Identification of 2-methylthio cyclic N⁶-threonylcarbamoyladenosine (ms²ct⁶A) as a novel RNA modification at position 37 of tRNAs

Abstract

Transfer RNA modifications play pivotal roles in protein synthesis. N⁶-threonylcarbamoyladenosine (t⁶A) and its derivatives are modifications found at position 37, 3΄-adjacent to the anticodon, in tRNAs responsible for ANN codons. These modifications are universally conserved in all domains of life. t⁶A and its derivatives have pleiotropic functions in protein synthesis including aminoacylation, decoding and translocation. We previously discovered a cyclic form of t⁶A (ct⁶A) as a chemically labile derivative of t⁶A in tRNAs from bacteria, fungi, plants and protists. Here, we report 2-methylthio cyclic t⁶A (ms²ct⁶A), a novel derivative of ct⁶A found in tRNAs from Bacillus subtilis, plants and Trypanosoma brucei. In B. subtilis and T. brucei, ms²ct⁶A disappeared and remained to be ms²t⁶A and ct⁶A by depletion of tcdA and mtaB homologs, respectively, demonstrating that TcdA and MtaB are responsible for biogenesis of ms²ct⁶A.

INTRODUCTION

RNA modifications are a type of qualitative information embedded in RNA molecules (1). To date, about 140 species of modified nucleosides have been identified in various RNAs from all domains of life (2). tRNAs contain a number of chemical modifications that are required for accurate translation of the genetic code and stabilization of the tRNA tertiary structure (3–5). In particular, a wide variety of modifications is present in the anticodon loop, especially at the first position of the anticodon (position 34) and position 37, which is 3΄-adjacent to the anticodon. These modifications play critical roles in modulating codon recognition and ensuring accurate translation (6).

N ⁶-threonylcarbamoyladenosine (t⁶A) (Supplementary Data) and its derivatives are evolutionarily conserved essential modified bases at position 37 of tRNAs responsible for recognition of adenosine-starting codons (ANN codons) (7). The bulky side chain of t⁶A stabilizes the anticodon loop, promoting accurate decoding of ANN codons during protein synthesis (8,9). In addition, t⁶A is required for efficient aminoacylation of tRNA and efficient translocation, and it also prevents leaky scanning of initiation codons and read-through of stop codons. The biogenesis of t⁶A has been studied extensively. In E. coli, four enzymes, TsaC (YrdC), TsaD (YgjD), TsaB (YeaZ), and TsaE (YjeE), are required to synthesize t⁶A; L-threonine, adenosine triphosphate (ATP) and bicarbonate as substrates (8). In yeast, the YrdC homolog Sua5 and several components of the EKC-KEOPS complex, including Kae1, Pcc1, Gon7 and Bud32, are involved in t⁶A formation (10–12).

The presence of t⁶A in tRNAs from E. coli and yeast was first documented more than four decades ago (13). In 2013, however, we showed that most fraction of t⁶A in E. coli tRNAs is a hydrolyzed artifact of cyclic t⁶A (ct⁶A) (Supplementary Data) (9). ct⁶A is an additional modification of t⁶A that enhances tRNA decoding activity. ct⁶A is present in tRNAs from certain groups of bacteria, fungi, plants and some protists, but not in tRNAs of mammals, archaea or other bacteria. In E. coli, little t⁶A is present in tRNAs because almost all t⁶A is converted to ct⁶A via an ATP-dependent dehydration catalyzed by TcdA. We also identified two additional factors, CsdA and CsdE, which are required for efficient ct⁶A formation. CsdA is a cysteine desulfurase, and CsdE is a sulfur acceptor protein, implying that sulfur relay is involved in efficient formation of ct⁶A.

Initially, mass spectrometric and nuclear magnetic resonance (NMR) analyses determined the chemical structure of ct⁶A to be a cyclized active ester of the oxazolone ring (Supplementary Data) (9,14). Very recently, however, X-ray crystallography of synthetic ct⁶A nucleoside revealed the existence of a distinct isoform with a hydantoin structure (Supplementary Data) (15). LC/MS co-injection analyses showed that chemically synthesized ct⁶A nucleoside and natural ct⁶A in Escherichia coli tRNAs co-elute as a single peak by both reverse-phase and hydrophilic interaction liquid chromatography. They also exhibit identical patterns of product ions in collision-induced dissociation (CID) and a characteristic UV spectrum with maximum absorption at 269 nm. These observations strongly suggest that the hydantoin isoform of ct⁶A is actually present in E. coli tRNAs.

Another derivative of t⁶A, N⁶-methyl-N⁶-threonylcarbamoyladenosine (m⁶t⁶A) (Supplementary Data), is present in tRNAs from bacteria, fly, plants and mammals. In E. coli, m⁶t⁶A is present at position 37 of two species of tRNA^Thr responsible for translating ACY codons. We identified TrmO, a member of a novel class of AdoMet-dependent RNA methyltransferase, as the enzyme responsible for N⁶ methylation of m⁶t⁶A37 of bacterial tRNA^Thr (16). Its human homolog, TRMO, is responsible for formation of m⁶t⁶A37 in cytoplasmic tRNA^Ser. Lack of TrmO decreases attenuation activity of the thr operon, indicating that N⁶ methylation of m⁶t⁶A37 ensures efficient decoding of ACY.

2-Methylthio-N⁶-threonylcarbamoyladenosine (ms²t⁶A) (Figure 1A) is another derivative of t⁶A found in tRNA^Lys from Bacillus subtilis (Figure 1B), Trypanosoma (Figure 1B), plants and mammals (17–20). The methylthiolation of ms²t⁶A is required for the accurate decoding of Lys codons. B. subtilis MtaB (21,22) and human Cdkal1 (23) are the methylthiotransferases responsible for ms²t⁶A formation in tRNAs. Mutations in CDKAL1 are associated with risk of type 2 diabetes (24); consistent with this, pancreatic β-cell–specific knockout of mouse Cdkal1 results in a phenotype similar to that of type 2 diabetes (23). The presence of ct⁶A and TcdA homologs in B. subtilis, plants and Trypanosoma prompted us to speculate that the cyclic form of ms²t⁶A (ms²ct⁶A) (Figure 1A) would also be present in these organisms. As with ct⁶A, ms²ct⁶A must be hydrolyzed and converted to ms²t⁶A during tRNA preparation or conventional nucleoside analysis, which explains why it has not been previously detected. Indeed, a Tris-adduct of ms²t⁶A was detected in tRNA^Lys (Figure 1B) from Trypanosoma brucei (19), indicating the presence of ms²ct⁶A; primary amines easily react with ct⁶A to form amine adducts of t⁶A (25).

Figure 1.

Detection of 2-methylthio-cyclic-N⁶-threonylcarbamoyladenosine (ms²ct⁶A). (A) The chemical structure of ms²t⁶A (left) and ms²ct⁶A (right). The 2-methylthio modification is shown in red, and the N⁶-threonylcarbamoyl group and its hydantoin form are shown in blue. (B) Secondary structures of B. subtilis tRNA^Lys (left) and T. brucei tRNA^Lys (right) with post-transcriptional modifications: dihydrouridine (D), 5-carboxymethylaminomethyl-2-thiouridine (cmnm⁵s²U), 2-methylthio cyclic N⁶-threonylcarbamoyladenosine (ms²ct⁶A), pseudouridine (Ψ), 7-methylguanosine (m⁷G), 5-methyluridine (m⁵U), N²-methylguanosine (m²G), 3-(3-amino-3-carboxypropyl)uridine (acp³U), N², N²-dimethylguanosine (m^2,2G), 2΄-O-methylcytidine (Cm), 5-methoxycarbonylmethyl-2-thiouridine (mcm⁵s²U), 3-(3-amino-3-carboxypropyl)dihydrouridine (acp³D), 5-methylcytidine (m⁵C). The position numbers of the residues are displayed according to the nucleotide numbering system (44). Pairs of gray triangles in B. subtilis tRNA^Lys indicate the positions of cleavage by RNase T₁ that generate RNA fragments containing the anticodon region.

Here, we describe ms²ct⁶A as a novel derivative of ct⁶A in tRNAs from B. subtilis, plants, and T. brucei. The chemical structure of ms²ct⁶A in natural tRNA was confirmed by LC/MS co-injection with the chemically synthesized authentic nucleoside. We also confirmed that orthologs of TcdA and MtaB are responsible for biogenesis of ms²ct⁶A in B. subtilis and T. brucei. In addition, we observed slower growth of T. brucei when the TcdA ortholog was down-regulated in the presence of cycloheximide, indicating that cyclic form of t⁶A is involved in integrity of protein synthesis.

MATERIALS AND METHODS

Bacterial strains and cultivation

B. subtilis str. 168 (WT) was kindly provided by Akiko Soma (Chiba University). B. subtilis strains were grown overnight in LB medium at 37°C. The ΔyqeV strain harboring the erythromycin resistance marker (Em^r) was obtained from the National BioResource Project (National Institute of Genetics, Japan). The ΔyrvM strain was constructed from the WT strain, and the ΔyqeV/ΔyrvM double-deletion strain was constructed from the ΔyqeV strain by homologous recombination (26,27). The 5΄ upstream and 3΄ downstream regions (700–800 nt) of the yrvM gene were PCR-amplified from B. subtilis str. 168 genomic DNA with pairs of primers, 5΄-ggcacctattctgtatccattgatg-3΄ and 5΄-aagcgcagctcttgatgatcaaggctgtttttttg-3΄ and 5΄-tgaggatgaaggctgatccatgagcagccg-3΄ and 5΄-gagcatgatccggaagaaggc -3΄, respectively. Chloramphenicol resistant gene (Cm^r) gene was PCR-amplified from pCBB31 (28) using a set of primers, 5΄-gatcatcaagagctgcgcttttttgtgtc-3΄ and 5΄-tggatcagccttcatcctcatattataaaagccag-3΄. These 3 products were ligated by PCR using primers, 5΄-ggcacctattctgtatccattgatg-3΄ and 5΄-gagcatgatccggaagaaggc-3΄, and subjected to the nested PCR amplification using a pair of primers, 5΄-tccatgatgatcaggcgatgga-3΄ and 5΄-ggaagaaggccgtttttacgca-3΄. The resultant PCR fragment was used for transformation.

Total RNA extraction

B. subtilis cells were suspended with 5 ml of RNA extraction buffer [50 mM NaOAc (pH 5.2) and 10 mM Mg(OAc)₂ (pH 5.2)] and vigorously stirred for 10 min at room temperature. Next, 5 ml of water-saturated phenol was added and stirred for 10 min at room temperature. The mixture was frozen with liquid nitrogen and thawed in water; this process was repeated twice. The thawed solution was stirred for 50 min at room temperature. The aqueous phase was separated by centrifugation and washed once with chloroform, followed by re-extraction with 0.75 volumes of Trizol-LS (Life Technologies). Then, total RNA was precipitated with 2-propanol. The RNA pellet was dissolved in deionized water and subjected to ethanol precipitation; the resultant pellet was rinsed with 80% ethanol and dried. Thus, prepared RNA can be stored in pellet form for a long period of time without hydrolysis of ct⁶A and ms²ct⁶A. For use in all experiments, pellets were dissolved in ultrapure water.

Total RNA of spinach and Arabidopsis thaliana were extracted from plants as previously described (9). Nicotiana tabacum total RNA was obtained by the same procedure from tobacco BY-2 cells cultured for 1 week in modified Linsmaier and Skoog medium (29). Total RNA samples of spinach and tobacco were subjected to brief purification by weak anion exchange chromatography with DEAE Sepharose Fast Flow (GE Healthcare) to remove bulk contaminants and rRNA, as described (30).

Isolation of B. subtilis tRNA^Lys

B. subtilis tRNA^Lys was isolated from B. subtilis total RNA by reciprocal circulating chromatography, as described previously (9,31). The DNA probe, TGGTGAGCCATGAAGGACTCGAACCTTCGA with 5΄-terminal EC amino linker was covalently immobilized on NHS-activated Sepharose 4 Fast Flow (GE Healthcare). Fifty one micrograms of highly purified tRNA^Lys was obtained from 1.7 mg total RNA.

Nucleoside preparation

Before digestion, total RNA was pre-cleared by gel filtration on a Centri-Sep spin column (Princeton Separations) with deionized water or trimethylamine (TMA)-HCl (pH 7.0) buffer to remove contaminants that could interfere with the ionization efficiency of nucleosides. For this experiment, enzymes including nuclease P1 (Wako Pure Chemical Industries), phosphodiesterase I (PDase I, Worthington Biochemical Corporation) and bacterial alkaline phosphatase (BAP from E. coli C75, Wako Pure Chemical Industries) were dialyzed with deionized water and stored at −30°C until use. Phosphodiesterase II (PDase II, from bovine spleen, Sigma) was dissolved in 10 mM TMA-AcOH (pH 5.3) buffer and centrifuged. The supernatant was filtered through a 0.22 μm Ultrafree-MC unit (Merck-Millipore) and stored at −30°C.

For conventional digestion (32), 40 μg of total RNA was digested at 37°C for 1 h in a 25–50 μl reaction mixture consisting of 0.1 U nuclease P1 and 25 mM NH₄OAc (pH 5.3), followed by addition of 0.1 volume of 1 M ammonium bicarbonate (pH 8.2) and 0.08 U BAP, and then incubated at 37°C for 3 h.

For neutral digestion (9), 40 μg of total RNA was digested at 37°C for 1 h in a 25–50 μl reaction mixture consisting of 0.1 U nuclease P1 and 25 mM NH₄OAc (pH 5.3), followed by addition of 0.1 volume of TMA-HCl (pH 7.0) and 0.127 U PDase I, and then incubated at 37°C for 1 h. The prepared nucleotides were dephosphorylated with 0.08 U BAP at 37°C for 3 h at neutral pH.

Digestion of plant and several other RNAs was carried out by one-step acidic digestion. Specifically, a solution (typically 40 μl) containing 1 μg/μl total RNA, 20 mM TMA acetate (pH 5.3), nuclease P1 (0.1 units for 40 μg of RNA), PDase II (0.1 units for 40 μg of RNA) and BAP (0.16 units for 40 μg of RNA) was incubated at 37°C for 1 h. In this procedure, PDase II was used for complete digestion of hypermodified adenosines under acidic conditions. We observed conversion of adenosine and N⁶-methyladenosine (m⁶A) to inosine, indicating contamination by adenosine deaminase activity in the Sigma PDase II product (P9041). t⁶A derivatives remained intact under these conditions.

LC/MS analyses of total nucleosides and isolated tRNA

LC/MS analyses of total nucleosides were performed essentially as described previously (9,33–34), using an LCQ Advantage ion-trap (IT) mass spectrometer (Thermo Fisher Scientific) equipped with an ESI source and an HP1100 liquid chromatography system (Agilent Technologies) or a Q Exactive hybrid Quadrupole-Orbitrap mass spectrometer (Thermo Fisher Scientific) equipped with an ESI source and an Ultimate 3000 liquid chromatography system (Dionex).

For RPC/ESI-MS with an LCQ Advantage instrument, nucleosides were separated on an Inertsil ODS-3 column (2.1 mm × 250 mm, GL sciences) and analyzed as described previously (34). For RPC/ESI-MS with a Q Exactive instrument, digests were separated on a Sunshell C18 column (2.6 μm core-shell silica particle, 2.1 × 150 mm, ChromaNik Technologies). The mobile phase consisted of 5 mM ammonium acetate (pH 5.3) (solvent A) and acetonitrile (ACN) (solvent B). The gradient program was as follows: 0–40% B from 0 to 30 min, 40% B for 5 min and then 0% B at a flow rate of 75 μl/min. Nucleoside digest (8–12 μg) or synthetic ms²ct⁶A (100–500 pmol) dissolved in LC/MS grade ultrapure water (Wako) was injected.

For HILIC/ESI-MS, a ZIC-cHILIC column (3 μm particle size, 2.1 × 150 mm, Merck-Millipore) was used on a Q Exactive instrument (33). The mobile phase consisted of 5 mM ammonium acetate (pH 5.3) (solvent A) and ACN (solvent B). Total nucleosides (8–12 μg) or synthetic ms²ct⁶A (450–500 fmol) dissolved in 90% ACN was injected and chromatographed at a flow rate of 100 μl/min in a multistep linear gradient: 90–40% B from 0 to 30 min, 40% B for 10 min and then 0% B. Proton adducts of nucleosides were scanned in a positive polarity mode over a range of m/z 110–700 or 110–900.

For RNA fragment analysis of isolated tRNA, B. subtilis tRNA^Lys was digested by RNase T₁, followed by subjected to capillary liquid chromatography (LC) coupled to nano electrospray (ESI)/mass spectrometry (MS) on a linear ion trap-Orbitrap hybrid mass spectrometer (LTQ Orbitrap XL; Thermo Fisher Scientific) as described (9,34).

Chemical synthesis of ms²ct⁶A

The substrate nucleoside, 2-methylthio-N⁶-threonylcarbamoyladenosine (ms²t⁶A) was synthesized according to the procedures described previously (35–38). Cyclization of ms²t⁶A to form ms²ct⁶A was performed on the basis of carbodiimide chemistry. ms²t⁶A (10 mg, 0.022 mmol) was dissolved in anhydrous DMF (1 ml) and mixed with EDC·HCl (42 mg, 0.22 mmol). The reaction mixture was stirred at room temperature. After 3 h, consumption of all substrate was confirmed by TLC analysis (nBuOH/H₂O, 85/15, v/v, R_f values of ms²t⁶A and ms²ct⁶A are 0.15 and 0.43, respectively). The solvent was removed under reduced pressure, and the crude product was purified by reverse-phase chromatography (Ascentis C18 HPLC Column, 10 μm, 21.2 × 250 mm) at a flow rate of 7 ml/min with a linear gradient of acetonitrile in 0.1% acetic acid (B) and water (A) as follows: 2–30% B from 0 to 40 min, 30–50% from 40 to 45 min, 50–2% B from 45 to 47 min, 2% B for 3 min. The ms²ct⁶A fraction (22.35 min) was collected and evaporated to dryness. Yield of ms²ct⁶A nucleoside was 44% (4.2 mg).

The purity of ms²ct⁶A was checked by high performance liquid chromatography (HPLC) analysis using an XTerra^® Waters column (MS C8, 5 μm, 4.6 × 150 mm, 100 Å) (Supplementary Data). The mobile phase consisted of 5 mM sodium acetate (pH 7) in water (solvent A) and ACN (solvent B). Chemically synthesized ms²t⁶A (A) and ms²ct⁶A (B) were chromatographed at a flow rate of 1 ml/min with a dual-step linear gradient: 0–20% B from 0 to 30 min, and 20–40% B from 30 to 40 min. Isolated ms²ct⁶A was characterized by UV spectroscopy (Supplementary Data), IR spectroscopy (Supplementary Data), ¹H-NMR (Supplementary Data), ¹³C-NMR (Supplementary Data) and high resolution MS (Supplementary Data).

Cultivation and RNAi of T. brucei

Partial segments of the coding sequences of the Tb427tmp.02.2830 (TbTcdA) and Tb427.06.3510 (TbMtaB) from T. brucei were cloned into the tetracycline-inducible RNAi vector p2T7-177. These plasmids were then linearized by NotI digestion and introduced into procyclic T. brucei 29-13 cells for genomic integration; clonal lines were obtained by limiting dilution. The cell lines were grown in SDM-79 medium, and RNAi was induced by addition of 1 μg/ml tetracycline. Cell counts were taken every 24 h using a Beckman Z2 Coulter counter over the course of 12 days post-induction in the presence or absence of tetracycline (1 μg/ml) and cycloheximide (50 μg/ml).

Total RNA was extracted from uniduced and RNAi-induced T. brucei cells using a standard protocol (39). The steady-state levels of individual mRNAs were measured by RT-PCR. The cDNAs of TbMtaB (677 bps) and TbTcdA (443 bps) were amplified using the following primers: 5΄-attcacttaacttccctatttgc-3΄ and 5΄-gttctgacagcattcttcaaacc-3΄ for TbMtaB, and 5΄-gcctaccaaccgggaagccttcttg-3΄ and 5΄-ggattgcattgacagcgtcgagtgtaag-3΄ for TbTcdA.

PCR products were then analyzed by agarose gel electrophoresis. The same reaction performed without reverse transcriptase was used as a negative control (RT-).

RESULTS

Identification of ms²ct⁶A in B. subtilis tRNAs

We previously detected ct⁶A in total nucleosides of B. subtilis tRNAs digested under neutral conditions (9). Consistent with this finding, the B. subtilis YrvM is an ortholog gene of TcdA, which catalyzes ATP-dependent dehydration of t⁶A to form ct⁶A (9). In addition, ms²t⁶A is present in tRNA^Lys from B. subtilis (21,22). These observations strongly suggest that, in B. subtilis, ms²t⁶A in tRNA^Lys is converted to ms²ct⁶A by the TcdA homolog.

To confirm the presence of ms²ct⁶A, total RNA of B. subtilis was digested into nucleosides under conventional or neutral conditions, and then subjected to LC/MS analysis (Figure 2A). In the conventional conditions, we observed proton adducts of t⁶A and ms²t⁶A, but no ct⁶A, as reported previously (9). On the other hand, in the neutral conditions, ct⁶A was clearly detected. In addition, we detected a dehydrated form of ms²t⁶A (m/z 441). This modified species, tentatively named N⁴⁴⁰, was not detected in total nucleosides digested in the conventional conditions, indicating that it was susceptible to hydrolysis (Figure 2A). This observation strongly suggested that N⁴⁴⁰ is ms²ct⁶A.

Figure 2.

Mass spectrometric analyses of total RNA and isolated tRNA^Lys from B. subtilis. (A) Nucleoside analyses of total RNA from B. subtilis. Total nucleosides were prepared under conventional conditions (left panels) and neutral conditions (right panels). The panels second from the bottom show mass chromatograms corresponding to the proton adducts of t⁶A (m/z 413), ms²t⁶A (m/z 459), ct⁶A (m/z 395) and ms²ct⁶A (m/z 441), respectively. n.d., not detected. (B) Mass spectrum of the 12 mer-fragments containing anticodon region. Three peaks for the triply-charged negative ions of the RNA fragments having ms²A, ms²t⁶A and ms²ct⁶A at position 37 are indicated. (C) A collision-induced dissociation (CID) spectrum of the 12 mer-fragment of B. subtilis tRNA^Lys digested by RNase T₁. The triply-charged negative ion of the ms²ct⁶A37-containing fragment (m/z 1364.492) was used as a precursor ion for CID. The product ions were assigned according to the literature (45). Sequences of parent ion and assigned product ions are described upper side in this panel.

Considering that B. subtilis tRNA^Lys has ms²t⁶A at position 37 (Figure 1B) (17), ms²ct⁶A should be found in this tRNA. B. subtilis tRNA^Lys was isolated by the reciprocal circulating chromatography (9,31), digested by RNase T₁ and subjected to capillary LC coupled to ESI/MS. The 12 mer-fragments containing anticodon region were clearly detected. Judging from the m/z values of triply-charged negative ions of this fragment (Figure 2B), we clearly detected three fragments having different modifications at position 37, namely ms²A, ms²t⁶A and ms²ct⁶A. Although ms²t⁶A37 was present more abundant than ms²ct⁶A37 in the isolated tRNA^Lys, it is likely that a certain population of ms²t⁶A37 in this tRNA originates from ms²ct⁶A37 hydrolyzed during tRNA isolation by RCC. The 12 mer-fragment with ms²ct⁶A was further probed by collision-induced dissociation to map the modified residues (Figure 2C). By assignment of product ions in the CID spectrum, we unequivocally mapped cmnm⁵s²U at position 34 and ms²ct⁶A at position 37 (Figure 2C).

To determine the structure of N⁴⁴⁰, we chemically synthesized ms²ct⁶A from ms²t⁶A (Supplementary Data). As reported in the accompanying paper (15), activation of the carboxyl group of t⁶A by water-soluble carbodiimide (EDC), which facilitates cyclization of the side chain, predominantly generates the hydantoin isoform of ct⁶A. Because we employed the same procedure as for ct⁶A synthesis to cyclize the side chain of ms²t⁶A, the hydantoin isoform should be the predominant form in chemically synthesized ms²ct⁶A. Detailed spectroscopic analyses of the synthesized ms²ct⁶A using UV (Supplementary Data), IR (Supplementary Data), ¹H NMR (Supplementary Data) and ¹³C NMR (Supplementary Data) supported the hydantoin isoform as observed for ct⁶A in the accompanying paper (15). Especially, in the IR spectrum (Supplementary Data), two characteristic absorption bands at 1720 cm⁻¹ and 1788 cm⁻¹ are associated with C = O bond stretching in the hydantoin ring.

The synthetic ms²ct⁶A was mixed with total nucleosides of B. subtilis and subjected to LC/MS analyses using reverse-phase chromatography (Figure 3A), as well as hydrophilic interaction chromatography (Figure 3B). The synthetic ms²ct⁶A co-eluted with N⁴⁴⁰ as a single peak under both conditions. Next, we used CID to further probe the base-related ions (BH₂⁺, m/z 309) of N⁴⁴⁰ and the synthetic ms²ct⁶A (Figure 3C). The product ions of these two compounds exhibited identical patterns (Figure 3D), and assigned in the chemical structures of the ms²ct⁶A base (Figure 3E). Collectively, these observations indicated that N⁴⁴⁰ is ms²ct⁶A.

Figure 3.

Structural confirmation of ms²ct⁶A from B. subtilis. Co-injection analyses of synthetic ms²ct⁶A and total RNA from B. subtilis by (A) RPC/ESI-MS and (B) HILIC/ESI-MS. UV trace at 254 nm and mass chromatograms of the synthetic ms²ct⁶A, total nucleosides of B. subtilis and co-injection are shown in the top, middle and bottom panels, respectively. Conversion of adenosine to inosine is due to the contamination of PDase II (Sigma P9041) with adenosine deaminase activity. (C) The mass spectrum for the proton adduct of natural ms²ct⁶A (MH⁺, m/z 440.9) from B. subtilis. The base-related ion (BH₂⁺, m/z 309.0) was also detected. (D) The CID spectra for natural (upper panel) and synthetic (lower panel) BH₂⁺ of ms²ct⁶A. Parent ions for CID are indicated by arrows. (E) Assignment of the product ions in the CID spectrum of the natural ms²ct⁶A BH₂⁺ ion.

Biogenesis of ms²ct⁶A in B. subtilis

The yqeV and yrvM of B. subtilis are homologs of mtaB and tcdA, respectively (9,22). To confirm that these genes are involved in ms²ct⁶A formation, we constructed a knockout of yrvM (ΔyrvM) and the double-deletion strain (ΔyqeV/ΔyrvM) by homologous recombination. Total RNA from wild-type B. subtilis (WT) and a series of knockout strains were digested into nucleosides and subjected to LC/MS analysis (Figure 4). In the ΔyqeV strain, both ms²ct⁶A and ms²t⁶A disappeared as expected and probably remained to be ct⁶A and t⁶A, respectively. In fact, the level of ct⁶A increased slightly relative to that in the WT strain. In ΔyrvM, both ms²ct⁶A and ct⁶A disappeared and remained to be ms²t⁶A and t⁶A, respectively. Indeed, the levels of ms²t⁶A and t⁶A clearly increased in this strain. In the case of ΔyqeV/ΔyrvM, all three derivatives (ms²ct⁶A, ct⁶A and ms²t⁶A) disappeared, and relevant tRNAs only have t⁶A. These observations demonstrated that yqeV (mtaB) and yrvM (tcdA) are responsible for 2-methylthiolation and cyclization of t⁶A, respectively, resulting in the formation of ms²ct⁶A on tRNA^Lys.

Figure 4.

Identification of enzymes responsible for ms²ct⁶A formation in B. subtilis. RPC/ESI-MS nucleoside analyses of total RNAs from B. subtilis wild type (str. 168). Mass chromatograms from ΔyqeV, ΔyrvM and ΔyqeV/ΔyrvM are shown from left to right, respectively. Top panels show the UV traces at 254 nm. Panels second from the bottom show mass chromatograms corresponding to the proton adducts of ms²ct⁶A (m/z 441), ct⁶A (m/z 395), ms²t⁶A (m/z 459), t⁶A (m/z 413) and i⁶A (m/z 336), respectively. n.d., not detected. Mass chromatograms of t⁶A derivatives were normalized by that of i⁶A, and abundance of each peak is displayed relative to the highest peak (100%) among them.

Identification of ms²ct⁶A in plant tRNAs

According to phylogenetic analyses (9,40), homologs of mtaB and tcdA are encoded in plant genomes, indicating the presence of ms²ct⁶A in plant tRNAs. Indeed, ct⁶A and ms²t⁶A were previously detected in Spinach (9) and Eleusine coracana (20), respectively. Therefore, we prepared total RNA from Nicotiana tabacum, Spinacia oleracea and Arabidopsis thaliana, and enzymatically digested these samples into nucleosides under acidic conditions. LC/MS analyses clearly detected ms²ct⁶A along with other t⁶A derivatives in all three species of plant (Figure 5). Next, total nucleosides of spinach RNA were co-injected along with synthetic ms²ct⁶A by LC/MS using hydrophilic interaction chromatography (Supplementary Data), as well as reverse-phase chromatography (Supplementary Data). Plant ms²ct⁶A co-eluted with the synthetic molecule as a single peak under both conditions, demonstrating that ms²ct⁶A is present in plant tRNAs.

Figure 5.

Detection of ms²ct⁶A in plant tRNAs. HILIC/ESI-MS nucleoside analyses of total RNAs from N. tabacum, S. oleracea and A. thaliana. Top panels show UV traces at 254 nm. Panels second from the bottom show mass chromatograms corresponding to the proton adducts of t⁶A (m/z 413), ms²t⁶A (m/z 459), ct⁶A (m/z 359) and ms²ct⁶A (m/z 441), respectively. Unassigned peaks are indicated by asterisks. Conversion of adenosine to inosine observed in spinach and A. thaliana is due to the contamination of adenosine deaminase activity in PDase II (Sigma P9041).

Presence of ms²ct⁶A in T. brucei tRNAs

We previously speculated that ms²ct⁶A is present in T. brucei tRNA^Lys (9). Total RNA from T. brucei was digested into nucleosides under neutral conditions and subjected to LC/MS analysis (Figure 6A). Along with ct⁶A, N⁴⁴⁰ was clearly detectable. Next, the total nucleosides of T. brucei and B. subtilis were co-injected to LC/MS, revealing that T. brucei N⁴⁴⁰ co-eluted with B. subtilis ms²ct⁶A as a single peak (Figure 6B). Finally, we probed BH₂⁺ of T. brucei N⁴⁴⁰ by CID (Figure 6C). The product ions exhibited a pattern identical to those of synthetic ms²ct⁶A (Figure 3D). These results demonstrated that ms²ct⁶A is also present in T. brucei.

Figure 6.

Detection of ms²ct⁶A in T. brucei tRNAs. (A) RPC/ESI-MS nucleoside analyses of total RNAs from T. brucei. Top panel shows a UV trace at 254 nm. Panels second from the bottom show mass chromatograms corresponding to the proton adducts of t⁶A (m/z 413), ms²t⁶A (m/z 459), ct⁶A (m/z 395) and ms²ct⁶A (m/z 441), respectively. (B) Mass chromatograms showing the proton adduct of ms²ct⁶A (m/z 441) in total nucleosides of B. subtilis (top panel), T. brucei (middle panel) and the co-injection fraction (bottom panel) obtained by RPC/ESI-MS. (C) CID spectrum of BH₂⁺ of ms²ct⁶A from T. brucei. The parent ion for CID is indicated by an arrow.

Growth phenotype of T. brucei with hypomodified ms²ct⁶A

We hypothesized that TbMtaB and TbTcdA are responsible for the synthesis of ms²ct⁶A in T. brucei. To test this speculation, we generated a transgenic RNAi line for each gene and confirmed knockdown efficiency by RT-PCR. No transcript of each gene was detected in either strain, even 7 days after knockdown induced by tetracycline (Supplementary Data). Upon knockdown of TbMtaB, ms²ct⁶A and ms²t⁶A disappeared and remained to be ct⁶A and t⁶A, respectively (Figure 7A). In fact, ct⁶A accumulated slightly more than non-treated T. brucei (WT). Likewise, when TbTcdA was down-regulated, ms²ct⁶A and ct⁶A disappeared and remained to be ms²t⁶A and t⁶A, respectively (Figure 7A). To determine the physiological importance of this modification, we measured the growth rates of these strains after induction of RNAi (Figure 7B). Down-regulation of expression of either gene alone did not cause a major growth defect (data not shown). However, in the presence of non-inhibitory concentrations of cycloheximide (Chx), growth of cells with down-regulated TbTcdA was slowed (Figure 7B), whereas little growth phenotype was observed when TbMtaB was knocked down. Thus, hypomodification of ms²ct⁶A and ct⁶A results in sensitivity to an antibiotic that targets the ribosome, suggesting that cyclization of these modifications contributes to efficient cell growth and protein synthesis.

Figure 7.

Biogenesis of ms²ct⁶A and growth phenotype of T. brucei with hypomodified tRNAs. (A) RPC/ESI-MS nucleoside analyses of total RNAs from T. brucei wild type (left panels) and transgenic RNAi lines of TbMtaB (middle panels) and TbTcdA (right panels). Mass chromatograms of t⁶A derivatives were normalized by that of i⁶A, and abundance of each peak is displayed relative to the highest peak (100%) among them. Top panels show the UV traces at 254 nm. Panels second from the bottom show mass chromatograms detecting the proton adducts of ms²ct⁶A (m/z 441), ct⁶A (m/z 395), ms²t⁶A (m/z 459), t⁶A (m/z 413) and i⁶A (m/z 336), respectively. n.d., not detected. (B) Growth curves of T. brucei transgenic RNAi lines of TbMtaB (left) and TbTcdA (right). Cumulative cell counts indicated by logarithmic cell number (Log of cells/ml) were measured in the presence of non-inhibitory concentration of cycloheximide. RNAi was induced by adding (black circle) or not adding (black square) tetracycline (Tet) at 7 days (as indicated by arrow) after inoculation.

DISCUSSION

Because ct⁶A is a chemically labile derivative of t⁶A, it had never been detected by conventional nucleoside analysis for more than 40 years since t⁶A's discovery. ct⁶A can be detected only if total RNA is extracted from the cell under acidic conditions and digested into nucleosides under neutral conditions, explaining why ms²ct⁶A has not been detected previously. In this study, we performed nucleoside analysis under neutral or acidic conditions and successfully identified ms²ct⁶A in total RNA samples from B. subtilis, three plants and T. brucei. Because we also detected t⁶A and ms²t⁶A, we assume that ct⁶A and ms²ct⁶A are partial modifications in these organisms, indicating that cyclization of t⁶A and ms²t⁶A might be regulated under certain physiological conditions.

By referring to the synthetic nucleoside as a reference, the chemical structure of ms²ct⁶A was determined to be a hydantoin isoform, rather than an oxazolone isoform previously predicted. According to the structural analysis of ct⁶A nucleoside (15), two carbonyl oxygen atoms of the hydantoin ring are repulsive to the nitrogen atoms (N1 and N7) of the adenine base (Supplementary Data). Consistent with this, the C6-N6 bond length is longer than the normal C-N bond length. Therefore, the hydantoin ring adopts a twisted position against the adenine base with a torsion angle of −52.7°. In light of this observation, it is difficult to speculate how the hydantoin ring of ct⁶A contributes to the efficient decoding of tRNA on the ribosome. Structural studies of ribosomes in complex with tRNA containing ct⁶A or ms²ct⁶A will be necessary to reveal the functional and structural roles of these modifications in protein synthesis.

We now know of five species of t⁶A derivatives (Figure 1A and S1), of which ms²ct⁶A is the most chemically complex. Initially, t⁶A is formed at position 37 on tRNAs with NNU anticodons (i.e. those responsible for ANN codons) (Figure 8). This process is catalyzed by multiple enzymes (TsaB, TsaC, TsaD and TsaE in bacteria) using L-threonine, bicarbonate and ATP as substrates. In tRNA^Thr from certain species of γ-proteobacteria, m⁶t⁶A37 is formed via the methylation of t⁶A37 by TrmO using AdoMet as a substrate. In some species of bacteria, fungi, plants and protists, ct⁶A37 is formed via the cyclization of t⁶A37 by ATP-dependent dehydration by TcdA. In organisms that harbor an MtaB homolog, t⁶A37 and ct⁶A37 are 2-thiomethylated to form ms²t⁶A and ms²ct⁶A, respectively. According to the nucleoside analyses of B. subtilis knockout strains, 2-methylthiolation and cyclization of t⁶A are independent reactions involved in synthesis of ms²ct⁶A (Figure 8).

Figure 8.

Biosynthetic pathway of ms²ct⁶A and other t⁶A derivatives. In bacteria, A37 of tRNAs responsible for ANN codons is modified to t⁶A by multiple enzymes (TsaB, TsaC, TsaD and TsaE) using L-threonine, bicarbonate and ATP as substrates. In tRNA^Thr from γ-proteobacteria, t⁶A37 is methylated by TrmO using AdoMet to form m⁶t⁶A37. In certain populations of bacteria, fungi, plant and protists, t⁶A37 is cyclized by TcdA using ATP to form ct⁶A37. In organisms bearing an MtaB homolog, t⁶A37 and ct⁶A37 in particular tRNAs are 2-thiomethylated by MtaB to form ms²t⁶A37 and ms²ct⁶A37, respectively.

In terms of phylogenetic distribution, ms²ct⁶A should be present in organisms with MtaB and TcdA orthologs. Among bacteria, δ-proteobacteria and approximately half of the species that compose Firmicutes and Bacteroidetes encode these two enzymes in their genomes. In protists, ms²ct⁶A is found in Trypanosoma and Tetrahymena. In Archaeplastida, both MtaB and TcdA are present in Plants and Chlorophyta, but not in Rhodophyta or Glaucophyta.

In E. coli, ct⁶A37 is involved in the decoding activity of tRNA^Lys (9). tcdA engages in a genetic interaction with mnmA, as demonstrated by the observation that a synthetic growth reduction occurs when both genes are deleted simultaneously. Considering that mnmA encodes a 2-thiouridylase to form mnm⁵s²U at the wobble position of tRNA^Lys, which also contains ct⁶A37, cyclization of ct⁶A37 must play a functional role in the decoding activity of tRNA^Lys.

The functional role of the 2-methylthio modification was first studied in an in vitro E. coli translation system using tRNA containing ms²i⁶A37. E. coli tRNA^Phe containing the ms²-modification was more active in poly(U)-dependent poly(Phe) synthesis than tRNA lacking the ms²-modification (41). In Salmonella typhimurium, +1 frameshift activities of tRNA^Phe and tRNA^Tyr at the ribosomal P-site were significantly higher when they lacked the ms²-modification (42), indicating that 2-methylthiolation of ms²i⁶A37 plays a critical role in maintaining the reading frame during elongation. The structure of ms²i⁶A37 in tRNA has revealed that the 2-methylthio group stabilizes the codon–anticodon interaction through cross-strand stacking with the first base of the P-site codon (43). This stabilization effect might contribute to the prevention of +1 frameshifting.

In the case of ms²t⁶A modification in B. subtilis, the decoding ability of tRNA^Lys with or without the ms²-modification was examined using a luciferase reporter. The results showed that the ms²-modification is required to decode AAA and AAG codons efficiently (23). Given that a large fraction of ms²t⁶A37 must be converted to ms²ct⁶A37 in B. subtilis tRNA^Lys, this finding throws light on the function of the ms²-modification of ms²ct⁶A37 (but not of ms²t⁶A37). However, no significant growth or temperature-sensitivity phenotypes have been observed to date in ΔyqeV, ΔyrvM or the double-deletion strains (data not shown). Future studies should investigate whether phenotypes emerge under various stress or culture conditions. In T. brucei, we observed a growth delay after knockdown of TbTcdA in the presence of non-inhibitory concentration of cycloheximide, indicating defective decoding activity of tRNA^Lys with hypomodified ms²ct⁶A. To obtain a deeper understanding of the physiological role of this modification, future studies should be directed toward analyzing the phenotypic features of T. brucei knockdown lines.

SUPPLEMENTARY DATA

Supplementary Data are available at NAR Online.

FUNDING

National Science Centre in Poland [UMO-2014/13/N/ST5/01591 to M.M.]; National Institute of General Medical Sciences of the National Institues of Health [R01GM084065 to J.D.A.]; Ministry of Education, Culture, Sports, Science and Technology of Japan (MEXT) [to T.S.]; Japan Society for the Promotion of Science (JSPS) [to T.S.]. Funding for open access charge: Grants-in-Aid for Scientific Research on Priority Areas from the Ministry of Education, Science, Sports and Culture of Japan.

Conflict of interest statement. None declared.