Enzymatic Conjugation of Modified RNA Fragments by Ancestral RNA Ligase AncT4_2

ABSTRACT

Oligonucleotide therapeutics have great potential as a next-generation approach to treating intractable diseases. Large quantities of modified DNA/RNA containing xenobiotic nucleic acids (XNAs) must be synthesized before clinical application. In this study, the ancestral RNA ligase AncT4_2 was designed by ancestral sequence reconstruction (ASR) to perform the conjugation reaction of modified RNA fragments. AncT4_2 had superior properties to native RNA ligase 2 from T4 phage (T4Rnl2), including high productivity, a >2.5-fold-higher turnover number, and >10°C higher thermostability. One remarkable point is the broad substrate selectivity of AncT4_2; the activity of AncT4_2 toward 17 of the modified RNA fragments was higher than that of T4Rnl2. The activity was estimated by measuring the conjugation reaction of two RNA strands, 3′-OH (12 bp) and 5′-PO₄ (12 bp), in which the terminal and penultimate positions of the 3′-OH fragment and the first and second positions of the 5′-PO₄ fragment were substituted by 2′-fluoro, 2′-O-methyl, 2′-O-methoxyethyl, and 2′-H, respectively. The enzymatic properties of AncT4_2 allowed the enzyme to conjugate large quantities of double-stranded RNA coding for patisiran (>400 μM level), which was formed by four RNA fragments containing 2′-OMe-substituted nucleic acids. Structural analysis of modeled AncT4_2 suggested that protein dynamics were changed by mutation to Gly or indel during ASR and that this may positively impact the conjugation of modified RNA fragments with the enzyme. AncT4_2 is expected to be a key biocatalyst in synthesizing RNA therapeutics by an enzymatic reaction.

IMPORTANCE RNA therapeutics is one of the next-generation medicines for treating various diseases. Our designed ancestral RNA ligase AncT4_2 exhibited excellent enzymatic properties, such as high thermal stability, productivity, specific activity, and broad substrate selectivity compared to native enzymes. These advantages create the potential for AncT4_2 to be applied in conjugating the modified RNA fragments containing various xenobiotic nucleic acids. In addition, patisiran, a known polyneuropathy therapeutic, could be synthesized from four fragmented oligonucleotides at a preparative scale. Taken together, these findings indicate AncT4_2 could open the door to synthesizing RNA therapeutics by enzymatic reaction at large-scale production.

INTRODUCTION

Oligonucleotide therapeutics, such as antisense oligonucleotides (ASOs) and small interfering RNA (siRNA), are emerging as treatment approaches to genetic diseases (1, 2). For example, nusinersen, a type of ASO that modulates pre-mRNA splicing and affects the production of target proteins, is utilized to treat spinal muscular atrophy (1, 3). Patisiran, a type of siRNA that reduces the production of the transthyretin protein in the liver via the RNA interference (RNAi) pathway, is utilized to treat polyneuropathy in people with hereditary transthyretin-mediated amyloidosis (4). Many biopharmaceutical companies have been developing new therapeutics, and they will be launched in the near future (1). Several technical hurdles must be addressed before these therapeutics are applied in clinical use; challenges include improving stability against nucleases and large-scale synthesis of these therapeutics (5, 6). The former hurdle could be cleared by synthesizing modified oligonucleotides containing xenobiotic nucleic acids (XNAs); introduction of XNAs in which 2′-OH was replaced by fluoro (2′-F), O-methoxyethyl (2′-MOE), or O-methyl (2′-OMe) into oligonucleotides improves nuclease resistance (6). Introduction of XNAs into the therapeutics resulted in other improvements, including increased binding affinity and reduced immunogenicity (6). Simultaneously, the importance of developing a synthetic approach to obtain the modified oligonucleotides is recognized now; this connects to the latter hurdle. The modified oligonucleotides were chemically synthesized by phosphoramidite chemistry with solid- and liquid-phase syntheses (6–9). Solid-phase synthesis has been applied over the past few decades due to the simplified reaction scheme (7, 8, 10). Although solid-phase synthesis rapidly supplies oligonucleotides in high quality, both the cost of synthesis and challenges regarding scale-up are significant barriers. Researchers have investigated unique liquid-phase approaches that are not restricted by these limitations, such as the AJIPHASE method (11).

Although many technical advantages have been achieved through chemical synthesis, the production of large quantities of modified oligonucleotides with long chains remains challenging due to their low coupling efficiency (12). This problem could be solved if the modified oligonucleotide fragments with short chains were conjugated by an enzymatic reaction (12, 13). Nucleic acid ligases, such as DNA ligase (12, 14) and RNA ligase (15–17), are expected to perform the conjugation reaction. The ligases conjugate nicked oligonucleotides utilizing ATP as a cofactor; the oligonucleotides are hybridized to cDNA or RNA (14). DNA ligase conjugates these fragments under the condition that they contain cDNA as the substrate (12, 14, 18).

On the other hand, RNA ligase is suitable for conjugating RNA fragments. RNA ligase 2 from T4 phage (T4Rnl2), the main target in this study, can conjugate the nicked RNA efficiently, utilizing cRNA or DNA as the substrates (18, 19). Structural and functional analyses revealed the reaction mechanism and functionally important residues of T4Rnl2 (20). The application of T4Rnl2 was reported to synthesize oligonucleotides containing XNAs, but only small amounts of the products (on the order of 10⁻¹³mol) were obtained (19). The enzymatic properties of T4Rnl2 must be improved before it can be used to synthesize modified RNA on a large scale.

In this study, utilizing the T4Rnl2 sequence as a template, we attempted to design new RNA ligases with broad substrate selectivity and productivity by sequence-based protein design methods. The enzymes generated by full consensus design (FCD) and ancestral sequence reconstruction (ASR) were called FcT4_2 and AncT4_2, respectively. FCD generates a sequence by mutating all residues of the template to consensus residues, which were the most frequent residues, determined by multiple sequence alignment (MSA) (21–24). ASR reconstructs a progenitor sequence predicted by the analysis of MSA results and phylogenetic tree data (25–27); the evolutionary transition between two proteins bearing distinct physicochemical properties is predicted through biochemical and structural analyses of the resurrected proteins (28, 29). Both methods are currently utilized to design highly functional enzymes in both clinical applications (30) and industrial use, such as the biosynthesis of fine chemicals (31–33). Through biochemical and reaction analyses, we will prove that AncT4_2 is suitable to conjugate two RNA oligonucleotides containing XNAs.

RESULTS AND DISCUSSION

Library curation and design of FcT4_2 and AncT4_2.

The preparation of a sequence library is necessary to design artificial proteins with superior properties to native proteins by sequence-based protein design methods (22, 34, 35). To design highly functional RNA ligases, a curated sequence library was prepared through the following procedure (23, 36, 37). First, a sequence library was prepared by submitting the T4Rnl2 sequence (NCBI reference sequence accession no. NP_049790.1) to blastp by applying the following parameters: the database and E value were set to nonredundant and 1.0E⁻⁶, respectively. The library containing 425 sequences could be obtained after the blastp analysis (Fig. 1A). Next, the library was curated by applying two steps: (i) omission of sequences with >10% longer or shorter sequence length than T4Rnl2 and (ii) deletion of sequences sharing >90% identity with other sequences in the library (Fig. 1A). This resulted in a curated library containing 172 sequences. The curation can be performed using an original Python script, which is available from a GitHub repository (https://github.com/shognakano/Library_curation). The assignment of key residues (A32, F116, S170, and I274) was performed by adopting a procedure identical to that reported in previous studies (23, 36, 38). The classification of the curated library was completed by selecting sequences bearing the key residues; a total of 12 sequences were obtained (Fig. 1A). The amino acid frequency of 11 residues around the key residues is represented in Fig. S1 in the supplemental material, indicating that the sequence logo was clear after the curation. This suggests that conserved residues are accurately assigned by MSA utilizing the classified sequences; in this situation, there is a high probability of reconstructing highly functional proteins by sequence-based protein design methods (34). Using the classified sequences, FcT4_2 and AncT4_2 were designed with FCD and ASR. The protein and DNA sequences of the designed RNA ligases are shown in Table S1.

FIG 1.

Schematic view indicating how to design FcT4_2 and AncT4_2 from T4Rnl2 homolog sequences (A). Phylogenetic tree analysis of T4Rnl1, T4Rnl2, and AncT4_2 and their homolog sequences (B). The homolog sequences are represented in Table S2. The MSA was performed by MAFFT (51), and phylogenetic tree data were obtained by MEGA7 (52). The tree was depicted by iTOL software (54).

Preliminary biochemical experiments with the designed RNA ligases indicated that AncT4_2 could be expressed in a soluble form. Although FcT4_2 was also expressed in a soluble form, the expression level was significantly lower than those of AncT4_2 and T4Rnl2. AncT4_2 shared 92% of its sequence identity with T4Rnl2 (Table S3). The phylogenetic tree analysis also supported this point; they are located near each other (Fig. 1B). MSA results indicated that functionally important residues were conserved between T4Rnl2 and AncT4_2 (red box in Fig. S2). Thus, we performed biochemical and conjugation reaction analyses of AncT4_2.

Biochemical analysis of T4Rnl2 and AncT4_2.

Biochemical analysis of AncT4_2 and T4Rnl2 indicated that the enzymatic properties of AncT4_2 were improved more than those of T4Rnl2. The protein expression level was estimated by SDS-PAGE, indicating that the level of AncT4_2 was higher than that of T4Rnl2; the band size of AncT4_2 was confirmed at 40 kDa for the soluble (lane S in Fig. 2A) and purified (lane P in Fig. 2A) fractions, which are larger than those of T4Rnl2 (Fig. 2A). Because the enzymes are nucleic acid binding proteins, there is a possibility that the purified samples contain DNA/RNA from the expression host. The UV spectra of purified enzymes appeared to be identical between T4Rnl2 and AncT4_2 (A₂₆₀/A₂₈₀ ratios of 0.798 and 0.793, respectively) (Fig 2B). The ratio was slightly higher than the ratio of a standard sample that contains only protein, inferring that there is some nucleic acid contamination of the T4Rnl2 and AncT4_2 samples. However, the ratio was almost identical, so there is no concern about comparing enzymatic properties between T4Rnl2 and AncT4_2 in this study.

FIG 2.

(A) Expression test of T4Rnl2 and AncT4_2 by SDS-PAGE. Lanes M, S, I, and P represent the protein marker, soluble fraction, insoluble fraction, and purified fraction after affinity chromatography. (B) UV spectra of T4Rnl2 and AncT4_2. Abs, absorbance. (C) Enzyme kinetic plots of T4Rnl2 and AncT4_2. The data were fitted to a substrate uncompetitive inhibition model. The fitted line is solid. Error bars represent standard deviation (SD). (D) Thermal stability analysis of T4Rnl2 and AncT4_2. The data for T4Rnl2 and AncT4_2 are plotted as gray and solid black circles, respectively. Error bars denote 1 SD. (E) Conjugation reaction of double-strand RNA (dsRNA) containing nicks in the sense and antisense strands. The dsRNA, represented as a schematic image, was conjugated by T4Rnl2 and AncT4_2. Here, the time-dependent concentration changes of the produced sense (left) and antisense strand (right) are plotted.

Enzyme activity of T4Rnl2 and AncT4_2 was measured by quantifying the products generated by the nick-sealing reaction of double-stranded RNA (dsRNA); the reaction substrates are listed in Table 1. Enzyme kinetic plots of T4Rnl2 and AncT4_2 can be fitted to a classic uncompetitive substrate inhibition model (see equation 1 below) as well as the case of T4 DNA ligase (39) (Fig. 2C). The enzyme kinetic parameters were represented in Table 2, indicating that AncT4_2 had higher activity than T4Rnl2. The turnover number (k_cat) and enzyme efficiency (k_cat/K_m) value of AncT4_2 were >2.5-fold larger than those of T4Rnl2 (Table 2). The inhibition equilibrium constants (K_i) were almost comparable between the enzymes (Table 2). The remarkable point is that AncT4_2 had higher thermal stability than T4Rnl2 despite the fact that improvement of the k_cat value was achieved; the melting temperature (T_m) of AncT4_2 (45°C) was >10°C higher than that of T4Rnl2 (32°C) (Fig. 2D).

TABLE 1.

Reaction substrates (RNA sequences) to measure enzymatic activity

Oligonucleotide	Sequence (5′→3′)a
Complementary fragment	AUUCCGAUAGUGGGGUCGCAAUUGCAU
3′-OH fragment	CAAUUGCGACCC
3′-OH fragment with 6-FAM	FAM-CAAUUGCGACCC
5′-PO4 fragment	PO4-CACUAUCGGAAU

^aPO₄, 5′-phosphate; FAM, 6-carboxyfluorescein.

TABLE 2.

Kinetic parameters of T4Rnl2 and AncT4_2

Parameter	Result fora:
	T4Rnl2	AncT4_2
kcat (s−1)	0.037 ± 0.003	0.10 ± 0.01
Km (μM)	0.98 ± 0.16	1.0 ± 0.2
Ki (μM)	36 ± 9	30 ± 8

^aData are shown as estimated values with standard error.

The biochemical analysis indicated that AncT4_2 has suitable properties for application in the conjugation reaction of nicked RNA. To confirm this, we tried to perform the conjugation reaction of dsRNA, which is formed by four RNA fragments containing 2′-OMe-substituted nucleic acids (13); thus, the dsRNA was formed by sense and antisense strands containing two nicks as shown in the schematic view in Fig. 2E. Here, the conjugation reaction generates patisiran (4), consisting of 21-mer sense and antisense strands. In the RNA fragments, three and six nucleotides in the sense strands and one nucleotide in each antisense strand were replaced by 2′-OMe nucleotides (U_m and C_m in Fig. 2E). High-performance liquid chromatography (HPLC) analysis of the reaction products indicated that the RNA fragments were conjugated to two strands by enzymatic reaction (Fig. S3). Liquid chromatography-time of flight mass spectrometry (LC-TOF MS) analysis of the product represents monoisotopic peaks at m/z of 1,689.2656 and 1,663.2282, which correspond to the calculated value for the monoisotopic peaks of the sense strand (C₂₁₀H₂₆₈N₇₆O₁₄₃P₂₀, m/z = 1,689.27 [M-4H]⁴⁻) and antisense strand (C₂₀₂H₂₅₂N₇₂O₁₄₇P₂₀, m/z = 1663.23 [M-4H]⁴⁻), respectively (Fig. S4). The time-dependent change of product concentration was plotted in Fig. 2E, indicating that AncT4_2 (solid black circles in Fig. 2E) can generate more products than T4Rnl2 (gray circles in Fig. 2E). After 24 h, the concentrations of sense and antisense RNA products reached >400 μM both sense and antisense RNA products for synthesis using AncT4_2. In contrast, about 100 and 300 μM sense and antisense RNA products were synthesized using T4Rnl2.

AncT4_2 exhibits superior properties for conjugating a large amount of fragmented RNA containing XNAs compared with T4Rnl2. Here, the amount of the conjugated RNA was at least 3 orders of magnitude greater than those reported in previous studies utilizing other oligonucleotide ligases (12).

Analysis of substrate selectivity for T4Rnl2 and AncT4_2.

To apply oligonucleotide therapeutics in clinical use, their in vivo stability, such as nuclease resistance, must be improved (1). Therapeutics containing XNAs have potential but require the development of oligonucleotide ligases to conjugate RNA and DNA fragments. Many research groups reported that, compared with native enzymes, ASR could design artificial enzymes bearing broad substrate selectivity (40, 41), which is required for AncT4_2.

To confirm the substrate selectivity of T4Rnl2 and AncT4_2, we estimated their enzyme activities by measuring the conjugation of nicked RNA fragments represented in Fig. 3A. The 3′-OH and 5′-PO₄ RNA strands with the nucleotides at each of the −2, −1, +1, and +2 positions substituted by 2′-F, 2′-OMe, 2′-MOE, and 2′-H, respectively, were utilized as the substrates (Fig. 3A). HPLC analysis of the reaction products indicated that the 3′-OH and 5′-PO₄ RNA strands were conjugated to one strand by enzymatic reaction (Fig. S5), and LC-TOF MS analysis of the products showed a monoisotopic peak at the expected m/z (Fig. S6). The relative activities of T4Rnl2 (gray bars) and AncT4_2 (black bars) are represented in Fig. 3B. The specific activity of T4Rnl2, measured using unmodified RNA fragments as the substrates, was set to 100% (gray bar labeled “Control” in Fig. 3B). The activity of AncT4_2 toward the unmodified fragments was about 2-fold higher than that of T4Rnl2 (black bar labeled “Control” in Fig. 3B). The activities of T4Rnl2 and AncT4_2 toward the modified 5′-PO₄ fragment reached about 100% and 200% those of the control value, respectively (positions +1 and +2 versus the “Control” in Fig. 3B), suggesting that the modification of 5′-PO₄ had minimal impact on the activity. The low specificities toward the 5′-PO₄ fragment were observed in the previous study, in which T4Rnl2 demonstrated equivalent activities toward RNA and DNA (18, 19). On the other hand, the activities toward the modified 3′-OH fragment were different between T4Rnl2 and AncT4_2, with T4Rnl2 exhibiting more stringent substrate specificity than AncT4_2. The activity toward the 3′-OH fragment, of which the penultimate position was substituted by 2′-F, 2′-MOE, and 2′-H, was less than 30% of the activity of the control (gray bar at position −2 in Fig. 3B). The activity of T4Rnl2 was significantly weakened after inserting a 3′-OH fragment with a terminal position modified by XNAs; no activity toward 2′-MOE could be confirmed (gray bar at position −1 in Fig. 3B). The low activities toward the 3′-OH fragment with modifications at positions −1 and −2 were consistent with previous results (19). On the other hand, the activity of AncT4_2 toward the modified 3′-OH fragment was higher than those of the T4Rnl2 (black bars at positions −2 and −1 in Fig. 3B). In particular, AncT4_2 conjugated the modified 3′-OH fragment with the terminal position substituted by 2′-MOE (black bar at position −1 in Fig. 3B); for T4Rnl2, no activity was confirmed toward the corresponding fragments.

FIG 3.

(A) Oligonucleotide substrates to estimate substrate selectivity of T4Rnl2 and AncT4_2. Two oligonucleotides (3′-OH [blue] and 5′-PO₄ [orange] fragments) formed by 12 bp and cRNA were utilized as the substrates; with XNAs, the −2 and −1 positions of the 3′-OH fragment and +1 and +2 positions of the 5′-PO4 fragment were substituted with 2′-OH (control), 2′-F, 2′-OMe, 2′-MOE, and 2′-H, respectively. (B) Relative activity of T4Rnl2 and AncT4_2 toward oligonucleotides containing XNAs. The specific activity of AncT4_2 toward RNA (control condition) is normalized as 100%. Error bars denote 1 SD. (C) Overall model structure of the AncT4_2 RNA binding form. The 22 mutated residues are represented as red spheres.

This demonstrated that AncT4_2 exhibits broad substrate selectivity toward the modified 3′-OH fragment at the penultimate and terminal positions compared with T4Rnl2 (Fig. 3B). The enzymatic properties of AncT4_2 are suitable for conjugating various modified oligonucleotide fragments.

Structural analysis of AncT4_2.

AncT4_2 has suitable properties for high conjugate quantities of modified RNA fragments. Structural analysis of AncT4_2 may provide insight into what gives the enzyme these desirable properties at the molecular level. The overall structure for the RNA binding form of AncT4_2 was modeled by the Swiss-Prot web server (42) (Fig. 3C). Here, AncT4_2 and T4Rnl2 shared >90% sequence identity; we anticipated that the AncT4_2 model structure could be predicted accurately by homology modeling. Structural comparison at the active site between T4Rnl2 and AncT4_2 indicated that the active site structures are highly conserved, suggesting that AncT4_2 would conserve the reaction mechanism.

Sequence and structural analyses indicated that almost all of the mutations were located remote from the active site (red spheres on the right side of Fig. 3C). The remote mutations may remodel the active site structure of AncT4_2, as well as other enzymes, to recognize the modified RNA fragments (43). In addition, these mutations may affect protein dynamics that play a vital role in the catalytic reaction of enzymes, including substrate recognition, transition state stabilization, and product release (44–46). Protein dynamics between AncT4_2 and T4Rnl2 would vary, as many of the mutations that affect the dynamics were introduced by ASR. Totals of four residues and two residues in T4Rnl2 were mutated to Gly and an indel in AncT4_2, respectively, and it was reported that these mutations would increase protein flexibility (47) and have a potent effect on the dynamics (48). Another possibility that, like the case of T4 DNA ligase (39), lowering substrate or product inhibition enhances the activity of AncT4_2 can be denied from the comparison of K_i values: the values were comparable between T4Rnl2 (36 μM in Table 2) and AncT4_2 (30 μM in Table 2). The remote mutations may stabilize the transition state of AncT4_2 and increase the k_cat value of T4Rnl2 from the model of enzyme reaction (49). Rational design of the remote mutations providing positive effects remains a challenge (24, 44), and the successful design of AncT4_2 suggests that ASR is influential in introducing the remote mutations exhibiting positive effects into target proteins.

Conclusion.

Recent advances in biotechnology enable us to apply various enzymes to industrial processes. Simultaneously, highly functional enzymes must be designed by protein engineering methods because the low stability and narrower substrate selectivity of native enzymes are unsuitable for these applications. Sequence-based protein design methods, such as FCD and ASR, have the potential to reconstruct new enzymes utilizing protein sequence data registered in a rapidly expanding sequence database. In this study, we demonstrated that AncT4_2 could be reconstructed by ASR using a total of 12 classified sequences; the enzyme had desirable properties, such as high productivity, thermostability, and broad substrate selectivity, that would enable its application in the enzymatic conjugation of large quantities of RNA fragments (on the order of 10⁻⁴ M). AncT4_2 can potentially contribute to synthesizing siRNA and ASOs from oligonucleotide fragments synthesized by chemical reaction. The next challenge is to design XNA ligases that can conjugate two oligonucleotide fragments formed by XNA. Based on the results of molecular dynamics (MD) and docking simulation, XNA ligase was successfully designed from Chlorella virus DNA ligase by inserting Gly at position 189 (50), suggesting that combinations of structural and ASR design may be effective in generating new ligase. This use of ASR as an enzyme engineering strategy illustrates one example of its potential impact on industrial processes.

MATERIALS AND METHODS

Reagents.

HPLC-purified oligonucleotides were purchased from GeneDesign as lyophilized solids; the oligonucleotides were dissolved in RNase-free water (Nacalai Tesque) prior to starting the enzyme reaction. T4 RNA ligase 2 reaction buffer (50 mM Tris-HCl [pH 7.5], 2 mM MgCl₂, 1 mM dithiothreitol [DTT], 400 μM ATP [pH 7.5]) (New England Biolabs) was used as the reaction buffer to estimate enzyme activities of T4Rnl2 and AncT4_2. The ATP concentration was changed to 1.4 mM when we performed the conjugation reaction of double-strand RNA (dsRNA) containing nicks in the sense and antisense strands to complete the conjugation reaction.

Reconstruction of ancestral RNA ligase (AncT4_2).

Application of the classification method described in Fig. 1A yielded a total of 12 sequences that were aligned by the multiple-sequence alignment (MSA) software MAFFT (51). Utilizing the aligned sequences, a phylogenetic tree was generated by MEGA7 software based on the maximum likelihood method (52). The sequences of ancestral RNA ligases were obtained by submitting MSA results and phylogenetic tree data to the FastML web server (53); here, a JTT empirical model was adopted for the reconstruction. The progenitor sequence of the 12 sequences was called AncT4_2 as shown in the phylogenetic tree (Fig. 1B) generated by iTOL software (54). Protein and DNA sequences of AncT4_2 are shown in Table S1 in the supplemental material.

Overexpression and purification of AncT4_2.

Genes coding for T4Rnl2, FcT4_2, and AncT4_2 (Table S1) were synthesized with optimization of codon usage for Escherichia coli (Eurofins Genomics) and inserted at the NdeI and BamHI sites of pET-16b (Merck Millipore). Both constructs contain an N-terminal His₁₀ tag derived from the cloning vector. The plasmids were transformed into the E. coli BL21(DE3) strain (Nippon Gene). Each transformant was cultured in a Luria-Bertani (LB) medium containing 100 mg/L of ampicillin for 16 h at 37°C, and 2 mL of the broth was then inoculated into 150 mL of LB medium containing 100 mg/L of ampicillin in 500-mL flasks and cultured aerobically on a reciprocal shaker at 37°C. When the optical density at 600 nm (OD₆₀₀) reached 0.5, IPTG (isopropyl-β-d-thiogalactopyranoside) was added to reach a final concentration of 0.1 mM, and the cell culture continued for 3 h.

Cells were harvested by centrifugation at 10,000 × g for 10 min at 4°C and suspended in 10 mL of buffer A (50 mM Tris-HCl [pH 7.5], 0.25 M NaCl, 10% [wt/vol] sucrose). Lysozyme and Triton X-100 were added to the suspended solutions at final concentrations of 50 μg/mL and 0.1% (vol/vol), respectively, and the solutions were incubated for 30 min on ice. Following sonication, the supernatant was collected by centrifugation at 10,000 × g for 10 min at 4°C. The supernatant was applied to an immobilized metal affinity chromatography column (HisTALON Superflow cartridge, 5 mL; TaKaRa Bio), and the column was washed with 30 mL of buffer A. The sample was eluted with 10 mL buffer B (50 mM Tris-HCl [pH 8.0], 0.25 M NaCl, 10% [vol/vol] glycerol) containing 0.05 to 0.5 M imidazole. Eluted fractions containing RNA ligase were concentrated to 50 μL using an Amicon Ultra-4 10-kDa centrifugal filter unit (Merck Millipore). Purified enzymes were stored at −20°C in buffer C [10 mM Tris-HCl (pH 7.5), 50 mM KCl, 35 mM (NH₄)₂SO₄, 0.1 mM DTT, 0.1 mM EDTA, 50% (vol/vol) glycerol] until subsequent biochemical assays. Enzyme purity was analyzed by SDS-PAGE. The UV spectra of the purified samples were measured by a UV-1800 spectrophotometer (Shimadzu); the enzyme concentration was set to 4 mg/mL. Enzyme concentrations were determined by the Bradford method, using a protein assay Coomassie brilliant blue (CBB) solution (Nacalai Tesque; code 29449-44) with bovine serum albumin as a standard.

Thermal stability assay.

The specific activities of T4Rnl2 and AncT4_2 were measured by monitoring the production of the dsRNA under the condition containing the 3′-OH fragment (12-mer) and 5′-PO4 fragment (12-mer) as the substrates and the complementary fragment (27-mer) (Table 1). The concentration of the conjugated substrates (24-mer) was estimated by HPLC. The thermal stability assay was performed as follows. Before starting the enzymatic reaction, 9 μL of reaction mixture containing 13 μg/mL of T4Rnl2 and 3.2 μg/mL of AncT4_2 without the substrates was incubated at temperatures ranging from 4°C to 70°C for 60 min using ProFlex PCR system (Thermo Fisher Scientific) with the heated lid to prevent the evaporation. The contents of the reaction mixture were as follows: 50 mM Tris-HCl (pH 7.5), 2 mM MgCl₂, 1 mM DTT, 400 μM ATP, 10 μM RNA fragments, and the complementary fragment. The reaction was started by adding 1 μL of substrate oligonucleotide solution and incubating at 25°C for 15 min. The product concentration was estimated after quenching the reaction mixture by adding EDTA. Residual enzyme activity was calculated as a percentage of enzyme activity, where enzymes with incubation at 4°C served as a positive control for the 100% activity level. The reactions were repeated three times.

Estimation of kinetic parameters.

To estimate enzyme kinetic parameters, we prepared the following RNA fragments; the 5′-PO₄ fragment, the 3′-OH fragment, of which the 5′-terminal end was modified with 6-carboxyfluorescein (FAM), and the complementary fragment (Table 1). The components of the reaction buffer are as follows: 50 mM Tris-HCl (pH 7.5), 2 mM MgCl₂, 1 mM DTT, 400 μM ATP, and 30 nM to 20 μM concentration of each RNA fragments. The reaction was started by adding diluted enzyme solution into the reaction buffer. After 3 min of the reaction at 25°C, the reaction was quenched by adding EDTA. After the concentration of the reaction product was determined by HPLC, the initial velocities were estimated at different concentrations of the substrates. The reactions were repeated three times.

To estimate the kinetic parameters, the initial velocity was fitted to a substrate inhibition model as shown in equation 1:

$$\frac{V_0}{[\mathrm{E}]}\hspace{0.17em}=\hspace{0.17em}\frac{k_{\text{cat}}\times [\mathrm{S}]}{K_m+\left[\mathrm{S}\right]\times (1+\frac{\left[\mathrm{S}\right]}{K_i})}$$ (1)

[S] is the substrate concentration, [E] is the enzyme concentration, V₀ is the initial rate, k_cat is the turnover number, K_m is the Michaelis constant for substrate binding, and K_i is the inhibitor binding constant. The fitting was performed by adopting a nonlinear least-squares model implemented with R statistical software (version 4.0.2; the R Foundation for Statistical Computing, version 4.0.2).

Production of double-stranded RNA from four fragments.

The sequences of four sense or antisense strands are represented in Table 3. The components of the reaction mixture were as follows: 50 mM Tris-HCl (pH 7.5), 2 mM MgCl₂, 1 mM DTT, 1.4 mM ATP, and 500 μM four the concentration for each RNA strands. The reaction was started by adding a final concentration of 7.2 μg/mL RNA ligase into the reaction mixture at 25°C. Reaction mixtures were collected at time intervals of 0.5, 1, 2, 4, 6, and 24 h to evaluate long-term stability of the enzymes. The stability is required to apply the enzymes in practical usage. After quenching the mixtures immediately by adding EDTA, the product concentration was analyzed by HPLC.

TABLE 3.

RNA sequences to produce dsRNA from four fragments

Oligonucleotide	Sequence (5′→3′)a
Sense strand
3′-OH fragment	G-Um-A-A-Cm-Cm-A-A-G-A-G
5′-PO4 fragment	PO4-Um-A-Um-Um-Cm-Cm-A-Um-dT-dT
Antisense strand
3′-OH fragment	A-U-G-G-A-A-Um-A-C-U-C-U
5′-PO4 fragment	PO4-U-G-G-U-Um-A-C-dT-dT

^aPO₄, 5′-phosphate; m in U_m and C_m, 2′-OMe RNA; dT, thymidine.

Comparison of substrate selectivity.

Substrate selectivity was determined by nick sealing reaction using the modified RNA fragments as the substrates. The contents of the modified fragments are as follows: the −2 and −1 positions of the 3′-OH fragment and +1 and +2 positions of the 5′-PO₄ fragments were substituted by XNAs, including 2′-OH (control), 2′-F, 2′-OMe, 2′-MOE, and 2′-H, respectively. The components of the reaction solution were as follows: 50 mM Tris-HCl (pH 7.5), 2 mM MgCl₂, 1 mM DTT, 400 μM ATP, 10 μM substrates, and 10 μM complementary strand. After the diluted T4Rnl2 and AncT4_2 were added, the reaction progressed for 10 min at 25°C. The specific activity was estimated by following identical procedures adopted to determine the activities of T4Rnl2 and AncT4_2. The relative activities of T4Rnl2 (gray bars in Fig. 3B) and AncT4_2 (black bars in Fig. 3B) toward various RNA fragments were calculated by normalizing the specific activity of T4Rnl2 toward RNA to be 100%.

Analysis of reaction products by HPLC and LC-TOF MS.

The reaction mixtures to estimate enzyme kinetic parameters were analyzed by HPLC (GL7700 HPLC system) equipped with an XBridge oligonucleotide BEH C₁₈ column (2.5 μm, 4.6 by 50 mm) (Waters) and FL7753 fluorescence detector (GL Science). The columns were kept at 60°C conditions. The volume of the injected samples was 10 μL. The following two solutions were utilized in the mobile phase: 8 mM triethylamine in 100 mM hexafluoroisopropanol (solution A) and 100% (vol/vol) methanol (solution B). The flow rate was set to 0.4 mL/min, and the products were eluted by applying the following conditions: (i) a linear gradient of 0 to 20% solution B for 18 min and (ii) 72% solution B for 3 min. The product peaks were detected by measuring fluorescence with excitation at 495 nm and emission at 520 nm.

The other substrate and conjugated oligonucleotides were analyzed by HPLC using an ACQUITY ultraperformance liquid chromatography (UPLC) H-class system with TUV detectors (Waters) and ACQUITY UPLC oligonucleotide BEH C₁₈ column (1.7 μm, 2.1 by 50 mm) (Waters) with a linear gradient of 5 to 25% of 8 mM triethylamine in 100 mM hexafluoroisopropanol and methanol at 60°C, a flow rate of 0.4 mL/min, and an injection volume of 10 μL. The chromatography was monitored at 260 nm. The standard was also analyzed, and the concentration of the conjugation product was calculated based on the peak area of the product. The molecular weight of the conjugated oligonucleotides was analyzed by LC-TOF MS using an Agilent 6230 TOF LC/MS system (Agilent Technologies) with an oligonucleotide BEH C₁₈ column (1.7 μm, 2.1 by 150 mm; Waters).

Data availability.

The amino acid sequence of AncT4_2 and nucleotide sequence for AncT4_2 expression were deposited in DDBJ/ENA/GenBank under accession no. BDQ64000 and LC726495.