Functional ribosomal RNAs are generated from longer precursor species in every organism known. Maturation of the 5′ side of 16S rRNA in Escherichia coli is catalysed in a two-step process by the cooperative action of RNase E and RNase G. However, many bacteria lack RNase E and RNase G orthologues, raising the question as to how 16S rRNA processing occurs in these organisms. Here we show that the maturation of Bacillus subtilis 16S rRNA is also a two-step process and that the enzyme responsible for the generation of the mature 5′ end is the widely distributed essential ribonuclease YkqC/RNase J1. Depletion of B. subtilis of RNase J1 results in an accumulation of 16S rRNA precursors in vivo. The precursor species are found in polysomes suggesting that they can function in translation. Mutation of the predicted catalytic site of RNase J1 abolishes both 16S rRNA processing and cell viability. Finally, purified RNase J1 can correctly mature precursor 16S rRNA assembled in 70S ribosomes, showing that its role is direct.
Bacterial ribosomal RNAs (rRNAs) are generally co-transcribed as long 30S precursor molecules containing 16S, 23S and 5S rRNAs, and in some cases, tRNAs. At fast growth rates, transcription of rRNA operons (rrn) can account for up to 70% of total cellular transcription (Bremer and Dennis, 1996). Thus, it is easily understood how rapid and accurate processing of the different rRNA species is important for the efficient generation of mature functional ribosomes and optimal translational capacity. In Escherichia coli, processing of rRNA begins as the rrn operons are being transcribed and the finishing steps occur as late as on fully assembled ribosomes or even polysomes. The first step is the separation of the different rRNA species as they are being transcribed by the double-stranded specific ribonuclease, RNase III. RNase III cuts in the middle of long processing stalks formed by base-pairing between sequences on either side of both the mature 16S and 23S rRNA sequences, liberating precursor 16S, 23S, 5S and tRNA species from the 30S precursor. Further processing of these RNAs then occurs to generate the mature species (see Fig. 6, left). The 115 nucleotides (nts) at the 5′ side of the 16S precursor are removed cooperatively by RNase G and RNase E (Li et al., 1999a). RNase E is an essential enzyme that also plays a key role in mRNA degradation and tRNA maturation (Ono and Kuwano, 1979; Li and Deutscher, 2002; Ow and Kushner, 2002), while RNase G is a nonessential paralogue that shows 54% similarity to the N-terminal half of RNase E. The two enzymes have overlapping cleavage site specificities. RNase E first cleaves 66 nts upstream of the mature 5′ end of 16S rRNA, an event that creates a new 5′ end free from the processing stalk and increases the rate at which RNase G catalyses the final maturation step. Although these two steps occur subsequent to RNase III cleavage in wild-type cells, they can still take place in cells lacking RNase III in vivo, indicating that the order of cleavage events is not critical for accurate maturation of the 5′ end of 16S rRNA (Srivastava and Schlessinger, 1989). The enzyme responsible for removal of the 33 nts at the 3′ side of the 16S rRNA precursor is still unknown and, while its efficiency is somewhat dependent on maturation of the 5′ side by RNase E and RNase G, correct 16S 3′ ends are formed in the absence of these two enzymes in vivo (Li et al., 1999a).
Cleavage of the primary 30S transcript by RNase III in E. coli generates 23S rRNA precursors with 7 or 3 nts on the 5′ side and 7–9 nts on the 3′ side. The exoribonuclease RNase T removes the extra sequences on the 3′ side (Li et al., 1999b), while the 5′ maturase is still unknown. Like 16S rRNA, final maturation of the 5S species also occurs in two steps in E. coli. The precursor liberated by RNase III cleavage is cleaved by RNase E to produce a 5S rRNA species with 3 extra nts on either side (Ghora and Apirion, 1978). Final maturation of the 3′ side is then catalysed by RNase T (Li and Deutscher, 1995), while the enzyme that processes the 5′ side has not yet been identified.
Although ribosomal RNA is also transcribed as a long 30S precursor in Bacillus subtilis (Zingales and Colli, 1977), the maturation processes are much less well understood. Co-transcriptional separation of the different rRNA precursor species from the primary transcript by RNase III also occurs in this organism (Herskovitz and Bechhofer, 2000) and the sites of RNase III cleavage in the long 16S and 23S rRNA processing stalks can be inferred from the mapping of the extremities of precursor species (Loughney et al., 1983). However, none of the key enzymes of the final maturation steps identified so far in E. coli, RNase E, RNase G or RNase T is present in B. subtilis (Condon and Putzer, 2002). In fact, these three enzymes are lacking in many bacteria, and thus their rRNA maturation pathways are necessarily different from E. coli. Only the 5S rRNA maturation pathway has been resolved in B. subtilis to date. Mature 5S rRNA is generated in a single step by cleavage of the 5S rRNA precursor by RNase M5 on either side of a double-stranded stem (Sogin and Pace, 1974; Condon et al., 2001). The phylogenetic distribution of RNase M5, encoded by the rnmV gene, is largely restricted to the low G+C Gram-positive bacteria. Maturation of 5S rRNA by RNase M5 only requires the presence of one ribosomal protein (L18) in vitro. However, 5S rRNA precursors are assembled into functional ribosomes in rnmV mutants in vivo (Condon et al., 2001), leaving open the possibility that this step occurs much later in practice.
In this article, we identify the enzyme responsible for the generation of the mature 5′ extremity of B. subtilis 16S rRNA. YkqC, also known as RNase J1 (Even et al., 2005), is an essential enzyme that preferentially catalyses accurate processing of the 16S rRNA precursor in assembled ribosomes. As in E. coli, final maturation appears to occur through a two-step mechanism, with an intermediate product that contains 38 nts upstream of the mature 5′ end and that is no longer connected to the 16S rRNA processing stalk.
YkqC plays a role in 16S ribosomal RNA maturation
The systematic mutational analysis of genes of unknown function, the so-called ‘y’ genes, led to the prediction that 271 genes are essential for B. subtilis viability under standard growth conditions in the laboratory (Kobayashi et al., 2003). Eight of these genes (ykqC, ymdA, yloQ, ywlC, yacA, ydiB, ylaN, yqeI) coded for proteins either with no previously characterized domains or containing domains found in other ribonucleases. In an effort to identify new essential ribonucleases in B. subtilis, IPTG-dependent conditional mutants were made for each of these genes by Campbell integration of the pMUTIN-4M vector designed for this purpose (Vagner et al., 1998). Cells were depleted for the particular gene products by growth in the absence of IPTG and screened for defects in 16S rRNA maturation (Li et al., 1999a). Depletion of B. subtilis cells of one of these gene products, YkqC, in strain CCB034 led to an accumulation of four 16S rRNA precursor species compared with non-depleted strains (Fig. 1A). These were revealed by primer extension assay of total RNA using an oligonucleotide (CC058) that hybridizes to nts 20–39 of the mature 16S rRNA sequence. The 5′ ends of the precursor cDNAs were identified by sequencing a cloned rrnW 16S gene with the same oligo. The 5′ ends of the precursor species mapped to approximately 38, 76, 102 and 140 nts upstream of the predicted 5′ end of mature 16S rRNA. The 5′ ends of the +38 and +76 precursors are shown on the 16S precursor sequence in Fig. 1B. The +76 product corresponds well to a predicted RNase III cleavage site mapped previously (Loughney et al., 1983). The +102 and +140 species correspond to the predicted sizes of precursors equivalent to +38 and +76 from four of the 10 rRNA operons of B. subtilis (rrnB, D, E, O) which contain a 64 nt insertion immediately upstream of the mature 16S rRNA sequence (Fig. 1). These experiments suggest that YkqC plays an important role in the generation of the mature 5′ end of 16S rRNA.
Multiple bands were observed at each of the 5′ ends identified, most clearly resolved at the mature 16S rRNA 5′ end, M(0). This could be a result of some variation in the processing site for the enzyme(s) that generate these products. It could also be explained by heterogeneity in the mature 16S rRNA sequence, although the sequences of the 10 16S rRNA genes are identical in the first 40 nts according to the Subtilist database (http://genolist.pasteur.fr/SubtiList/). In an attempt to resolve this issue, we also mapped the 5′ ends of individual 16S rRNA precursors by 5′/3′ RACE. Total RNAs were circularized with RNA ligase, and cDNA synthesis was primed across the 5′/3′ junction and then amplified by PCR. Similar reactions were performed using divergent oligos specific for 23S rRNA. A reverse transcription polymerase chain reaction (RT-PCR) product of the expected size for mature 16S rRNA (160 nts) was seen with the +IPTG RNA sample, while the –IPTG sample yielded two major additional products that were larger in size (Fig. 2). The RT-PCR reaction specific for 23S rRNA only yielded products predicted for the mature species (177 nts), suggesting that YkqC does not participate in 23S rRNA maturation. DNA fragments 1, 2 and 3 were excised and purified from the agarose gel and cloned in pBluescript KS+ for sequencing. Sequence analysis of clones revealed 5′ ends at +76, +38 for fragments 1 and 2, respectively, while the 5′ end of clones of fragment 3 were primarily at the first of three U-residues at M(0). The 5′ end of one fragment-3 clone was mapped to +2 and the 2 extra nts corresponded to one of the four rrn operons with a 64 nt insertion, suggesting that the insertion has a minor effect on processing site selection. Some sequence heterogeneity was also seen at position 13 of mature 16S rRNA, where some RT-PCR products lacked an A residue at this position. These experiments suggest that the heterogeneity seen at M(0) is a combination of heterogeneity in both the sequence of the first 40 nts of the 10 rrn operons and in processing site specificity.
16S rRNA precursors are found in functional ribosomes in vivo
There has been some ambiguity about whether 16S rRNA precursors inhibit ribosome activity in E. coli (Mangiarotti et al., 1974; Wireman and Sypherd, 1974). We were therefore curious to know whether the 16S precursor species that accumulated under conditions of YkqC depletion in B. subtilis could be assembled into functional ribosomes. We performed sucrose gradient analyses of whole-cell extracts of strain CCB034 grown in the presence and absence of IPTG, and treated with chloramphenicol before harvesting to allow stabilization of polysomes. No major differences in polysome profile were observed in strains depleted of YkqC compared either with the same strain grown in the presence of IPTG or with the wild-type strain (Fig. S1 in Supplementary material). RNAs were isolated from the peak sucrose gradient fractions and analysed for the presence of precursor RNAs by primer extension assay as described above. All four 16S rRNA precursor species (+38, +76, +102 and +140) were clearly detected in the 30S and 70S subunit peaks, and in the disome (P2) and trisome (P3) peaks of the polysome portion of the gradient (Fig. 3A). Thus, 16S rRNA precursors can be found in functioning ribosomes, suggesting that they do not completely inhibit translation in B. subtilis in vivo. Interestingly, the amount of precursor species was significantly higher in 30S compared with 70S subunits isolated from YkqC-expressing cells (Fig. 3B), suggesting that 16S rRNA processing occurs after association of the 30S and 50S subunits. Under conditions of YkqC depletion, the proportion of mature 16S rRNA in ribosomal subunits and in polysomes was significantly higher than that found in total RNA (Fig. 3B), again suggesting that maturation follows subunit assembly and that these cells contain a pool of free 16S precursors that is not assembled into 30S subunits.
Bacillus subtilis contains a 46% identical, nonessential paralogue of YkqC, called YmfA. To determine whether YmfA might also play a role in 16S rRNA processing, we constructed a strain containing a deletion in the ymfA gene (CCB078) and a strain in which this mutation was combined with the Pspac-ykqC conditional mutation (CCB079). Whole-cell extracts of both of these strains were also subjected to sucrose gradient analysis. Deletion of the ymfA gene alone had no effect on 16S rRNA maturation and further depletion of YkqC in cells lacking YmfA did not significantly aggravate the defect in rRNA processing compared with cells depleted of YkqC alone (Fig. 3). Thus, YmfA appears to have a limited effect on 16S rRNA maturation in B. subtilis under these conditions, despite its similarity to YkqC.
YkqC catalytic activity is essential in B. subtilis and is required for 16S rRNA maturation
As the final steps of rRNA maturation are believed to occur after subunit assembly in E. coli, mutations in genes involved in ribosome assembly per se often lead to indirect defects on rRNA processing (Charollais et al., 2003; 2004; Hage and Alix, 2004). Thus, it was possible that the effects of YkqC on 16S rRNA maturation were indirect. YkqC is a member of the β-lactamase family of metallo-hydrolases that includes endoribonucleases such as RNase Z, involved in maturation of the 3′ end of tRNA (Pellegrini et al., 2003). Indeed, this was the initial justification for including YkqC in the screen for enzymes involved in 16S rRNA processing. A recent study has since shown that YkqC has endoribonuclease activity that is functionally similar to RNase E and the enzyme was renamed RNase J1 (Even et al., 2005). Enzymes of the β-lactamase family have a signature HxHxDH motif involved in the co-ordination of a pair of metal ions required for catalysis (Li de la Sierra-Gallay et al., 2005). To confirm that the effect of YkqC/RNase J1 depletion on 16S rRNA maturation was directly related to its ribonuclease activity rather than an indirect ribosome assembly defect, we mutated the central histidine residue (H76) of the HxHxDH motif to alanine and placed the mutant gene on the chromosome under control of the xylose-dependent promoter, PxylA (Bhavsar et al., 2001). Mutations in this residue of RNase Z severely reduce its catalytic activity (Minagawa et al., 2004; Li de la Sierra-Gallay et al., 2006) and we expected the same would hold true for YkqC/RNase J1. The Pxyl-ykqC (H76A) construct was integrated into the amyE locus of CCB034, creating strain RB571, and allowed independent induction of the wild-type and mutant copies of the ykqC gene by IPTG and xylose respectively. A control strain, RB569, containing a wild-type copy of the ykqC gene under the control of each promoter was also constructed. While strain RB569 could grow on plates containing either IPTG or xylose, strain RB571, which contains the Pxyl-ykqC (H76A) construct, could only form colonies in the presence of IPTG (Fig. 4A). This experiment shows not only that YkqC/RNase J1 is essential in B. subtilis, but, more precisely, that its vital function is its catalytic activity.
To confirm that the H76A mutation had an effect on 16S rRNA processing, we isolated total RNA from RB569 and RB571 cells grown in xylose-containing medium in the presence and absence of IPTG. We then performed primer extension assays to analyse the nature of the 16S rRNAs under these conditions. While depletion of RB569 cells of IPTG-inducible YkqC/RNase J1 had no effect on 16S rRNA maturation, depletion of RB571 cells of wild-type YkqC/RNase J1 resulted in an accumulation of the characteristic 16S rRNA precursors seen in strains lacking this enzyme (Fig. 4B). Thus, the defect in 16S rRNA processing seen under conditions of YkqC depletion is indeed due to a lack of its ribonuclease activity and for the rest of the article we will thus use the RNase J1 nomenclature.
Maturation of 16S rRNA by RNase J1 in vitro
Initial attempts to demonstrate maturation of an in vitro transcribed 16S rRNA precursor RNA by purified B. subtilis RNase J1 resulted in inaccurate precursor processing (data not shown). In addition, the in vivo data described above suggested that RNase J1 preferentially processes 16S rRNA precursors after association of the 30S and 50S ribosomal subunits. Therefore, to demonstrate 16S rRNA maturation activity in vitro, we isolated 70S ribosomes from strains depleted for RNase J1 and thus contain a significant proportion of 16S rRNA in its precursor form. These immature ribosomes were then incubated with increasing concentrations of whole-cell extracts of CCB034 grown either in the presence or in the absence of IPTG, and assayed for 16S rRNA maturation by primer extension assay. RNase J1-depleted extracts were incapable of further maturation of 16S rRNA in immature 70S ribosomes (Fig. 5, lanes 6–9), while extracts of cells grown in the presence of IPTG showed clear rRNA processing activity (lanes 2–5), indicated by the accumulation of the M(0) product. Incubation of purified RNase J1 with immature 70S ribosomes produced a similar increase in the amount of 16S rRNA processing, showing that the effect of RNase J1 on maturation was direct (lanes 10–13). Addition of purified RNase J1 to depleted CCB034 extracts showed slightly higher levels of 16S rRNA maturation than RNase J1 alone (compare lane 12 to 16 and 13 to 17), suggesting that the whole-cell extract helps stabilize RNase J1 activity.
In this article we show that the maturation of the 5′ end of 16S rRNA is catalysed by RNase J1 in B. subtilis. First, depletion of RNase J1 in B. subtilis cells in vivo results in an accumulation of four 16S rRNA precursor molecules from two different classes of rrn operons. The two smaller species correspond in size to precursor transcripts from the rrnA, G, H, I, J and W operons, generated by a predicted cleavage by RNase III at +76 relative to the mature 5′ end of 16S rRNA and a processing intermediate, generated by an unknown enzyme at +38. The two larger species correspond in size to the equivalent products derived from transcripts of rrnB, D, E and O, which are 64 nts longer due to an insertion just upstream of the mature 16S sequence. Second, we show that RNase J1 ribonucleolytic activity is required for B. subtilis cell viability and correct 16S rRNA processing in vivo. Finally, we show that purified RNase J1 can process 16S rRNA in immature 70S ribosomes in vitro, indicating that the effect of RNase J1 depletion on 16S rRNA maturation is direct.
RNase J1 is a member of the metallo-β-lactamase family that includes a rapidly growing number of ribonucleases, such as RNase Z, involved in maturation of the 3′ end of tRNAs (Pellegrini et al., 2003), CPSF-73, a subunit of the eukaryotic cleavage and polyadenylation specificity factor (Dominski et al., 2005), and Int11, responsible for the generation of the 3′ end of U snRNAs (Baillat et al., 2005). RNase J1 and its paralogue, RNase J2 (YmfA), were shown to catalyse the RNase E-like cleavage of the B. subtilis thrS leader mRNA that originally suggested the presence of an analogous enzyme in B. subtilis (Condon et al., 1997). Like RNase E, RNases J1 and J2 showed sensitivity to the phosphorylation state of the 5′ end of the mRNA and together had a modest effect on global mRNA stability (Even et al., 2005). They thus share many of the properties of RNase E and have been proposed to be the functional equivalents of this enzyme in B. subtilis. Although these observations would suggest that RNase J1 generates the mature 5′ end of 16S by endonucleolytic cleavage, we have been unable to detect a 5′ cleavage product in vivo by Northern blot (data not shown).
RNase J1/J2 is widely distributed throughout the bacterial and archaeal kingdoms, and is also present in some eukaryotes (Even et al., 2005). Two hundred and nine of the 377 completed prokaryotic genomes contain RNase J homologues with an expect value of 2e−27 or better. In many cases, a single orthologue exists, with about equal homology to RNase J1 and J2. However, multiple copies are often found in the low G+C Gram-positive bacteria, with Bacillus anthracis having three, and Bacillus thuringiensis as many as four RNase J paralogues. We thus suspect that the RNase J-dependent pathway will prove to be a major pathway for 16S rRNA maturation throughout the bacterial kingdom. This highlights the importance of using B. subtilis as an alternative or complementary model organism to E. coli to get a more global view of bacterial mechanisms and pathways.
The 5′ processing pathways of 16S rRNA in E. coli and B. subtilis show some similarities that are intriguing. Final maturation appears to occur in a two-step mechanism in both cases (Fig. 6). In E. coli, the 115 nt 5′ extension is first cleaved at +66 by RNase E creating a new 5′ end free from the processing stalk and increasing the efficiency of the final maturation steps at both the 5′ and 3′ extremities. Depletion of RNase J1 in B. subtilis results in the accumulation of a +38 precursor, suggesting that an intermediate processing step occurs at this site that would also serve to sever the connection to the processing stalk. The +38 and +102 precursors are preferentially converted to mature 16S rRNA by RNase J1 in the in vitro processing assays on immature 70S ribosomes (Fig. 5); the RNase III-generated precursors at +76 and +140 are an order of magnitude more resistant to RNase J1 activity. As RNase J1 is expected to have similar single-stranded 5′ end-dependence to RNase E in E. coli, it presumably has difficulty gaining access to the 5′ end of the rRNA precursor when it is in a double-stranded conformation. This would suggest that generation of the +38 intermediate and cutting the connection with the double-stranded processing stalk is an important step in 16S rRNA maturation by RNase JI. The analogy to the E. coli processing pathway appears to extend only partially to the enzymes involved. Although one of the final maturation steps, in this case at M(0), is catalysed by an essential functional homologue of RNase E, the other step does not appear to be catalysed by its nonessential paralogue RNase J2 (YmfA), at least under the conditions tested. The enzyme responsible for cleavage at +38 therefore remains to be identified. We also do not know whether maturation of the 5′ side of 16S rRNA has an effect on the efficiency of the 3′ processing reaction (step 4). Although the RT-PCR products synthesized from the +38 and +76 precursors species both had mature 3′ ends, we cannot rule out the possibility that these are derived from concatemers of mature and immature species, rather than circular products of RNA ligation of a single precursor molecule.
While 23S rRNA precursors can be assembled into functional ribosomes in E. coli, there has been some debate about whether the same is true for 16S precursors. Although 16S rRNA precursors can be found in polysomes in vivo (Mangiarotti et al., 1974), 30S subunits reconstituted using precursor RNA are not functional in vitro (Wireman and Sypherd, 1974). It has been proposed that the presence of 16S precursors in polysomes could be explained if the final processing steps occurred during formation of the 70S translation initiation complex before the first elongation step occurs (Srivastava and Schlessinger, 1990). If this were the case, one would expect to see progressively less 16S rRNA precursor in monosomes, disomes and trisomes, as only an initiating ribosome in each case would account for the presence of precursor. However, B. subtilis cells depleted of RNase J1 have the same amount of mature 16S rRNA in these three fractions (Fig. 3), suggesting that, in this organism, 16S rRNA precursors are at least partially functional in vivo. The apparent functional defect in E. coli ribosomes assembled with 16S rRNA precursor in vitro may therefore reflect a limitation in in vitro reconstitution methods or in the ability of in vitro translation assays to detect low levels of ribosome activity. We do not yet know whether the lack of 16S rRNA processing explains the lethality of RNase J1-deficient strains. As 16S precursor rRNA can be assembled into functional ribosomes, it may be that RNase J1 is involved in the maturation of some other essential substrate RNA. However, we cannot rule out the possibility that the translation efficiency of ribosomes containing immature 16S rRNA is simply too poor to support bacterial growth.
We were able to demonstrate maturation of 16S rRNA on 70S ribosomes with purified RNase J1 in vitro, but processing of in vitro transcribed primary transcripts is inaccurate. Furthermore, 16S rRNA precursors were detectable in 30S ribosomal subunits even in cells expressing RNase J1, but were no longer present in 70S ribosome fraction (Fig. 3). This suggests that the preferred in vivo substrate for RNase J1 is assembled ribosomes. The first purification step for RNases J1 and J2 by Even et al. (2005) was a high salt ribosome wash, consistent with an association between this enzyme and the ribosome. We have also observed an association of RNase J1 with purified 30S subunits (data not shown). Finally, Hunt et al. (2006) have recently exploited a GFP–YkqC fusion protein to show that YkqC/RNase J1 colocalizes with ribosomes at the poles of B. subtilis cells. These observations are also consistent with a role for RNase J1 late in the ribosome assembly process.
Bacterial strains used in this study
Strain CCB034 (Pspac-ykqC mls, pMAP65 km) was constructed by amplifying a 314 nt fragment corresponding to the N-terminal portion of the ykqC coding region using oligos CC013 and CC014 (Table S1 in Supplementary material). The resulting PCR fragment was cut with EcoRI and BamHI and cloned into pMUTIN-4M (Vagner et al., 1998) for Campbell integration into the chromosome of wild-type B. subtilis W168 to make strain CCB031 (Pspac-ykqC mls). This strain was transformed with plasmid pMAP65 (Petit et al., 1998), providing extra copies of the LacI repressor, to create strain CCB034.
Strain CCB078 (ymfA::spc) was constructed by amplifying a 553 nt fragment upstream and a 428 nt fragment downstream of ymfA using oligo pairs CC175/CC176 and CC190/CC191 (Table S1 in Supplementary material) respectively. The upstream fragment was cut with BamHI and HindIII; the downstream fragment with XhoI and KpnI. These fragments were cloned sequentially on either side of the spectinomycin resistance (spc) cassette in plasmid pBS-Spc to create plasmid pBS-YmfA::Spc. Plasmid pBS-Spc contains the spc cassette from pDG1727 (Guerout-Fleury et al., 1995) between the HindIII and XhoI sites of pBluescript KS+ (Stratagene). Transformation of wild-type B. subtilis W168 with pBS-YmfA::Spc digested with BamHI yielded CCB078 by double cross-over.
Strain CCB079 (Pspac-ykqC mls, ymfA::spc, pMAP65 km) was made by transforming strain CCB034 with chromosomal DNA isolated from strain CCB078.
Strain RB569 (Pspac-ykqC mls, PxylA-ykqC::amyE cm, pMAP65 km) was made by integration (double cross-over) of plasmid pCT5 in strain CCB034. Plasmid pCT5 (PxylA-ykqC) was made by cloning the ykqC gene (amplified using oligos CT20 and CT21) into the PacI/NheI sites of pSWEET (Bhavsar et al., 2004) under the control of the PxylA promoter. pCT5 was linearized with PstI prior to transformation.
Strain RB571 [Pspac-ykqC mls, PxylA-ykqC(H76A)::amyE cm, pMAP65 km] was made by integration (double cross-over) of plasmid pCT6 in strain CCB034. Plasmid pCT6 [PxylA-ykqC(H76A)] was derived from pCT5 by site-directed mutagenesis using a QuikChange II XL Site-Directed Mutagenesis kit (Stratagene) and primers CT28 and CT29. pCT6 was linearized with PstI prior to transformation.
Culture conditions for YkqC depletion
Overnight cultures (maximum 14 h) were typically grown in 5 ml of 2× YT containing 0.5% glucose, 1 mM IPTG and relevant antibiotics and final OD600 was measured. Cells were pelleted, washed twice in 2× YT and resuspended in the same volume of 2× YT. Parallel cultures containing or lacking 1 mM IPTG were then inoculated at a starting OD600 of 0.1 in 2× YT containing 0.5% glucose and relevant antibiotics. IPTG-containing cultures were grown to an OD600 of around 0.6 before harvesting. YkqC-depleted cultures typically started to plateau at OD600 between 0.4 and 0.6 after about 4–4.5 h growth under these conditions. Cells were harvested at this point for RNA isolation or application to sucrose gradients.
Testing the viability of the H76A mutation in YkqC was done on Luria–Bertani (LB) agar plates containing 5 μg ml−1 kanamycin (km), 0.5 μg ml−1 lincomycin and 12.5 μg ml−1 erythromycin (mls). IPTG (1 mM) and 2% xylose were added where indicated. Plates were incubated overnight at 37°C.
Primer extension assays
Total RNA was isolated by Qiagen Midiprep kit according to the manufacturer's instructions. And 0.5 pmol of 5′-labelled (32P) oligo CC058 (Table S1) was added to 5 μg of RNA in 5 μl of final volume RT buffer (50 mM Tris-HCl, pH 8.3, 10 mM MgCl2, 80 mM KCl) and denatured at 75°C for 4 min. The denatured mixture was frozen in a dry ice/ethanol bath for 2 min before being transferred to ice. A 5.2 μl mix of 2 mM each dNTP, 8 mM dithiothreitol (DTT), 4 units AMV reverse transcriptase in RT buffer was then added. Reaction mixtures were incubated for 30 min at 45°C, stopped with 5 μl of 95% formamide, 20 mM EDTA, 0.05% bromophenol blue, 0.05% xylene cyanol and loaded on 5% sequencing gels.
Total RNA (5 μg) was circularized with RNA ligase (Biolabs) and subjected to RT-PCR across the 5′/3′ junction, using oligo pairs CC254/CCC255 (16S rRNA) and CC256/CC257 (23S rRNA) according to the manufacturer's (Invitrogen) instructions. PCR fragments were isolated from agarose gels, digested with XbaI and SacI and cloned in pBluescript KS+ for sequencing. Sequencing was performed by MWG (Germany).
The sucrose gradient protocol for polysome separation was adapted from Gould et al. (1969). Two hundred millilitres of wild-type or YkqC-depleted cultures at OD600 around 0.6 were treated with freshly prepared chloramphenicol (100 μg ml−1) for 5 min. Cells were spun down, washed in 20 ml of Buffer A (10 mM Tris-HCl pH 7.5, 60 mM KCl, 10 MgCl2, 6 mM β-mercaptoethanol) containing 1 mM PMSF at 4°C and pelleted again. Cells were stored as frozen pellets at −20°C. Pellets were resuspended in 1 ml of Buffer A containing 20% sucrose and 5 μg ml−1 fresh lysozyme and incubated at 25°C for 15 min. Protoplasts were spun down for 15 min at 13 200 r.p.m. in a bench-top centrifuge (Eppendorf) and resuspended in 1 ml of Buffer A (4°C). Then 22.5 μl of 10% deoxycholate were added and (RNase free) DNase I to 3 μg ml−1 and mixed by pipetting through a cut-off 1 ml pipette tip. The mixture was incubated on ice for 30 min, spun down (13 200 r.p.m. in a bench-top centrifuge) and the supernatant was loaded on a 29 ml 10–40% sucrose gradient in Buffer A in SW28 (Beckman) centrifuge tubes. Gradients were centrifuged for 15 h at 17 000 r.p.m. at 4°C and collected at 2 ml min−1 using an ISCO (Lincoln, NE) apparatus. Individual fractions were phenol extracted, ethanol precipitated and resuspended in water for analysis of rRNAs.
Overexpression and purification of YkqC
The wild-type B. subtilis ykqC gene was amplified by PCR using oligos CC071 and CC237. The resulting PCR fragment was digested with BamHI and HindIII and cloned in pET28 (Novagen) to yield plasmid pET28-YkqC. Plasmid pET28-YkqC was transformed into BL21 CodonPlus cells (Stratagene) for overexpression in E. coli. Cells were grown to OD600 of 0.6 in 2× YT, and 1 mM IPTG added for 3 h. YkqC was purified as described in Even et al. (2005) without the ribosome wash step and concentrated to 5.6 mg ml−1.
In vitro processing assays
70S ribosomes were isolated from YkqC-depleted CCB034 cells on sucrose gradients as described above. One microgram of 70S ribosomes was incubated with either dilutions of purified YkqC or dilutions of YkqC-depleted extracts in 5 μl reaction volume containing 20 mM Tris-HCl pH 8.8 mM MgCl2, 100 mM NH4Cl, 0.1 mM DTT. Reactions were incubated at 37°C for 30 min, phenol extracted, ethanol precipitated and subjected to primer extension assays as described above. Extracts of YkqC-depleted or YkqC-sufficient cells were prepared from 20 ml of cultures. Cells were pelleted, washed and resuspended in 500 μl of Buffer A containing 1 mM PMSF for sonication. Lysates were clarified by centrifugation at 13 200 r.p.m. in a bench-top centrifuge at 4°C. One microlitre of dilutions of these extracts was used in in vitro processing assays.
We thank Lionel Bénard for helpful discussion and H. Putzer for communicating results prior to publication. C.C. was supported by funds from the CNRS (UPR 9073), Université Paris VII-Denis Diderot AND PRFMMIP 2001/2003, ACI Jeunes Chercheurs from the Ministère de l'Education Nationale, and the Agence Nationale de la Recherche (ANR). R.A.B. was supported by start-up funds and a New Investigator Award (IRGP) from Michigan State University.