YebU is a m⁵C Methyltransferase Specific for 16 S rRNA Nucleotide 1407

The rRNAs in Escherichia coli contain methylations at 24 nucleotides, which collectively are important for ribosome function. Three of these methylations are m⁵C modifications located at nucleotides C967 and C1407 in 16 S rRNA and at nucleotide C1962 in 23 S rRNA. Bacterial rRNA modifications generally require specific enzymes, and only one m⁵C rRNA methyltransferase, RsmB (formerly Fmu) that methylates nucleotide C967, has previously been identified. BLAST searches of the E. coli genome revealed a single gene, yebU, with sufficient similarity to rsmB to encode a putative m⁵C RNA methyltransferase. This suggested that the yebU gene product modifies C1407 and/or C1962. Here, we analysed the E. coli rRNAs by matrix assisted laser desorption/ionization mass spectrometry and show that inactivation of the yebU gene leads to loss of methylation at C1407 in 16 S rRNA, but does not interfere with methylation at C1962 in 23 S rRNA. Purified recombinant YebU protein retains its specificity for C1407 in vitro, and methylates 30 S subunits (but not naked 16 S rRNA or 70 S ribosomes) isolated from yebU knockout strains. Nucleotide C1407 is located at a functionally active region of the 30 S subunit interface close to the P site, and YebU-directed methylation of this nucleotide seems to be conserved in bacteria. The yebU knockout strains display slower growth and reduced fitness in competition with wild-type cells. We suggest that a more appropriate designation for yebU would be the rRNA small subunit methyltransferase gene rsmF, and that the nomenclature system be extended to include the rRNA methyltransferases that still await identification.

Introduction

The ribosomal RNAs implement and monitor the majority of key interactions that occur during the course of protein synthesis on the ribosome. To discriminate between functional and non-functional interactions, the rRNAs make use of their complexly folded structures and their ability to switch conformation.¹ The rRNAs require a broad structural repertoire to carry out these tasks competently, and their structures are therefore supplemented with a range of post-transcriptional modifications.² In bacterial rRNAs, modifications consist mainly of base and sugar methylations and pseudo-uridylation,³ and these have been comprehensively mapped in the enterobacterium Escherichia coli, where there are 11 modified nucleotides in 16 S rRNA and 25 in 23 S rRNA (Table 1). Charting the spatial locations of these modifications on the ribosome crystal structures4, 5, 6, 7, 8 reveals that they cluster within several discrete regions9, 10, 11, 12 that are concerned with essential ribosomal functions including mRNA decoding and peptide bond formation. The collective importance of the rRNA modifications for efficient protein synthesis has been demonstrated by the superior performance of authentic rRNAs compared to their unmodified 16 S¹³ and 23 S counterparts.14, 15

The mechanisms for post-transcriptionally modifying rRNAs in bacteria are fundamentally different from those in archaea and eukaryotes. Pseudo-uridinylations and 2′-O-methylations, which make up the bulk of eukaryotic rRNA modifications, are guided by a variety of small nucleolar (sno)RNAs that function together with a limited set of enzymes,16, 17, 18 and similar mechanisms are used by archaea.¹⁹ In contrast, bacteria generally require a specific enzyme for each rRNA modification.¹² Our knowledge of the enzymes that are responsible for these modifications (Table 1) has unfortunately lagged considerably behind the comprehensive mapping of the rRNA locations and chemical nature of the modifications.3, 20, 21 For example, in the E. coli rRNAs there are three m⁵C modifications, at C967 and C1407 in 16 S rRNA and at C1962 in 23 S rRNA and, of these, only the enzyme that modifies C967 (the methyltransferase RsmB, formerly Fmu) has been characterized.22, 23 RsmB is a protein of 47 kDa and contains motifs typical for S-adenosyl methionine (AdoMet) dependent methyltransferases24, 25, 26 with additional concise motifs that are distinctive for the m⁵C RNA methyltransferase subfamily.²⁷ Using the RsmB sequence in an iterative database search revealed one other E. coli open reading frame, yebU, with sufficient similarity to warrant its classification as a potential m⁵C RNA methyltransferase gene;²⁷ repeating the search in the current databases has not uncovered new candidates in the E. coli genome. Thus, YebU is presently the only obvious candidate for catalyzing m⁵C methylation at the E. coli rRNA nucleotides C1407 and/or C1962.

The crystal structures of RsmB,²⁸ and most recently, YebU²⁹ have been solved. These studies represent important steps towards understanding how m⁵C RNA methyltransferases might recognize and modify their specific targets. These studies also indicate what additional information is needed. In the case of YebU, an unambiguous identification of the methylation target is required, together with the target's structural context (unassembled rRNA or ribosomal particles) that is recognized by the enzyme.

Here, we address these matters by inactivating the yebU gene in E. coli, and comparing the rRNA methylation patterns in YebU⁺ and YebU⁻ strains using matrix assisted laser desorption/ionization (MALDI) mass spectrometry (MS). After defining the rRNA target methylated by YebU in vivo, we determined the substrate in which YebU recognizes this target. A recombinant version of the YebU enzyme was constructed, purified and tested in vitro for its ability to methylate naked 16 S rRNA, 30 S subunits or 70 S ribosome couples from the YebU⁻ strain. The results unambiguously establish the substrate required for recognition and methylation by YebU, as well as the identity of the methylated nucleotide.