7GDP), in the presence of an activator.⁴⁾ The complete chemical synthesis method for mRNA differs from the IVT-derived mRNA production process in that it does not require the production of a DNA template and directly synthesizes mRNA, which leads to cost reduction and shortened manufacturing time. Furthermore, the chemical synthesis method allows the introduction of position-specific chemical modifications into RNA. Therefore, chemically synthesized mRNA has garnered attention and is expected to be a new therapeutic approach for cancer treatment and immunotherapy.^5,6) On the other hand, while the synthesis length of mRNA by IVT can be sufficient up to about 10000 bases, the maximum length for chemically synthesized mRNA is about 150 bases. Therefore, when designing untranslated regions (UTRs) and protein-coding regions in chemically synthesized mRNA, it is necessary to design the minimal functional sequences. Thus, aiming to establish a sequence design for chemically synthesized mRNA, we planned to develop optimization techniques for non-translated regions.

There have been no reports of optimization attempts for non-translated regions in about 150-base mRNA synthesized by chemical synthesis, so we examined whether sequence optimization techniques for long mRNA synthesized by IVT could be applied. In addition to selection methods utilizing expressed proteins,^7–10) selection methods that do not depend on reporters and focus on ribosomes involved in translation have also been reported. The polysome profiling method, which falls into the latter category, allows the analysis of translation states by separating and recovering mRNA based on the number of ribosome bindings. Several examples of the selection of UTR sequences using this method have been reported.^11–13) Also, ribosome profiling, which investigates the sequences bound by ribosomes in mRNA, has been used to predict mRNA translation efficiency.^14,15) Therefore, in this study, we applied selection methods used for long mRNA to short mRNA to experimentally determine whether useful sequences could be obtained. Specifically, using ribosome profiling and polysome profiling, we recovered mRNA bound by ribosomes and analyzed the 5′-UTR sequences of this mRNA. As a result, we identified a 9-base 5′-UTR sequence that showed about twice the translation efficiency compared to the sequence containing the Kozak sequence used as a positive control. This method is expected to be useful for the sequence design of chemically synthesized mRNA.

Results

mRNA Design and in Vitro Translation

In this study, a full-length 68-nucleotide (nt) RNA sequence encoding the HiBit peptide, which is a short peptide consisting of 11 amino acids, was used as a model mRNA. The HiBit peptide can bind to the LgBit protein to reconstitute Nano Luciferase, which emits luminescence in the presence of a substrate. This experimental system was chosen for several reasons, including the ease of evaluating translation activity after selection and the fact that the activity of HiBit with a 9-nt short 5′-UTR containing the Kozak sequence has been confirmed in previous studies.⁴⁾ We synthesized 2 short mRNAs of 68 nt, encoding HiBit and containing either a 9-nt 5′-UTR with the 6 nt of the Kozak sequence (PC, RNA1) or a randomized 5′-UTR (random, RNA2) (Fig. 1 and Supplementary Fig. S1). To avoid any influence from transfection efficiency on the results of this study, the translation activity of the synthesized mRNA was evaluated using HeLa.S3 extract. After a certain period of translation reaction, the translation was stopped by adding cycloheximide and chloramphenicol on ice. mRNA selection was performed using the halted translation solution, and the 5′-UTR sequences of the collected mRNA were analyzed by next-generation sequencing (NGS).

Fig. 1. Schematic Image of Short mRNA

A 68-nt short mRNA is composed of 5′ cap, 9 nt of 5′-UTR, 33 nt CDS coding the HiBit peptide, a stop codon, and poly A20. The translation was quenched by the addition of cycloheximide. The quenched solution was used for mRNA selection. The recovered mRNA was analyzed by next-generation sequencing.

mRNA Selection Using Ribosome Profiling

In a typical ribosome profiling method, translation is halted by adding cycloheximide, which prevents ribosomes from dissociating from the mRNA, thus forming mRNA–ribosome complexes during translation.¹⁴⁾ This method analyzes only the regions where ribosomes are bound. To achieve this, the unprotected portions of the ribosome-bound mRNA are first digested. The remaining sample is then layered onto a sucrose solution and subjected to ultracentrifugation. mRNA fragments with 1 or more ribosomes attached can be recovered as a pellet, which allows for the investigation of mRNA sequences with ribosome occupancy.¹⁴⁾ This enables the analysis of the ribosome’s position and the ribosome density on mRNA. On the other hand, because the mRNA length used in this study is short, it is expected that only 1 ribosome will bind at most. Therefore, we attempted to separate mRNA based on the presence or absence of ribosomes by placing the translation mixture onto a sucrose solution and performing ultracentrifugation, without nuclease digestion step (Fig. 2A).

Fig. 2. Workflow and Result from Ribosome Profiling

(A) Schematic image of mRNA selection by ribosome profiling. The quenched translation mixture was placed onto the sucrose solution and ultracentrifuged. After ultracentrifugation, mRNA from the pellet was recovered. (B) The difference in mRNA recovery ratio from pellets between cap (+) and cap (−) mRNA was increased over time. After ribosome profiling, mRNA from both supernatant and pellet were recovered and quantified. The mRNA recovery rate was calculated as follows: the amount of RNA recovered in the pellet divided by the RNA amount recovered in both the pellet and supernatant.

Because there were no previous examples of applying ribosome profiling to full-length mRNA, this study first aimed to confirm whether short mRNAs could be separated based on the presence of ribosomes. For this, a no-cap mRNA (cap [−], RNA3) was prepared as a negative control, and a capped mRNA (cap [+], RNA4), which is crucial for translation, was prepared as a positive control (Supplementary Figs. S2A–S2C and Supplementary Table S3). Translation was performed using these mRNAs, and their translation level over time was monitored using a luciferase assay (Supplementary Fig. S2D). As a result, HiBit protein expression reached a plateau 1 h after translation initiation. The translation level of cap (+) after 1 h of incubation was about 2.6 times higher than that of cap (−) (Supplementary Fig. S2D).

Next, ribosome profiling was performed using either cap (+) or cap (−) mRNAs, and the amount of mRNA recovered from each was compared. Cap (+) and cap (−) mRNAs were translated, and translation was quenched by adding cycloheximide at 5, 15, and 60 min. Ribosome profiling was then conducted. The amounts of mRNA in the supernatant and pellet after ultracentrifugation were quantified using RT-quantitative (q)PCR. The results showed that for cap (−) mRNA, the amount of mRNA recovered from the pellet remained constant, even as the translation time changed from 5 to 60 min (Fig. 2B). On the other hand, for the positive control cap (+) mRNA, the mRNA recovery from the pellet increased from 25 to 45% along with translation time (Fig. 2B). Although some translation occurred even with cap (−) mRNA, as shown in Supplementary Fig. S2D, this likely led to the recovery of mRNA from the pellet as well. However, the difference in recovery amounts between cap (−) and cap (+) mRNA increased over. time, reflecting the difference in translation efficiency between cap (−) and cap (+). These results indicate that ribosome profiling is useful not only for analyzing general ribosome-protected mRNA sequences but also for selecting short mRNAs based on translation levels. We applied the ribosome profiling method to select highly translated sequences from the mRNA library containing randomized 9 nt as 5′-UTRs (random, RNA2) following translation.

mRNA Selection Using Polysome Profiling

Polysome profiling, a method commonly used in mRNA selection, is a technique that separates mRNAs based on ribosome density^11,13) (Fig. 3A). Similar to ribosome profiling, this method utilizes mRNA–ribosome complexes in the translation mixture. These complexes are then loaded onto a sucrose solution with a concentration gradient and separated by ultracentrifugation, allowing the mRNA to be isolated based on the number of ribosomes bound (Fig. 3A). mRNAs with higher translation efficiency tend to have multiple ribosomes (polysomes) bound to a single mRNA molecule, and by recovering and sequencing them, mRNA selection focused on translation efficiency becomes possible. However, unlike long mRNAs, the mRNAs used in this study are short, with lengths under 100 nt. It has been reported that the distance between ribosomes during translation is approximately 270 nt,¹⁶⁾ and for the 68-nt short mRNA used in this study, it is unlikely that more than 1 ribosome would bind to a single mRNA molecule due to the length constraint. Therefore, after performing polysome profiling, we recovered short mRNAs from the fraction containing a single ribosome (monosome) (Fig. 3B). We applied the polysome profiling method to select highly translated sequences from an mRNA library containing randomized 9-nt sequences as 5′-UTRs (random, RNA2) following translation.

Fig. 3. Workflow and Results from Polysome Profiling

(A) Schematic image of mRNA selection by polysome profiling. mRNA containing a randomized 5′-UTR was translated and quenched by the addition of cycloheximide. The quenched translation mixture was placed onto the sucrose gradient solution and ultracentrifuged. After the fractionation of the solution, mRNA from the monosome fraction was collected. mRNA was converted into DNA through reverse transcription and PCR, followed by sequencing. (B) shows the fractionation profile of polysome profiling. The navy and gray line show the profile and fraction, respectively. Fractions 3 and 4, indicated by a red box, were collected as the monosome fraction. The monosome peak is indicated by a red arrow.

Results from NGS Analysis

After performing selection using ribosome profiling or polysome profiling, the recovered mRNA was prepared for NGS analysis. Since the mRNA used in this study has a random sequence in the 5′-UTR from the cap structure to the start codon, there is no region for hybridizing PCR primers. Therefore, the cap structure was removed by decapping enzyme to generate a 5′ monophosphate, adapters were ligated, and then reverse transcription and PCR amplification were performed to prepare the NGS library (Supplementary Fig. S3). Sequence analysis was conducted using NGS, and the read occupancy rate for each 9-nt sequence was calculated by dividing the read count of each sequence by the total number of reads. The enrichment score was then calculated by dividing the read occupancy rate after selection by the read occupancy rate before selection, which was used as an indicator of translation efficiency improvement for the 9-nt UTR sequences. After excluding sequences with 0 read counts in the data before selection, the enrichment score for each sequence was calculated. The enrichment scores obtained from each selection method were visualized as histograms (Supplementary Figs. S4A, S4B). The results showed that the enrichment scores varied between sequences, but when comparing the results of polysome profiling and ribosome profiling, similar sequences were obtained (Supplementary Figs. S4A, S4B). Next, the results from the NGS analysis for each method were compared using a scatter plot (Supplementary Fig. S4C). Although the correlation coefficient was not high, sequences with a high enrichment score tended to show relatively high enrichment in both methods (Supplementary Fig. S4C). The enrichment scores were ranked, and sequence logos were depicted for sequences that showed higher enrichment than the previously reported short 5′-UTR as standard.⁴⁾ As a result, while there was little bias between nucleotides, 5′-UTRs containing relatively many consecutive A bases were obtained from both selection methods (Supplementary Figs. S4D, S4E). High enrichment scores suggest that the mRNA is more likely to have ribosomes bound, indicating high translation efficiency. Therefore, the top 5 sequences with the highest enrichment scores and the bottom 5 sequences with the highest read counts before selection but an enrichment score of 0 were selected and designated as the high translation efficiency (TOP) and low translation efficiency (BOTTOM) sequences, respectively. The TOP5 and BOTTOM5 sequences for each method are summarized in Table 1.

Table 1. List of 5′-UTR Sequence Used for Luciferase Assay

Sample ID	5′-UTR sequence (5′–3′)	Used in
RNA1	AGAGCCACC	mRNA selection as positive control (PC)
RNA2	NNNNNNNNN	mRNA selection as randomized mRNA
RNA5	CUCGUAAAA	Polysome profiling-TOP1
RNA6	CCUUAAUAA	Polysome profiling-TOP2; Ribosome profiling-TOP2
RNA7	CAUAAAAAU	Polysome profiling-TOP3; Ribosome profiling-TOP3
RNA8	CACAAAAAU	Polysome profiling-TOP4; Ribosome profiling-TOP1
RNA9	CUCUUAAAC	Polysome profiling-TOP5; Ribosome profiling-TOP5
RNA10	CUCUAUUAA	Ribosome profiling-TOP4
RNA11	AGCUGGCAU	Polysome profiling-BOTTOM1
RNA12	GUUAAAACA	Polysome profiling-BOTTOM2
RNA13	AAAACCUUA	Polysome profiling-BOTTOM3
RNA14	GACCUAUAA	Polysome profiling-BOTTOM4
RNA15	GGGGUAUAU	Polysome profiling-BOTTOM5
RNA16	GGUCAAACC	Ribosome profiling-BOTTOM1
RNA17	AAUCGUUCA	Ribosome profiling-BOTTOM2
RNA18	ACAGGCGCA	Ribosome profiling-BOTTOM3
RNA19	ACUGCGUUA	Ribosome profiling-BOTTOM4
RNA20	CCACCCCUU	Ribosome profiling-BOTTOM5

Synthesis and Evaluation of Sequences Showing High or Low Enrichment Score

The sequences with high and low enrichment scores (Table 1) were individually synthesized using a nucleic acid synthesizer, and after performing a chemical capping reaction based on previous studies,⁴⁾ the purity was confirmed by gel electrophoresis and LC-MS analysis (Supplementary Fig. S5 and Supplementary Table S3). Next, the translation efficiency of the synthesized mRNA was individually evaluated using a luciferase assay. The results showed that, although there were exceptions, sequences with higher enrichment scores tended to show higher translation efficiency than sequences with lower enrichment scores, regardless of the method used for selection (Fig. 4 and Supplementary Fig. S6). Furthermore, when the translation efficiency of high-enrichment mRNA was compared to that of the PC mRNA containing the Kozak sequence (RNA1), a short 5′-UTR sequence was found to show approximately twice the translation activity of the PC mRNA (Table 2). On the other hand, there were sequences with high enrichment scores that exhibited lower luciferase activity. Additionally, when only the TOP5 and BOTTOM5 sequences were used to investigate the correlation of enrichment scores, a strong correlation with a correlation coefficient of 0.9 was observed between polysome profiling and ribosome profiling (Supplementary Fig. S7). Since sequences with high enrichment scores were recovered from both polysome profiling and ribosome profiling, it was suggested that they readily form complexes with ribosomes.

Fig. 4. Luciferase Assay of mRNAs with 5′-UTRs of TOP5 or BOTTOM5 Enrichment Score from Polysome Profiling Selection of Randomized mRNA

Experiments were performed in triplicate, and data are presented as the mean ± standard deviation (S.D.). One-way ANOVA with Tukey’s multiple comparison test was used to determine the significance between the different groups. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 vs. Kozak.

Table 2. Selected 5′-UTR Sequences with Statistically Significant Difference Compared to PC (Kozak)

Sequence		Translation	Samples
CUCGUAAAA	RNA5	1.98	Polysome profiling random only-TOP1
GUUAAAACA	RNA12	1.95	Polysome profiling random only-BOTTOM2

Discussion

In this study, we selected short mRNA sequences with high translation efficiency using 2 selection methods, ribosome profiling and polysome profiling. Ribosome profiling broadly collects mRNA with 1 or more ribosomes bound, while polysome profiling is a method that fractionates mRNA based on the number of ribosomes bound. In fact, the number of sequence types recovered from polysome profiling was fewer than that from ribosome profiling, suggesting that the recovered sequences likely contain mRNA with high translation efficiency (Supplementary Fig. S4). The strong correlation of the enrichment ratio between ribosome profiling and polysome profiling methods, with a correlation coefficient exceeding 0.9 (Supplementary Fig. S7), confirmed that both methods effectively select mRNA that readily forms complexes with ribosomes. However, since polysome profiling requires specialized equipment, ribosome profiling has the advantage of being simpler and more suitable for experiments in terms of ease of use. Although each method has its strengths and limitations, both were shown to be effective tools for seeking mRNA with high translation efficiency. Next, as a main result of this study, we successfully identified a short 5′-UTR sequence exhibiting higher luciferase activity than the Kozak sequence used as a positive control. According to Dikstein and colleagues, the translation initiator of short 5′-UTR (TISU) sequence present in short 5′-UTRs is known to reduce the probability of leaky scanning and enhance translation efficiency.^17,18) However, it had not been sufficiently verified whether these sequences are still effective for short mRNA sequences. The short 5′-UTR sequence identified in this study, which showed high translation efficiency, did not contain Kozak or TISU sequences, suggesting the presence of a unique sequence motif different from known translation efficiency-enhancing sequences. These sequences seem to contain stretches of consecutive sequences of adenines (Table 2). In eukaryotic organisms, short stretches of adenines in the 5′-UTR have been observed. It has been reported that such sequences can interact with initiation factors and enhance translation efficiency.^19,20) Furthermore, in circular mRNAs, sequences containing short polyA stretches have been discovered to exhibit Internal Ribosome Entry Site-like activity.²¹⁾ Similarly, the sequences identified in this study may also recruit ribosomes and enhance translation efficiency. Next, the applicability of this sequence to mRNAs with different open reading frame (ORF) sequences was considered based on the 2 mRNA sequences with high translation efficiency (Table 2). The mRNA structure was predicted using the mfold web server.²²⁾ The 5′-UTR of RNA5, which exhibited the highest translation efficiency, does not seem to interact with the ORF sequence (Supplementary Fig. S8). While it might be influenced by the ORF sequence, this effect seems minimal. In contrast, the 5′-UTR of RNA12 hybridized with its ORF sequence (Supplementary Fig. S8), suggesting that this 5′-UTR sequence is more susceptible to the ORF sequence. Sequences with higher enrichment ratios generally showed higher translation efficiency, confirming ribosome enrichment as a valid indicator. However, some sequences with high enrichment in both ribosome profiling and polysome profiling exhibited low luciferase activity. These sequences may contain specific motifs or structures that interact non-translationally with ribosomes. This phenomenon has also been observed in long non-coding RNAs.^23–25)

A limitation of this study is that unstable mRNA may have been degraded during selection due to the use of cell extracts. However, since mRNA stability is not necessarily correlated with translation efficiency,²⁶⁾ it is speculated that high luciferase activity was observed in some sequences with lower enrichment ratios. Nevertheless, sequences with high enrichment ratios tended to show higher luciferase activity than those with lower enrichment ratios, indicating that these selection methods are useful for screening highly translated short mRNAs.

Conclusion

In this study, we aimed to identify short 5′-UTR sequences that exhibit high translation efficiency in chemically synthesized mRNAs. To achieve this, we modified ribosome profiling and polysome profiling methods for the selection of short mRNAs, and performed the selection of short mRNAs with a 9-nt random sequence in the 5′-UTR. The mRNAs containing the obtained UTR sequences were individually synthesized, and their translation efficiency was evaluated using a luciferase assay. As a result, we identified a 5′-UTR sequence that demonstrated a 2-fold higher translation efficacy compared to the mRNA containing the Kozak sequence, which was used as a positive control. By partially modifying conventional methods used for the exploration of long mRNAs, we demonstrated that they can also be applied to the exploration of short mRNAs, making this approach a useful tool for the functional design of short mRNAs in the future.

Experimental

mRNA Synthesis

Sixty-eight-nt RNAs were chemically synthesized using an automated oligonucleotide synthesizer (nucleotide reactor NRs-4A10R7 [Nihon Techno Service] or MerMade 48X synthesizer [LGC Biosearch Technologies, Middlesex, U.K.]) with commercially available phosphoramidites (5′-DMT-2′-TOM-ribo guanosine(n-acetyl) OP [ChemGenes], 5′-DMT-2′-TOM-ribo adenosine (n-acetyl) OP [ChemGenes], 5′-DMT-2′-TOM-ribo cytidine (n-acetyl) OP [ChemGenes], 5′-DMT-2′-TOM-ribo uridine OP [ChemGenes], rA (Bz) CPG 3000A [LGC Biosearch Technologies], and rA (Bz) CPG Column [2000A, 200 nmol for MerMade; LGC Biosearch Technologies]). A 5′-monophosphate group was added using 100 mM bis(2-cyanoethyl) N,N-diisopropylphosphoramidite (Angene International Limited) on an automated oligonucleotide synthesizer. The RNA was cleaved from the solid support, and base deprotection was performed with ammonium hydroxide/methylamine for 20 min at 65°C. A speed-vac evaporator (CVE-3000; Tokyo Rikakikai, Tokyo, Japan) completely dried the solvent. The TOM-protecting group was eliminated with 100 μL of dimethyl sulfoxide (DMSO), 50 μL of triethylamine trihydrofluoride, and 50 μL of triethylamine for 2.5 h at 65°C. The deprotection was quenched by the addition of 0.5 M triethylamine acetate (pH 7.0). The quenched solution was desalted by Presep DNA/RNA Type A column (255 mg/3 mL) (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan). RNA was eluted with 50% acetonitrile, and acetonitrile was evaporated by a speed-vac evaporator. The RNA was mixed well with a calcium chloride solution and lyophilized. Then, 10 μM RNA, 10 mM Im-m7GDP, and 10 mM 2-nitroimidazole in DMSO were incubated for 3 h at 55°C. After isopropanol precipitation with 0.3 M sodium acetate (pH 5.2) and 0.2 mg/mL glycogen, RNA was purified by HPLC (Shimadzu, Kyoto, Japan; detection: 260 nm; column: Triart Bio C4 [YMC, 250 mm × 4.6 mm I.D., S-5 μm, 30 nm]; column temperature: 50°C; gradient: 0% B, 5 min; 0–50% B, 20 min; 80% B, 5 min; 0% B, 15 min; solvent A: 50 mM triethylammonium acetate [pH 7.0], 5% acetonitrile; solvent B: acetonitrile). RNA was recovered by ethanol precipitation with 0.3 M sodium acetate (pH 5.2). The RNAs were analyzed by denaturing polyacrylamide gel electrophoresis (PAGE) and LC-MS analysis (UPLC System; Agilent 1290 Infinity II; MS System: Agilent 6530 LC/Q-TOF; column; ACQUITY UPLC Oligonucleotide BEH C18 Column, 130A, 1.7 μm, 2.1 mm × 50 mm; Sol A: 100 mM 1,1,1,3,3,3-hexafluoro-2-propanol [pH 8.3], 8.6 mM triethylamine; Sol B: 100% MeOH; column temperature: 60°C; detection: 260 nm; flow rate: 0.3 mL/min; gradient: 5–10% B, 0.5 min; 10–30% B, 19.5 min; 80% B, 1 min; 5% B, 4 min).

In Vitro Translation

The following preparation method was for a 10 μL reaction scale. The amount of reaction was applied at 10 times the usual amount for mRNA selection. The overall procedure was based on previous studies.²⁷⁾ A measure of 2.43 μL of Mixture-2 without dithiothreitol (DTT) (97.16 mM N-(2-hydroxyethyl)piperazine-N′′-2-ethanesulfonic acid [HEPES]-KOH [pH 7.5], 85.44 mM potassium acetate, 5.42 mM magnesium acetate), and 0.82 μL of 50 mM DTT were mixed with 4.35 μL of HeLa.S3 extract (lab-made) by pipetting. After incubation at room temperature for 10 min, 0.9 μL of Mixture-3 (12.56 mM ATP, 1.22 mM GTP, 200 mM creatine phosphate, 0.6 mg/mL creatine kinase, and a 20 amino acids mix) was added, as well as 10 ng (for luciferase assay) or 0.125 μM (for selection) of random mRNA. The incubation time was 60 min, unless otherwise noted. Also, translation was quenched by the addition of lysis buffer (20 mM Tris–HCl [pH 7.5], 150 mM sodium chloride, 5 mM magnesium chloride, 1 mM DTT, 100 μg/mL cycloheximide, 100 μg/mL chloramphenicol, and 1% Triton X-100) on ice. The quenched translation mixture was centrifuged at 15000 rpm for 10 min at 4°C, and the supernatant was used for the subsequent experiments.

mRNA Selection with Ribosome Profiling

A measure of 0.9 mL of sucrose cushion solution (20 mM Tris–HCl [pH 7.5], 50 mM sodium chloride, 5 mM magnesium chloride, 1 mM DTT, 0.1 mg/mL cycloheximide, 0.1 mg/mL chloramphenicol, and 1 M sucrose) was placed under the quenched sample. An ultracentrifuge (100000 rpm, 1 h, 4°C) using the Optima MAX-TL Ultracentrifuge and TLA-110 rotor (Beckman, Brea, CA, U.S.A.) separated ribosome-associated mRNA from free mRNA by precipitation. The supernatant was carefully removed, and the pellet was used for mRNA recovery.

mRNA Selection with Polysome Profiling

A measure of 10 and 50% sucrose solution containing 20 mM Tris–HCl (pH 7.5), 5 mM magnesium chloride, 150 mM NaCl, 1 mM DTT, and 100 μg/mL cycloheximide were prepared in advance. A 10–50% sucrose gradient was prepared with Gradient Master (Biocomp, San Antonio, TX, U.S.A.). The supernatant of the quenched translation mixture was placed onto the sucrose gradient and centrifuged using an L-70K ultracentrifuge with an SW-41Ti swing rotor (Beckman) at 37000 rpm for 1.5 h at 4°C. The solution was fractionated using a piston gradient fractionator system (Biocomp). The fractions containing monosomes (fractions 3–4 in Fig. 3B) were collected.

mRNA Purification after mRNA Selection

Three hundred microliters of TRI reagent (MOR) was used for the extraction of total RNA. After mixing by vortex and incubating for 5 min at room temperature, 50 μL of chloroform was added, mixed, and incubated for 5 min. Then, the RNA/TRI reagent mixture was centrifuged at 12000 × g for 15 min at 4°C. A certain amount of the supernatant was recovered and moved to further purification. The second purification was conducted using a Monarch RNA purification column (New England Biolabs [NEB], Ipswich, MA, U.S.A.). The purification was conducted following the official procedure. For the final RNA elution step, 50 μL of autoclaved water was used twice.

Quantification of Recovered mRNA Amount

The recovered mRNA was reverse-transcribed with ReverTra Ace RT qPCR kit (Toyobo, Osaka, Japan) and primer (ON1). mRNA was quantified by RT-qPCR (CFX Connect Real-Time System; Bio-Rad, Hercules, CA, U.S.A.) with THUNDERBIRD® NEXT SYBR® qPCR Mix (Toyobo) and qPCR primers (ON2 and ON3). The thermal cycle was conducted as follows: 95°C, 30 s, followed by 40 cycles of 95°C for 5 s and 60°C for 10 s. To calculate the mRNA concentration, PCR products generated from the same RNA using a specific primer at a given concentration were simultaneously quantified as standards. The starting quantity (SQ) was calculated using real-time PCR analysis software, Bio-Rad CFX Manager (Bio-Rad). The volumes of the supernatant and pellet differed significantly, and since only a portion of each was used for purification and quantification, the unused portions were considered for correcting the SQ values.

mRNA Preparation and NGS Analysis

The cap structure of mRNA was removed by an mRNA decapping enzyme (NEB) for 1 h at 37°C, following the provided protocol. The decapped RNA was purified by Agencourt AMPure XP magnetic beads (Beckman Coulter, Brea, CA, U.S.A.). An adapter RNA was ligated to the 5′ end of the decapped RNA. Adapter ligation was conducted with 1× RNA ligase buffer, 2 μM adapter RNA (ON4), 0.07 μL/μL T4 RNA ligase (TaKaRa, Shiga, Japan), and 0.01% BSA for 2 h at 15°C. Magnetic bead purification was conducted again. Reverse transcription was conducted with (1× first-strand buffer, 2.5 mM DTT, 0.5 mM dNTPs, 2.5 μM primer (ON1), and 0.05 μL/μL SuperScript® III Reverse Transcriptase [Thermo Fisher, Waltham, MA, U.S.A.]) for 50 min at 50°C and for 10 min at 70°C. Then, the original RNA strand was digested by RNase A (NIPPON GENE, Tokyo, Japan) for 30 min at 37°C. The cDNA was purified by magnetic beads using the same protocol as before and amplified by PCR to convert it to dsDNA. Based on cDNA strand, dsDNA was amplified by PCR. PCR was performed with the thermal cycle (95°C for 2 min, followed by 20 or 25 cycles of 98°C for 10 s, 50°C for 30 s, and 68°C for 20 s) in a PCR reaction mixture (1× KOD plus neo buffer, 0.2 mM dNTPs, 1.5 mM magnesium sulfate, 0.3 μM PCR primers (ON5 and ON6), and 0.02 μL/μL KOD plus neo [Toyobo]) using the cDNA. The PCR products were analyzed on an 8% native PAGE. After magnetic beads purification, the concentration of PCR products was measured using a NanoDrop (Thermo Fisher). A 2nd PCR was conducted to attach barcode sequences for NGS analysis. The 2nd PCR mixture consisted of 1× Platinum II PCR buffer, 0.2 mM dNTPs, 0.2 μM NEBNext Multiplex Oligos for Illumina (Dual Index Primers Set 1) (NEB)) as 2nd PCR primers, 0.04 U/μL Platinum II Taq Hot-start DNAP (Thermo Fisher), and 0.5 ng/μL PCR products. This mixture was subjected to the following thermal cycle: 95°C, 10 s, followed by 8 cycles of 62°C for 20 s and 72°C for 30 s. Amplified DNA was analyzed on an 8% native PAGE. Magnetic beads purification was performed again. DNA concentration was measured by Qubit dsDNA High sensitivity (Thermo Fisher). DNAs at 0.5 ng/μL in 15 μL nuclease-free water were prepared and submitted to Novogene.

We extracted reads that met the following criteria: the first 30 bases matched the sequence “TAATACGACTCACTATAGGGAGACCCAAGC” with no more than 2 mismatches, and bases 40–69 matched the sequence “ATGGTGAGCGGCTGGCGGCTGTTCAAGAAG” with no more than 2 mismatches. From these reads, we identified the 9-mers at positions 31–39 and counted their occurrence frequencies. The sequences containing undetected bases “N” were eliminated before the sequence analysis. The enrichment score of each sequence was calculated using the formula below. Sequences were ranked by enrichment ratio, excluding those with a start codon in the 5′-UTR and those likely to cause frameshifts. From the remaining sequences, the TOP 5 and BOTTOM 5 were selected, respectively,

  

$$\begin{array}{l}enrichmentscore=\\ \frac{(readcountoftargetsequenceafterselection)}{(readcountoftargetsequencebeforeselection)}\\ \frac{/(readcountoftotalsequenceafterselection)}{/(readcountoftotalsequencebeforeselection)}\end{array}$$

The histograms showing the distribution of enrichment scores were drawn using data that excluded sequences with NGS read counts of 0 before the selection. The sequence logos were generated using the same data as the histograms, including only those samples that exhibited higher enrichment rates than the positive control. Scatter plots between polysome profiling and ribosome profiling were depicted using the same data as the histograms and sequence logos.

Luciferase Assay

A luciferase assay was conducted using the Nano-Glo® HiBiT Lytic Detection System (Promega, Madison, WI, U.S.A.). The luciferase assay mixture was prepared with 20 μL of Nano-Glo HiBit Lytic Buffer, 0.4 μL of Nano-Glo HiBit Lytic Substrate, and 0.2 μL of LgBit protein. A measure of 1 μL of translation mixture was mixed with 20 μL of the luciferase assay mixture and incubated for 5 min with shaking at 500 rpm. Luminescence was measured using a TriStar5 plate reader (Berthold, Vienna, Austria).

Prediction of RNA Structure

RNA structures were predicted using the mfold web server.²²⁾ A structure plot of the first structure was obtained and compared.

Acknowledgments

This work was supported by a Grant-in-Aid for JSPS Fellows from JSPS KAKENHI Grant Number: JP23KJ1106.

Conflict of Interest

The authors declare no conflict of interest.

Supplementary Materials

This article contains supplementary materials.

References

• 1)  Polack F. P., Thomas S. J., Kitchin N., Engl N., et al., N. Engl. J. Med., 383, 2603–2615 (2020).
• 2)  Baden L. R., El Sahly H. M., Essink B., et al., N. Engl. J. Med., 384, 403–416 (2021).
• 3)  Wang Y. S., Kumari M., Chen G. H., Hong M. H., Yuan J. P. Y., Tsai J., Wu H., J. Biomed. Sci., 30, 84 (2023).
• 4)  Abe N., Imaeda A., Inagaki M., Li Z., Kawaguchi D., Onda K., Nakashima Y., Uchida S., Hashiya F., Kimura Y., Abe H., ACS Chem. Biol., 17, 1308–1314 (2022).
• 5)  Husseini R. A., Abe N., Hara T., Abe H., Kogure K., Biol. Pharm. Bull., 46, 301–308 (2023).
• 6)  Imani S., Tagit O., Pichon C., NPJ Vaccines, 9, 14 (2024).
• 7)  Evfratov S. A., Osterman I. A., Komarova E. S., Pogorelskaya A. M., Rubtsova M. P., Zatsepin T. S., Semashko T. A., Kostryukova E. S., Mironov A. A., Burnaev E., Krymova E., Gelfand M. S., Govorun V. M., Bogdanov A. A., Sergiev P. V., Dontsova O. A., Nucleic Acids Res., 45, 3487–3502 (2017).
• 8)  Cao J., Novoa E. M., Zhang Z., Chen W. C. W., Liu D., Choi G. C. G., Wong A. S. L., Wehrspaun C., Kellis M., Lu T. K., Nat. Commun., 12, 4138 (2021).
• 9)  Mie M., Shimizu S., Takahashi F., Kobatake E., Biochem. Biophys. Res. Commun., 373, 48–52 (2008).
• 10)  Barendt P. A., Shah N. A., Barendt G. A., Sarkar C. A., PLoS Genet., 8, e1002598 (2012).
• 11)  Castillo-Hair S. M., Seelig G., Acc. Chem. Res., 55, 24–34 (2022).
• 12)  Leppek K., Byeon G. W., Kladwang W., et al., Nat. Commun., 13, 1536 (2022).
• 13)  Castillo-Hair S., Fedak S., Wang B., Linder J., Havens K., Certo M., Seelig G., Nat. Commun., 15, 5284 (2024).
• 14)  Ingolia N. T., Brar G. A., Rouskin S., McGeachy A. M., Weissman J. S., Nat. Protoc., 7, 1534–1550 (2012).
• 15)  Nakahigashi K., Takai Y., Shiwa Y., Wada M., Honma M., Yoshikawa H., Tomita M., Kanai A., Mori H., BMC Genomics, 15, 1115 (2014).
• 16)  Tomuro K., Mito M., Toh H., Kawamoto N., Miyake T., Chow S. Y. A., Doi M., Ikeuchi Y., Shichino Y., Iwasaki S., Nat. Commun., 15, 7061 (2024).
• 17)  Elfakess R., Dikstein R., PLoS One, 3, e3094 (2008).
• 18)  Elfakess R., Sinvani H., Haimov O., Svitkin Y., Sonenberg N., Dikstein R., Nucleic Acids Res., 39, 7598–7609 (2011).
• 19)  Xia X., MacKay V., Yao X., Wu J., Miura F., Ito T., Morris D. R., Genetics, 189, 469–478 (2011).
• 20)  Gallie D. R., Tanguay R., J. Biol. Chem., 269, 17166–17173 (1994).
• 21)  Fan X., Yang Y., Chen C., Wang Z., Nat. Commun., 13, 3751 (2022).
• 22)  Zuker M., Nucleic Acids Res., 31, 3406–3415 (2003).
• 23)  Ruiz-Orera J., Messeguer X., Subirana J. A., Alba M. M., eLife, 3, e03523 (2014).
• 24)  Zeng C., Fukunaga T., Hamada M., BMC Genomics, 19, 414 (2018).
• 25)  Guttman M., Russell P., Ingolia N. T., Weissman J. S., Lander E. S., Cell, 154, 240–251 (2013).
• 26)  Siwiak M., Zielenkiewicz P., PLOS Comput. Biol., 6, e1000865 (2010).
• 27)  Kobayashi T., Machida K., Imataka H., Methods Mol. Biol., 1118, 149–156 (2014).