Identification of the 3-amino-3-carboxypropyl (acp) transferase enzyme responsible for acp3U formation at position 47 in Escherichia coli tRNAs

Abstract

tRNAs from all domains of life contain modified nucleotides. However, even for the experimentally most thoroughly characterized model organism Escherichia coli not all tRNA modification enzymes are known. In particular, no enzyme has been found yet for introducing the acp3U modification at position 47 in the variable loop of eight E. coli tRNAs. Here we identify the so far functionally uncharacterized YfiP protein as the SAM-dependent 3-amino-3-carboxypropyl transferase catalyzing this modification and thereby extend the list of known tRNA modification enzymes in E. coli. Similar to the Tsr3 enzymes that introduce acp modifications at U or m1Ψ nucleotides in rRNAs this protein contains a DTW domain suggesting that acp transfer reactions to RNA nucleotides are a general function of DTW domain containing proteins. The introduction of the acp3U-47 modification in E. coli tRNAs is promoted by the presence of the m7G-46 modification as well as by growth in rich medium. However, a deletion of the enzymes responsible for the modifications at position 46 and 47 in the variable loop of E. coli tRNAs did not lead to a clearly discernible phenotype suggesting that these two modifications play only a minor role in ensuring the proper function of tRNAs in E. coli.

Figure 1.

RNA acp modifications. (A) Modified nucleosides m¹acp³Ψ (eukaryotic 18S rRNA), acp³U (bacterial and eukaryotic tRNA and archaeal 16S rRNA) and m⁷G together with the responsible modification enzymes discussed in this work. (B) Variable loop sequences of all acp³U containing tRNAs from E. coli. (C) Secondary structure of E. coli tRNA^Arg-ICG with its posttranscriptional modifications. The acp³U (X) at position 47 in the variable loop is highlighted in red.

Figure 2.

Clustal Omega alignment of eukaryotic and archaeal Tsr3 proteins and bacterial YfiP homologs with the DTW domain and the conserved cysteines predicted to bind Zn²⁺. Ec: Escherichia coli; Vp: Variovorax paradoxus; Cs: Clostridium saccharoperbutylacetonicum; Vd: Vulcanisaeta distributa; Ss: Sulfolobus solfataricus; Hs: Homo sapiens; Sc: Saccharomyces cerevisiae.

Figure 3.

YfiP is necessary for E. coli tRNA acp modification. (A) RP-HPLC elution profiles of tRNA nucleosides from wild type and ΔyfiP E. coli cells. acp³U elutes at 20.9 min and is missing in total tRNA from the ΔyfiP strain. (B) UV and MS spectra of purified acp³U. Left: UV spectrum of acp³U in 5 mM NH₄OAc (pH 5.3) as detected in the diode array detector coupled to LC–MS/MS. Right: MS spectrum of acp³U at 180 V fragmentor voltage. (C) Primer extension analysis of tRNA^Arg-ICG acp modification including a sequencing ladder. The reverse transcriptase arrest at C48 is missing in the ΔyfiP strain and can be restored by complementation with plasmid encoded YfiP (ΔyfiP + YfiP; vc: empty vector control).

Figure 4.

Primer extension analysis of acp³U-47 modification of different putative target tRNAs using specific oligonucleotides hybridized to tRNA samples from wild type (wt) and ΔyfiP (Δ) E. coli cells. YfiP depending arrest signals of varying intensities are detectable for five different tRNAs, but not for the control tRNA^Val-VAC carrying no acp³U-47.

Figure 5.

In vitro activity of the E. coli YfiP protein. (A) Representative ITC thermograms for titrations of purified YfiP protein with either SAM (left), 5-MTA (center) or SAH (right). The resulting dissociation constants are indicated. The cofactor SAM is bound with the highest affinity. (B) Abundance of acp³U after in vitro modification of total tRNA isolated from a ΔyfiP strain with overexpressed and purified YfiP. Equal amounts of total tRNA from mock-incubated (black) and YfiP incubated samples (red) were analyzed by nucleoside LC-MS/MS. On the left the abundance of acp³U (mass transition 346 → 214) is shown and on the right uridine (mass transition 245 → 113) is shown. (C) Detection of acp³U-47 in tRNA^Arg-ICG by a primer extension stop at position C48 after in vitro incubation of total tRNA from a ΔyfiP strain with purified YfiP and SAM. The reverse transcriptase arrest at C48 is missing if the protein or the cofactor SAM are excluded from the reaction.

Figure 6.

Determinants of acp transferase activity in E. coli YfiP. (A) Truncated or point mutated YfiP variants were expressed in E. coli ΔyfiP cells, and the degree of acp³U-47 modification of tRNA^Arg-ICG was analyzed by primer extension in comparison to the wild type (wt) and ΔyfiP cells with the empty vector (vc). (B) As shown by primer extension analysis, the YfiP homolog from Variovorax paradoxum (Vp), but not from Clostridium saccharoperbutylacetonicum (Cs), is functional as an acp³U-47 transferase in E. coli ΔyfiP.

Figure 7.

Dependency of the acp³U-47 modification on the presence of the m⁷G-46 methyltransferase TrmB. (A) HPLC chromatograms of nucleosides in total tRNA from single and double deletion strains compared to the E. coli wild type. The amount of acp³U is significantly reduced in the ΔtrmB mutant, whereas a ΔyfiP deletion has no influence on m⁷G formation. (B) Primer extension analysis comparing the level of the acp³U-47 modification in wild type and ΔyfiP or ΔtrmB deleted cells. The acp³U dependent primer extension stop signal is strongly reduced by the ΔtrmB deletion for tRNA^Ile-GAU, tRNA^Lys-SUU and tRNA^Val-GAC. (C) Absolute quantification of modified nucleosides in purified tRNA isoacceptors. E. coli wild type (WT) and ΔtrmB were grown in LB medium, total tRNA was isolated and tRNA isoacceptors were purified as previously published (52). Absolute quantification of modified nucleosides was performed by LC–MS/MS using stable isotope labeled internal standards following published protocols (53). D, dihydrouridine (n = 5).

Figure 8.

Influence of the growth medium on the level of acp³U modification. (A) RP-HPLC elution profiles of tRNA nucleosides from wild type E. coli cells cultivated in LB (black) or M9 (red) medium. Q, queuosine; ς, epoxyqueuosine. (B) Absolute quantification by LC–MS/MS of modified nucleosides in purified tRNA isoacceptors from wild type cells grown in LB and M9 medium at OD 0.4 (n = 3).

Figure 9.

Phenotypic analysis of deletion mutants lacking variable loop modifications. (A) Growth curves (n = 3) for E. coli ΔyfiP lacking acp³U-47, ΔtrmB lacking m⁷G-46 and ΔyfiP ΔtrmB lacking both modifications in comparison to the wild type strain at 20°C in M9 minimal medium, at 37°C and 42°C in LB medium and in the presence of paromomycin (2.5 μg/ml) or 10 mM H₂O₂. The deletion mutants do not show differences in growth rate compared to the wild type under all conditions tested. (B) The stability of the tRNA^Arg-ICG from E. coli wild type, the ΔyfiP and the ΔyfiP ΔtrmB deletion strains was analyzed by northern blotting after cultivation with the transcription inhibitors rifampicin and nalidixic acid for the indicated amounts of time (hours). The comparison of the tRNA^Arg-ICG amounts present at equivalent time points in the different strains shows that the relative stability of this tRNA is similar in all tested strains. (C) In order to compare the aminoacylation efficiency for modified and unmodified tRNA^Arg-ICG, total RNAs of the wild type and the ΔyfiP strains were isolated and gel-separated under acidic conditions (pH 4.5), and aminoacylated and deaminoacylated tRNA^Arg-ICG was detected by northern blotting. A part of each RNA sample was incubated separately with buffer at pH 9.0 for complete deaminoacylation.