Unexpected Accumulation of ncm⁵U and ncm⁵s²U in a trm9 Mutant Suggests an Additional Step in the Synthesis of mcm⁵U and mcm⁵s²U

Abstract

Background

Transfer RNAs are synthesized as a primary transcript that is processed to produce a mature tRNA. As part of the maturation process, a subset of the nucleosides are modified. Modifications in the anticodon region often modulate the decoding ability of the tRNA. At position 34, the majority of yeast cytosolic tRNA species that have a uridine are modified to 5-carbamoylmethyluridine (ncm⁵U), 5-carbamoylmethyl-2′-O-methyluridine (ncm⁵Um), 5-methoxycarbonylmethyl-uridine (mcm⁵U) or 5-methoxycarbonylmethyl-2-thiouridine (mcm⁵s²U). The formation of mcm⁵ and ncm⁵ side chains involves a complex pathway, where the last step in formation of mcm⁵ is a methyl esterification of cm⁵ dependent on the Trm9 and Trm112 proteins.

Methodology and Principal Findings

Both Trm9 and Trm112 are required for the last step in formation of mcm⁵ side chains at wobble uridines. By co-expressing a histidine-tagged Trm9p together with a native Trm112p in E. coli, these two proteins purified as a complex. The presence of Trm112p dramatically improves the methyltransferase activity of Trm9p in vitro. Single tRNA species that normally contain mcm⁵U or mcm⁵s²U nucleosides were isolated from trm9Δ or trm112Δ mutants and the presence of modified nucleosides was analyzed by HPLC. In both mutants, mcm⁵U and mcm⁵s²U nucleosides are absent in tRNAs and the major intermediates accumulating were ncm⁵U and ncm⁵s²U, not the expected cm⁵U and cm⁵s²U.

Conclusions

Trm9p and Trm112p function together at the final step in formation of mcm⁵U in tRNA by using the intermediate cm⁵U as a substrate. In tRNA isolated from trm9Δ and trm112Δ strains, ncm⁵U and ncm⁵s²U nucleosides accumulate, questioning the order of nucleoside intermediate formation of the mcm⁵ side chain. We propose two alternative explanations for this observation. One is that the intermediate cm⁵U is generated from ncm⁵U by a yet unknown mechanism and the other is that cm⁵U is formed before ncm⁵U and mcm⁵U.

Introduction

Transfer RNAs are adapter molecules, which decode mRNA into protein and thereby play a central role in gene expression. The primary tRNA transcript is processed by different endo and exonucleases, and tRNA modifying enzymes to produce a mature tRNA [1], [2], [3]. In this maturation process, a subset of the four normal nucleosides adenosine (A), guanosine (G), cytidine (C) and uridine (U) are modified [2], [3]. The modifications are introduced post-transcriptionally, and the formation of a modified nucleoside may require one or several enzymatic steps [2], [3]. Of the 50 modified nucleosides so far identified in eukaryotic tRNAs, 25 are present in cytoplasmic tRNAs from S. cerevisiae [2], [4], [5]. In the anticodon region, especially in positions 34 (wobble position) and 37, nucleosides are frequently modified. Modified nucleosides in these positions are important for reading frame maintenance and efficient decoding during translation [2], [3]. In yeast, there are in total 42 cytosolic tRNA species, of which 11 have a uridine at position 34 modified to 5-carbamoylmethyluridine (ncm⁵U), 5-carbamoylmethyl-2′-O-methyluridine (ncm⁵Um), 5-methoxycarbonylmethyl-uridine (mcm⁵U) or 5-methoxycarbonylmethyl-2-thiouridine (mcm⁵s²U) [6]. The formation of these nucleosides requires addition of mcm or ncm side chains at the 5-position of the uracil moity and a subset of these tRNAs also have a thio (s²) group at the 2-position of U₃₄ or a methylation at the 2′ position of the ribose.

The common step in synthesis of ncm⁵ and mcm⁵ side chains at U₃₄ in tRNAs requires at least 11 gene products (Figure 1). Deletion strains missing one of ELP1-ELP6, KTI11, KTI12, KTI14 or SIT4 genes, or both SAP185and SAP190 genes completely lack the mcm⁵U, mcm⁵s²U and ncm⁵U nucleosides, whereas a kti13 deletion mutant show dramatically reduced levels of these nucleosides [7], [8]. In strains with these genes mutated, no intermediates of mcm⁵U or ncm⁵U have been detected, whereas s²U is detected in tRNAs normally containing mcm⁵s²U [7], [8], [9], [10], [11], [12]. Thus, these gene products are required for an early step in synthesis of mcm⁵ and ncm⁵ groups (Figure 1). The earliest intermediate in the synthesis of mcm⁵U and ncm⁵U that has been detected is cm⁵U, and there is evidence that it originates from a metabolite related to acetyl-CoA [13] (Figure 1).

Figure 1. Model for formation of mcm⁵ side chain at wobble uridines.

The Elongator complex (Elp1-Elp6) and its potential regulators are required for the formation of cm⁵U. A methyl group is added to cm⁵U by Trm9p/Trm112p complex in tRNA species that in their mature form should have a mcm⁵ side chain. The cm⁵U in other tRNA species are converted to ncm⁵U by an unknown enzyme. For tRNAs that should contain a s² group, presence of a mcm⁵ or ncm⁵ side chain is a prerequisite for efficient thiolation.

The ELP1-ELP6 gene products form the Elongator complex that consists of a core complex Elp1-Elp3 and a sub complex Elp4-Elp6 [14], [15], [16]. In the C-terminal part of Elp3p there is a potential acetyl-CoA binding domain [17], and the central region shares homology to the Radical SAM superfamily [18]. Members of this family contain an iron-sulphur (FeS) cluster and use S-adenosylmethionine (SAM) to catalyze a variety of radical reactions. The presence of a FeS cluster and ability to bind SAM has been verified for the M. jannaschii Elp3p homologue [18], whereas no binding of SAM to S. cerevisiae Elongator complex was observed [19]. At least Elp1 and Elp3 of Elongator core complex are in intimate contact with tRNA that is modified with a mcm side chain at U₃₄ [7]. The KTI11-KTI14, SIT4 or SAP185 SAP190 gene products seem to regulate the activity of Elongator complex [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30].

The last step in formation of mcm⁵ side chain of U₃₄ is a methyl esterification of cm⁵ [13], and requires Trm9p/Trm112p in yeast and ALKBH8/TRM112 in mammalians [31], [32], [33]. We confirm that Trm112p is also required for the last step of mcm⁵ side chain formation at position 34 in a subset of tRNAs. In vivo, Trm112p is essential for the methyl esterification to mcm⁵U₃₄, and in vitro Trm112p improves the methyltransferase activity of Trm9p. The observation that the major intermediates accumulating in trm9 and trm112 mutants are ncm⁵U and ncm⁵s²U and not the expected cm⁵U and cm⁵s²U raises the question; what is the order of intermediates formed in biosynthesis of the mcm⁵ side chain of U₃₄?

Materials and Methods

Yeast strains, media and genetic procedures

Strains used in this report, except those from the yeast deletion collection (Open Biosystems), are listed in Table S1A. Yeast media, genetic procedures and yeast transformation have been described previously [34]. To construct mtq2::KanMX6 and trm112::KanMX6 deletions, oligonucleotides (2104 and 2015, 1391 and 1392) in Table S1B containing 45nt sequence homology flanking the MTQ2 and TRM112 genes were used to amplify the KanMX6 cassette [35]. To delete TRM9, TRM11 and LYS9 in W303 strains, chromosomal DNA from the corresponding null mutants in the yeast deletion collection (Open Biosystems) were used as templates. The KanMX6 cassette together with 300–500 base pair flanking sequences to each gene were amplified with specific primers (1035 and 1036 for TRM9, 1950 and 1951 for TRM11, and 2059 and 2060 for LYS9) listed in Table S1B. The PCR products were introduced into diploid yeast strain UMY3104 and transformants were selected on YEPD plates containing 200 µg/ml Geneticin (G418). Transformants were sporulated and tetrad analysis verified a 2∶2 segregation of mating type and G418 resistance. Deletions were confirmed by PCR. The double mutants trm9Δ trm11Δ, trm9Δ lys9Δ, trm9Δ mtq2Δ, trm11Δ lys9Δ, trm11Δ mtq2Δ and lys9Δ mtq2Δ were generated by crossing the single mutants. The quadruple mutant was generated in a cross between trm9Δ lys9Δ and trm11Δ mtq2Δ.

Plasmid constructions

To generate the expression vector for the Trm9 protein, TRM9 gene was amplified by PCR using oligos 2015 and 2016 (Table S1B) and W303-1A genomic DNA as template. The PCR product was digested with BamH1 and HindIII, and subcloned to the corresponding sites of the expression vector pRSF-Duet1 (Novagen), generating an in frame fusion with the histidine tag. To construct the Trm9p-Trm112p co-expression vector, the TRM112 gene was amplified from W303-1A genomic DNA using oligos 2013 and 2014 (Table S1B) and cloned into the pRSF-Duet1-TRM9 vector using NdeI and XhoI.

Protein purification

The expression vectors were introduced into BL21(DE3)pLysS competent cells. Overnight cultures of transformed cells were grown in LB media containing 50 µg/ml Kanamycin at 37°C. Cultures were diluted to OD₆₀₀ 0.05 and grown to OD₆₀₀ 0.5 at 37°C. Cultures were placed on ice for 10 minutes. IPTG was added to a final concentration of 120 µg/ml and protein expression was induced at 15°C overnight. Harvested cell pellets were washed once by 0.9% NaCl and resuspended in breaking buffer (20 mM Tris pH 8.0, 10 mM imidazole, 150 mM NaCl, 0.2% NP-40, 2 mM β-mercaptoethanol) in the presence of proteinase inhibitor cocktail (Roche). Cells were broken by sonication and the cell extract was clarified by centrifugation at 16,000 g for 1 hour. The supernatant was mixed with TALON resin, equilibrated with breaking buffer and incubated at 4°C for 2 hours. The protein bound TALON resin was first washed with buffer 1 (20 mM Tris pH 8.0, 10 mM imidazole, 150 mM NaCl, 2 mM β-mercaptoethanol) and then with buffer 2 (20 mM Tris pH 8.0, 10 mM imidazole, 500 mM NaCl, 2 mM β-mercaptoethanol). Proteins were eluted with 330 mM imidazole and dialyzed overnight against storage buffer (25 mM Tris pH 8.0, 150 mM NaCl, 5 mM DTT, 10% glycerol) and kept at 4°C for future use.

Methyltransferase reaction

In the methyltransferase reaction, 50 µl of 2X reaction buffer (200 mM Tris 7.5, 0.2 mM EDTA, 20 mM MgCl₂, 20 mM NH₄Cl) was mixed with 20 µl [³H]AdoMet (0.55 mCi/ml, Perkin Elmer) and 20 µg tRNA, incubated at 37°C for 5 minutes. The methyltransferase reaction was initiated by adding 10 µg Trm9p or Trm9p-Trm112p. Aliquots of the reaction was withdrawn at different time points and mixed with 1 ml of 5% ice cold trichloroacetic acid (TCA). The tubes were incubated on ice for 10 minutes and samples were vacuum filtered through nitrocellulose filter (Millipore 0.45 µm). The [³H] incorporation was measured using a Wallac 1409 scintillation counter. To analyze [³H] incorporation in total tRNA by HPLC, 200 µg of tRNA was used. After 30 minutes of methyltransferase reaction, 2.5 volume of 99% ice cold ethanol was added into the reaction and samples were centrifuged for 30 minutes in eppendorf tubes at maximum speed. The pellet was resuspended in MQ water, digested with nuclease P1 and analyzed by HPLC [36]. The [³H] incorporation was monitored by a flow scintillation analyzer (Packard Bioscience).

Single tRNA isolation

Yeast cells were grown in 2L YEPD at 30°C to OD600 = 1.5. Total tRNA was prepared as described [36]. Single tRNA species were isolated from total tRNA by hybridizing to biotinylated complementary oligonucleotides [36] and separated from total tRNA by attachment to streptavidin coated Dynabeads M-280 (Invitrogen). The single tRNAs were digested to nucleosides with nuclease P1 followed by bacterial alkaline phosphatase (BAP) treatment [0.2 M (NH₄)₂SO₄ pH 8.3], and analyzed by HPLC [37].

Results and Discussion

Trm112p is required for the methyl esterification of mcm⁵U and mcm⁵s²U

In a global analysis of protein complexes in yeast, Trm112p was found to interact with three methyltransferases Trm9p, Trm11p and Mtq2p [38], [39], [40], [41]. In addition, Trm112p interacts with the saccharopine dehydrogenase Lys9p, the essential DEAH-box ATP-dependent RNA helicase Ecm16p and an essential component of the RSC chromatin remodeling complex Sfh1p [38], [39], [40], [41]. The N²-Monomethylguanosine-10 (m²G₁₀) methyltransferase Trm11p, as well as the eRF1 methyltranferase Mtq2p, has to be in complex with Trm112p to be active [42], [43]. Trm9p is required for the methyl esterification of modified uridine nucleosides, resulting in the formation of 5-methylcarbonylmethyluridine (mcm⁵U₃₄) and 5-methylcarbonylmethyl-2-thiouridine (mcm⁵s²U₃₄) present in a subset of tRNA species in yeast, including and [31]. In the methyl esterification reaction of these tRNAs, cm⁵U₃₄ and cm⁵s²U₃₄ were suggested to be the substrates [13], [31], [32,].

Both Trm9 and Trm112 are required for methyl esterification to mcm⁵U and mcm⁵s²U [31], [32], [33]. To analyze the tRNA modification status in these two mutants, total tRNA from trm9Δ, trm112Δ and wild type strains were isolated, digested to nucleosides and analyzed by HPLC. Similar to previous reports [31], [32], [33], total tRNA isolated from trm9 and trm112 deletion mutants lacked mcm⁵U and mcm⁵s²U nucleosides (data not shown). In order to provide a more detailed analysis of all possible nucleoside intermediates in trm9Δ and trm112Δ mutants, single tRNA species, , and , were isolated from wild type, trm9Δ and trm112Δ strains and the purified tRNAs were digested to nucleosides and analyzed by HPLC (Figure 2, Table 1, data not shown). As expected, the ncm⁵U nucleoside was present in independent if the tRNA was isolated from trm9Δ, trm112Δ or wild type strains (data not shown). The mcm⁵U and mcm⁵s²U nucleosides were present in and isolated from wild type but not from trm9Δ and trm112Δ cells (Figure 2, Table 1). In isolated from trm9Δ and trm112Δ strains, we observed the appearance of ncm⁵U and cm⁵U (Figure 2B, Table 1). Interestingly, the major intermediate of the mcm⁵U nucleoside generated in the trm9Δ and trm112Δ mutants is ncm⁵U (Figure 2, Table 1). The presence of ncm⁵U and cm⁵U has also been observed in , and isolated from an alkbh8 ^−/− mice [32], [44]. In isolated from the trm9Δ and trm112Δ strains, there was a complete lack of mcm⁵s²U and a concomitant increase of cm⁵U, ncm⁵U and ncm⁵s²U (Table 1). The presence of cm⁵U supports the earlier observation that formation of a completed mcm⁵ side chain appears to be a prerequisite for efficient and complete thiolation of position 2 in mcm⁵s²U containing tRNAs [10], [11], [12], [32]. An unexpected observation was that the major species accumulating in and isolated from trm9Δ and trm112Δ strains were ncm⁵U and ncm⁵s²U, respectively (Table 1). We considered the possibility that the ncm⁵ side chain was spontaneously generated from cm⁵ by amidation during the bacterial alkaline phosphatase (BAP) treatment in the digestion step of tRNA to nucleosides for HPLC analysis. To test this hypothesis, synthetic cm⁵U nucleoside was treated in the same way as in the digestion step of tRNA and analyzed by HPLC (Figure 3). We did not detect any conversion of cm⁵U to ncm⁵U (Figure 3) indicating that formation of ncm⁵U is enzymatically catalyzed and not an artifact of the sample preparation procedure.

Figure 2. trm9 and trm112 mutants are lacking the mcm⁵ side-chain in at wobble uridines.

HPLC analysis of modified tRNA nucleosides from wild-type (UMY3169, left panels), trm9::KanMX4 (Open Biosystems, middle panels) and trm112::KanMX4 (UMY3330, right panels). Arrows in red and black indicate expected retention time of mcm⁵U and cm⁵U, respectively. Arrow heads in red and black indicate expected retention time of m²G and ncm⁵U, respectively. (A), Part of the chromatogram between retention times 34 and 44 min is shown. (B), Part of the chromatogram between retention times 7 and 17 min is shown. The small peak in wild-type at 14 min represents an unrelated compound with a spectrum different from ncm⁵U. The chromatograms were monitored at 254 nm.

Table 1. Relative amounts of various modified nucleosides in and isolated from wild type, trm9Δ and trm112Δ strains.

	cm5U/Ψ	ncm5U/Ψ	mcm5U/Ψ	cm5U/Ψ	ncm5U/Ψ	mcm5U/Ψ	cm5s2U/Ψ	ncm5s2U/Ψ	mcm5s2U/Ψ
WT	0.016	0.044	0.183	ND	ND	ND	0.029	ND	0.220
trm9Δ	0.051	0.199	ND	0.065	0.028	ND	0.013	0.191	ND
trm112Δ	0.046	0.149	ND	0.044	0.044	ND	0.022	0.161	ND

Pseudouridine (Ψ) was used as the internal control. The numbers displayed are the ratios (modified nucleoside/Ψ). ND: not detected. The modified nucleosides cm⁵U, ncm⁵U, mcm⁵U and Ψ were monitored at 254 nm, and cm⁵s²U, ncm⁵s²U and mcm⁵s²U were monitored at 314 nm as thiolated nucleosides absorb well at this wavelength, while nonthiolated nucleosides do not.

Figure 3. Nucleoside ncm⁵U is not generated by amidation of cm⁵U during conversion of tRNA into nucleosides.

Synthetic cm⁵U (A and B) or a mixture of synthetic cm⁵U and ncm⁵U (C and D) were treated with nuclease P1 for 16 hours, followed by a 2 hours incubation with either water (A and C) or bacterial alkaline phosphatase (BAP) (B and D). Parts of the chromatogram of HPLC analysis between 5 and 20 min are shown. The chromatograms were monitored at 254 nm.

In addition to Trm9p, Trm112p also interacts with Trm11p, Lys9p and Mtq2p encoded by non-essential genes, and Ecm16p and Sph1p encoded by essential genes [38], [39], [40], [41]. Therefore, we also analyzed single tRNA species , and from trm11Δ, lys9Δ and mtq2Δ strains. Trm11p and Trm112p are essential for formation of the m²G nucleoside [42]. Consistently, isolated from trm11Δ or trm112Δ strains does not have the m²G modified nucleoside, whereas the same tRNA from wild-type has m²G (Figure 2 and S1). In single tRNAs from lys9Δ and mtq2Δ strains, there was no notable change in modified nucleosides as assessed by HPLC analysis (Figure S1, data not shown). A deletion of the TRM112 gene causes a dramatic reduction in growth and a mtq2Δ strain also shows a clear reduction in growth, whereas trm11Δ, lys9Δ or trm9Δ strains show mild growth defects in YEPD medium at both 30°C and 37°C (Figure 4). We considered the possibility that strains with multiple null alleles of genes encoding Trm112p interacting proteins would show additive growth defects, possibly mimicking a trm112Δ null allele. Since two Trm112p associated proteins, Ecm16 and Sfh1, are encoded by essential genes, we were only able to make strains with combinations of the trm11Δ, lys9Δ, trm9Δ, and mtq2Δ alleles. We first made the double mutants trm11Δ lys9Δ, trm11Δ trm9Δ, trm11Δ mtq2Δ, lys9Δ trm9Δ, lys9Δ mtq2Δ and trm9Δ mtq2Δ. No additive growth reduction was observed in any of the constructs at both 30°C and 37°C (Figure 4, data not shown), in contrast to the previously observed growth defect of the trm9Δ mtq2Δ mutant [33]. Further we made a trm11Δ lys9Δ trm9Δ, mtq2Δ quadruple mutant strain that grew like a mtq2Δ strain at both 30°C and 37°C (Figure 4). These data show that the poor growth of trm112Δ cells is not entirely caused by defects in tRNA modification, eRF1 methylation and dehydrogenase activity in the quadruple mutant. Possibly it is caused by reduced function of Ecm16p or Sfh1p which might require the interaction with Trm112p to be fully active.

Figure 4. Growth phenotypes.

Wild type (UMY2067), trm112Δ (UMY3679), trm9Δ (UMY3267), trm11Δ (UMY3677), lys9Δ (UMY3650), mtq2Δ (UMY3675), trm9Δ mtq2Δ (UMY3673) and trm9Δ trm11Δ lys9Δ mtq2Δ (UMY3680) strains were cultivated in YEPD at 30°C and 37°C.

Trm112p/Trm9p complex efficiently incorporates methyl groups into trm9 substrate tRNA in vitro

Trm9p has been shown to catalyze the methyl esterification to mcm⁵U and mcm⁵s²U in vitro [31]. We cloned the TRM9 gene into the expression vector pRSF duet to produce 6xHis-Trm9p recombinant protein in E. coli. We also made a pRSF duet vector construct, simultaneously expressing the 6xHis-Trm9p recombinant protein and a non-tagged Trm112p. When Trm9p was expressed alone, the majority of Trm9p recombinant protein was insoluble (Figure 5A), and the solubility of Trm9p dramatically improved when Trm112p was co-expressed with Trm9p. Purification of Trm9p by virtue of its 6xHis tag resulted in co-purification of Trm112p (Figure 5A), indicating that Trm9p forms a stable complex with Trm112p.

Figure 5. Trm9p/Trm112p complex efficiently catalyzes the methyl incorporation into trm9 substrate tRNA.

(A) SDS-PAGE analysis of histidine tagged Trm9p expressed alone or co-expressed with Trm112p and purified from E. coli. The gel was stained with Colloidal Blue (Invitrogen). Lane 1: Molecular weight standard (PageRuler prestained, Fermentas). Lane 2: Soluble fraction of extract from E. coli strains expressing Trm112p and histidine tagged Trm9p. Lane 3: Soluble fraction of extract from E. coli strains expressing histidine tagged Trm9p. Lane 4: Pellet from crude extract of E. coli strains expressing Trm112p and histidine tagged Trm9p. Lane 5: Pellet from crude extract of E. coli strains expressing histidine tagged Trm9p. Lane 6: Trm112p co-purified with histidine tagged Trm9 protein. Lane 7: Purified histidine tagged Trm9 protein. (B) [³H] methyl incorporation into tRNA as a function of time. Substrates were total tRNA preparations from strain UMY2067 (wild-type) and UMY3267 (trm9Δ). (▪) and (▴) are methyl incorporation reactions into wild-type tRNA by using Trm9p or Trm9p/Trm112p as enzyme. (○) is methyl incorporation reaction into trm9 tRNA by using Trm9p as enzyme. (C). The methyl incorporation into trm9 tRNA using Trm9p/Trm112p as enzyme (◊), in addition to the reactions in (B).

Purified Trm9p and Trm9/Trm112p complex was used to methylate total tRNA isolated from wild type and a trm9 deletion strains in vitro. Saponification of total tRNA with sodium hydroxide leads to the production of cm⁵U and cm⁵s²U from mcm⁵U and mcm⁵s²U, and this method has previously been used to generate substrates for Trm9p or ALKBH8 [31], [32]. However, saponification also efficiently degrades tRNA and we found that tRNA isolated from the trm9 deletion strain was a superior substrate in the methyl esterification assay (data not shown). To track methylation of tRNA substrates in vitro, S-adenosylmethionine containing a tritiated methyl donor group was used together with tRNA and purified enzyme. When total tRNA from wild type was used as a substrate, there was a small increase in incorporation of radioactive methyl groups with time using either Trm9p or the Trm9p/Trm112p complex (Figure 5B). In contrast, use of total tRNA from the trm9Δ strain and Trm9p leads to a clear but modest increase in the incorporation of radioactive methyl groups (Figure 5B). Moreover, the incorporation of radioactive methyl groups was 20-fold more efficient using Trm9p/Trm112p over Trm9p alone (Figure 5C). Thus, Trm112p is required for Trm9p to methylate its substrate tRNA more efficiently in vitro and is a prerequisite in vivo as no mcm⁵ nucleosides are formed in a trm112Δ mutant (Figure 2, Table 1). In the reaction using tRNA from the trm9Δ strain and Trm9p/Trm112p, there was a rapid incorporation of [³H] methyl groups in the first 5 minutes that entered to a plateau after 30 minutes (Figure 5C). The reduced incorporation was not due to enzyme inactivation with time as adding more enzyme at 30 minutes did not improve incorporation of radioactivity (data not shown).

Based on HPLC analysis, there is an accumulation of cm⁵U, ncm⁵U, and ncm⁵s²U in total tRNA from a trm9Δ strain compared with a wild-type strain [33] (data not shown). When tRNA isolated from a trm9Δ strain was used as substrate in vitro, we observed a reduction of the cm⁵U nucleoside and appearance of mcm⁵U (Figure 6A-D, Table 2) consistent with cm⁵U being the substrate of Trm9 [31], [32], [33]. Furthermore, the relative amounts of ncm⁵U and ncm⁵s²U did not change after the methylation reaction, showing that these two nucleosides are not substrates of Trm9p/Trm112p under these conditions (Table 2) [33]. By using saponified tRNA, cm⁵s²U was suggested to be a substrate for Trm9p or ALKBH8/Trm112 [31], [32]. However, cm⁵s²U was not detected in total tRNA isolated from trm9 or trm112 mutants [33]. In our analysis of trm9 total tRNA, we observed a very small peak migrating in the position of cm⁵s²U, which was absent after the methylation reaction (Figure 6G-H, Table 2). When [³H]-CH₃ was monitored by flow scintillation analyzer coupled to the HPLC, we found that the incorporated radioactivity migrated with retention times identical to those known for mcm⁵U and mcm⁵s²U nucleosides (Figure 6D, F and H). As the signal for the tentative cm⁵s²U is very weak, we cannot exclude the possibility that mcm⁵s²U originated from another species. These observations are consistent with those shown by Kalhor and Clarke [31], [32] and fully support the assertion that Trm9p is the methyltransferase catalyzing the formation of mcm⁵U from cm⁵U. Why and how ncm⁵U and ncm⁵s²U accumulates in tRNAs from strains lacking Trm9p or Trm112p, remains to be elucidated.

Figure 6. HPLC analysis of total trm9 tRNA after methyl incorporation by using Trm9p/Trm112p as enzyme.

(A–B) Part of the chromatogram between retention time 10 and 19 min is shown. The arrow in B indicates the expected retention time of cm⁵U. (C–D). Part of the chromatogram between retention time 34 and 45 min is shown. The arrow in C indicates the expected retention time of mcm⁵U. (E–F). Part of the chromatogram between retention time 47 and 52 min is shown. The arrow in E indicates the expected retention time of mcm⁵s²U. (G–H). Part of the chromatogram between retention time 26 and 52 min is shown. Open and closed arrowheads in G and H indicate the expected retention time of mcm⁵s²U and cm⁵U, respectively. Chromatograms in A–F were monitored at 254 nm and at 314 nm in G–H. The dashed line in D, F and H indicates the migration of isotope labeled nucleoside which overlaps with mcm⁵U and mcm⁵s²U, respectively. The Y axis to the left corresponds to absorbance units and the Y axis to the right shows the [³H] incorporation in cpm.

Table 2. Relative amounts of various modified nucleosides of total tRNA isolated from the trm9Δ strain before and after methylation reaction.

	cm5U/Ψ	ncm5U/Ψ	mcm5U/Ψ	cm5s2U/Ψ	ncm5s2U/Ψ	mcm5s2U/Ψ
Before reaction	0.01752	0.04634	ND	0.00081	0.00571	ND
After reaction	0.00227	0.04707	0.01720	ND	0.00521	0.00132

Pseudouridine (Ψ) was used as the internal control. The numbers displayed are the ratios (modified nucleoside/Ψ). ND: not detected. The modified nucleosides cm⁵U, ncm⁵U, mcm⁵U and Ψ were monitored at 254 nm, and cm⁵s²U, ncm⁵s²U and mcm⁵s²U were monitored at 314 nm as thiolated nucleosides absorb well at this wavelength, while nonthiolated nucleosides do not.

Alternative mechanisms in formation of the mcm⁵ side chain at wobble position

In trm9Δ or trm112Δ strains, the major species generated are ncm⁵U and ncm⁵s²U instead of the expected cm⁵U or cm⁵s²U. According to the model proposed in Figure 1, Elongator complex is required for and might directly catalyze the formation of cm⁵U. In the presence of Trm9 and Trm112p, cm⁵U is rapidly converted to mcm⁵U in tRNAs destined to contain a mcm⁵U nucleoside. Those tRNAs destined to contain ncm⁵U are not recognized by Trm9p/Trm112p and ncm⁵U is formed by an uncharacterized enzyme. In order to account for the presence of ncm⁵U and ncm⁵s²U in tRNAs that normally should contain mcm⁵U and mcm⁵s²U, one has to postulate that in the absence of Trm9p/Trm112p the uncharacterized enzyme responsible for amidation also recognizes these tRNA substrates (Figure 2). For tRNAs that should contain a s² group, the presence of a mcm⁵ side chain has been suggested to be a prerequisite for efficient thiolation [10], [12]. We suggest that the presence of ncm⁵U, but not cm⁵U, in these tRNAs also promotes efficient thiolation, resulting in accumulation of ncm⁵s²U (Table 1).

The observation that the major U₃₄ intermediates in and are ncm⁵U and ncm⁵s²U in trm9 and trm112 mutants also supports an alternative model, i. e. ncm⁵U is generated before cm⁵U (Figure 7). Such a model would require a conversion of ncm⁵U to cm⁵U before the Trm9p/Trm112p complex finally can form mcm⁵U. A similar mechanism has been described in Eubacteria that have mnm⁵ instead of mcm⁵ side chains and the first intermediate in its synthesis is cmnm⁵U [45]. The bi-functional MnmC demodifies cmnm⁵U to nm⁵U and thereafter methylates nm⁵U to form mnm⁵U [45], [46], [47]. By analogy, the Trm9p/Trm112p complex may be involved in two reactions; deamination of ncm⁵U to cm⁵U, and then catalyzing formation of mcm⁵U. The deaminase activity is not necessarily part of Trm9p or Trm112p. In the absence of Trm9p or Trm112p, ncm⁵U accumulates in tRNAs destined to contain mcm⁵s²U, like. As postulated in model 1, the presence of an ncm⁵ side chain in these tRNAs promotes thiolation, generating ncm⁵s²U. MnmC requires flavin adenine dinucleotide (FAD) as co-factor in the de-modification reaction and SAM in the methylation reaction. We performed an in vitro reaction with [³H]AdoMet in the presence or absence of FAD. We assumed if ncm⁵U is converted to cm⁵U in the presence of FAD, more [³H]-methyl groups would be incorporated into total tRNA isolated from trm9 deletion strain when FAD is included in the reaction. Reactions conducted in the presence of FAD did not increase the incorporation of [³H]-methyl into trm9 deletion tRNA, nor did it decrease the overall amount of ncm⁵U as analyzed by HPLC (data not shown). We also investigated the potential use of other cofactors in the conversion of ncm⁵U to cm⁵U such as NAD⁺ and NADP⁺ without success (data not shown). It remains to be elucidated which of these two alternative pathways for formation of mcm⁵ side chains is used.

Figure 7. An alternative model for formation of mcm⁵ side chain at wobble uridines.

Elongator complex (Elp1-Elp6) and its potential regulators catalyzes the formation of ncm⁵U. The ncm⁵U is converted to cm⁵U by an unknown mechanism in tRNA species that in their mature form should have a mcm⁵ side chain. This unknown mechanism requires Trm9p/Trm112p. In the last step, a methyl group is added to cm⁵U by Trm9p/Trm112p complex in these tRNA species. For tRNAs that should contain a s² group, presence of a mcm⁵ or ncm⁵ side chain is a prerequisite for efficient thiolation.

Supporting Information

Figure S1

HPLC analysis of modified nucleosides in isolated from wild-type, trm11Δ, lys9Δ and mtq2Δ strains. Arrows in red and black indicate expected retention time of mcm⁵U and cm⁵U, respectively. Arrow heads in red and black indicate expected retention time of m²G and ncm⁵U, respectively. (A–D), Part of the chromatogram between retention times 34 and 44 min is shown. (B), Part of the chromatogram between retention times 7 and 17 min is shown. The small peak in wild-type at 14 min represents an unrelated compound with a spectrum different from ncm⁵U. Absorbance at 254 nm (AU) was used to create the chromatograms.

(EPS)

Table S1

Strains and primers used in this study (see also [48] and [49]).

(DOC)