16S ribosomal RNA modification drives transcript-specific translation efficiency

SUMMARY

Bacterial ribosomal RNAs (rRNAs) are decorated with conserved nucleotide modifications, but the functionality of these modifications is often underexplored. MraW (RsmH) is a 16S rRNA methyltransferase. Here, we report that deletion of mraW corrects a late-stage sporulation defect in Bacillus subtilis by bypassing a sporulation checkpoint. Ribosomes purified from ΔmraW cells display a ~2-fold decrease in translation efficiency; in vivo, ΔmraW cells produced decreased levels of the sporulation checkpoint protein CmpA. Reduced production of CmpA is mediated by mRNA sequences that form a stem-loop which occludes early cmpA codons. Proteomic analysis revealed that MraW mediates production of multiple proteins, some of whose mRNA form similar structures as the cmpA transcript. We propose that MraW modification of 16S rRNA enhances translation efficiency in general, and that specific transcripts have evolved structural features that fine-tune protein levels. This type of control may be prevalent in bacteria which exhibit uncoupled transcription and translation.

Graphical abstract

In brief

Park et al. show that a conserved 16S rRNA modification in bacteria enables transcript-specific translation by overcoming inhibitory mRNA secondary structures to fine-tune the production of key regulatory proteins. Their findings establish ribosome modification as a mechanism for controlling protein dosage during development.

INTRODUCTION

Ribosomes are composed of myriad proteins and three RNA species called ribosomal RNAs (rRNA)^1–5 that synthesize proteins from mRNA templates. In bacteria, the active ribosome is composed of the 30S small subunit, 50S large subunit, and various translation factors that participate in initiation, elongation, and eventual termination of translation.^1,3–5 The essentiality of translation has made it a successful target for antibiotics, but many aspects regarding target specificity during translation remain underexplored. For example, instead of cells harboring uniform populations of ribosomes, studies have suggested that ribosomes in a cell may be differentially modified, and that this ribosomal heterogeneity may be instrumental in the ability of bacteria to rapidly respond to changing environments.^6–12 In any case, chemical modifications of the ribosome may represent a mechanism that dictates protein synthesis levels of key factors during stress responses and developmental programs.

One common ribosomal modification target is rRNAs. The 16S rRNA (in the 30S subunit) and 23S rRNA (in the 50S subunit) contain widely conserved nucleotide modifications that impact ribosome assembly, translation fidelity, and antimicrobial resistance.^13–17 Removal of any single modification typically does not impact cell survival, suggesting that these modifications have more nuanced or condition-specific roles in modulating ribosome activity.¹³ One such rRNA modification is the dual methylation of C1402/C1410 (positions of the modified nucleotides in Escherichia coli or Bacillus subtilis, respectively) in the 16S rRNA. This modification is present in almost all bacteria and consists of methylation at the N4 and 2′O positions (m⁴Cm), catalyzed by the methyltransferases MraW (RsmH) and RsmI, respectively.^5,13,18–22 Previous reports have shown that both modifications occur after the 16S rRNA has assembled into the 30S subunit.^20,23 Although deletion of either mraW or rsmI results only in minor growth defects in E. coli^20,24 and B. subtilis²⁵ in laboratory media, it does result in attenuated virulence in E. coli O157:H7²⁶ and Staphylococcus aureus.²⁷ Additional work in E. coli showed that almost 100% of 16S rRNAs contain the dual-modified C1402 residue²⁸ and that loss of MraW or RsmI impaired start codon selection by increasing mis-initiation at AUU start codons.²⁰ Consistent with this in vivo observation, structural studies revealed that the methylation at N4 of C1402 engages in direct interactions with the phosphate backbone of the mRNA start codon and participates in interactions that likely stabilize the P site of the ribosome.^5,20 Despite the conservation of C1402 and widespread conservation of the methyltransferases responsible for its modification, the absence of a strong growth phenotype upon loss of these modifications suggests a condition- or transcript-specific role for C1402 methylation.

Previous studies have indicated that, in addition to the well-studied regulation of transcription,^29,30 regulation of translation also ensures the proper completion of spore formation (sporulation) in the Gram-positive bacterium B. subtilis.³¹ The proper initiation of sporulation itself requires translational regulation via the activity of translation elongation factors that enhance the ability of ribosomes to synthesize proteins with polyproline stretches.³² While proline codons themselves are in general difficult for the ribosome to decode, polyproline tracts augment this problem and can lead to ribosome stalling. Accordingly, several factors have evolved to aid the ribosome in synthesizing proteins with these sequences.^32–35

Sporulation is a developmental program that transforms a vegetative cell into a dormant spore when the cell senses extreme nutrient depletion^36,37 (Figure 1A); spores then germinate and continue vegetative growth when favorable environmental conditions resume.³⁸ Spores are highly resistant to environmental insults,³⁹ and this resistance is partially mediated by two cell surface structures that encase the spore: a thick proteinaceous “coat” and a peptidoglycan “cortex” that is sandwiched between two membranes.⁴⁰ The basement layer of the spore coat is composed of the ATPase SpoIVA,^41–43 a cytoskeletal protein that forms static homopolymers upon ATP hydrolysis, which serves as a platform for the subsequent recruitment and assembly of all other coat proteins.^44–46 Disrupting the polymerization of SpoIVA impairs not only coat assembly, but also cortex assembly via activation of the coat assembly checkpoint⁴⁷ that contains at least two independent pathways: the SpoVID-mediated coat assembly monitoring system^48,49 and the CmpA-mediated degradation system, which promotes the proteolysis of defective SpoIVA subunits that do not polymerize.⁵⁰ Together, both systems ensure that spores with mis-assembled coats do not progress to the cortex assembly step and can therefore be cleared from the population.⁵¹

Figure 1. Deletion of mraW and its partner rRNA modifying enzyme suppresses mutant SpoIVA sporulation defects.

(A) Schematic of the major steps of sporulation. Membranes depicted in yellow; cell wall, gray. MC: mother cell; FS: forespore.

(B) Sporulation efficiencies (determined as resistance to wet heat at 80°C for 20 min) of strains harboring the indicated alleles of spoIVA in the presence or absence (Δ) of cmpA, spoVID, or mraW. thrC is a chromosomal locus into which the indicated allele of spoIVA is integrated to complement the spoIVA deletion. Bars represent the mean (n = 3 biological replicates); errors: SD.

(C) ATPase activity (measured as generation of free phosphate) of purified WT SpoIVA (blue) and SpoIVA^L59P (pink). Data were fit to the Michaelis-Menten enzyme saturation model. Data points represent the mean (n = 3 independent trials); errors: S.E.

(D) Subcellular localization of GFP-SpoIVA (left) and GFP-SpoIVA^L59P (right) using fluorescence microscopy. Top row: fluorescence from GFP (green); bottom row: overlay, GFP and membranes visualized using FM4–64 (red). Scale bars, 1 μM.

(E–G) Sporulation efficiencies of cells harboring (E) spoIVA^L* in the presence or absence of mraW or mraW* (D52A and D100A, respectively), (F) the indicated allele of spoIVA in the presence or absence of mraW, or (G) the indicated allele of spoIVA in the presence or absence of yabC. amyE is a chromosomal locus into which the indicated allele of mraW is integrated to complement the mraW deletion. Bars represent the mean (n = 3 biological replicates); errors: SD.

Here, we report that inactivation of mraW bypasses the coat assembly checkpoint and suppresses sporulation defects associated with defective alleles of spoIVA. We show that ribosomes isolated from ΔmraW cells are slightly defective for translation in vitro. In vivo, MraW modification of 16S rRNA facilitates the efficient translation of cmpA mRNA. We observed that cmpA mRNA harbors a conserved stem-loop structure that necessitates MraW modification of C1410 for efficient translation. Prote- omic analysis revealed other MraW-dependent transcripts in stationary phase that also harbor a similar stem-loop structure. We propose that certain mRNA transcripts, particularly those that encode inhibitory proteins like CmpA that are only required in limiting quantities, evolved stem-loop structures that are intrinsically difficult to translate and require methylation of C1410 in 16S rRNA to serve as a mechanism to modulate cellular protein levels.

RESULTS

SpoIVA^L59P is defective for ATP hydrolysis and sporulation

The major structural component of the spore coat basement layer in B. subtilis is a cytoskeletal protein called SpoIVA.⁴¹ SpoIVA is a TRAFAC family NTPase that hydrolyzes ATP to drive its irreversible polymerization to form a platform at the forespore surface on which subsequent spore coat proteins assemble.^45,46 The SpoIVA^L59P variant was previously identified using PCR mutagenesis as a variant that confers a sporulation defect.⁵² Whereas other well-characterized loss-of-function SpoIVA variants harbored disruptions in the nucleotide-binding site of SpoIVA, position L59 is located far away from the active site. We therefore sought to explore the defect associated with the L59P substitution. Cells harboring spoIVA^L59P (hereafter, simply “spoIVA^L*”) displayed a ~1000-fold reduction in sporulation efficiency compared to wild-type (WT) (Figure 1B), despite the variant being produced at similar levels as the WT SpoIVA (Figure S1A). We next purified recombinant SpoIVA^L* and measured ATP hydrolysis at increasing nucleotide concentrations in vitro to produce saturation curves that revealed the substrate turnover rate (k_cat). Compared to WT SpoIVA, which displayed a k_cat of 0.83 pmol min⁻¹ pmol⁻¹ SpoIVA,⁴³ SpoIVA^L* did not specifically hydrolyze ATP, as evidenced by the failure of the reaction to reach saturation and display Michaelis-Menten kinetics (Figure 1C). Examination of the subcellular localization of SpoIVA^L*-GFP revealed that, similar to SpoIVA variants harboring disruptions in the nucleotide binding site, SpoIVA^L*-GFP displays a slight localization defect in that it fails to uniformly polymerize around the forespore and instead localizes as caps on the mother cell-distal and -proximal sides of the forespore (Figure 1D).⁴⁸ We conclude that the L59P substitution likely causes a structural defect in SpoIVA that disrupts its ability to hydrolyze ATP and therefore inhibits sporulation by activating the coat assembly checkpoint.⁴⁷

Deletion of mraW suppresses several spoIVA mutants

We took advantage of the severe sporulation defect of spoIVA^L* to isolate spontaneous mutants that would correct this defect. Cells were induced to sporulate, after which nonsporulating and poorly sporulating cells were killed by exposure to 80°C for 20 min. Surviving cells were enriched by repeated dilution in fresh sporulation media where they could germinate and re-sporulate. Whole-genome sequencing revealed a guanidyl insertion in the mraW (rsmH) gene at position 597, resulting in a frameshift that altered the next three amino acids and introduced a premature stop codon that would truncate the protein by 109 residues (approximately one-third of the protein). To test if this mutation resulted in a loss of function in MraW, we examined the sporulation efficiencies of cells harboring spoIVA^L* in the presence and absence of mraW. Deletion of mraW did not affect the sporulation efficiency of otherwise WT cells (Figure 1E, lane 3) but did increase the sporulation efficiency of cells harboring spoIVA^L* ~10-fold (Figure 1B, lane 10; Figure 1E, lane 8). Suppression of the spoIVA^L* defect was reversed upon complementation of mraW at an ectopic genomic locus (Figure 1E, lane 9), indicating that the suppression is due to the loss of mraW. Deletion of mraW also partially corrected the sporulation defects of other spoIVA alleles that disrupted residues in the nucleotide-binding pocket, including disruption of the SpoIVA Walker A motif (spoIVA^A*), sensor threonine (spoIVA^T*), and Walker B motif (spoIVA^B*) (Figure 1F), indicating that the suppressor was not allele-specific.

MraW is a methyltransferase that modifies 16S rRNA by methylating the conserved cytidine at position 1402 in E. coli that sits in the ribosomal peptidyl tRNA binding site (P site).²⁰ To determine if loss of the methyltransferase activity specifically of MraW suppresses the sporulation defect caused by defective spoIVA alleles, we identified putative residues involved in binding S-adenyl-methionine guided by the co-crystal structure of E. coli MraW¹⁹ and substituted those residues with Ala. Disruption of either D52 or D100 did not affect accumulation of the protein in vivo (Figure S1B), but did partially suppress the sporulation defect caused by spoIVA^L* (Figure 1E, lanes 10–11; mraW* represents D52A and D100A variants, respectively). Whereas MraW methylates the cytosine base of C1402 in E. coli,^19,20 C1402 is also methylated on the deoxyribose moiety by RsmI,²⁰ which in B. subtilis is predicted to be encoded by yabC. Deleting yabC also partially suppressed the sporulation defects of spoIVA^L* and spoIVA^T*, albeit not to the same extent as deleting mraW (Figure 1G). These varying impacts of deleting either mraW or yabC are consistent with the epistatic effects of these mutations reported for E. coli (wherein the double mutant phenocopies the deletion of rsmH/mraW alone).²⁰ Together, the data suggest that the loss of methylation of the 16S rRNA permits increased sporulation despite the presence of coat assembly defects.

MraW impacts translation in vitro

Several studies have reported the effects of deleting mraW on the expression of various constructs in vivo in E. coli,²⁰ but a direct effect on translation has not been reported. To determine if the loss of MraW directly affects the functionality of the ribosome, we purified ribosomes from either WT or ΔmraW cells and assayed their ability to produce GFP from a transcript expressed from the T7 promoter in a combined transcription-translation reaction. WT ribosomes synthesized GFP, but ribosomes from ΔmraW produced > 2-fold less GFP as measured by GFP fluorescence (Figure 2A). This reduction in activity was not due to reduced levels of ribosomes, since ribosomes purified from ΔmraW cells contained similar quantities of rRNA (Figure 2B) and total protein (which displayed a similar banding pattern; Figure 2C), suggesting that reduced GFP levels by the ΔmraW-derived ribosomes were due to compromised translation efficiency. Consistent with the reduced amount of GFP fluorescence, immunoblotting the contents of the in vitro reaction revealed that ribosomes isolated from ΔmraW cells produced a reduced amount of GFP protein (Figure 2G). Moreover, the addition of ribosomes isolated from ΔmraW cells did not impact the amount of gfp mRNA produced by T7 polymerase in the in vitro reaction (Figure 2H). The relatively large decrease in translation efficiency suggested that deleting mraW may result in a growth defect in vivo. However, we observed no difference in growth rate between WT and ΔmraW in two different rich media and one more defined medium (Figures S2A–S2C). We therefore suspect that the in vitro translation conditions present a stringent environment where a strong defect can be observed, whereas rapid growth in nutrient-rich conditions largely masks these effects. Consistent with this hypothesis, we observed similar production of GFP from a construct expressed by the T7 promoter in vivo in the presence and absence of MraW (Figure S2D).

Figure 2. MraW-modified ribosomes promote robust translation in vitro and are required early during sporulation to ensure spore quality control.

(A–F) In vitro transcription/translation assay using ribosomes purified from WT or ΔmraW B. subtilis (strains PY79 or ZP1) after (A–C) 3 h (exponentially growing cells), or (D–F) 7 h (sporulating cells) of growth in LB 10/10/10 medium. (A and D) GFP fluorescence generated from in vitro transcription/translation using ribosomes purified from WT or ΔmraW cells (A) during exponential growth or (D) sporulation using a gfp template that contains a generic upstream untranslated region. Bars represent the mean (n = 3 independent trials using 3 separate sets of ribosomes purified from each strain); errors: SD p values were calculated using an unpaired t test with the Bonferroni-Dunn method of correction for multiple statistical tests. (B and E) non-denaturing agarose gel of ribosomes purified from WT or ΔmraW cells (B) during exponential growth or (E) sporulation stained with propidium iodine. Predicted bands for the 23S and 16S rRNA based on electrophoretic mobility are labeled. Lane 1: DNA size marker with length (kilobase pairs) indicated to the left. (C and F) Coomassie-stained PAGE of the same ribosomal samples from (B and E), respectively. Lane 1: molecular weight markers with sizes (kDa) indicated to the left.

(G) Immunoblot (left) using anti-GFP antiserum to probe the amount of total GFP produced during a representative in vitro transcription/translation assay using ribosomes from either WT (PY79) or ΔmraW (ZP1) during exponential growth; Coomassie-stained SDS-PAGE gel (center) of the samples used for the immunoblot analysis; fluorescence of the GFP produced (right) by the in vitro transcription/translation assay in the SDS-PAGE gel before staining with Coomassie. Reactions were performed in the presence and absence of template DNA as indicated; relative mobilities of size markers (kDa) are indicated to the left of the gels.

(H) Agarose gel stained with propidium iodide showing the production of the GFP mRNA template from the in vitro transcription/translation assays in (G) using ribosome extracts from both WT or ΔmraW cells in the presence or absence of the DNA template as indicated. Lane 1: molecular weight markers (size in kDa are indicated to the left of the gel).

(I) Schematic of the construct used to measure ribosomal frameshifting. The gfp ORF is fused to the mCherry ORF (containing a stop codon that is in-frame relative to the gfp ORF where indicated) via a linker sequence that contains either a single or double nucleotide insertion to generate a +1 or −1 frameshift, respectively, in the mCherry ORF. Ribosomal frameshifting would therefore permit proper translation of mCherry.

(J) Ratio of fluorescence intensity from mCherry to fluorescence intensity from GFP in WT (purple) and ΔmraW (red) cells after 3 h or 7 h of growth in LB 10/10/10 harboring a construct with a linker encoding either a +1 (left) or −1 (right) frameshift. Bars represent mean (n = 3 biological replicates); errors: SD.

(K) Schematic depicting the timing of expression of two mother cell-specific sigma (σ) factors, σ^E and σ^K. σ^E is the first mother cell-specific sigma factor, and σ^K is the final mother cell-specific sigma factor.

(L) Sporulation efficiencies of indicated strains that express mraW constitutively or under control of a σ^E-controlled promoter (_PspoIVA) or σ^K-controlled promoter (P_cotR). Bars represent mean (n = 3 biological replicates); errors: SD.

Since the loss of MraW suppressed a sporulation defect, we next examined if the loss of MraW differently impacts the functionality of ribosomes harvested from sporulating cells. 2 h after entry into stationary phase, B. subtilis cultures displayed activation of the sporulation-specific spoIIE promoter, which drives the expression of an early sporulation gene (Figures S3A and S3B).⁵³ Ribosomes isolated from ΔmraW cells at this stage of sporulation displayed a similar defect in the production of GFP in vitro as ribosomes isolated from exponentially growing cells (Figure 2D); similar to ribosomes isolated from exponentially growing cells, ribosomes isolated from sporulating ΔmraW cells were present at similar quantities of rRNA and total protein as those obtained from WT cells (Figures 2E and 2F). The results therefore indicate that loss of methylation of the 16S rRNA by MraW abrogates the ability of ribosomes to efficiently translate, especially under growth-limiting conditions, and that this loss of activity is exhibited regardless of the growth phase of the bacterium.

MraW does not impact ribosomal frameshifting

A previous study reported that deletion of mraW in E. coli results in a modestly increased level of ribosomal frameshifting that results in mistranslation of an open reading frame (ORF).²⁰ To determine if frameshifting could cause the reduced translation efficiency we observed in B. subtilis upon deletion of mraW, we employed an in vivo reporter construct inspired by the one that was used previously in E. coli wherein a transcript encoding GFP was fused to the ORF of mCherry via a linker sequence that contained either a single or double nucleotide insertion to generate a +1 or −1 frameshift, respectively, in the ORF of mCherry, thereby preventing mCherry production (Figure 2I). Impacting the reading frame of the translating ribosome would therefore result in increased mCherry fluorescence relative to GFP fluorescence as compared to WT. Deletion of mraW in E. coli resulted in the modestly increased production of the downstream gene upon −1 frameshifting,²⁰ suggesting increased frameshifting in this mutant. In contrast, in B. subtilis, we observed a slight decrease in the production of mCherry in the absence of MraW in both frameshifted contexts relative to WT (Figure 2J), indicating that the loss of MraW modification of the 16S rRNA does not impact the propensity of the ribosome to frameshift in this organism.

Early presence of MraW during sporulation is required to activate the coat assembly checkpoint

Given the relatively reduced translation efficiency of ribosomes that were unmodified by MraW, we reasoned that deletion of mraW may suppress the sporulation defect caused by SpoIVA variants by specifically impacting the translation of a sporulation factor (likely a negative regulator of sporulation). To perform a candidate-based approach to identify this factor, we first assessed the point during the sporulation program at which MraW is required to block sporulation in the presence of a defective SpoIVA variant. SpoIVA is produced in the mother cell shortly after asymmetric division, under control of the σ^E sigma factor (Figure 2K). We therefore expressed mraW under the control of either a σ^E-driven promoter or a promoter driven by a late mother cell sigma factor (σ^K; Figure 2L) and monitored if either construct would complement the deletion of mraW. Expression of mraW from a σ^E-driven promoter complemented the deletion of mraW similar to the complementation of ΔmraW by mraW under its native promoter (Figure 2L, lanes 6–9), but expression of mraW from a σ^K-driven promoter did not (Figure 2L, lanes 10–11). Consistent with this observation, transcriptional fusions to lacZ revealed that the promoter for mraW (as well as the promoter for yabC) remains active for several hours after induction of sporulation (Figure S4). We therefore conclude that MraW is required relatively early during sporulation to block cells harboring defective SpoIVA variants from forming heat-resistant spores.

MraW is required for the efficient translation of CmpA

There are two known negative regulators of sporulation that participate in the coat assembly checkpoint that prevents cells from proceeding through the sporulation program if the cell harbors a defective SpoIVA variant: SpoVID⁴⁸ and CmpA.⁵⁰ We therefore investigated, using β-galactosidase fusions, if deletion of mraW affects the production of either protein. Expression of spoVID was not appreciably affected by deletion of mraW (Figures S5A and S5B). However, a translational fusion to the cmpA ORF revealed reduced translation of cmpA over time in the absence of mraW in cells producing a defective allele of spoIVA (Figure 3A) or in otherwise WT cells (Figure 3C). This effect of MraW was not observed when we employed a transcriptional fusion to the cmpA promoter only (Figures 3B and 3D). Moreover, the in vivo half-life of CmpA-LacZ was similar in a spectinomycin chase experiment in the presence and absence of mraW, as was the expression of lacZ alone when fused to a constitutive promoter (Figures S5C and S5D), indicating that the turnover of CmpA and the synthesis of LacZ alone were unaffected by the absence of MraW. Additionally, deletion of mraW also reduced production of CmpA-GFP from an overexpressed transcript encoding a translational fusion of cmpA-gfp driven by a constitutive promoter (instead of the relatively weak P_cmpA promoter), suggesting that MraW specifically affects the translation of cmpA (Figure S5E). Finally, similar to deletion of mraW, deletion of yabC also reduced the translation of a cmpA fusion to lacZ, measured at 5 h after induction of sporulation, albeit to a lower extent, and deleting both mraW and yabC did not result in a further decrease in cmpA translation (Figure 3E). Overexpression of cmpA from a constitutive promoter reduces sporulation efficiency even in an otherwise WT cell⁵⁰ (Figure 3F, lanes 2 and 6). This overexpression effect is more pronounced in the presence of a defective spoIVA allele (Figure 3F, lane 10). However, consistent with the previous observation, deletion of mraW also suppressed this sporulation defect arising from cmpA overexpression, either in the presence of WT SpoIVA (Figure 3F, lane 3, 7) or in the presence of SpoIVA^T* (Figure 3F, lane 11). We therefore hypothesize that deletion of mraW (and to a lesser extent yabC) and the resulting loss of 16S rRNA methylation negatively impacts translation of cmpA mRNA, which in turn suppresses the sporulation defect of defective SpoIVA variants.

Figure 3. CmpA translation, but not transcription, is reduced in ΔmraW.

(A–D) β-galactosidase activity arising from (A and C) translational fusions of the cmpA ORF to lacZ expressed from the cmpA promoter or (B and D) transcriptional fusions of the cmpA promoter fused directly to lacZ in cells harboring (A–B) spoIVA^T* or (C and D) WT spoIVA, in the (black) presence or (gray) absence of mraW at indicated time points after induction of sporulation. Data points represent mean (n = 3 biological replicates; errors: SD).

(E) β-galactosidase activity arising from translational fusions of the cmpA ORF to lacZ expressed from the cmpA promoter in WT (purple), ΔmraW (red), ΔyabC (teal), and ΔmraW ΔyabC (green). Bars represent mean (n = 3 biological replicates); errors: SD.

(F) Sporulation efficiencies of strains in the presence of cmpA that is constitutively produced (P_sigA-cmpA) harboring the indicated allele of spoIVA in the presence or absence of mraW. Bars represent mean (n = 3 biological replicates); errors: SD.

MraW mediates CmpA protein levels via a feature of the cmpA mRNA sequence

Next, we investigated if an element within the cmpA transcript is responsible for the MraW-mediated control of CmpA protein levels. The cmpA transcript harbors an additional AUG sequence between the ribosome binding site (RBS) and the AUG start codon, and translation initiation from this out-of-frame AUG would result in a truncated protein. We therefore wondered if this AUG could function as an alternative start site that might drive reduced synthesis of full-length CmpA. However, mutating this upstream AUG to AAA did not abrogate the ΔmraW impact on cmpA translation (Figure S5F), suggesting that this AUG sequence is not involved in MraW control of CmpA protein levels. We then took a systematic approach and truncated the cmpA ORF fused to lacZ to identify the minimal region responsible for this control. Sequential 3′ truncations of the 37-codon-long cmpA ORF showed that MraW-mediated translational control persisted when the 5′ untranslated region was fused to just the first 10 codons of cmpA (Figures 4A–4D). To test if specific codon usage was important for the ΔmraW effect, we fused the upstream untranslated region and codons 1–20 to lacZ, introduced a −1 frameshift immediately after the cmpA start codon and corrected the reading frame by adding a nucleotide at the fusion to lacZ to restore the proper reading frame (Figure 4E). This disruption largely preserved the nucleotide sequence of cmpA^1–20 but disrupted the amino acid sequence in this region. Interestingly, deletion of mraW abrogated the production of the frameshifted construct, similar to the abrogation of production of cmpA^1–20-lacZ (Figure 4F). We therefore conclude that a property intrinsic to the cmpA nucleotide sequence, and not codon usage, is likely required for MraW control of cmpA translation.

Figure 4. The nucleotide sequence of the first 20 codons of cmpA is required for MraW regulation.

(A–D) β-galactosidase activity arising from translational fusions of (A) full-length cmpA, (B) cmpA^1-30, (C) cmpA^1-20, or (D) cmpA^1-10 fused to lacZ in the (black) presence or (gray) absence of mraW at indicated time points after induction of sporulation. Data points represent mean (n = 3 biological replicates; errors: SD).

(E) Illustration of the sequence changes to create a −1-frame shift within the first 20 codons of the cmpA sequence. Top: cmpA^1-20 sequence; bottom: cmpA^1-20(−1) sequence. Red indicates a deleted nucleotide in the first cmpA codon; green indicates an added nucleotide at codon 20 to restore the ORF with the lacZ fusion. Amino acids encoded by each codon in either construct are displayed above the corresponding codons.

(F) β-galactosidase activities relative to WT of cmpA^1-20-lacZ or cmpA^1-20(−1)-lacZ in (purple) presence or (red) absence of mraW measured 4 h after induction of sporulation. Bars represent mean (n = 3 biological replicates); errors: SD.

A predicted early stem-loop structure in cmpA mRNA contributes to MraW-mediated control of translation

RNA structure prediction models revealed that the nucleotide sequence between the RBS and the AUG start codon (a sequence that we term the “intervening sequence”), along with the first 6 codons of cmpA, forms a stem-loop structure (Figure 5A; Figure S6). The nucleotides contributing to this structure are relatively well conserved across cmpA orthologs (Figure S7) and, with the exception of the B. anthracis ortholog, the different orthologs are predicted to form a stem-loop structure near the beginning of the ORF in which the intervening sequence participates at least partially in base pairs with early codons in the cmpA ORF (Figure S8).

Figure 5. Nucleotides involved in stem-loop formation are required for MraW-mediated translation of cmpA mRNA in vitro and in vivo.

(A) mRNA structure prediction of the 20 nucleotides upstream of start plus the first 35 nucleotides of the B. subtilis cmpA ORF generated by Mfold (B and C) CmpA-GFP fluorescence generated from an in vitro transcription/translation assay of a construct containing either the native cmpA IS sequence (“AAU-GAGCCU”) or a mutant (IS*) sequence (“AAAAAAU”) using ribosomes purified from WT (purple) or ΔmraW (red) B. subtilis (strains PY79 or ZP1) (B) after 3 h (exponentially growing cells) or (C) 7 h (sporulating cells) of growth in LB 10/10/10 media. Bars represent mean (n = 3 independent trials using 3 separate sets of ribosomes purified from each strain); errors: SD p value was calculated using an unpaired t test with the Bonferroni-Dunn method of correction for multiple statistical tests.

(D) Sporulation efficiencies of strains in the presence of either wild-type (WT) or IS mutant (IS*) cmpA that is constitutively produced (P_sigA-cmpA) in the presence or absence of mraW. Bars represent mean (n = 3 biological replicates); errors: SD.

(E and F) Mfold predictions of the 5′UTR and either the first (E) four or (F) six codons of cmpA.

(G) β-galactosidase activity arising from the translational fusion of the 5′UTR and either the first four or six codons of cmpA to lacZ in the (purple) presence or (red) absence of mraW at indicated time points after induction of sporulation. Data points represent mean (n = 3 biological replicates; errors: SD).

(H) Mfold structure prediction in which several nucleotides comprising the stem were swapped.

(I) β-galactosidase activity arising from the translational fusion of the swapped sequence in the presence (purple) and absence (red) of mraW at indicated time points after induction of sporulation. Data points represent mean (n = 3 biological replicates; errors: SD).

(J and K) Mfold predictions of the 5′ UTR and first six codons of cmpA with (J) codon 4 mutated to “ACC” and codon 5 mutated to “GAA” to disrupt the stem-loop structure, and (K) a stronger stem structure created by mutation of nucleotides in (F).

(L) β-galactosidase activities from the constructs in (J–K) in the presence (purple) or absence (red) of mraW at 5 h post-induction of sporulation. Data points represent mean (n = 3 biological replicates; errors: SD).

To test if this structure is important for MraW-mediated control of cmpA translation, we constructed several mutants that disrupted or restored this structure and assessed the impacts of these changes on CmpA production in vitro and in vivo. First, we replaced the entire intervening sequence (5′-AAUGAGCCU) with 5′- AAAAAAU to disrupt the stem structure and measured the production of CmpA-GFP in vitro in a combined transcription-translation reaction. As seen previously (Figure 2A), ribosomes isolated from ΔmraW cells produced ~2-fold lower levels of CmpA-GFP compared to ribosomes isolated from WT cells (Figure 5B). In addition, the production of CmpA-GFP was much lower than GFP production (Figure 2A) alone, even by WT ribosomes, suggesting that the presence of the CmpA sequence is inhibitory to even WT ribosomes. However, in vitro translation of the mRNA harboring a disrupted stem (CmpA^IS*) resulted in an increase of CmpA-GFP production by ΔmraW ribosomes, similar to levels produced by WT ribosomes that translated the mRNA containing the stem (Figure 5B). Interestingly, WT ribosomes displayed even more increased production of CmpA-GFP from the stem-disrupted transcript. We observed a similar pattern of CmpA-GFP production when we instead used ribosomes isolated from sporulating cells (Figure 5C). However, we also observed that ribosomes isolated from sporulating cells displayed reduced overall CmpA-GFP production compared to those isolated from exponentially growing cells, suggesting either that ribosomes during sporulation are less active or that additional factors may be present during sporulation to reduce the activity of ribosomes, perhaps to maintain resources during nutrient deprivation.

To test if the increased production of CmpA from the mutant transcript produced higher quantities of CmpA that are physiologically relevant, we tested the effect of this construct in vivo during sporulation. Overexpression of cmpA from a constitutive promoter resulted in a sporulation defect that was corrected by deletion of mraW (Figure 5D, lanes 2–3). Overexpression of the cmpA transcript containing the stem disruption also resulted in a similar sporulation defect (Figure 5D, lane 4), but deletion of mraW did not correct this sporulation defect (Figure 5D, lane 5). These findings suggest that the increase in CmpA-GFP production by ΔmraW ribosomes in vitro from a template in which the stem was disrupted translates to an observable phenotype in vivo.

Next, we sought to further probe the importance of the stem-loop structure in vivo. Based on the structure prediction, the 5′ untranslated region and first 6 codons of cmpA, but not codons 1–4, are sufficient to form the stem-loop (Figures 5E and 5F). Consistent with this prediction, deleting mraW reduced synthesis of CmpA^1–6-LacZ, but not CmpA^1–4-LacZ (Figure 5G). To test the sufficiency of stem formation for MraW control, we swapped 4 nucleotides in the intervening sequence with the corresponding 4 nucleotides in the cmpA coding region so that the predicted stem would not be disrupted (Figure 5H). Deleting mraW resulted in reduced synthesis of this stem-swapped variant of CmpA-LacZ (Figure 5I), consistent with the model that the presence of the early stem structure in the mRNA requires MraW-modified ribosomes for robust translation.

We next examined if disrupting the stem-loop structure in vivo would abrogate the need for MraW-modified ribosomes for efficiently translating cmpA mRNA. Using the cmpA^1–6-lacZ template, we mutated codons 4 and 5 from “TGG CTT” to “ACC GAA” to disrupt the stem-loop structure (Figure 5J). Translation of this mutant cmpA transcript (cmpA^1–6(stem*)-lacZ) showed reduced dependence on the presence of MraW-modified ribosomes and resulted in similar levels of β-galactosidase activity as the native CmpA^1–6-LacZ produced by WT ribosomes (Figure 5L). Interestingly, similar to what we observed in vitro (Figure 5B), disrupting the stem-loop also increased expression by WT ribosomes in vivo, again suggesting that the presence of the stem-loop limits CmpA production even in WT cells (Figure 5L). Finally, we noticed that the native stem structure in cmpA^1–6-lacZ contained two instances of imperfect or no base pairing (either U-G or U-U; Figure 7B). We therefore introduced additional mutations to change the U-U non-base pair to A-U and the U-G base pair to A-U to construct cmpA^1–6(stem^strong)-lacZ (Figure 5K). Surprisingly, the resulting stable stem structure completely blocked translation, even in otherwise WT cells (Figure 5L). The data are therefore consistent with a model in which cmpA can be more efficiently translated by MraW-modified ribosomes due to the enhanced ability of these ribosomes to engage mRNAs with secondary structures that occlude early codons in the ORF. However, the inability to fully restore WT expression levels in ΔmraW cells even after mutating the stem-loop suggests that MraW may be affecting translation efficiency via additional mechanisms that are still undetermined.

Figure 7. MraW is required for the efficient translation of other transcripts with early stem-loop structures.

(A) Mfold predicted structure of the 5’ UTR and start codon of isp with six codons from lacZ.

(B) β-galactosidase activity arising from the fusion of the promoter and 5’UTR of isp to lacZ at indicated time points after induction of sporulation in the presence (purple) and absence (red) of MraW. Data points represent mean (n = 3 biological replicates; errors: SD).

(C) Mfold predicted structure of the 5’UTR and start codon of isp with six codons from lacZ in which the nucleotides that form the stem structure have been swapped with each other.

(D) β-galactosidase activity arising from the stem-swapped mutant of isp at indicated time points after induction of sporulation in the presence (purple) and absence (red) of MraW. Data points represent mean (n = 3 biological replicates; errors: SD).

(E) Mfold predicted structure of the 5’UTR and start codon of isp in which the IS has been mutated from “GCUUUUUU” to “AAAAAAU” fused to lacZ.

(F) β-galactosidase activity arising from the IS mutant of isp at 1.5 h after induction of sporulation in the presence (purple) and absence (red) of MraW. Data points represent mean (n = 3 biological replicates; errors: SD).

(G) Mfold predicted structure of the 5’ UTR and start codon of isp in which the A–T base pairs were substituted with G–C base pairs, fused to lacZ.

(H) β-galactosidase activity arising from the G–C mutant of isp at 1.5 h after induction of sporulation in the presence (purple) and absence (red) of MraW. Data points represent mean (n = 3 biological replicates; errors: SD).

Taken together, we conclude that MraW-modified ribosomes facilitate the efficient production of CmpA to ensure proper activation of the coat assembly checkpoint (Figure 6A). Since excessive amounts of CmpA inhibit sporulation, proper (relatively low) CmpA levels are presumably achieved in a WT cell via the presence of a difficult-to-translate stem-loop structure near the 5′ end of the cmpA ORF. The data suggest that in the absence of MraW, this stem-loop structure impedes ribosome progression and reduces CmpA production, thereby inactivating the coat assembly checkpoint and permitting progression through sporulation even when coat assembly defects are present (Figure 6B).

Figure 6. Model for how the stem-loop structure in cmpA mRNA promotes production of the proper levels of CmpA to maintain the coat assembly checkpoint.

(A) A stem-loop structure near the 5’ end of the cmpA ORF results in a difficult-to-translate message. WT ribosomes modified by MraW and YabC methylation promote sufficient translation of cmpA mRNA to produce enough CmpA to maintain an active coat assembly checkpoint, but not too much CmpA that it inhibits sporulation.

(B) In the absence of mraW, unmodified ribosomes cannot efficiently translate past the stem-loop structure in cmpA mRNA, resulting in reduced CmpA production that is insufficient to activate the coat assembly checkpoint, permitting the sporulation program to continue despite coat assembly defects.

Translation of other mRNAs with predicted stem-loops is impacted in ΔmraW

We next examined the proteome of B. subtilis, either WT or cells harboring a mutant allele of spoIVA, during vegetative growth and sporulation using mass spectrometry to assess the impact of MraW on the production of other proteins besides CmpA. The raw data revealed scores of proteins that were differentially produced in the presence and absence of MraW (Dataset S1). Of note, loss of MraW led to ~ WT levels of SpoIVA^T*, consistent with our model of lower CmpA levels in ΔmraW inhibiting mutant SpoIVA degradation (Figure S9). To identify proteins whose levels were altered directly because of MraW, we employed three criteria to narrow the list of impacted proteins, based on the assumption that the loss of MraW should impact translation of a given transcript to a similar extent regardless of the allele of spoIVA the strain harbored. First, we considered only target proteins whose production changed at least 2-fold; second, the protein had to be detected in all analyzed strain backgrounds; third, the change in production had to be reproducible and statistically significant across three replicates (p < 0.05).

Using these relatively stringent criteria, we identified two proteins that we further analyzed: Isp, a protease involved in the transition to stationary phase⁵⁴; and YukJ, a protein of unknown function that is likely upregulated in stationary phase.⁵⁵ Like the cmpA transcript, the 5′ end of the isp mRNA is predicted to harbor a stem-loop structure, but unlike cmpA (whose stem occludes the untranslated intervening sequence and several early codons) the isp stem-loop exclusively comprises sequences upstream of the RBS, the intervening sequence, and just the start codon (Figure 7A). Fusion of just the isp promoter and this 5′ untranslated region of isp to lacZ resulted in the MraW-dependent production of β-galactosidase at 0 h, 1.5 h, and 4 h after induction of sporulation (Figure 7B). This dependence on MraW persisted when we swapped the nucleotides in the predicted stem (Figures 7C and 7D), suggesting that the stem-loop structure in the 5′ UTR of isp inherently reduces translation efficiency of this transcript and that MraW-modified ribosomes are capable of efficiently translating this mRNA. Consistent with this model, disrupting the stem-loop by exchanging the intervening sequence (normally GCTTTTTT) with the short intervening sequence that we used in the cmpA experiments (AAAAAAT) (Figure 7E) resulted in enhanced translation in ΔmraW cells and, like what we observed with the similarly mutated cmpA transcript, also enhanced expression in otherwise WT cells (Figure 7F). Conversely, altering the stem so that it became more stable by substituting the A-T base pairs to G-C resulted in drastically reduced expression in both WT and ΔmraW cells (Figures 7G and 7H), similar to what we observed when we enhanced the stability of the cmpA stem (Figures 5K and 5L).

The yukJ transcript is also predicted to form an early stem-loop structure, but here, the structure comprises exclusively codons 2–11 (and partially codon 12) and excludes the start codon and any upstream untranslated sequences (Figure S10A). YukJ protein levels were reduced in cells lacking MraW at each time point analyzed (Figure S10B), but fusing the yukJ promoter alone to lacZ showed no MraW-dependent effect, indicating that the impact of ΔmraW on YukJ expression occurs at the translation step (Figure S10C). We conclude that the fine-tuning of translation via secondary structures that limit translation efficiency, which evolved in select transcripts, is a general method in B. subtilis by which precise cellular concentrations of certain proteins can be maintained, and that this control is dependent on ribosomes whose 16S rRNA has been methylated by MraW.

DISCUSSION

In this study, a selection for spontaneous suppressor mutations that corrected the sporulation defect of cells producing a defective variant of the spore coat protein SpoIVA revealed a potential role for rRNA modifications in mediating the proper completion of the sporulation developmental program. A loss-of-function mutation in the widely conserved 16S rRNA methyltransferase, MraW, enhanced sporulation efficiency even in the presence of defective SpoIVA alleles, suggesting that these suppressor mutants bypassed the coat assembly checkpoint.⁴⁷ Using a candidate approach, we determined that the transcript encoding the negative sporulation regulator CmpA was poorly translated in the absence of MraW, thereby suggesting a model by which cells harboring mistakes in coat assembly could nonetheless proceed through the sporulation program. We also observed reduced CmpA translation in the absence of YabC, the predicted homolog of RsmI in E. coli, which methylates the same residue in 16S rRNA as MraW, but at a different location.¹³ Together with data presented in a recent preprint,⁵⁶ we propose that YabC likely performs an orthologous function in B. subtilis and should be renamed ‘‘RsmI.’’ We discovered that the cmpA transcript harbors a stem-loop structure that occludes several early codons in the cmpA ORF, resulting in a transcript that requires MraW-modified ribosomes to fine-tune the translation (likely translation initiation) of the transcript encoding this negative regulator: not so robustly as to inhibit sporulation in normal cells, but simultaneously producing enough CmpA to arrest sporulation if a coat assembly mistake arises during sporulation. A high-throughput analysis revealed other transcripts harboring similar stem-loop structures near the 5′ end of the ORF that also mediate translation of these mRNAs in an MraW-dependent manner. We propose that specific transcripts, perhaps especially those that encode proteins whose cellular levels should be limited (such as CmpA), exploit this rRNA modification to control protein production to ensure the faithful progression of bacterial development. Evolution of these mechanisms may have been facilitated by the uncoupling of transcription and translation observed in B. subtilis and predicted for other species.⁵⁷ While a similar stem-loop mediated mechanism could also be present in E. coli, for example, the tight coupling of transcription and translation in this organism^58,59 may preclude a mechanism by which structural elements in the mRNA could fold before the engagement with the ribosome.

Consistent with our model of secondary structure-mediated translational control of select transcripts, even though deletion of mraW does not result in an obvious growth defect, we observed in vitro that purified ribosomes not modified by MraW displayed an inherently lower efficiency in translating even optimized transcripts, and that this impairment was magnified in cmpA transcripts that harbor the translation-inhibitory stem-loop structure. We therefore speculate that the requirement for MraW may be more evident when cells are exposed to growth conditions that may stabilize translation-inhibitory structural elements present in select mRNAs. Interestingly, a recent report showed that B. subtilis lacking MraW exhibits defective growth at cold temperatures (18°C) and in recovery from cold shock.⁶⁰ ΔmraW cells did not display obvious defects in ribosome assembly after cold shock, but instead a reduction in the presence of polysomes, leading us to speculate that this defect is perhaps due to the stabilization of translation-inhibitory mRNA structural elements at reduced temperatures. The deletion of mraW resulted in a less severe defect in E. coli,⁶⁰ indicating that this level of MraW-mediated control is species-specific.

The transcript specificity for the requirement of MraW was evidenced by the relative scarcity of proteins whose production was reproducibly affected in vivo in the absence of MraW (consistent again with the lack of an obvious growth phenotype of an mraW deletion mutant). RNA secondary structure is a common mechanism for the fine-tuning of post-transcriptional regulation, with repressive effects of the structured regions often relieved by small RNAs, riboswitches, and RNA thermometers.^61–67 We propose that the presence of such secondary structures in the ORFs of select B. subtilis mRNAs represents a different regulatory mechanism that perhaps requires rRNA modifications for efficient engagement of the trapped codons to facilitate efficient translation initiation in cases where the start codon is occluded, and to facilitate efficient translation elongation in cases where the stem-loop forms immediately after the start codon, as exemplified in the yukJ transcript. Once the stable interaction between the modified rRNA and mRNA codons is formed, the intrinsic helicase activity of ribosomes⁶⁸ is likely more efficient at unwinding the mRNA. Indeed, structural data indicate that m4mC1402 directly engages with the mRNA backbone and forms key chemical interactions to stabilize the P site of the small ribosomal subunit.^5,20 This role in increased translation efficiency for C1402 methylation is distinct from, but complementary to, another phenotype that was previously described for an mraW deletion mutant wherein translation from GUG and UUG start codons was impaired, coupled with mis-initiation at AUU codons.²⁰ Combined with the observations we report here, we hypothesize that loss of MraW impairs the ability of the small subunit to efficiently engage start codons and/or codons immediately following the start codon that are present in the context of these structured regions, which could explain why CmpA production is not completely abrogated in the ΔmraW mutant and why there are not severe growth defects in ΔmraW strains. In sum, given the widespread conservation of mraW in bacterial genomes,²¹ combined with the presence of orthologous methyltransferases and rRNA modifications in chloroplasts and mitochondria,^69,70 it appears that rRNA modifications may broadly play a key role in accurate start site selection by the ribosome, and that secondary structures in the mRNA that encode sensitive regulatory proteins may be a general strategy to control the production of those factors.

Limitations of the study

Our results establish that mRNAs harboring stem-loop structures that occlude early codons reduce translation of those transcripts and that MraW-dependent 16S rRNA modification can enhance translation of those structured mRNAs (in addition to generally increasing the translation fitness of ribosomes). However, the molecular mechanism by which methylation of C1410 rRNA in B. subtilis leads to more robust translation remains to be determined. Our unexpected finding that ribosomes from sporulating cells are intrinsically less active also suggests either the existence of a factor that promotes reduced activity or the loss of a ribosomal factor that leads to reduced translation efficiency during sporulation which remains to be identified. Finally, using a proteomics approach, we identified several MraW-sensitive transcripts, but employing more sensitive approaches such as ribosome profiling may reveal the full scope of transcripts at different stages of growth and across different physiological conditions whose efficient translation requires modification of the 16S rRNA by MraW.

RESOURCE AVAILABILITY

Lead contact

Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Kumaran Ramamurthi (ramamurthiks@mail.nih.gov).

Materials availability

All plasmids, bacterial strains, and other unique reagents generated in this study are available from the lead contact with a completed materials transfer agreement.

STAR★METHODS

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

All Bacillus subtilis strains were derived from strain PY79 and are listed in the Key Resources Table.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit polyclonal anti-MraW	This paper	N/A
Rabbit polyclonal anti-SpoIVA	Castaing et al.45	N/A
Rabbit polyclonal anti-GFP	Castaing et al.45	N/A
Rabbit polyclonal anti-σA	Ebmeier et al.71	N/A
Bacterial and virus strains
Prototrophic derivative of B. subtilis 168	Youngman et al.72	PY79
B. subtilis ΔspoIVA::kan	Price et al.73	KP73
B. subtilis ΔspoIVA::kan ΔcmpA::erm	Tan et al.50	IT895
B. subtilis ΔspoIVA::erm ΔspoVID::kan	This study	JPC282
B. subtilis ΔspoIVA::kan ΔmraW::erm	This study	ZP44
B. subtilis ΔspoIVA::kan thrC::spoIVA spec	Ramamurthi and Losick42	KR394
B. subtilis ΔspoIVA::kan thrC::spoIVAL59P spec	This study	ZP47
B. subtilis ΔspoIVA::kan ΔcmpA::erm thrC::spoIVAL59P spec	This study	ZP50
B. subtilis ΔspoIVA::erm ΔspoVID::kan thrC::spoIVAL59P spec	This study	ZP53
B. subtilis ΔspoIVA::kan ΔmraW::erm thrC::spoIVAL59P spec	This study	ZP56
B. subtilis thrC::gfp-spoIVA spec	Ramamurthi and Losick42	KR160
B. subtilis ΔspoIVA::kan amyE::spoIVAL59P cat thrC::gfp-spoIVAL59P spec	This study	ZP35
B. subtilis ΔspoIVA::kan thrC::spoIVA spec ΔmraW::erm	This study	ZP77
B. subtilis ΔspoIVA::kan thrC::spoIVAK30A spec	Tan et al.50	KR367
B. subtilis ΔspoIVA::kan ΔmraW::erm thrC::spoIVAK30A spec	This study	ZP62
B. subtilis ΔspoIVA::kan thrC::spoIVAT70A,T71A spec	Castaing et al.45	JPC221
B. subtilis ΔspoIVA::kan ΔmraW::erm thrC::spoIVAT70A,T71A spec	This study	ZP65
B. subtilis ΔspoIVA::kan thrC::spoIVAD97A spec	Delerue et al.48	JB103
B. subtilis ΔspoIVA::kan ΔmraW::erm thrC::spoIVAD97A spec	This study	ZP59
B. subtilis ΔspoIVA::kan thrC::spoIVA spec ΔmraW::erm amyE::mraW cat	This study	ZP134
B. subtilis ΔspoIVA::kan thrC::spoIVA spec ΔmraW::erm amyE::mraWD52A cat	This study	ZP231
B. subtilis ΔspoIVA::kan thrC::spoIVA spec ΔmraW::erm amyE::mraWD100A cat	This study	ZP233
B. subtilis ΔspoIVA::kan ΔmraW::erm thrC::spoIVAL59P spec amyE::mraW cat	This study	ZP137
B. subtilis ΔspoIVA::kan thrC::spoIVAL59P spec ΔmraW::erm amyE::mraWD52A cat	This study	ZP239
B. subtilis ΔspoIVA::kan thrC::spoIVAL59P spec ΔmraW::erm amyE::mraWD100A cat	This study	ZP241
B. subtilis ΔspoIVA::kan thrC::spoIVA spec ΔyabC::erm	This study	ZP221
B. subtilis ΔspoIVA::kan ΔmraW::erm thrC::spoIVA spec ΔyabC::erm::cat	This study	ZP287
B. subtilis ΔspoIVA::kan thrC::spoIVAL59P spec ΔyabC::erm	This study	ZP215
B. subtilis ΔspoIVA::kan ΔmraW::erm thrC::spoIVAL59P spec ΔyabC::erm::cat	This study	ZP283
B. subtilis ΔspoIVA::kan thrC::spoIVAT70A,T71A spec ΔyabC::erm	This study	ZP249
B. subtilis ΔspoIVA::kan ΔmraW::erm thrC::spoIVAT70A,T71A spec ΔyabC::erm::cat	This study	ZP285
B. subtilis ΔmraW::erm	This study	ZP1
B. subtilis ΔspoIVA::kan ΔmraW::erm thrC::spoIVA spec amyE::PspoIVA-mraW cat	This study	ZP89
B. subtilis ΔspoIVA::kan ΔmraW::erm thrC::spoIVAL59P spec amyE::PspoIVA-mraW cat	This study	ZP83
B. subtilis ΔspoIVA::kan ΔmraW::erm thrC::spoIVA spec amyE::PcotR-mraW cat	This study	ZP107
B. subtilis ΔspoIVA::kan ΔmraW::erm thrC::spoIVAL59P spec amyE::PcotR -mraW cat	This study	ZP104
B. subtilis ΔspoIVA::kan sacA::spoIVAT70A,T71A cat::spec amyE::PcmpA-cmpA-lacZ cat	This study	ZP385
B. subtilis ΔspoIVA::kan 1‘:spoIVAT70A,T71A cat::spec B. subtilis ΔmraW::erm amyE::PcmpA-cmpA-lacZ cat	This study	ZP389
B. subtilis ΔspoIVA::kan sacA::spoIVAT70A,T71A cat::spec amyE::PcmpA-lacZ cat	This study	ZP383
B. subtilis ΔspoIVA::kan sacA::spoIVAT70A,T71A cat::spec ΔmraW::erm amyE::PcmpA-lacZ cat	This study	ZP387
B. subtilis amyE::PcmpA-cmpA-lacZ cat	Ebmeier et al.71	SE230
B. subtilis amyE::PcmpA-cmpA-lacZ cat ΔmraW::erm	This study	ZP313
B. subtilis amyE::PcmpA-lacZ cat	Ebmeier et al.71	SE222
B. subtilis amyE::PcmpA-lacZ cat ΔmraW::erm	This study	ZP311
B. subtilis sacA::PsigA-cmpA cat	Tan et al.50	IT686
B. subtilis ΔmraW::erm sacA::PsigA-cmpA cat	This study	ZP116
B. subtilis ΔspoIVA::kan thrC::spoIVA spec sacA::PsigA-cmpA cat	This study	ZP315
B. subtilis ΔspoIVA::kan thrC::spoIVA spec amyE::PsigA-cmpA cat ΔmraW::erm	This study	ZP321
B. subtilis ΔspoIVA::kan thrC::spoIVAT70A,T71A spec amyE::PsigA-cmpA cat	This study	ZP317
B. subtilis ΔspoIVA::kan thrC::spoIVAT70A,T71A spec amyE::PsigA-cmpA cat ΔmraW::erm	This study	ZP319
B. subtilis amyE::PcmpA-cmpA1-30-lacZ cat	This study	ZP441
B. subtilis amyE::PcmpA-cmpA1-30-lacZ cat ΔmraW::erm	This study	ZP446
B. subtilis amyE::PcmpA-cmpA1-20-lacZ cat	This study	ZP439
B. subtilis amyE::PcmpA-cmpA1-20-lacZ cat ΔmraW::erm	This study	ZP444
B. subtilis amyE::PcmpA-cmpA1-10-lacZ cat	This study	ZP437
B. subtilis amyE::PcmpA-cmpA1-10-lacZ cat ΔmraW::erm	This study	ZP442
B. subtilis amyE::PcmpA-cmpA1-20 (−1 frameshift)-lacZ cat	This study	ZP463
B. subtilis amyE::PcmpA-cmpA1-20 (−1 frameshift)-lacZ cat ΔmraW::erm	This study	ZP469
B. subtilis sacA::PsigA-IS*-cmpA cat	This study	ZP560
B. subtilis sacA::PsigA-IS*-cmpA cat ΔmraW::erm	This study	ZP562
B. subtilis amyE::PcmpA-cmpA1-4-lacZ cat	This study	ZP451
B. subtilis amyE::PcmpA-cmpA1-4-lacZ cat ΔmraW::erm	This study	ZP457
B. subtilis amyE::PcmpA-cmpA1-6-lacZ cat	This study	ZP473
B. subtilis amyE::PcmpA-cmpA1-6-lacZ cat ΔmraW::erm	This study	ZP481
B. subtilis amyE::PcmpA-cmpAstem-swap-lacZ cat	This study	ZP503
B. subtilis amyE::PcmpA-cmpAstem-swap-lacZ cat ΔmraW::erm	This study	ZP497
B. subtilis amyE::PcmpA-cmpA1-6 (codons 4*/5*)-lacZ cat	This study	ZP532
B. subtilis amyE::PcmpA-cmpA1-6 (codons 4*/5*)-lacZ cat ΔmraW::erm	This study	ZP534
B. subtilis amyE::PcmpA-cmpA1-6 (stem strong)-lacZ cat	This study	ZP568
B. subtilis amyE::PcmpA-cmpA1-6 (stem strong)-lacZ cat ΔmraW::erm	This study	ZP570
B. subtilis amyE::Pisp-lacZ cat	This study	ZP507
B. subtilis amyE::Pisp-lacZ cat ΔmraW::erm	This study	ZP513
B. subtilis amyE::Pisp (stem-swap)-lacZ cat	This study	ZP509
B. subtilis amyE::Pisp (stem-swap)-lacZ cat ΔmraW::erm	This study	ZP515
B. subtilis amyE::Pisp (IS*)-lacZ cat	This study	ZP577
B. subtilis amyE: Pisp (IS*)-lacZ cat ΔmraW::erm	This study	ZP583
B. subtilis amyE::PspoIIQ-T7 pol spec sacA::PT7-NEB5′UTR-gfp cat	This study	ZP572
B. subtilis amyE::PspoIIQ-T7 pol spec sacA::PT7-NEB5′UTR-gfp cat ΔmraW::erm	This study	ZP574
B. subtilis ΔspoIVA::kan sacA::spoIVAT70A,T71A cat::spec amyE::PspoVID-spoVID-lacZ cat	This study	ZP405
B. subtilis ΔspoIVA::kan sacA::spoIVAT70A,T71A cat::spec amyE::PspoVID-spoVID-lacZ cat ΔmraW::erm	This study	ZP410
B. subtilis amyE::PspoVID-spoVID-lacZ cat	This study	ZP339
B. subtilis amyE::PspoVID-spoVID-lacZ cat ΔmraW::erm	This study	ZP345
B. subtilis amyE::PsigA-lacZ cat	This study	ZP363
B. subtilis amyE::PsigA-lacZ cat ΔmraW::erm	This study	ZP361
B. subtilis amyE::PsigA-cmpA-gfp cat	This study	ZP523
B. subtilis amyE::PsigA-cmpA-gfp cat ΔmraW::erm	This study	ZP525
B. subtilis ΔspoIVA::kan sacA::spoIVA cat::spec	This study	ZP429
B. subtilis ΔspoIVA::kan sacA::spoIVA cat::spec ΔmraW::erm	This study	ZP435
B. subtilis ΔspoIVA::kan sacA::spoIVAT70A,T71A cat::spec	This study	ZP379
B. subtilis ΔspoIVA::kan sacA::spoIVAT70A,T71A cat::spec ΔmraW::erm	This study	ZP381
B. subtilis amyE::PyukJ-lacZ cat	This study	ZP546
B. subtilis amyE::PyukJ-lacZ cat ΔmraW::erm	This study	ZP550
B. subtilis amyE::Pisp (stem strong)-lacZ cat	This study	ZP511
B. subtilis amyE: Pisp (stem strong)-lacZ cat ΔmraW::erm	This study	ZP517
B. subtilis amyE:: PcmpA-(upstream AUG to AAA)cmpA-lacZ cat	This study	ZP401
B. subtilis amyE:: PcmpA-(upstream AUG to AAA)cmpA-lacZ cat ΔmraW::erm	This study	ZP403
B. subtilis amyE::PcmpA-cmpA-lacZ cat ΔyabC::erm	This study	ZP663
B. subtilis amyE::PcmpA-cmpA-lacZ cat ΔyabC::erm ΔmraW::erm::spec	This study	ZP665
B. subtilis amyE::PyabC-lacZ cat	This study	ZP667
B. subtilis amyE::PmraW-lacZ cat	This study	ZP367
B. subtilis amyE::PspoIIE-GFP spec	Fujita and Losick74	RL3520
B. subtilis sacA::PmraW-GFP-+1frameshift linker-mCherry cat ΔmraW::erm	This study	ZP669
B. subtilis sacA::PmraW-GFP−1frameshift linker-mCherry cat ΔmraW::erm	This study	ZP671
B. subtilis sacA::PmraW-GFP-+1frameshift linker-mCherry cat	This study	ZP673
B. subtilis sacA::PmraW-GFP−1frameshift linker-mCherry cat	This study	ZP675
Chemicals, peptides, and recombinant proteins
ONPG	Thermo Scientific	34055
FM4-64	Themo Scientific	H3570
DNAse I (RNase-free)	NEB	M0303L
RNase Inhibitor	NEB	M0314L
Difco Sporulation Medium	KD Medical	CUS-0801
CH I + II Medium	KD Medical	CUS-1150
CH 3	KD Medical	CUS-0804
CH 4	KD Medical	CUS-0805
Resuspension Medium A + B	KD Medical	CUS-1151
Critical commercial assays
PURExpress Δ Ribosome Kit	NEB	E3313S
GTPase Assay Kit	Abcam	Ab270553
Deposited data
Sporulation mass spectrometry data (ProteomeXchange; (https://proteomecentral.proteomexchange.org/)	This study	PXD074358
Experimental models: Organisms/strains
Bacillus subtilis PY79	–	–
Oligonucleotides
Primer: cmpA in vitro transcription/translation forward: GCG AAT TAA TAC GAC TCA CTA TA G G GC TTA AGT ATA AGG AGG A ATGAGCCT ATGCCAAATTGGCTTAAAA	This study	N/A
Primer: cmpA IS* in vitro transcription/translation forward: GCG AAT TAA TAC GAC TCA CTA TAG G GC TTA AGT ATA AGG AGG AAA AAA T ATGCCAAATTGGCTTAAAAAGCAGATG	This study	N/A
Primer: GFP in vitro transcription/translation forward: GCG AAT TAA TAC GAC TCA CTA TA G G GC TTA AGT ATA AGG AGG AAA AAA TATGAGTAAAGGAGAAG AACTTTTCACT	This study	N/A
Primer: GFP in vitro transcription/translation reverse: AAA CCC CTC CGT TTA GAG AGG GGT TAT GCT AG TTA TTATTTGTAT AGTTCATCCATGCC	This study	N/A
Recombinant DNA
amyE::PcmpA-cmpA-gfp cat (plasmid used for generating in vitro transcription/translation templates)	Ebmeier et al.71	pSE24
his6-spoIVA in pET28a vector for protein expression	–	pKR145
his6-spoIVAL59P in pET28a vector for protein expression	This study	pZP7
mraW-his6 in pET28a vector for protein expression	This study	pARD23
Software and algorithms
mFold	Zuker75 Zuker76	https://www.unafold.org/mfold/applications/rna-folding-form.php
Vienna Fold	Lorenz et al.77	http://rna.tbi.univie.ac.at/cgi-bin/RNAWebSuite/RNAfold.cgi
SubtiWiki	Elfmann et al.78	https://subtiwiki.uni-goettingen.de/v5/welcome
Prism 10 ver. 10.6.1	GraphPad	N/A
NCBI Sequence Viewer 3.35.1	Rangwala et al.79	N/A
Clustal Omega 1.2.4	Madeira et al.80	N/A
JalView 2.11.4.1	Waterhouse et al.81	N/A
ImageJ	NIH	N/A
Other
Vanquish Neo UHPLC system	Thermo Fisher Scientific	VN-S10-A-01
Orbitrap Exploris 480 mass spectrometer	Thermo Fisher Scientific	BRE725533
EASY-Spray™ Neo UHPLC column	Thermo Fisher Scientific	ES75500PN

METHOD DETAILS

Strain Construction

All strains used in this study were derived from the parental B. subtilis strain PY79.⁷² The ΔmraW::erm deletion was obtained from an ordered deletion library⁸² whose chromosomal DNA was transformed into PY79 to generate strain ZP1. Insertion of spoIVA at the sacA locus was achieved by cloning the spoIVA CDS plus 450 bp upstream and 360 bp downstream into RL4165 before transformation into B. subtilis. SpoIVA variants were derived from that plasmid (pZP30) using QuikChange mutagenesis. Fusions to lacZ were generated using either synthesized gene fragments (IDT) or QuikChange mutagenesis of previously constructed plasmids derived from RL5037 for insertion of the fusions at the amyE locus. The constitutively expressed cmpA-GFP fusion was constructed by using QuikChange mutagenesis to insert two early in-frame stop codons into the lacI gene in pSE24.⁷¹

General methods

SpoIVA, MraW, and σ^A levels were measured using immunoblotting of sporulating cell extracts prepared as previously described.⁴⁶ All proteins were detected using rabbit antiserum raised against recombinant SpoIVA (1:10,000 dilution),⁸³ MraW (1:10:000 dilution), and σ^A (1:5000 dilution)⁸⁴ (Covance) as the primary antibody and goat StarBright Blue 700 (1:10,000 dilution) (Bio-Rad) as the secondary antibody. Immunoblots were imaged on a ChemiDoc MP Imaging System (BioRad). GFP and mCherry production was quantified from 100 μL of cultures grown in indicated conditions (vegetative of sporulating) at the indicated time points grown in a 96-well plate using a Synergy H1 microplate reader (BioTek) (GFP: 488 nm excitation, 530 nm emission; mCherry: 579 nm excitation, 616 nm emission). For growth curves, overnight cultures were grown in lysogeny broth (LB) at 25°C and diluted (1:100) into the indicated media in a 96-well plate and grown at 37°C with shaking in a Synergy H1 microplate reader (BioTek) plate reader with O.D._600nm readings taken every 15 min. Statistical analyses of the data were conducted using GraphPad Prism 10. To measure induction of sporulation via activation of the spoIIE promoter, overnight cultures of strains PY79 or MF277⁷⁴ (which harbors a fusion of the high-threshold Spo0A-activated promoter of the spoIIE gene to gfp) grown at 25°C were diluted 1:100 into 10 mL LB (10/10/10 formulation) and grown at 37°C. O.D._600nm and fluorescence from GFP (488 nm excitation, 530 nm emission) were measured at 1 h intervals for 9 h in a 96-well plate using a Synergy H1 microplate reader (BioTek).

Sporulation efficiency

Sporulation efficiencies were calculated by heating cultures at 80°C for 20 min after 24 h of growth in Difco Sporulation Media at 37°C (DSM) to remove non-sporulating or defective sporulating cells from the population. Serial dilutions of the heat killed cultures were then plated on LB to determine the number of colony forming units (cfu); sporulation efficiency is reported relative to heat-resistant cfu calculated for a parallel culture of WT (PY79).

SpoIVA purification and ATP hydrolysis assay

Plasmid expressing his₆-spoIVA^L59P was generated by QuikChange mutagenesis of pKR145.⁴³ His₆-SpoIVA^L59P was purified as described previously.⁸⁵ Purified protein (0.3 μM) was then incubated with varying concentrations of ATP at 37°C for 1 h. The reactions were stopped by addition of 50 μL of 0.02 M EDTA. Freed phosphate from SpoIVA hydrolysis was measured using the colorimetric GTPase Assay Kit (abcam). A standard curve generated from known [P_i] solution was used to determine the levels of free phosphate generated from SpoIVA hydrolysis.⁴⁵ Reactions were measured at A_620nm and the data was fitted to a Michaelis-Menten curve using Prism software (GraphPad).

Epifluorescence microscopy

Cells expressing GFP-SpoIVA were induced to sporulate using the resuspension method.⁸⁶ Samples were taken at 2.5 h after induction of sporulation, collected by centrifugation, and resuspended in a 5 μg mL⁻¹ solution of FM4-64 in water to visualize membranes. Stained samples were placed on a glass bottom culture dish and covered with a 1% agarose pad. Cells were visualized with a DeltaVision Core microsocpe (Applied Precision/GE Healthcare) as described previously,⁸⁷ and images were processed with SoftWorx and ImageJ software.

Ribosome purification

B. subtilis ribosomes were purified as previously described.^70,88 Briefly, overnight cultures of WT (PY79) or ΔmraW (ZP1) grown at 25°C in LB medium (made using the 10/10/10 formulation; hereafter, simply “LB 10/10/10”) were diluted into 500 mL of fresh LB. Cells were harvested after 3 h of growth at 37°C in LB 10/10/10 and cell pellets were stored at −80°C. To purify ribosomes from sporulating cells, 2.5 mL of overnight cultures of PY79 or ZP1 grown at 25°C in LB 10/10/10 were diluted 1:100 into 250 mL LB 10/10/10 and grown for 7 h at 37°C with shaking to allow for activation of sporulation. Cells were harvested and stored at −80°C. Cells were then resuspended in 2 mL of lysis buffer (20 mM Tris pH 7.6, 10 mM magnesium acetate, 150 mM potassium acetate, 0.1% Triton X-100, 1 mg mL⁻¹ lysozyme, 100 U ml⁻¹ DNase I, 400 U ml⁻¹ RNase Out ribonuclease inhibitor) and incubated at 30°C for 30 min. Lysed samples were then cleared by centrifugation at ~8,500 × g (14,000 RPM in an Eppendorf MiniSpin Plus rotor) for 10 min ~500 μL of supernatant were gently layered on a 500 μL sucrose cushion (50 mM Tris-acetate at pH 7.5, 1 M potassium acetate, 15 mM magnesium acetate, 1.44 M sucrose, 1 mM dithiothreitol (DTT), 0.1% Complete Protease Inhibitor) in 11 × 34 mm PC tube (Beckman) and centrifuged for 1 h at ~430,000 × g (100,000 rpm in a Beckman TLA 120.2 rotor). The ribosome-containing pellet was resuspended in 50 μL Buffer B (50 mM Tris-acetate at pH 7.5, 50 mM potassium acetate, 5 mM magnesium acetate, 1 mM DTT). Ribosome samples were analyzed using a Nanodrop spectrophotometer to measure RNA concentration, by separation on a non-denaturing agarose gel to assess rRNA integrity, and by Coomassie-PAGE to assess relative abundance of ribosomal proteins. Quality of purified ribosomes was further analyzed using the TapeStation system (Agilent) to determine RIN values. Purified ribosomes were used immediately for the in vitro transcription and translation assay (described below) and stored at −80°C.

In vitro transcription and translation

In vitro transcription/translation was performed using the PURExpress In Vitro Protein Synthesis kit (NEB) using ribosomes purified from WT or ΔmraW cells. DNA templates for the reaction were generated using PCR with primers designed using kit guidelines and including sequences from pSE24 to amplify the GFP sequence. For the experiments in Figures 2A and 2D, the gfp gene is preceded by a sequence that we predict is not sensitive to MraW (AAAAAAT, as suggested by the kit manufacturer). Reactions (25 μL) were run in biological triplicate largely following the manufacturer’s protocol using ~250 ng of template DNA, 4.5 μL of the purified ribosome preparation (~200 ng μL⁻¹ of 30S as measured by TapeStation analysis), and 1 μL of RNase Out ribonuclease inhibitor (Invitrogen). Reactions were incubated at 37°C for 4 h and were stopped by incubation on ice. Reactions were then mixed with 25 μL of sterile diH₂O and fluorescence from the synthesized GFP was measured using a Synergy H1 microplate reader (BioTek) (488 nm excitation, 530 nm emission).

β-galactosidase assays

Cells expressing lacZ fusions were induced to sporulate using the resuspension method. At the indicated times, the O.D._600nm of the culture was measured and 1 mL of culture was harvested and stored at −80°C. Assays were conducted largely as described.⁷¹ Briefly, harvested cells were resuspended in 1 mL of Z-buffer (0.06M Na₂HPO₄-7H₂O, 0.04M Na₂HPO₄-H₂O, 0.01M KCl, 0.001M MgSO₄, 0.05M β-mercaptoethanol) and 50 μL of each sample were aliquoted into a well of a sterile 96-well plate. These samples were diluted with 50 μL of Z-buffer with 0.4 mg mL⁻¹ lysozyme and incubated at 37°C for 30 min. After incubation, 20 μL of 4 mg mL⁻¹ o-Nitrophenyl β-D-galactopyranoside in Z-buffer were added to each sample and the absorbance at 420 nm was measured every minute for one hour using a Synergy H1 microplate reader (BioTek) plate reader. Slopes were determined using Excel and arbitrary units were calculated.

mRNA folding

mRNA structure predictions were conducted using the Mfold RNA secondary structure prediction program^75,76 and verified with Vienna Fold.⁷⁷ Sequence identities of the 5′ UTR were determined using the SubtiWiki genome browser,⁷⁸ which identifies upshifts indicating the transcription start site. The standard parameters were used for both modeling platforms.

Sequence conservation

The DNA sequences from 100 nucleotides upstream to the first 10 codons of the cmpA genes in various spore forming bacteria¹ were obtained from NCBI Sequence Viewer 3.35.1.⁷⁹ Sequences were aligned with Clustal Omega 1.2.4⁸⁰ first, then gaps were adjusted manually to improve the alignment in JalView 2.11.4.1.⁸¹

Mass spectrometry

B. subtilis cell lysis was performed as follows. Frozen cell pellets from 20 mL of cultures induced to sporulate via the resuspension method, harvested at 0 h, 1.5 h, and 4 h after induction of sporulation, were suspended in 250 μL modified RIPA buffer (150 mM NaCl, 50 mM Tris-HCl at pH7.4, 1% NP-40, 0.25% Sodium deoxycholate, and 1 mM EDTA) containing protease inhibitors (Roche). To facilitate lysis, 25 μg lysozyme were added to each sample, and the tubes were incubated at 37°C for 30 min, shaking at 750 rpm. Samples were mixed 1:1 with urea lysis buffer (6 M Urea, 25 mM HEPES at pH 8.0, 5 mM EDTA) containing the protease inhibitors and lysed using a TissueLyser bead mill homogenizer (Qiagen) for 2 min, repeated three times to ensure complete lysis. Samples were then centrifuged at 12000 × g for 15 min at 10°C. After centrifugation, supernatants were transferred and protein concentrations measured using the BCA method (Thermo Scientific). Next, 100 μg of each protein sample was adjusted to a volume of 50 μL with urea lysis buffer, mixed with an equal volume of 50 mM TEAB, and reduced with 10 mM TCEP at 56°C for 45 min in a shaker (750 rpm). Samples were then cooled and alkylated with 20 mM iodoacetamide at room temperature in the dark. Samples were further diluted with 200 μL 50 mM TEAB to reduce the urea concentration to below 2 M and proteins were digested with trypsin (1:50) and Lys-C (1:125) at 37°C overnight. Following digestion, samples were acidified to a final concentration of 0.5% trifluoroacetic acid and desalted on a C18 spin column (Thermo Scientific).

Mass spectrometry analysis was performed using a Vanquish Neo UHPLC system (Thermo Fisher Scientific) coupled to an Orbitrap Exploris 480 mass spectrometer (Thermo Fisher Scientific) operating in data-independent acquisition (DIA) mode. Peptides were separated on a 75 μm × 50 cm EASY-Spray Neo UHPLC column (Thermo Fisher Scientific). The mobile phases consisted of (A) 0.1% formic acid (FA); (B) 0.1% FA, 80% acetonitrile. Peptides were separated using a 130 min gradient at a constant flow rate of 300 nL min⁻¹: 2% B for 1 min, 2–32% B for 121 min, 32–90% B for 20 min. The Orbitrap Exploris 480 mass spectrometer was operated in positive ion mode with full-scan precursor spectra (MS1) acquired at a resolution of 120,000 over the m/z scan range of 420–680. The AGC target was set to 300% (3e6 ions absolute) with a maximum injection time of 100 ms. Fragment ion scans (DIA MS2) were collected using 60 isolation windows with a 4 m/z window width, covering a precursor range of 430–670 m/z. DIA windows were placed using the instrument’s automatic window-placement optimization. Fragment spectra were recorded in the Orbitrap at a resolution of 90,000, using stepped higher-energy collisional dissociation (HCD) with 25%, 27.5%, and 30% normalized collision energies. The DIA AGC target was set to 3000% (3e6 ions absolute) and the injection time was automatically determined. The scan range for fragment ions was 200–1800 m/z with DIA cycle time of 3 s. DIA raw files were analyzed using DIA-NN 2.2.0 using an in-silico DIA-NN predicted spectral library from the UniProt Bacillus subtilis database (downloaded in 2025, 4191 sequences), allowing for cysteine carbamidomethylation and N-terminal methionine excision and 1 missed cleavage. The DIA-NN search included the following settings: Protein inference = Genes, Neural network classifier = Single-pass mode, Quantification strategy = Robust LC (high precision), Cross-run normalization = RT-dependent, Library Generation = IDs, RT and IM Profiling and Speed and RAM usage = Optimal results. Mass accuracy and MS1 accuracy were set to 0 for automatic inference. No share spectra, Heuristic protein inference and MBR were checked.

QUANTIFICATION AND STATISTICAL ANALYSIS

All experiments were performed using at least three independent biological replicates. Details of all statistical analyses (mean, median, SD, p value determination) are found in the corresponding figure legends. Statistical significance is defined as p < 0.05.

Supplementary Material

Supplemental information can be found online at https://doi.org/10.1016/j.celrep.2026.117887.

Highlights.

• A conserved 16S rRNA modification enhances translation of structured mRNAs

• Early mRNA stem-loops impose translational control dependent on ribosome modification

• mRNA structure and rRNA modifications likely co-evolved to fine-tune protein dosage

Data and code availability

The mass spectrometry proteomics data have been deposited in the ProteomeXchange Consortium via the MassIVE partner repository (https://proteomecentral.proteomexchange.org/) with the dataset identifier PXD074358. All other data supporting the findings of this study are available within the article and the supplemental information. This paper does not report original code. All data reported in this paper will be shared by the lead contact upon request. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.