Simultaneous in vitro expression of minimal 21 transfer RNAs by tRNA array method

Miyachi, Ryota; Masuda, Keiko; Shimizu, Yoshihiro; Ichihashi, Norikazu

doi:10.1038/s41467-025-62588-y

Download PDF

Article
Open access
Published: 26 August 2025

Simultaneous in vitro expression of minimal 21 transfer RNAs by tRNA array method

Nature Communications volume 16, Article number: 7418 (2025) Cite this article

7083 Accesses
48 Altmetric
Metrics details

Subjects

Abstract

Transfer RNA (tRNA) plays a central role in translation. The simultaneous in vitro synthesis of minimal yet sufficient tRNA species (at least 21) poses a challenge for constructing a self-reproducible artificial cell. A key obstacle is the processing of the 5’ and 3’ ends, which requires a multi-step reaction in the natural cells. In this study, we develop a simplified processing method that allows simultaneous expression of all 21 tRNAs in a reconstituted transcription/translation system (PURE system). We test two methods for 5’-end processing (the leader and 5’-G variants methods) and two methods for 3’-end processing (the direct tRNA linkage and HDVR attachment methods). Finally, by combining the direct tRNA linkage and HDVR attachment methods (newly termed the tRNA array method), we succeed in simultaneously expressing all 21 tRNAs from a single polycistronic DNA template in the PURE system. The tRNA mixture produced by the tRNA array method supports a similar level of translation to the individually synthesized tRNA mixture for luciferase and GFP. This study represents a step toward the realization of self-reproducible artificial cells and also provides an easy method for preparing all tRNAs useful for genetic code engineering.

Continuous in situ synthesis of a complete set of tRNAs sustains steady-state translation in a recombinant cell-free system

Article Open access 05 July 2025

Predictable control of RNA lifetime using engineered degradation-tuning RNAs

Article 21 June 2021

A multiplex platform for small RNA sequencing elucidates multifaceted tRNA stress response and translational regulation

Article Open access 05 May 2022

Introduction

In the field of bottom-up synthetic biology, researchers have focused on creating artificial molecular systems in vitro to deepen our understanding of living systems^{1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16}. One of the major goals is to construct artificial systems with self-reproductivity, a universal ability of living organisms, using minimal components. To date, many efforts have been made to reconstitute molecular systems with functions that contribute to self-reproduction, such as DNA replication utilizing natural^17,18,19,20 or artificial schemes^{21,22,23,24,25}, membrane synthesis^8,26, metabolic reactions^27,28,29, and membrane division^30,31. However, the construction of a transcription/translation (TxTL) system capable of self-regeneration remains a significant challenge³². To realize such systems, it is necessary to establish a TxTL system that regenerates its components from the DNA encoding all the components. A well-studied reconstituted TxTL system is the PURE system³³, which consists of the E. coli translation machinery. In the PURE system, translation proteins^7,34,35,36, ribosomal proteins^37,38, and rRNA^39,40 are partially synthesized.

tRNA is an essential component of the translation system⁴¹, but its in vitro expression scheme is still in progress. In the last few years, in vitro synthesis of unmodified tRNAs has been demonstrated, enabling translation with a reconstituted genetic code^{42,43,44,45,46,47}. At least 21 types of tRNAs are required for the translation of arbitrary proteins (20 for each amino acid and one for formyl methionine). In previous studies, each of the 21 tRNA was chemically synthesized or transcribed in vitro and purified individually before use. The next challenge toward a self-reproducible translation system is to simultaneously express all 21 tRNA in their functional forms in the PURE system.

A major hurdle for the simultaneous expression of the 21 tRNA in the PURE system is 5’- and 3’-end processing. In E. coli, many tRNAs are transcribed as polycistronic premature RNA that also encodes either identical tRNAs, unrelated tRNAs, proteins or rRNA. Premature RNA undergoes complicated processing by various RNases, including RNase E, III, T, PH, II, D, P, and PNPase, to produce mature tRNAs with the correct 5′ and 3′ ends^41,48. This processing reaction has not been fully reconstituted in vitro, and the necessity for such a complicated process remains unclear. We think that simpler processing may be possible, at least in vitro.

In our previous study, we reported the expression of 15 of 21 tRNAs from monocistronic DNA in a PURE system⁴⁹. In the experiment, 5’-end processing was not needed because these 15 tRNAs possess 5’-G, which can be directly transcribed from the 5’ end with T7 RNA polymerase. The 3’-end processing is also not needed because the 3’-end of tRNA was produced by run-off transcription by using template DNA that possesses the 3’-end matching with the 3’-end of the tRNA gene⁴². However, the 5’-end method is not applicable to the other six non-G-start tRNAs. In addition, the run-off method for the 3’-end poses a limitation to the 3’-terminal sequence of the template DNA. This 3’-end limitation should be removed to achieve a self-reproducible artificial system because with this limitation, the DNA template is difficult to replicate by currently reconstituted DNA replication schemes, which require circular DNA or linear DNA that contains specific sequences at the termini^18,21,22. In addition, for efficient DNA replication, all 21 tRNA should be encoded in a single DNA in a polycistronic manner. Consequently, the next critical challenge in tRNA synthesis for self-reproducing systems is to develop a method for producing all 21 tRNAs with correct 5’- and 3’-ends from a single template DNA with an arbitrary 3’-end sequence in the PURE system.

In this study, we tested five methods (leader, 5’-G variant, HDVR, linkage of tRNAs, and tRNA array method) for the expression of 21 tRNAs with correct ends using a simpler scheme than that used in E. coli. For the six non-G-start tRNAs (Glu, Pro, Ile, Asn, Trp, and fMet), we employed two methods: attachment of a leader sequence, which was removed with RNase P, and mutating the 5’ nucleotide to G to generate 5’G variants. Using these methods, we succeeded in simultaneously expressing all six non-G-start tRNAs coupled with translation from six monocistronic DNA templates in the PURE system. For the 3’-end, we employed two methods: using a self-cleaving ribozyme (HDVR) and direct linkage of tRNAs separated by RNase P. We then found that the combination of these two 3’-end methods (named tRNA array method) also solves the problem of 5’-end processing. Using this tRNA array method, we succeeded in simultaneously expressing all 21 tRNAs from a single polycistronic DNA template in the PURE system. We also demonstrated that tRNA prepared using the tRNA array method is useful for genetic code engineering. We reassigned five leucine codons in Nanoluc to threonine codons (ACG) and recovered the activity by introducing a tRNA^LeuCGU variant.

Results

Preparation of tRNA-free PURE system (tfPURE system)

In this study, we aimed to express tRNAs from DNA and use them for translation in the PURE system (Fig. 1A). For this purpose, a tRNA-free PURE system (tfPURE system) is required. We previously found that two components of the PURE system, EF-Tu and ribosomes, contain non-negligible levels of tRNA⁴⁹. To remove residual tRNA, we re-purified EF-Tu and ribosomes. However, the repurification process for the ribosome, the separation of 50S and 30S subunits through sucrose gradient ultracentrifugation, was both laborious and low in yield. To solve this problem, we used a new and simpler method using a size-separation spin column, allowing the preparation of tRNA-free ribosomes at a higher yield (Supplementary Fig. 1).

**Fig. 1: Leader method for 5’-end processing.**

Leader method for 5’-end processing: attachment of leader sequence processed with RNase P

To enable the expression of six non-G-start tRNAs that cannot be directly synthesized using T7 RNA polymerase, we first evaluated the method of attaching a 5’-G leader sequence to the 5’-end and cleaving it with RNase P⁵⁰ (Fig. 1B). Although this method has been previously used for in vitro transcription of tRNA^42,46, it has not yet been applied to tRNA synthesis in the PURE system. To examine whether this method works for the translation-coupled reaction in the PURE system, we performed the reaction in the tfPURE system containing a tRNA template DNA encoding one of the six non-G start tRNAs, the other 20 in vitro synthesized (IVS) tRNA, RNase P, and a luciferase DNA template at 30 °C for 16 h. The 3’-terminus of the tRNA template matched the 3’-end of each tRNA for the run-off transcription. If this method works as expected, a premature tRNA with the leader sequence is synthesized from the tRNA template and cleaved by RNase P to produce a mature tRNA with the correct 5’-end. The mature tRNA works with the other 20 tRNAs to translate luciferase, which is used as a reporter for translation activity (Fig. 1C). Native RNase P consists of M1 RNA and C5 protein, but we used only the M1 RNA component in this experiment because M1 RNA is sufficient for RNase P activity in vitro^51,52.

First, we conducted this reaction using a tRNA template encoding tRNA^Gln in the tfPURE system that contains 20 other IVS tRNAs. The concentrations of NTP, Mg(OAc)₂, and spermidine were optimized before the experiment (Supplementary Fig. 2). Luciferase activity increased to 10⁷, depending on the tRNA template (Fig. 1D). We then performed experiments with the remaining five tRNAs (tRNA^Pro, tRNA^Ile, tRNA^Asn, tRNA^Trp, and tRNA^fMet) and detected similar levels of luciferase activity in the presence of the tRNA template. These results indicate that functional tRNAs were produced by this method. We also observed higher background luciferase activity for tRNA^fMet in the absence of the tRNA template. This might be attributed to undetectable amounts of tRNA^fMet remaining in the tfPURE system⁴⁹ or other elongator tRNA (mMet or other) involved in translation initiation as previously reported⁵³.

Next, we directly checked the cleavage of the leader sequence in the tfPURE system without adding other 20 IVS tRNAs, using electrophoresis. For tRNA^Gln, we detected a band corresponding to mature tRNA depending on RNase P (Fig. 1E). Similar bands were detected for the other five tRNAs (Fig. 1F, yellow arrowheads). Quantification based on band intensity showed that synthesized tRNAs ranged from 16 to 35 ng/µL (Supplementary Fig. 3), similar to or greater than the average tRNA concentration used for IVT tRNA (11.7 ng/µL) in a previous study⁴⁹. For tRNA^fMet, a relatively large amount of uncleaved premature tRNA was detected (Fig. 1F, blue arrowheads), indicating that the cleavage efficiency varies depending on the tRNA sequence. One possible factor affecting 5’-end processing efficiency is the base pairing at position 1–72 in the tRNA acceptor stem⁵⁴.

5’ variant method: introducing G mutations to the 5’-end

As a second method to prepare tRNA with a correct 5’-end, we substituted the non-G nucleotide at the 5’-end with G to generate 5’-G variants (Fig. 2A). We also designed 5’-A variants for tRNA^fMet because T7 RNA polymerase can initiate RNA synthesis from A when using the class II T7 promoter⁵⁵. Since the structure of the acceptor stem is known to be critical for tRNA functionality, we also introduced mutations into the complementary base sequences (Fig. 2A, Supplementary Data 4). Some of these variants (Trp: G | U and G | C, Gln: G | A and G | C, Ile: G | C, Pro: G | C and G | G, fMet: G | A and G | U) have been analyzed in previous studies on aminoacylation activity^56,57,58,59 or formylation activity⁶⁰. Furthermore, tRNA^Asn (G | C), tRNA^Ile (G | C), and tRNA^fMet (G | A) have been used for translation in reconstituted systems^43,46,47. To evaluate the translation activities of these tRNAs in our system, we first added each purified IVS tRNA variant to the tfPURE system containing the other 20 IVS tRNAs and a luciferase template. After incubation at 30 °C for 16 h, luciferase activity was measured. Most tRNA variants exhibited similar or higher luciferase activity than the wild-type tRNA (WT) (Fig. 2B), with the exception of one variant, tRNA^Pro (G | C). The lack of activity for this variant is consistent with the previous aminoacylation result⁵⁹. Next, we selected the tRNA variant with the highest activity for each amino acid (Fig. 2B, red triangles), and conducted tRNA synthesis coupled with luciferase translation in the tfPURE system (Fig. 2C). Luciferase activity was detected for all the variants tested, depending on the tRNA template (Fig. 2D). For tRNA^fMet, a higher background luciferase activity was detected without the tRNA template. The luciferase activity of tRNA^Asn was much lower than those of the other tRNAs.

**Fig. 2: 5’-G variant method for 5’-end of tRNA.**

The synthesized tRNAs were analyzed by electrophoresis. We detected a band of the expected size for all the tRNAs (Fig. 2E). Quantification based on band intensity showed that synthesized tRNAs, except for tRNA^Asn, ranged from 22 to 35 ng/µL (Supplementary Fig. 4), which was comparable to the amount obtained using the previous method with the leader sequence. However, the yield of tRNA^Asn (6.6 ng/µL) was significantly lower than that of the other tRNAs. To explore the low luminescence observed with tRNA^Asn in the 5’-G scheme, we expressed the G | C variant using the leader method with RNase P (Supplementary Fig. 5). The resulting luminescence was similar to that of the wild-type tRNA^Asn, indicating that the G | C variant is functional when properly processed. Reduced activity in the 5’-G scheme may result from suboptimal 5’-end formation or misfolding during transcription.

To determine which method to use in the next simultaneous expression experiment, we compared the translation activity of tRNAs prepared by the leader and 5’-G variant methods described above (Supplementary Fig. 6). These two methods were sufficiently useful for most tRNAs, except for tRNA^Asn, for which the 5’G variant method produced a much lower expression level than the leader method. We also found that a relatively large amount of uncleaved premature tRNA^fMet remained when using the leader method. Based on these results, we decided to use the leader method for tRNA^Asn, the 5’G variant method for tRNA^fMet, and both methods for the other tRNAs in the subsequent simultaneous expression experiment.

At this point, we examined the effect of the 5’-end phosphate on translation. tRNAs synthesized using the 5’-G variant method are expected to contain a triphosphate group at the 5’-end. In contrast, tRNAs synthesized by the leader method are expected to contain 5’-monophosphate because of the RNase P treatment. To investigate whether this difference affects translation, we prepared 5’-triphosphate or monophosphate tRNA^Ala and tRNA^His using simple in vitro transcription or the leader method. A comparison of luciferase activity with these tRNAs revealed that the translation activities of 5’-monophosphates were 83% and 77% of those of 5’-triphosphates for tRNA^Ala and tRNA^His, respectively (Supplementary Fig. 7).

Simultaneous six tRNA synthesis from monocistronic DNAs

We simultaneously synthesized all six non-G-start tRNAs from the six tRNA templates in the tfPURE system and evaluated their functionality in luciferase translation (Fig. 3A). The leader and 5’-G variant methods were adopted for tRNA^Asn and tRNA^fMet, respectively, and both methods were compared for the remaining four tRNAs (Fig. 3B). The luminescence with tRNA templates for both methods were significantly higher than that without tRNA templates, reaching more than 0.6×10⁷. This value indicated that all five tRNAs (for Gln, Ile, Pro, Trp, and Asn) were sufficiently expressed. Notably, using the 5’-G variant method for the four tRNAs (Gln, Ile, Pro, and Trp) yielded higher luciferase activity (1.3 × 10⁷) than the leader method, indicating that the 5’-G variant method is more useful for maximizing tRNA expression from monocistronic DNA templates in the PURE system.

**Fig. 3: Simultaneous expression of tRNAs from monocistronic tRNA templates.**

HDVR method for 3’-end processing: introducing self-cleavage ribozyme (HDVR)

In the previous experiment for simultaneous 21 tRNA synthesis, the 3’-terminus of the RNA template must match the 3’-end of the tRNA gene for run-off transcription. Accordingly, the tRNA template must be monocistronic (i.e., one DNA template encodes only one tRNA gene). To overcome this limitation and allow tRNA synthesis from polycistronic DNA, as in the cell, we explored two 3’-end processing methods. The first method utilized the self-cleaving ribozyme HDVR, as previously reported⁶¹. We placed the HDVR immediately downstream of the tRNA gene, followed by a strong T7 terminator, T7 hyb10⁶² (Fig. 4A). During transcription, the premature tRNA connected with HDVR and terminator is synthesized and then undergoes self-cleavage of HDVR, generating a tRNA with a 2’,3’ cyclic phosphate at its 3’-end. Since this cyclic phosphate cannot be aminoacylated by aminoacyl-tRNA transferases, it is dephosphorylated by T4 polynucleotide kinase (PNK)⁶³.

**Fig. 4: HDVR and linked tRNA methods for 3’-end processing.**

To test the self-cleavage ability of premature tRNA^Ser-HDVR, we first performed in vitro transcription in standard buffer (see Methods) and subjected the transcript to PAGE analysis. The transcript exhibits two main bands (Fig. 4B, lane 1), each of which corresponds to the HDVR + terminator (lane 2) and mature tRNA^Ser (lane 3), respectively, supporting successful self-cleavage. Next, we investigated whether the tRNA^Ser generated through self-cleavage is functional for translation in the tfPURE system. We conducted the reaction shown in Fig. 4A in the tfPURE system containing the tRNA^Ser-HDVR DNA template, the other 20 IVS tRNAs, and a luciferase template at 30 °C for 16 h, and luciferase activity was measured. The luminescence increased to ~ 10⁷ depending on both the tRNA template and T4 PNK (Fig. 4C), supporting the functionality of tRNA^Ser generated through the self-cleavage of HDVR and dephosphorylation with T4 PNK. The luminescence increased up to 24 h (Supplementary Fig. 8). For this reason, we extended the incubation period in later experiments.

Linked tRNA method for 3’- and 5’-ends processing

As the second method for 3’-end processing, we devised another strategy utilizing RNase P for the simultaneous processing of the 3’-end of a tRNA and the 5’-end of another tRNA. RNase P recognizes the structure of a tRNA⁶⁴ and cleaves the specific site between the 5’-leader sequence and the subsequent tRNA. Based on the weak sequence requirement for the 5’-leader sequence^65,66, we hypothesized that if the 5’-leader RNA was replaced with another tRNA, RNase P could cut between the two tRNAs, allowing simultaneous processing of the 3’-end of the former tRNA and the 5’-end of the latter tRNA (Fig. 4D). To test this hypothesis, we prepared a substrate RNA (tRNAs^Ala-Ser) composed of directly linked tRNA^Ala (76 nt) and tRNA^Ser (88nt) and then treated it with RNase P (M1) in standard buffer for RNase P⁵¹ (see legend). The RNA product (Fig. 4E and Supplementary Fig. 9) exhibited four bands, including two major bands (lane 1) that correspond to mature tRNA^Ala (lane 2) or tRNA^Ser (lane 3), supporting the successful cleavage of tRNAs, while the cleavage efficiency seemed to be lower than the method with HDVR shown in Fig. 4C.

Encouraged by this result, we expressed linked tRNAs^Ala-Ser from the DNA template in the tfPURE system and processed it in the same reaction mixture. The reaction mixture contained a DNA template that encoded tRNAs^Ala-Ser, the other 19 tRNAs, RNase P (M1), a luciferase template, and the tfPURE system. In this experiment, the 3’-end of tRNAs^Ala-Ser was determined by run-off transcription. After incubation at 30 °C for 16 h, luciferase activity was measured. The luminescence increased to ~ 10⁷ depending on both the tRNA^Ala-Ser template and RNase P, supporting the production of functional tRNAs through digestion with RNase P (Fig. 4F). The reaction mixture was subjected to PAGE (Supplementary Fig. 10). The products exhibited two bands corresponding to each mature tRNAs, although a thick band was detected at the position of uncut tRNA (tRNAs^Ala-Ser), indicating that there is still room for improvement in digestion efficiency, which was addressed later.

Next, we combined this linked tRNA method (Fig. 4D) with the HDVR method (Fig. 4A) for the simultaneous expression of the two tRNAs without run-off transcription (Fig. 4G). In this experiment, the premature linked tRNA attached to HDVR (tRNAs^Ala-Ser-HDVR) was expressed from a DNA template and then processed to mature tRNA^Ala and tRNA^Ser through digestion by RNase P and HDVR self-cleavage, followed by dephosphorylation by T4 PNK. We conducted this reaction in the tfPURE system containing the other 19 IVS tRNAs and the luciferase template at 30 °C for 16 h and measured luciferase activity (Fig. 4H). The luminescence increased to ~ 10⁷ depending on the tRNA template, RNase P (M1), and T4 PNK, indicating that functional tRNAs were produced according to the expected processes.

Multicistronic tRNA expression using tRNA array methods

The combination of the linked tRNA and HDVR methods can be extended to a greater number of tRNA. We named this combined method for multiple tRNAs as “tRNA array method.” Next, we attempted to express all 21 tRNAs using this method. First, we divided the 21 tRNAs into 3–5 groups (Fig. 5A). The four tRNAs, Ile, Phe, Glu, and Asn (IPEN), are grouped as single operons because they are required at particularly high concentrations for efficient translation⁴². The remaining 17 tRNAs were divided into four operons, each containing 3–5 tRNAs, arranged to maintain similar operon sizes. The order of each array was determined as follows. We presumed that the most upstream tRNA would affect transcription because it directly attaches to the promoter. To find tRNAs suitable for this position, we compared the expression levels of some tRNAs (Gly, Asp, Ser, Leu, Ala, and His) (Supplementary Fig. 11), which were expressed at higher levels in our previous study⁴⁹. Four tRNAs (Gly, Asp, Ser, and Leu) that exhibited higher transcription levels were assigned to the most upstream positions in each array. The remaining tRNAs were arranged for no specific reason, but mainly alphabetically.

**Fig. 5: Simultaneous expression 3–5 tRNAs from multicistronic DNA template by the tRNA array method.**

To evaluate the functionality of tRNA expressed from these tRNA arrays, we performed a translation-coupled reaction in the tfPURE system. During the reaction, multicistronic DNA is transcribed to produce premature tRNA composed of connected 3–5 tRNAs and HDVR, which are then separated into mature tRNAs through digestion by RNase P and self-cleavage of HDVR, followed by dephosphorylation by T4 PNK (Fig. 5B). The resultant tRNAs are used for luciferase translation.

We conducted this translation-coupled reaction with each one of the four groups (#1 GARCQ, #2 DfMHKV, #3 SmMFT, and #4 LWY) in the tfPURE system containing RNase P (M1 RNA) at 30 °C for 24 h. The results are shown in Fig. 5C (yellow bars indicated as “M1 RNA”). The luminescence for #3 and #4 reached approximately 10⁶, whereas those for #1 and #2 were less than 10². PAGE analysis of the synthesized tRNAs revealed that the cleaved products corresponding to the expected tRNA sizes were faint for templates #1 and #2 (Fig. 5D, “RNase P (M1)”). We considered the possibility that these low yields might be caused by insufficient digestion with RNase P from M1 RNA only. We then supplemented the reaction system with C5 protein, another subunit of E. coli RNase P. Although the C5 protein itself does not catalyze the cleavage reaction, it facilitates the proper binding to its substrate⁶⁵. As expected, C5 addition increased both luminescence (Fig. 5C, “M1 + C5”) and the expected band intensities (Fig. 5D, “RNase P (M1 + C5)”) for all templates.

Next, we evaluated the expression of the highly demanding IPEN tRNA from a multicistronic template. To determine the optimal arrangement for transcription and cleavage, we prepared four DNA templates encoding the IPEN tRNA genes in different orders (IPEN, PIEN, EIPN, and NIPE). The translation-coupled assay for these templates showed that higher luminescence was detected for the IPEN (10⁶) and PIEN (10⁷) templates with RNase P (M1 + C5) (Fig. 5E). Consistently, PAGE analysis showed that clear bands were detected at the expected sizes for the IPEN and PIEN templates only in the presence of C5 protein (Fig. 5F). The PIEN template (#5 PIEN, Fig. 5A) was used for subsequent experiments.

To examine why RNase P (M1 only) inefficiently cleaved longer arrays (#1 and #2), we analyzed shorter derivatives of array #1. While a 3-tRNAs array (GAR) showed partial expression, a 2-tRNAs array (GA) showed full expression (Supplementary Fig. 12A). Secondary structure prediction implied that GA retained a more clover-leaf-like structure, which was probably favorable for RNase P recognition (Supplementary Fig. 12B). These results indicate that processing efficiency may be affected by array length and tRNA arrangement, likely due to structural constraints.

To investigate whether unintended transcription termination contributed to the variation in tRNA yield, we next transcribed pre-tRNAs from each template under standard T7 transcription buffer conditions without RNase P treatment, thereby eliminating cleavage effects between tandem tRNAs (Supplementary Fig. 13). Both uncleaved and self-cleaved products were observed, along with additional smaller bands, likely representing prematurely terminated transcripts. These results suggest that intrinsic termination events may partially account for reduced tRNA yields in certain constructs.

Polycistronic 21 tRNA expression using tRNA array method

To advance toward self-replicating artificial cells, it is ideal for all tRNAs to be encoded in a single genomic DNA. We then constructed a single polycistronic DNA containing all the five tRNA arrays shown in Fig. 5A and evaluated the function of the expressed tRNAs (Fig. 6A). In this experiment, the five transcripts were synthesized from five T7 promoters and underwent a series of processes: self-cleavage of HDVR, dephosphorylation by T4 PNK, and digestion by RNase P to synthesize 21 tRNAs, which are used for luciferase translation. After optimizing NTP and magnesium concentrations (Supplementary Fig. 14), we detected a luciferase activity of 4.7 × 10⁶, dependent on the DNA template (Fig. 6B). This luminescence value was comparable to those of the 21 IVS tRNAs, which are 7.9 × 10⁵ under the same conditions and 9.1 × 10⁶ under the conditions used in the above experiments (Fig. 2B, WT). These results demonstrate that a sufficient amount of tRNAs was produced in the PURE system using this method. Consistently, PAGE analysis of the reaction product revealed that the total amount of synthesized tRNA (0.82 µg/µl) was similar level to the amount of 21 IVS tRNA (0.6 µg/µl) usually included in the PURE system⁴⁹ (Fig. 6C). We further investigated the composition of the synthesized tRNAs (Fig. 6D) by direct RNA sequencing using an Oxford Nanopore sequencer according to a previous study^67,68. In this method, adapters were attached to RNAs with CCA termini, and direct RNA sequencing revealed the relative abundance of tRNAs of the correct size. We found that all 21 tRNAs were present at a concentrations of greater than 1% of the total synthesized tRNA (7.9 ng/µl).

**Fig. 6: Simultaneous expression of 21 tRNAs from a single polycistronic DNA template.**

We further tested whether additional tRNAs could be included in a single array. First, we constructed four arrays consisting of 8–9 tRNA genes and evaluated the luciferase activity in the translation-coupled reaction (Supplementary Fig. 15A). All four constructs exhibited similar luminescence at approximately 10⁶. We then chose two constructs (GL and DS) and integrated them into a single array consisting of 17 tRNA genes of two different orders. Luciferase activity of the GLDS construct was comparable to that of the negative control without tRNA templates, whereas the DSGL construct exhibited luciferase activity at a level of approximately 10⁵ (Supplementary Fig. 15B). When an additional PIEN unit was appended to the 3′ end of the DSGL to construct an all 21-tRNA array, luciferase activity was higher than that without the tRNA template, but the level was significantly low (3.6 × 10²) (Supplementary Fig. 15C). These results suggest that while the integration of 21 tRNAs into a single operon is feasible, the translation activity is much lower than that of the separated operons used in Fig. 6.

To further examine the versatility of our tRNA array method, we tested its ability to support translation of proteins other than luciferase. Specifically, we expressed sfGFP and β-glucuronidase (GUS), two reporter proteins that differ in size and structure (Supplementary Fig. 16). sfGFP is a small fluorescent protein commonly used in reconstituted systems⁴², whereas GUS is a larger, tetrameric enzyme that hydrolyzes β-D-glucuronide substrates⁶⁹. Both proteins were successfully synthesized in the presence of the single 21 tRNA template, although their activity levels were approximately half (sfGFP) or 1/6 (GUS) of those observed with the IVS 21 tRNA mixture. These results support the applicability of the tRNA array approach to the translation of diverse protein types, where the activity level depends on the target proteins. The lower activity of GUS was probably because GUS is active as a tetramer, and its concentration can decrease nonlinearly by decreasing GUS expression. The GUS expression by the tRNA array method decreased less than half, as shown in the next subsection.

Quality check of the luciferase produced by the tRNA array method

The tRNA mixture synthesized using the tRNA array method lacks any modifications and has a different tRNA composition from the 21 IVS tRNAs used in previous studies^42,49. These differences may reduce the fidelity of translation and, consequently, lower the activity of the synthesized protein. To investigate this, we compared the activity per molecule of luciferase translated by three different tRNA mixtures: native tRNAs purified from E. coli, 21 in vitro synthesized (IVS) tRNAs, and 21 tRNAs expressed by the tRNA array method. Protein concentrations of the translated luciferases were quantified using western blotting (Supplementary Fig. 17A, B), and their activities were assessed by measuring their luminescence. The synthesized protein concentrations were similar for the 21 IVS (53 nM) and array methods (61 nM). Luciferase activity per molecule (luminescence/fmol) was compared (Supplementary Fig. 17C). The activity per molecule for the 21 tRNAs expressed using the tRNA array method was approximately 46% and 56 % of that of the native E. coli tRNAs and 21 IVS tRNAs, respectively, indicating that the current tRNA array method reduced the luciferase activity per molecule by approximately half.

To validate these findings, we additionally quantified luciferase protein using the HiBiT system⁷⁰ (Supplementary Fig. 18A), which provides a linear readout of protein abundance via complementation with LgBiT. The expression levels obtained using HiBiT closely matched those determined using western blotting, and the calculated activity per molecule showed similar trends (Supplementary Fig. 18B, C), confirming that the specific activity of the synthesized luciferase was slightly decreased using the tRNA assay method. Similar results were observed for sfGFP and GUS expressions (Supplementary Fig. 18D–G), supporting the generality of this effect across different proteins.

An application of tRNA array method for genetic code modification

In the experiment described above, we demonstrated the simultaneous expression of all 21 tRNAs using a single polycistronic DNA template in the tfPURE system. If this tRNA array method is used in conventional in vitro transcription (IVT) reactions instead of in the PURE system, it can be used as an easy method to prepare the minimum tRNA set, which has been previously prepared by laborious individual IVT^42,43,46,49 or chemical synthesis^45,49. The minimum tRNA set is useful for genetic code engineering in vitro. As a demonstration, we conducted an IVT using the single 21 tRNA template shown in Fig. 6A, T7 RNA polymerase, and T4 PNK at 37 °C for 12 h, and then processed with RNase P at 37 °C for 12 h to generate the 21 tRNAs (Fig. 7A). PAGE analysis of the reaction mixture after RNase P treatment revealed bands corresponding to the correct tRNAs (Fig. 7B). We obtained 33 µg of total tRNA from 50 µL of the in vitro transcription reaction. After purification, tRNA mixture was added to the tfPURE system at 0.6 µg/µl. The translational activity of the 21 tRNA prepared by the tRNA array method (Fig. 7C, Array) was comparable to that for the same total concentration of conventional 21 tRNA prepared individually (Conv), in which each tRNA ratio follows previous studies^42,49. Note that, in this experiment, we changed the reporter gene from firefly luciferase to Nanoluc⁷¹, a shorter variant of a deep see shrimp luciferase, for ease of codon rearrangement conducted below.

**Fig. 7: Simultaneous in vitro transcription of 21 tRNAs by the tRNA array method and an application for genetic code reassignment.**

This method allows for easy customization of the genetic code by introducing additional tRNAs. To demonstrate genetic code reassignment, we replaced five Leu residues (CUU) in Nanoluc with ACG, which is assigned to Thr in the native codon table but is not used in the minimum codon table used here (Fig. 7D). This reassignment has not been used in previous genetic code-engineering studies^42,43,46,47. If this mutant Nanoluc is used, highly active Nanoluc should be translated only in the presence of an engineered tRNA (tRNA^Leu_CGU) that assigns Leu to the ACG codon (Fig. 7E). To validate this hypothesis, we performed translation reactions of the mutant Nanoluc in the tfPURE system containing the 21 tRNA set prepared by the tRNA array method, and either tRNA^Thr_CGU (native anticodon) or tRNA^Leu_CGU (engineered anticodon) (Fig. 7F). As expected, high Nanoluc activity (> 10⁸) was detected when engineered tRNA^Leu_CGU was present, whereas activity decreased by approximately 1/100 when normal tRNA^Thr_CGU was present. Furthermore, experiments using additional mutant Nanoluc with 2 or 10 ACG substitutions showed similar trends, and the background activity with tRNA^Thr_CGU decreased with increasing substitution numbers (Supplementary Fig. 19).

To confirm that the observed luminescence changes reflected amino acid reassignment, we performed LC-MS/MS analysis of translated Nanoluc proteins. The results confirmed the incorporation of threonine at the reassigned ACG codon in the presence of tRNA^Thr_CGU, and leucine incorporation when using tRNA^Leu_CGU (Supplementary Fig. 20). These results support the applicability of the minimal 21 tRNAs set for genetic code reassignment.

Discussion

A major hurdle in constructing a tRNA synthesis system for self-reproducing artificial cells is the processing of the 5’ and 3’ ends of tRNA, which is achieved through complicated multistep processes in the cell⁴⁸. In this study, we developed simpler processing methods that work in the PURE system. For the 5’-end, we tested two methods for six tRNA species that could not be directly synthesized from the T7 promoter: the leader method (Fig. 1) and 5’-G variant method (Fig. 2). Using these methods, we demonstrated that all six non-G-start tRNAs could be simultaneously expressed in the PURE system from six monocistronic templates (Fig. 3). For the 3’-end, we developed two methods: the HDVR and linked tRNA methods (Fig. 4). By combining these two methods (termed the tRNA array method), all 21 tRNAs were simultaneously expressed from a single polycistronic DNA template (Fig. 6). The translation ability of the 21 tRNA produced was comparable to the conventional 21 IVS tRNAs that were prepared and purified individually for luciferase (Fig. 6) and sfGFP (Supplementary Fig. 18). These results demonstrate that all minimum tRNA sets can be simultaneously synthesized from multicistronic or single polycistronic DNA templates in the PURE system and used for translation, providing a step toward self-reproducing artificial cells.

The tRNA array method established in this study also provides an easy means for the simultaneous in vitro synthesis of multiple tRNAs in a conventional IVT buffer, instead of the PURE system (Fig.7). In vitro-synthesized tRNA have been utilized for several applications, such as peptide drug screening through the incorporation of non-canonical amino acids^43,72,73 and biological containment by constructing a swapped genetic code^42,46,47. To date, each tRNA has been individually synthesized through in vitro transcription or chemical synthesis, which requires significant effort, time, and cost. Using the method developed in this study, all 21 tRNAs were synthesized from a single DNA template through IVT in the presence of T4 PNK, followed by RNase P processing (Fig. 7A). As shown in Fig. 7D, E, and F, this system also allows the incorporation of additional tRNA to customize the genetic code. The tRNA array method allows easier preparation of the minimum tRNA set and facilitates further development in the field of genetic code reprogramming⁷⁴.

One of the remaining challenges in the tRNA array method is the regulation of the expression levels of individual tRNAs. The required amount of each tRNA is expected to vary depending on the protein to be translated, but the current expression level of each tRNA is unregulated. Composition analysis using nanopore sequencing revealed that the expression levels of the tRNAs depended on their position in the arrays (Fig. 6D). In particular, central tRNA (e.g., A, fM, H, W, and I) in each array tended to be more highly expressed. This might have been caused by unexpected transcriptional termination in the middle of the tRNA array. Exploring and utilizing RNA polymerases with a higher processivity⁷⁵ may address this issue. In addition, the tRNA composition of each array appeared to influence the transcriptional efficiency, as observed in the lower expression levels of the LWY group. Rearranging the order of tRNAs within the array or incorporating transcriptional enhancement sequences as 5’-leader RNAs⁷⁶ or internal promoters^77,78, may help optimize expression balance. The importance of the order of the array is also supported by the results shown in Fig. 5E, where the expression levels varied significantly depending on the arrangement of IPEN. These differences could be due to the different secondary structures caused by different orders. This structure can also affect RNase P recognition⁷⁹. Understanding the relationship between linked tRNA structure, transcriptional efficiency, and cleavage efficiency is important for the regulation and improvement of tRNA expression. Furthermore, the absence of post-transcriptional modifications may contribute to decreased translation efficiency, especially for tRNA^Ile and tRNA^Glu, whose activities are affected by modifications⁴². Important future research directions include introducing modification enzymes or selecting tRNA variants that function without modifications.

In this study, we demonstrated that when tRNAs were directly linked, both 5’- and 3’-end processing were achieved with RNase P. In contrast, extant living systems use various multistep processing for the 3’-end. For example, in E. coli, the 3’-end of tRNAs is processed by RNase E, leaving several nucleotides that are further trimmed by multiple exonucleases⁴⁸. In eukaryotes, tRNA genes do not generally encode CCA at the 3’-end; instead, extra sequences at the 3’-end are removed by tRNase, and CCA residues are subsequently added by nucleotidyltransferase^80,81. This complexity in nature raises a fundamental question: why do natural organisms use such intricate processing systems, even though correct processing can be done by only RNase P, at least in vitro? One possible explanation is its role in ensuring regulation and quality control⁸². Sequential processing steps may act as checkpoints to exclude improperly folded or defective pre-tRNAs, thereby maintaining translational fidelity. On the other hand, a processing scheme using a self-cleaving ribozyme to 3’-end of a tRNA, as utilized in this study, has recently been found in many viruses⁸³, which suggests that such a simpler processing mechanism may also work in nature. Further investigation into the differences between our simplified processing and natural processing would provide useful information for understanding the biological significance of natural processing systems.

Recently, Li et al. reported a similar tRNA expression from a single DNA template in the PURE system⁸⁴. While their study highlights important advances, such as continuous tRNA expression using a microfluidic chemostat, our approach also offers advantages in the six points described below. (1) Li et al. used a nicked DNA template for 3’-end processing, but nicking in DNA is known to reduce the efficiency of DNA replication⁴⁹. Therefore, we believe that the use of intact DNA proposed in this study is ideal for coupling with DNA replication in the future. (2) Li et al. reported another processing method using tRNase Z, but this was not performed in the PURE system, whereas we have shown that processing using RNase P and HDVR works in the PURE system. (3) The background translation level without the addition of tRNA was high in Li et al.. Therefore, the dependency of each tRNA template on translation, shown in Figs. 1D and 2D in this study, was not shown in Li et al. The high background is probably due to the contamination of tRNA in the PURE system, which we eliminated by further purification of EF-Tu and ribosome (Supplementary Fig. S1). 4) In Li et al., the 5’-G variant method was only tested for Asn and fMet, whereas we tested it for all six non-G-start tRNAs (Fig. 2) and also decided not to use it for Asn due to its lower transcription level. 5) We measured both the total amount of tRNA by PAGE analysis (Fig. 6C) and the frequency of each of 21 tRNAs by nanopore sequencing (Fig. 6D) when 21 tRNAs were expressed simultaneously, neither of which was done in Li et al. 6) We have demonstrated that our method is used for genetic code reassignment (Fig. 7).

The 21 tRNA synthesized by the tRNA array method enabled approximately 3.7 µg/mL translation, comparable to that using 21 IVS tRNAs; however, it remained significantly lower than that with E. coli native tRNAs (56 µg/mL). Although this yield is sufficient for in vitro reporter assays, it is still below the requirement for constructing a self-reproducing central dogma system, which would require a protein output comparable to the total concentration of translation proteins in the PURE system (the order of several mg/mL³⁶). The translation ability should be improved to realize self-reproducible artificial cells^32,36. The limitation in translation ability is likely caused by a lack of modification, which reduces aminoacylation efficiency and translational fidelity^42,85, particularly for tRNAs, such as tRNA^IPEN. One possible strategy for overcoming this limitation is in vitro evolution. The tRNA array method established in this study can be used for a circular DNA template. If tRNA is encoded in a circular DNA template and the translation of phi29 DNA polymerase is coupled, the tRNA sequence can be subjected to in vitro evolution using our previous rolling-circle replication scheme²¹. Through in vitro evolution, tRNA sequences that exhibit a higher translation ability, even in the absence of modifications, would be selected.

Methods

Reporter DNA preparation

The luciferase template DNA was prepared by PCR amplification using primers 1 and 2 (Supplementary Data 1 and 2) and a plasmid (pUC-T7p-Fluc_21tRNA) constructed in our previous study⁴⁹. DNA fragments encoding Nanoluc sequences used in Fig. 7 (WT and genetic code reprogrammed version) were obtained using Twist Bioscience’s artificial gene synthesis service. The sequences are listed in Supplementary Data 3. DNA templates for the translation experiments in the PURE system were prepared by PCR amplification using primers 1 and 2. DNA concentrations were quantified based on the A₂₆₀. To prepare the luciferase template fused with a HiBiT tag, the HiBiT tag fragment was first amplified using primers 63 and 64. This fragment was then fused to a luciferase PCR fragment (amplified using primers 1 and 65) by overlap extension PCR using primers 1 and 66.

tRNA and M1 RNA preparation

The sequences of the 21 tRNA were based on a previous study⁴². The 21 IVS tRNAs were prepared as previous described⁴⁹. The 15 IVS tRNAs that has G in the 5’ end (Ala, Arg, Asp, Cys, Gly, Glu, His, Leu, Lys, mMet, Phe, Ser, Thr, Tyr Val) and 5’-G or 5’-A variants shown in Fig. 2A were prepared by in vitro transcription utilizing tRNA template prepared by PCR using primer sets shown Supplementary Data 1 and 2. Six non-G-start tRNAs were chemically synthesized (Invitrogen). The reaction mixture for in vitro transcription included 5 mM DTT, 1 μ/μl T7 RNA polymerase (Takara, Japan), 2 mM each NTPs(ATP, GTP, CTP, and UTP), 3 mM GMP, 40 mM Tris-HCl pH 8.0, 10 mM magnesium chloride, 2 mM spermidine, 5 μg/μl template DNA, 2 μ/ml inorganic pyrophosphatase (New England Biolab), 0.8 μ/μl RNasin (Promega). The mixture was incubated at 37 °C for 12 h, and the RNA product was purified using the PureLink RNA Mini Kit (Invitrogen). Both tRNAs and M1 RNA were dissolved in water and stored at − 80 °C until use. RNA concentrations were determined based on the A260. The template DNAs for in vitro transcription were prepared by PCR using a plasmid encoding each tRNAs as a template and the primers shown in Supplementary Data 1. Each reverse primer contained a 2’-O-methylation at the second nucleotide from the 5’ terminus to prevent nucleotide addition by T7 RNA polymerase. The template DNA for M1 RNA was prepared by PCR using a plasmid encoding M1 RNA⁸⁶ (Supplementary Data 3) as a template, and primers 53 and 54.

To prepare tRNA^Thr with a CGU anticodon, plasmid DNA containing tRNA^Thr GGU was mutated at the anticodon region by PCR using primers 61 and 62, followed by cloning in E. coli. Similarly, tRNA^Leu variants were prepared by substituting only the anticodon of tRNA^LeuCAG with primers 63 and 64. The resultant plasmids were used as PCR templates to prepare the template DNA for in vitro transcription.

tRNA template preparation for expression in PURE system

The tRNA templates used in Fig. 1 (those attached to a leader sequence) were prepared by PCR using the plasmids encoding each tRNA⁴² and the primer sets listed in Supplementary Data 1, 2. The tRNA templates used in Fig. 2 (5’-G or-A variants) were the same as those used for in vitro transcription, as described above. The tRNAs templates used in Fig. 4 (those attached to the HDVR and linked tRNAs) were prepared as follows. Plasmids encoding each sequence (tRNA^Ala-Ser, tRNA^Ala-Ser-HDVR, and tRNA^Ala-Ser-HDVR-T7hyb10) were constructed using PCR to prepare the vector and insert sequences, followed by ligation by using the In-Fusion cloning kit (Takara, Japan). The tRNAs templates were prepared by PCR using primers 3 and 56 and each of the plasmids. The tRNA templates used in the tRNA array method were prepared as follows. Each plasmid encoding GARCQ, DfMHKV, SmMFT, LWY, or IPEN was obtained from Twist Bioscience’s Artificial Gene Synthesis Service (ex. pTwist_tRNAs^GARCQ). The plasmids used for the larger tRNA arrays are shown in Supplementary Fig. 15 were constructed sequentially. The plasmid encoding all 21 tRNAs used in Fig. 6 was constructed by PCR amplification of each fragment and assembled using the In-Fusion cloning kit (Takara, Japan). The tRNAs templates were prepared by PCR using primer 3 and 56 and each of the plasmid. All tRNA sequences used in this study are listed in Supplementary Data 4.

Strains

E. coli strain A19 was used for ribosome purification. For other protein components, an E. coli strain, BW25113ΔuidA(DE3)/pREP4, was used after introducing an expression plasmid for the target gene. The plasmids were previously reported³³. The strain was constructed from JW1609 in the KEIO collection⁸⁷ by removing the Km marker and then introducing λDE3 prophage, which expresses T7 RNA polymerase, and a plasmid (pREP4), which expresses lacI. The reason for using the ΔuidA strain is because we previously used glucuronidase (uidA) as a reporter gene.

Culture for ribosome purification

For the ribosome, fresh colonies of E. coli A19 strain were inoculated in 5 mL of LB medium and incubated at 37 °C for 8 h. The culture was added to 100 mL LB medium and incubated at 37 °C overnight. The culture was poured into a fermenter (Bioneer-500, B. E. MARUBISHI Co., Ltd., Japan) that contained 2.5 L of a prewarmed fermenter medium for ribosome and incubated at 37 °C for 4–5 h with mixing at 900 rpm and an air flow (4) until OD600 reached 7.5. During incubation, the pH was maintained at approximately 7 using 6 N NaOH. The cells (~ 60 g) were collected by centrifugation, quickly frozen in liquid nitrogen, and stored at − 80 °C.

The fermenter medium for ribosome was prepared as follows. Solution 1 was prepared by mixing yeast extract (17.5 g), NaCl (2.5 g), (NH₄)₂SO₄ (2.5 g), CaCl₂·2H₂O (2.1 g), FeSO₄·7H₂O (1.25 g), K₂SO₄ (1.5 g), CoCl₂·6H₂O (12.5 mg), CuSO₄·5H₂O (75 mg), Na₂MoO₄·2H₂O (12.5 mg), Mn(OAc)₂·4H₂O (60 mg), and ZnSO₄·7H₂O (163 mg) in 2200 mL tap water. Solution 2 was prepared by dissolving D(+)-glucose (25 g) and thiamine-HCl (25 mg) in 62.5 mL tap water. Solution 3 was prepared by dissolving KH₂PO₄ (4.25 g) and (NH₄)₂HPO₄ (27.5 g) in distilled water (188 mL). Solution 4 was prepared by dissolving MgSO₄·7H₂O (6.25 g) in 62.5 mL tap water. After autoclaving (121 °C for 20 min) each solution, all solutions were mixed in the fermenter.

Culture for the translational proteins

For other protein components, after introducing each expression plasmid into the E. coli strain, the colonies were inoculated into 5 mL of LB medium containing ampicillin (50 µg/mL) and kanamycin (50 µg/mL) and incubated at 37 °C for 8 h. The culture was added to 100 mL of LB medium containing ampicillin and kanamycin at the same concentrations and incubated at 37 °C overnight. The cells were collected by centrifugation and resuspended in 100 mL fresh LB medium. The cell suspension was poured into a fermenter (Bioneer-500, B. E. MARUBISHI Co., Ltd., Japan) containing 2 L of prewarmed fermenter medium and incubated at 37 °C for 6 h with mixing at 900 rpm and an air flow of 4. Next, 600 µL of 500 mg/ml ampicillin was added to the culture. The feeding medium 1 was attached to the fermenter, and the culture was further incubated at 37 °C overnight with supplying the feeding medium 1 at a rate of a few drops per 10 s. During incubation, the pH was maintained at approximately 7 using 6 N NaOH. The next day, 1.125 mL of 500 mg/ml ampicillin and 1 M IPTG (2.5 mL) was added to the culture. The feeding medium 2 was attached to the fermenter, and the culture was further incubated at 37 °C for 4 h with supplying the feeding medium 2 at a rate of a few drops per 4 s. If the Do value was greater than 10% during incubation, the feeding speed decreased. Then, the cells (approximately 300 g) were collected by centrifugation, quickly frozen in liquid nitrogen, and stored at − 80 °C.

The fermenter medium for translational proteins was prepared as follows: Solution 1 was prepared by mixing yeast extract (14 g), NaCl (2 g), (NH₄)₂SO₄ (2 g), CaCl₂ (1.28 g), FeSO₄·7H₂O (1 g), K₂SO₄ (1.2 g), CoCl₂·6H₂O (10 mg), CuSO₄·5H₂O (60 mg), Na₂MoO₄·2H₂O (10 mg), Mn(OAc)₂·4H₂O (52 mg), and ZnSO₄·7H₂O (130 mg) in 1750 mL of tap water. Solution 2 was prepared by dissolving D(+)-glucose (20 g) and thiamine-HCl (20 mg) in 50 mL of tap water. Solution 3 was prepared by dissolving KH₂PO₄ (3.4 g) and (NH₄)₂HPO₄ (22 g) in distilled water (150 mL). Solution 4 was prepared by dissolving MgSO₄ · 7H₂O (5 g) in tap water (50 mL). After autoclaving (121 °C for 20 min) each solution, all solutions were mixed in the fermenter with 0.4 mL of 500 mg/ml ampicillin and 2 mL of 50 mg/ml kanamycin.

The feeding medium 1 was prepared as follows. Solution 1 was prepared by dissolving yeast extract (20 g) in 100 mL of tap water. Solution 2 was prepared by dissolving D(+)-glucose (100 g) and thiamine-HCl (10 mg) in tap water (150 mL). After autoclaving (121 °C for 20 min) each solution, the two solutions, 175 µL of 500 mg/ml ampicillin, and 250 µL of 50 mg/ml kanamycin were mixed just before use.

The feeding medium 2 was prepared as follows. Solution 1 was prepared by dissolving yeast extract (100 g) in 200 mL of tap water. Solution 2 was prepared by dissolving D(+)-glucose (80 g) and thiamine-HCl (10 mg) in tap water (50 mL). After autoclaving (121 °C for 20 min) each solution, the two solutions, 350 µL of 500 mg/ml ampicillin, and 250 µL of 50 mg/ml kanamycin were mixed just before use.

Purification of ribosome

The frozen cells (~ 10 g) were mixed with two volumes (~ 20 mL) of S20 buffer. Three times the amount of 0.1 mm glass beads (~ 30 g) was added to the cell mixture and vigorously shaken with Multi-beads Shoker (Yasui Kikai, Japan) at 2500 rpm for 30 s for 3 cycles at 0 °C to disrupt the cells. After removing cell debris and beads by centrifugation at 6 k × g for 20 min at 4 °C, the supernatant was further centrifuged at 18 krpm for 30 h at 4 °C with R10A2 rotor (Himac). The supernatant was centrifuged again at 18 krpm for 30 min at 4 °C with R20A2 rotor (Himac). Ammonium sulfate (0.222 g per 1 mL of the supernatant) was dissolved in the supernatant with stirring at 4 °C, stand for 20 min at at 4 °C, and centrifuged at 18 krpm for 60 min at 4 °C with R20A2 rotor (Himac). The supernatant was filtered through a 0.45 µm filter and applied to a HiTrap Butyl FF column (5 mL ×4, Cytiva) at 2 mL/min using Akta go (Cytiva). The column was washed at 1 mL/min with 30 mL of butyl-column buffer A. Then, the column was washed at 2 mL/min with 60 mL of a solution (80% butyl-column buffer A and 20% butyl-column buffer B), followed by elution with a solution at 2 mL/min with 40 mL of a solution (50% butyl-column buffer A and 50% butyl-column buffer B). The fractions around the A260 peak were collected. The collected fractions ( ~ 10 mL) were placed on 30 mL of 30% sucrose buffer and centrifuged at 30 krpm for 17 h at 4 °C with 45Ti rotor (Beckman). The precipitate was dissolved in 70S buffer and concentrated using an Amicon Ultra (Millipore, 100 kDa cut). The final solution was quickly frozen in liquid nitrogen and stored at − 80 °C. The protein concentration was calculated using the A260 value.

S20 buffer contained 10 mM HEPES-KOH (pH7.6), 10 mM MgCl₂, 10 mM KCl, and 1 mM DTT. Butyl-column buffer A contains 20 mM HEPES-KOH (pH7.6), 10 mM Mg(OAc)₂, 1.5 M (NH₄)₂SO₄, and 7 mM 2-mercaptoethanol. Butyl column buffer B contained 20 mM HEPES-KOH (pH7.6), 10 mM Mg(OAc)₂, and 7 mM 2-mercaptoethanol. The 30% sucrose buffer contained 20 mM HEPES-KOH (pH7.6), 10 mM Mg(OAc)₂, 30 mM (NH₄)Cl, 30% sucrose, and 7 mM 2-mercaptoethanol. 70S buffer contained 20 mM HEPES-KOH (pH7.6), 6 mM Mg(OAc)₂, 30 mM KCl, and 7 mM 2-mercaptoethanol.

Purification of the translational proteins

The frozen cells (20–40 g) were dissolved in three volumes (60–120 mL) of Ni-NTA buffer A. The same amount of glass beads 0.1 mm (20–40 g) was added to the cell suspension and vigorously shake with Multi-beads Shoker (Yasui Kikai, Japan) at 2500 rpm 30 sec for 3 cycles at 0 °C to disrupt the cells. After removing cell debris and the beads by centrifugation at 8 k × g for 10 min at 4 °C, the supernatant was further centrifuged at 40 krpm for 1 h at 4 °C with a 45Ti rotor (Beckman). The supernatant was diluted up to 100 mL with Ni-NTA buffer A and centrifuged at 40 krpm for 1 h at 4 °C with 45Ti rotor again. After filtering with a 0.45 µm filter, the supernatant was applied to a HisTrap column (5 mL × 3, Cytiva) at 2.5 mL/min using Akta go (Cytiva). The column was washed at 2 mL/min with 60 mL of a solution (95% Ni-NTA buffer A and 5% Ni-NTA buffer B). The target protein was eluted at 2 mL/min with a linear gradient of 5% to 50% Ni-NTA buffer B for 30 min. Fractions containing the target proteins (typically ~ 12 mL) were collected and concentrated using an Amicon Ultra (Millipore). The pore size of Amicon Ultra was determined based on the size of the target protein. The protein solution was then applied to HiLoad® 16/600 Superdex 200 pg (Cytiva) using Akta Go and eluted with 150 mL of gel-filtration buffer. Fractions containing the target proteins (typically ~12 mL) were collected and concentrated using an Amicon Ultra (Millipore). The final solution was mixed with 2 × glycerol stock buffer, quickly frozen in liquid nitrogen, and stored at − 80 °C. The protein concentration was calculated from the A280 value, and the absorbance was estimated using ProtParam⁸⁸.

The Ni-NTA buffer A contained 50 mM HEPES-KOH (pH7.6), 10 mM MgCl₂, 1 M NH₄Cl, 15% glycerol, and 7 mM 2-mercaptoethanol. When purifying EF-Tu, 10 µM GDP was additionally supplied at 10 µM. The Ni-NTA buffer B contained 50 mM HEPES-KOH (pH7.6), 10 mM MgCl₂, 100 mM KCl, 500 mM imidazole, 15% glycerol, and 7 mM 2-mercaptoethanol. When purifying EF-Tu, 10 µM GDP was additionally supplied at 10 µM. The gel filtration buffer contained 50 mM HEPES-KOH (pH7.6), 10 mM MgCl₂, 100 mM KCl, and 7 mM 2-mercaptoethanol. The 2 × glycerol stock buffer contained 50 mM HEPES-KOH (pH7.6), 10 mM MgCl₂, 100 mM KCl, 60% glycerol, and 7 mM 2-mercaptoethanol.

RNase P digestion in buffer R

RNase P digestion in Figs. 4E and 7B was performed in buffer R, which consisted of 50 mM Tris-HCl (pH 7.6), 60 mM NH₄Cl, 10 mM Mg(OAc)₂, and 5 mM spermidine. Transcripts (250 nM) were incubated with M1 RNA (500 nM) in buffer R at 37 °C for 16 h.

Preparation of tfPURE system and C5 protein

All components of the lab-made PURE system were prepared using affinity chromatography with a histidine tag, followed by gel-filtration chromatography based on a previous study⁸⁹. EF-Tu was further purified by two times of affinity chromatography to remove residual tRNA as previously reported⁴⁹. Purified EF-Tu was 10-fold diluted with a stringent buffer [50 mM HEPES–KOH (pH 7.6), 1 M KCl, 10 mM MgCl2, 15% glycerol, 1 mM DTT, and 1% Triton X-100] and allowed to stand on ice for 1 h. The solution was injected into to the HisTrap column (Thermo Fisher Scientific) and washed with the same buffer, omitting Triton X-100. EF-Tu was eluted using a 0–250 mM imidazole gradient. This purification process was repeated once. Protein-containing fractions were collected and stored following buffer exchange with the stock buffer [50 mM HEPES–KOH (pH 7.6), 100 M KCl, 10 mM MgCl2, 30% glycerol, and 7 mM 2-mercaptoethanol) using a 30 kDa Amicon Ultra centrifugal filter.

The ribosome was further purified by using a size-separation spin column as follows. Ribosomes purified by the previous method were diluted to a concentration of 200 nM using Buffer D (10 mM Tris-OAc, pH 7.5, 1 mM Mg(OAc)₂, 60 mM NH₄Cl, 0.5 mM EDTA, 6 mM 2-mercaptoethanol), which had the same composition as the buffer used for subunit separation⁹⁰. Centrifugation was performed using a 100 kDa cut-off Amicon Ultra centrifugal filter until the ribosome concentration reached approximately 1 μM. The sample was then diluted again to ~ 200 nM with Buffer D, followed by repeated centrifugation. This cycle was repeated 20 times. Subsequently, buffer exchange was performed by diluting the sample five-fold in ribosome stock buffer (20 mM HEPES-KOH, pH 7.6, 6 mM Mg(OAc)₂, 30 mM KCl, and 7 mM 2-mercaptoethanol) and centrifuging once, followed by two consecutive 50-fold dilutions and centrifugation. The tfPURE systems were prepared using further purified EF-Tu and ribosome, with other proteins purified by the conventional method. Their compositions are listed in Supplementary Tables 1 and 2.

C5 protein was purified from E. coli harvoring C5 expressing plasmid, pET-C5⁸⁶, using affinity chromatography for the histidine tag by the same method as the component of the PURE system as described above.

PAGE analysis

Electrophoresis was carried out using 8% (acrylamide:bis = 19:1) polyacrylamide gel containing 8 M urea, 0.1% ammonium persulfate, and 0.1% N,N,N’,N’-tetramethylethylene-diamine in Tris–borate EDTA buffer. The samples were prepared by mixing with a stripping buffer composed of 50 mM EDTA, 90% formamide, and 0.025% bromophenol blue. RNA was stained with SYBR Green II (Takara, Japan). RNA size marker (DynaMarker RNA Low II Easy Load, BioDynamics Laboratory Inc., Japan) was loaded alongside the samples for all PAGE analyses. While the cropped gel images shown in the figures focus on the relevant bands for clarity, full uncropped images including the size marker are provided in the source data.

Translation assays with the 21 IVS tRNAs conducted in Fig. 2B

The reaction mixture containing the 21 IVS tRNAs, luciferase DNA (1 nM), T7 RNA polymerase (0.42 µ/µL, Takara, Japan), NTP (0.88 mM), magnesium acetate (7.9 mM), and the laboratory-made tfPURE system was incubated at 30 °C for 16 h. The total concentration of the IVS tRNAs was 600 ng/μL (100 ng/μL each for IPEN and 11.7 ng/μL each for the others), as previously reported⁴⁹. For assays involving tRNA variants, PURE systems lacking the respective tRNA were prepared, and each tRNA variant was added at the same concentration as the WT tRNA in the IVS tRNA mixture. After the incubation, an aliquot (1 μL) of the reaction mixture was added to 30 μL of Luciferase assay reagent (Promega), and luminescence was measured with GloMax Luminometer (Promega).

tRNA expression-coupled translation in the PURE system

The reaction mixture containing reporter DNA encoding luciferase, Nanoluc, sfGFP, or GUS (1 nM), tRNA template DNA, T7 RNA polymerase (Takara, Japan), T4 PNK (New England Biolabs), RNase P (M1 RNA and C5 protein), and the tfPURE system was incubated at 30 °C. The concentrations of the tRNA template DNA, T7 RNA polymerase, T4 PNK, and RNase P varied depending on the experiment (see each legend). The reaction time was set to 16 h or 24 h, depending on the experiment (see each legend). For each tRNA expression scheme, the concentrations of NTPs, Mg²⁺, spermidine, and T7 RNA polymerase were optimized (Supplementary Table 2). After incubation, an aliquot (1 μL) of the reaction mixture was added to 30 μL of Luciferase assay reagent (Promega) for luciferase or Nanoluc assay reagent (Promega) for Nanoluc, and luminescence was measured using GloMax Luminometer (Promega). sfGFP fluorescence was measured 10 min for 24 h (Mx3005P, Agilent Technologies). To measure GUS activity, a 2 μL aliquot of the translation reaction was added to 8 μL of GUS reaction buffer⁹¹ (final concentrations: 50 mM HEPES, pH 7.4; 0.01% Triton X-100; 5 μM TokyoGREEN-βGlcU(Na) (Sekisui Medical, Japan)). Fluorescence was monitored every 1 min for 60 min using an Mx3005P with FAM settings. The apparent k_cat was calculated from the slope of the linear portion of the fluorescence time course and the amount of synthesized GUS quantified using the HiBiT assay.

Western blotting

After incubation, the reaction mixture (1 μL) was mixed with SDS buffer (17 mM Tris-HCl, pH 7.4, 0.7% SDS, 0.3 M 2-mercaptoethanol, and 3% glycerol), boiled at 95 °C for 5 min, and subjected to 10% SDS-PAGE. Protein transfer was performed using the WSE-4115 system (Atto, Japan), followed by blocking with a primary antibody (Host/Isotyle: Rabbit/IgG, anti-luciferase antibody; PGI Proteintech Group. Inc., 27986-1-AP, 1:1000 dilution), and secondary antibodies (Host: Goat anti-rabbit IgG, PGI Proteintech Group. Inc., SA00001-2, 1:1000 dilution) using an iBind Western System (Thermo Fisher). Pierce ECL Western Blotting Substrate (Thermo Fisher) was used as a light-emitting chemiluminescent substrate, incubated at room temperature, and detected using a FUSION-SL4 imaging system (Vilber-Lourmat). A dilution series of recombinant luciferase (L9420; Sigma-Aldrich) was prepared and used as a standard for quantification.

Protein quantification using HiBiT system

Luminescence from HiBiT-tagged proteins synthesized in the PURE system was measured using the Nano-Glo HiBiT Lytic Detection System (Promega). A dilution series of HiBiT Control Protein (Promega) was used as the standard.

Nanopore tRNA direct sequencing

After the tRNA expression-coupled translation reaction in the PURE system, the total RNA was purified using the PureLink RNA Mini Kit (Invitrogen). To remove the 5’-end triphosphate from specific tRNAs (Gly, Asp, Ser, Leu, and Pro), RppH was added to NEBuffer 2, followed by incubation at 37 °C for 30 min and 65 °C for 5 min. The RNA was purified using AMPure RNAClean XP. Subsequently, direct tRNA sequencing was performed as previously reported^67,68 using RNA Direct Sequencing Kit (Oxford Nanopore). Purified RNAs were incubated with 5’ (5’ P-GGCUUCUUCUUGCUCUUAGGAAAAAAAAAAaaa-3’, where lowercase “a” denotes DNA) and 3’ (5’-CCUAAGAGCAAGAAGAAGCCUGGN-3’, where “N” represents A, U, G, or C) splint adapters for the first ligation. The RNA was purified again using AMPure RNAClean XP. RNA was then ligated to the RTA adapter. The RNA was linearized by reverse transcription using Maxima H Minus Reverse Transcriptase (Life Technologies, EP0751). Linearized tRNAs were cleaned using 2× AMPure RNAClean XP beads. Finally, ONT RMX sequencing adapters were ligated and applied to a MinION flow cell (FLO-MIN-106) using the standard ONT SQK-RNA002 protocol. Reads were mapped using minimap2 version 2.24-r1122 with sensitive parameters (-ax map-on -k10).

Simultaneous 21 tRNAs synthesis for codon reprograming

In vitro transcription was carried out with the single 21 tRNAs template (25 nM) in the presence of T7 RNA polymerase (1.0 µ/μL) and T4 PNK (0.1 U/μL, New England Biolab), NTPs (2 mM each: ATP, GTP, CTP, and UTP), GMP (3 mM), Tris-HCl pH 8.0 (40 mM), magnesium chloride (10 mM), spermidine (2 mM), 0.2 µ/ml inorganic pyrophosphatase (New England Biolab), and 0.8 µ/μl RNasin (Promega) at 37 °C for 12 h. The premature tRNA product was purified using PureLink RNA Mini Kit (Invitrogen). The purified premature tRNA was incubated in buffer R (see RNase P digestion in buffer R section) with RNase P (4 μM M1 RNA and 6 μM C5 protein) at 37 °C for 12 h for digestion. The mature tRNA product was purified using the PureLink RNA Mini Kit.

Mass spectrometry analysis of synthesized Nanoluc

Nanoluc with Strep tag II (Supplementary Data 3) was synthesized using the tfPURE system under each experimental condition (Fig. 7C, F). The strep-tagged Nanoluc was concentrated using MagStrep-type 2HC beads (IBA) and then digested on the beads. The reaction mixture (15 μL) was mixed with 85 μL of binding buffer (50 mM Tris-HCl, pH 7.6, 150 mM NaCl) and 5 μL of the beads slurry for 30 min. The beads were washed three times with washing buffer (50 mM Tris-HCl, pH 7.6, 500 mM NaCl) and then digested with 5 ng/μL Lys-C/Trypsin Protease Mix (Thermo Scientific) in 50 mM NH₄HCO₃ at 37 °C for 2 h. The digest was purified using GL-Tip SDB (GL Sciences) according to the manufacturer’s instructions, dried under reduced pressure, re-dissolved in 10 μL of 0.1% TFA, and 1 μL of the sample was analyzed by LC-MS. Each biological replicate (n = 1) was injected once. The peptides were concentrated and separated using a nano-LC system (EASY-nLC, Thermo Scientific, Bremen, Germany) equipped with a trap column (C18, 0.075 × 20 mm, 3 µm, Acclaim PepMap 100, Thermo Scientific) and a nanocapillary analytical column (C18, 0.075 × 150 mm, 3 µm, Nikkyo Technos, Tokyo, Japan) at a flow rate of 200 nl/min. Mobile phases A (0.1% formic acid) and B (90% acetonitrile and 0.1% formic acid) were combined in a gradient as follows: 5% B for 5 min, 5–40% B for 75 min, 40–90% B for 1 min, and 90% B for 4 min. Mass spectrometry (MS) analysis was performed using an Orbitrap mass spectrometer (Velos Pro, Thermo Scientific) equipped with a nanospray ion source (Nanospray Flex, Thermo Scientific) with the following parameters: spray voltage of 2.2 kV, positive mode, scan range of m/z 310–2000, and 70,000 resolution. The ten most intense multiply charged ions (z = 2–3) were fragmented in the collision cell by collision-induced dissociation (CID). The raw data files obtained from mass spectrometry analysis were converted into ms1 format using MSConvertGUI, included in ProteoWizard version 3.0.25148⁹². Based on the converted files, extracted ion chromatograms were generated using an in-house script written in Java (version 1.8.0_311), where the sum of intensities within a window of twice the theoretical resolution (2xFWHM) at each retention time was normalized and plotted.

Statistics and reproducibility

No statistical method was used to predetermine sample size. No data was excluded from the analysis. The investigators were not blinded to allocation during experiments and outcome assessment. Each experiment was independently repeated at least three times. Biological replicates refer to independent experiments conducted on separate days using freshly prepared reagents.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

All data supporting the findings of this study are available within the Article and its Supplementary Information files. Source data underlying the figures are provided in the accompanying Source Data file. The Source Data file includes raw measurements, quantified values, and uncropped scans of gels and blots used to generate the main figures and Supplementary Figs. Nanopore data generated in this study have been deposited in the NCBI Sequence Read Archive under accession code PRJNA1233667. MS data are available in the ProteomeXchange Consortium via the jPOST partner repository with the dataset identifiers PXD066302 and JPST003947. Source data are provided in this paper.

References

Buddingh’, B. C. & Van Hest, J. C. M. Artificial cells: Synthetic compartments with life-like functionality and adaptivity. Acc. Chem. Res. 50, 769–777 (2017).
Ivanov, I. et al. Bottom-Up synthesis of artificial cells: Recent highlights and future challenges. Annu. Rev. Chem. Biomol. Eng. 12, 287–308 (2021).
Olivi, L. et al. Towards a synthetic cell cycle. Nat. Commun. 12, 4531 (2021).
Biner, O., Fedor, J. G., Yin, Z. & Hirst, J. Bottom-Up construction of a minimal system for cellular respiration and energy regeneration. ACS Synth. Biol. 9, 1450–1459 (2020).
Libicher, K., Hornberger, R., Heymann, M. & Mutschler, H. In vitro self-replication and multicistronic expression of large synthetic genomes. Nat. Commun. 11, https://doi.org/10.1038/s41467-020-14694-2 (2020).
Libicher, K. & Mutschler, H. Probing self-regeneration of essential protein factors required for: In vitro translation activity by serial transfer. Chem. Commun. 56, 15426–15429 (2020).
Doerr, A., Foschepoth, D., Forster, A. C. & Danelon, C. In vitro synthesis of 32 translation-factor proteins from a single template reveals impaired ribosomal processivity. Sci. Rep. 11, 1898 (2021).
Blanken, D., Foschepoth, D., Serrão, A. C. & Danelon, C. Genetically controlled membrane synthesis in liposomes. Nat. Commun. 11, 4317 (2020).
Wang, C., Yang, J. & Lu, Y. Modularize and Unite: Toward creating a functional artificial cell. Front. Mol. Biosci. 8, https://doi.org/10.3389/fmolb.2021.781986 (2021).
Liu, Y., Davis, R. G., Thomas, P. M., Kelleher, N. L. & Jewett, M. C. In vitro-constructed ribosomes enable multi-site incorporation of noncanonical amino acids into proteins. Biochemistry 60, 161–169 (2021).
Article CAS PubMed Google Scholar
Vibhute, M. A. et al. Transcription and translation in cytomimetic protocells perform most efficiently at distinct macromolecular crowding conditions. ACS Synth. Biol. 9, 2797–2807 (2020).
Amy Yewdall, N., Mason, A. F. & Van Hest, J. C. M. The hallmarks of living systems: Towards creating artificial cells. Interface Focus 8, https://doi.org/10.1098/rsfs.2018.0023 (2018).
Tror, S., Jeon, S. M., Nguyen, H. T., Huh, E. & Shin, K. A self-regenerating artificial cell, that is one step closer to living cells: Challenges and perspectives. Small Methods 7, 2300182 (2023).
Article CAS Google Scholar
Kosaka, Y. et al. Autonomous ribosome biogenesis in vitro. Nat. Commun. 16, 1–14 (2025).
Article Google Scholar
Ichihashi, N. What can we learn from the construction of in vitro replication systems?. Ann. N. Y. Acad. Sci. 1447, 144–156 (2019).
Article PubMed Google Scholar
Nishikawa, S. et al. Amino acid self-regenerating cell-free protein synthesis system that feeds on PLA plastics, CO2, ammonium, and α-Ketoglutarate. ACS Catal. 14, 7696–7706 (2024).
Article CAS Google Scholar
Fujiwara, K., Katayama, T. & Nomura, S. I. M. Cooperative working of bacterial chromosome replication proteins generated by a reconstituted protein expression system. Nucleic Acids Res. 41, 7176–7183 (2013).
Article CAS PubMed PubMed Central Google Scholar
Su’etsugu, M., Takada, H., Katayama, T. & Tsujimoto, H. Exponential propagation of large circular DNA by reconstitution of a chromosome-replication cycle. Nucleic Acids Res. 45, 11525–11534 (2017).
Article PubMed PubMed Central Google Scholar
van Nies, P. et al. Self-replication of DNA by its encoded proteins in liposome-based synthetic cells. Nat. Commun. 9, 1583 (2018).
Article PubMed PubMed Central Google Scholar
Abil, Z. et al. Darwinian evolution of self-replicating DNA in a synthetic protocell. Nat. Commun. 15, 1–15 (2024).
Article Google Scholar
Sakatani, Y., Yomo, T. & Ichihashi, N. Self-replication of circular DNA by a self-encoded DNA polymerase through rolling-circle replication and recombination. Sci. Rep. 8, 13089 (2018).
Sakatani, Y., Ichihashi, N., Kazuta, Y. & Yomo, T. A transcription and translation-coupled DNA replication system using rolling-circle replication. Sci. Rep. 5, 10404 (2015).
Okauchi, H. & Ichihashi, N. Continuous cell-free replication and evolution of artificial genomic DNA in a compartmentalized gene expression system. ACS Synth. Biol. 10, 3507–3517 (2021).
Grasemann, L., Thiel Pizarro, P. & Maerkl, S. J. C2CAplus: A one-pot isothermal circle-to-circle DNA amplification system. ACS Synth. Biol. 12, 3137–3142 (2023).
Article CAS PubMed PubMed Central Google Scholar
Okauchi, H., Sakatani, Y., Otsuka, K. & Ichihashi, N. Minimization of elements for isothermal DNA replication by an evolutionary approach. ACS Synth. Biol. 9, 1771–1780 (2020).
Article CAS PubMed Google Scholar
Eto, S. et al. Phospholipid synthesis inside phospholipid membrane vesicles. Commun. Biol. 5, 1016 (2022).
Article CAS PubMed PubMed Central Google Scholar
Bailoni, E. et al. Synthetic vesicles for sustainable energy recycling and delivery of building blocks for lipid biosynthesis. ACS Synth. Biol. 13, 1549–1561 (2024).
Article CAS PubMed PubMed Central Google Scholar
Sato, G. et al. Metabolic tug-of-war between glycolysis and translation revealed by biochemical reconstitution. ACS Synth. Biol. 13, 1572–1581 (2024).
Article CAS PubMed Google Scholar
Giaveri, S. et al. Integrated translation and metabolism in a partially self-synthesizing biochemical network. Science 385, 174–178 (2024).
Article CAS PubMed Google Scholar
Kohyama, S., Merino-Salomón, A. & Schwille, P. In vitro assembly, positioning and contraction of a division ring in minimal cells. Nat. Commun. 13, 1–14 (2022).
Article Google Scholar
Chang, M. Y., Ariyama, H., Huck, W. T. S. & Deng, N. N. Division in synthetic cells. Chem. Soc. Rev. 52, 3307–3325 (2023).
Article CAS PubMed Google Scholar
De Capitani, J. & Mutschler, H. The long road to a synthetic self-replicating central dogma. Biochemistry 62, 1221–1232 (2023).
Article PubMed Google Scholar
Shimizu, Y. et al. Cell-free translation reconstituted with purified components. Nat. Biotechnol. 19, 751–755 (2001).
Lavickova, B., Laohakunakorn, N. & Maerkl, S. J. A partially self-regenerating synthetic cell. Nat. Commun. 11, 6340 (2020).
Schwarz-Schilling, M. et al. Autonomous biogenesis of the entire protein translation machinery excluding ribosomes. Preprint at https://doi.org/10.1101/2024.10.20.619270 (2024).
Hagino, K., Masuda, K., Shimizu, Y. & Ichihashi, N. Sustainable regeneration of 20 aminoacyl-tRNA synthetases in a reconstituted system toward self-synthesizing artificial systems. Sci. Adv. 11, 6269 (2025).
Article Google Scholar
Levy, M., Falkovich, R., Daube, S. S. & Bar-Ziv, R. H. Autonomous synthesis and assembly of a ribosomal subunit on a chip. Sci. Adv. 6, https://doi.org/10.1126/sciadv.aaz6020 (2020).
Li, J. et al. Cogenerating synthetic parts toward a self-replicating system. ACS Synth. Biol. 6, 1327–1336 (2017).
Article CAS PubMed Google Scholar
Shimojo, M. et al. In vitro reconstitution of functional small ribosomal subunit assembly for comprehensive analysis of ribosomal elements in E. coli. Commun. Biol. 3, 142 (2020).
Article CAS PubMed PubMed Central Google Scholar
Aoyama, R. et al. In vitro reconstitution of the Escherichia coli 70S ribosome with a full set of recombinant ribosomal proteins. J. Biochem. 171, 227–237 (2022).
Article CAS PubMed Google Scholar
Shepherd, J. & Ibba, M. Bacterial transfer RNAs. FEMS Microbiol. Rev. 39, 280 (2015).
Article CAS PubMed PubMed Central Google Scholar
Hibi, K. et al. Reconstituted cell-free protein synthesis using in vitro transcribed tRNAs. Commun. Biol. 3, 350 (2020).
Iwane, Y. et al. Expanding the amino acid repertoire of ribosomal polypeptide synthesis via the artificial division of codon boxes. Nat. Chem. 8, 317–325 (2016).
Article CAS PubMed Google Scholar
Li, J., Tang, M. & Qi, H. Codon-reduced protein synthesis with manipulating tRNA components in cell-free system. Front. Bioeng. Biotechnol. 10, 891808 (2022).
Article PubMed PubMed Central Google Scholar
Calles, J., Justice, I., Brinkley, D., Garcia, A. & Endy, D. Fail-safe genetic codes designed to intrinsically contain engineered organisms. Nucleic Acids Res. 47, 10439–10451 (2019).
Fujino, T., Tozaki, M. & Murakami, H. An amino acid-swapped genetic code. ACS Synth. Biol. 9, 2703–2713 (2020).
Article CAS PubMed Google Scholar
Fujino, T. et al. Ser/Leu-swapped cell-free translation system constructed with natural/in vitro transcribed-hybrid tRNA set. Nat. Commun. 15, 1–10 (2024).
Article Google Scholar
Mohanty, B. K. & Kushner, S. R. New insights into the relationship between tRNA processing and polyadenylation in Escherichia coli. Trends Genet. 35, 434–445 (2019).
Article CAS PubMed PubMed Central Google Scholar
Miyachi, R., Shimizu, Y. & Ichihashi, N. Transfer RNA synthesis-coupled translation and DNA replication in a reconstituted transcription/translation system. ACS Synth. Biol. 11, 2791–2799 (2022).
Article CAS PubMed Google Scholar
Cobaleda, C. & Sánchez-García, I. RNase P: from biological function to biotechnological applications. Trends Biotechnol. 19, 406–411 (2001).
Article CAS PubMed Google Scholar
Guerrier-Takada, C., Gardiner, K., Marsh, T., Pace, N. & Altman, S. The RNA moiety of ribonuclease P is the catalytic subunit of the enzyme. Cell 35, 849–857 (1983).
Article CAS PubMed Google Scholar
Baer, M. F. et al. The recognition by RNase P of precursor tRNAs. J. Biol. Chem. 263, 2344–2351 (1988).
Article CAS PubMed Google Scholar
Samhita, L., Virumäe, K., Remme, J. & Varshney, U. Initiation with elongator tRNAs. J. Bacteriol. 195, 4202 (2013).
Article CAS PubMed PubMed Central Google Scholar
Meinnel, T. & Blanquet, S. Maturation of Pre-tRNAfMet by Escherichia coli RNase P is specified by a guanosine of the 5′-flanking sequence. J. Biol. Chem. 270, 15908–15914 (1995).
Article CAS PubMed Google Scholar
Coleman, T. M., Wang, G. & Huang, F. Superior 5’ homogeneity of RNA from ATP-initiated transcription under the T7 phi 2.5 promoter. Nucleic Acids Res.32, e14 (2004).
Article PubMed PubMed Central Google Scholar
Himeno, H., Hasegawa, T., Asahara, H., Tamura, K. & Shimizu, M. Identity determinants of E. coli tryptophan tRNA. Nucleic Acids Res. 19, 6379 (1991).
Article CAS PubMed PubMed Central Google Scholar
Jahn, M., Rogers, M. J. & Söll, D. Anticodon and acceptor stem nucleotides in tRNA(Gln) are major recognition elements for E. coli glutaminyl-tRNA synthetase. Nature 352, 258–260 (1991).
Article CAS PubMed Google Scholar
Nureki, O. et al. Molecular recognition of the identity-determinant set of isoleucine transfer RNA from Escherichia coli. J. Mol. Biol. 236, 710–724 (1994).
Article CAS PubMed Google Scholar
Liu, H., Peterson, R., Kessler, J. & Musier-forsyth, K. Molecular recognition of tRNA(Pro) by Escherichia coli proline tRNA synthetase in vitro. Nucleic Acids Res. 23, 165 (1995).
Article CAS PubMed PubMed Central Google Scholar
Guillon, J. M. et al. Nucleotides of tRNA governing the specificity of Escherichia coli methionyl-tRNAfMet formyltransferase. J. Mol. Biol. 224, 359–367 (1992).
Article CAS PubMed Google Scholar
Schürer, H., Lang, K., Schuster, J. & Mörl, M. A universal method to produce in vitro transcripts with homogeneous 3′ ends. Nucleic Acids Res. 30, e56 (2002).
Article PubMed PubMed Central Google Scholar
Calvopina-Chavez, D. G., Gardner, M. A. & Griffitts, J. S. Engineering efficient termination of bacteriophage T7 RNA polymerase transcription. G312, jkac070 (2022).
Article CAS PubMed PubMed Central Google Scholar
Das, U. & Shuman, S. Mechanism of RNA 2′,3′-cyclic phosphate end healing by T4 polynucleotide kinase–phosphatase. Nucleic Acids Res. 41, 355 (2012).
Article PubMed PubMed Central Google Scholar
Zhang, J. & Ferré-D’Amaré, A. R. Trying on tRNA for Size: RNase P and the T-box Riboswitch as molecular rulers. Biomolecules 6, 18 (2016).
Article PubMed PubMed Central Google Scholar
Niland, C. N., Anderson, D. R., Jankowsky, E. & Harris, M. E. The contribution of the C5 protein subunit of Escherichia coli ribonuclease P to specificity for precursor tRNA is modulated by proximal 5′ leader sequences. RNA 23, 1502–1511 (2017).
Article CAS PubMed PubMed Central Google Scholar
Sun, L., Campbell, F. E., Zahler, N. H. & Harris, M. E. Evidence that substrate-specific effects of C5 protein lead to uniformity in binding and catalysis by RNase P. EMBO J. 25, 3998–4007 (2006).
Article CAS PubMed PubMed Central Google Scholar
Lucas, M. C. et al. Quantitative analysis of tRNA abundance and modifications by nanopore RNA sequencing. Nat. Biotechnol. 42, 72–86 (2023).
Article PubMed PubMed Central Google Scholar
Thomas, N. K. et al. Direct nanopore sequencing of individual full length tRNA strands. ACS Nano 15, 16642–16653 (2021).
Article CAS PubMed PubMed Central Google Scholar
Matsuura, T., Hosoda, K., Ichihashi, N., Kazuta, Y. & Yomo, T. Kinetic analysis of β-galactosidase and β-glucuronidase tetramerization coupled with protein translation. J. Biol. Chem. 286, 22028 (2011).
Article CAS PubMed PubMed Central Google Scholar
Sato, W., Rasmussen, M., Deich, C., Engelhart, A. E. & Adamala, K. P. Expanding luciferase reporter systems for cell-free protein expression. Sci. Rep. 12, 1–10 (2022).
Article Google Scholar
Hall, M. P. et al. Engineered luciferase reporter from a deep sea shrimp utilizing a novel imidazopyrazinone substrate. ACS Chem. Biol. 7, 1848–1857 (2012).
Article CAS PubMed PubMed Central Google Scholar
Ohta, A. et al. Validation of a new methodology to create oral drugs beyond the rule of 5 for intracellular tough targets. J. Am. Chem. Soc. 145, 24035–24051 (2023).
Article CAS PubMed Google Scholar
Goto, Y. & Suga, H. The RaPID platform for the discovery of Pseudo-natural macrocyclic peptides. Acc. Chem. Res. 54, 3604–3617 (2021).
Article CAS PubMed Google Scholar
Costello, A., et al. Genetic code expansion history and modern innovations. Chem. Rev. 124, 11962–12005 (2024).
Article CAS PubMed Google Scholar
Zhu, B., Tabor, S. & Richardson, C. C. Syn5 RNA polymerase synthesizes precise run-off RNA products. Nucleic Acids Res. 42, e33–e33 (2014).
Article CAS PubMed Google Scholar
Conrad, T., Plumbom, I., Alcobendas, M., Vidal, R. & Sauer, S. Maximizing transcription of nucleic acids with efficient T7 promoters. Commun. Biol. 3, 1–8 (2020).
Article Google Scholar
Komura, R., Aoki, W., Motone, K., Satomura, A. & Ueda, M. High-throughput evaluation of T7 promoter variants using biased randomization and DNA barcoding. PLoS ONE 13, e0196905 (2018).
Article PubMed PubMed Central Google Scholar
Deich, C. et al. T7Max transcription system. J. Biol. Eng. 17, 1–14 (2023).
Article Google Scholar
Krupp, G., Kahle, D., Vogt, T. & Char, S. Sequence changes in both flanking sequences of a Pre-tRNA influence the cleavage specificity of RNase P. J. Mol. Biol. 217, 637–648 (1991).
Article CAS PubMed Google Scholar
Xiong, Y. & Steitz, T. A. A story with a good ending: tRNA 3′-end maturation by CCA-adding enzymes. Curr. Opin. Struct. Biol. 16, 12–17 (2006).
Article CAS PubMed Google Scholar
Mohan, A., Whyte, S., Wang, X., Nashimoto, M. & Levinger, L. The 3’ end CCA of mature tRNA is an antideterminant for eukaryotic 3’-tRNase. RNA 5, 245–256 (1999).
Article CAS PubMed PubMed Central Google Scholar
Mohanty, B. K., Maples, V. F. & Kushner, S. R. Polyadenylation helps regulate functional tRNA levels in Escherichia coli. Nucleic Acids Res. 40, 4589–4603 (2012).
Article CAS PubMed PubMed Central Google Scholar
Kienbeck, K. et al. Identification of HDV-like theta ribozymes involved in tRNA-based recoding of gut bacteriophages. Nat. Commun. 15, 1–10 (2024).
Article Google Scholar
Li, F., Baranwal, A. K. & Maerkl, S. J. Continuous in situ synthesis of a complete set of tRNAs sustains steady-state translation in a recombinant cell-free system. Nat. Commun. 16, 1–12 (2025).
CAS PubMed PubMed Central Google Scholar
Cui, Z., Stein, V., Tnimov, Z., Mureev, S. & Alexandrov, K. Semisynthetic tRNA complement mediates in vitro protein synthesis. J. Am. Chem. Soc. 137, 4404–4413 (2015).
Article CAS PubMed Google Scholar
Fukunaga, J. I. et al. Use of RNase P for efficient preparation of yeast tRNATyr transcript and its mutants. J. Biochem. 139, 123–127 (2006).
Article CAS PubMed Google Scholar
Baba, T. et al. Construction of Escherichia coli K-12 in-frame, single-gene knockout mutants: The Keio collection. Mol. Syst. Biol. 2, https://doi.org/10.1038/msb4100050 (2006).
Walker, J. M. et al. The Proteomics Protocols Handbook.(2005).
Fujii, S. et al. Liposome display for in vitro selection and evolution of membrane proteins. Nat. Protoc. 9, 1578–1591 (2014).
Article CAS PubMed Google Scholar
Jewett, M. C., Fritz, B. R., Timmerman, L. E. & Church, G. M. In vitro integration of ribosomal RNA synthesis, ribosome assembly, and translation. Mol. Syst. Biol. 9, https://doi.org/10.1038/msb.2013.31 (2013).
Ahmad, S. et al. A high Throughput Assay for Discovery of Bacterial β-Glucuronidase Inhibitors. Curr. Chem. Genomics 5, 13 (2011).
Article CAS PubMed PubMed Central Google Scholar
Chambers, M. C. et al. A cross-platform toolkit for mass spectrometry and proteomics. Nat Biotechnol 30, 918–920 (2012).

Download references

Acknowledgements

We thank Ms. Kayo Aoyama and Ayu Saito for technical support. This work was supported by JST, CREST (Grant Number JPMJCR20S1 to N.I.), Japan, and Kakenhi (Grant Numbers 22H05402 to N.I., 24H01111 to N.I., and 23KJ0815 to R.M.).

Author information

Authors and Affiliations

Department of Life Science, Graduate School of Arts and Science, The University of Tokyo, Meguro, Tokyo, Japan
Ryota Miyachi & Norikazu Ichihashi
Laboratory for Cell-Free Protein Synthesis, RIKEN Center for Biosystems Dynamics Research (BDR), IIB Bldg. S304, 6-7-1, Minatojima-minamimachi, Chuo-ku, Kobe, Hyogo, Japan
Keiko Masuda & Yoshihiro Shimizu
Komaba Institute for Sciences, The University of Tokyo, Meguro, Tokyo, Japan
Norikazu Ichihashi
Universal Biology Institute, The University of Tokyo, Meguro, Tokyo, Japan
Norikazu Ichihashi

Authors

Ryota Miyachi
View author publications
Search author on:PubMed Google Scholar
Keiko Masuda
View author publications
Search author on:PubMed Google Scholar
Yoshihiro Shimizu
View author publications
Search author on:PubMed Google Scholar
Norikazu Ichihashi
View author publications
Search author on:PubMed Google Scholar

Contributions

R.M., K.M., N.I., and Y.S. designed the study and wrote the manuscript. R.M., K.M., and N.I. performed experiments and analyses.

Corresponding author

Correspondence to Norikazu Ichihashi.

Ethics declarations

Competing interests

The authors declare no competing interest.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Supplementary Data 1

Supplementary Data 2

Supplementary Data 3

Supplementary Data 4

Reporting Summary

Transparent Peer Review file

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Miyachi, R., Masuda, K., Shimizu, Y. et al. Simultaneous in vitro expression of minimal 21 transfer RNAs by tRNA array method. Nat Commun 16, 7418 (2025). https://doi.org/10.1038/s41467-025-62588-y

Download citation

Received: 23 February 2025
Accepted: 25 July 2025
Published: 26 August 2025
DOI: https://doi.org/10.1038/s41467-025-62588-y