mscL and arfA co-localisation is enriched in gammaproteobacteria

Because of the important roles that mscL and arfA display in responding to osmotic and translational cell stress both individually (3, 28) and collectively (17) and their genomic co-localisation in E. coli, we sought to explore the degree to which their presence is conserved across bacterial and archaeal kingdoms. Hidden Markov models (HMMs) of the MscL and ArfA proteins were obtained from the Pfam database (29), and these were used to query NCBI prokaryote representative and reference complete proteomes. A cladogram was constructed using NCBI taxonomic information (Fig 1A). The presence of MscL-coding sequence is shown to be predominately conserved across the bacteria kingdom present in 72% (2,575/3,594) of the genomes queried and was identified in a small subset of archaea genomes (10%, 22/228). In contrast, ArfA-coding sequence is shown to be present in a smaller number of bacterial genomes (10%, N = 375), within these 75% also contain mscL (N = 282). Taxonomic analysis of the resultant clustered genomes indicates that the cluster containing only mscL (mscL-only) is composed of species from across a wide range of phyla present in the dataset (Fig S1). Notably, actinobacteria, bacteroidetes, and firmicutes are highly enriched with mscL-only species comprising 93%, 92%, and 77% of the entire phylum, respectively (Table S1). In contrast, the arfA-only cluster is almost exclusively composed of Gammaproteobacteria (95%), and the exception to this is a small subset of Betaproteobacteria, including clinically important N. meningitidis and N. gonorrhoeae, indicating probable horizontal gene transfer (Fig S1). The mscL–arfA cluster is almost entirely restricted to Gammaproteobacteria (94%), enriched with species belonging to the Enterobacterales (49%) and Pseudomonadales (28%) orders (Fig 1A). A small number of Betaproteobacteria exceptions were identified within the mscL–arfA cluster (N = 14), including a number of Neisseria species. In contrast to the mscL-only cluster, which is much more taxonomically diverse, the mscL–arfA cluster displays narrower taxa indicative of an earlier ancestry of the mscL gene and suggesting that arfA was acquired later after the proteobacteria division.

Intriguingly, the intergenic distance of mscL and arfA is not conserved, and three major sub-clusters can be assigned based on where the two genes are (i) co-located and the 3′ CDS of mscL is overlapping with the 3′ CDS of arfA (overlap), (ii) co-located but are not overlapping (proximal), and (iii) not co-located (distal) (Fig 1B). The overlap cluster (N = 25) is predominately composed of bacterial species belonging to the Enterobacteriaceae family (72%), including important clinical and model genera Salmonella and Klebsiella. The proximal cluster (N = 103) is composed of Enterobacteriaceae (38%), Pectobacteriaceae (19%), and Yersiniaceae (19%) and includes important clinical and model organisms E. coli and Citrobacter koseri. The distal cluster (N = 154) is predominately composed of Pseudomonadaceae including important clinical and model organisms Pseudomonas aeruginosa and Pseudomonas putida. As an interesting side observation, we also noticed that in some instances (N = 71) arfA is co-located with other genes instead of mscL, displaying both CDS overlap (20%) and proximal (80%) arrangements (Table S2). These observations led us to question whether the enriched co-localisation of mscL and arfA within Enterobacteriaceae genomes has a regulatory function. To probe this, we sought to explore the intergenic regulation between these genes in E. coli (Fig 2A).

arfA negatively regulates mscL expression

To evaluate whether arfA regulates mscL expression, we performed transcriptional studies using qRT-PCR to monitor transcript abundance of these genes in E. coli wt and gene-specific deleted strains. Carefully designed single-gene deletions were generated to avoid polar effects (30) and direct disruption of the overlapping genes (Fig S2). Analysis of the steady-state transcript level of mscL in the arfA-deleted E. coli strain (ΔarfA) indicated an increase of >twofold in the absence of arfA (Fig 2B). Upon episomal expression of full-length arfA (pFL), the mscL transcript abundance was restored to wt levels (Fig 2B), confirming a role of arfA in the negative regulation of mscL.

To explore at which level this regulation occurs, promoter–reporter plasmids (P_mscL-sfGFP and P_arfA-sfGFP) were created by fusing the mscL or arfA promoters (11, 15) with a reporter gene. First, promoter activity of both genes was assessed in a knockout mutant of rpoS (∆rpoS) during both exponential and stationary phases (Fig 2C and D). In agreement with previous findings (15, 31), mscL expression was associated with the activity of RpoS, as we observed a reduction in mscL promoter activity in the ∆rpoS strain relative to the wt in both exponential (1.5-fold) and stationary phases (4.0-fold) (Fig 2C). Moreover, the enhanced mscL promoter activity observed during stationary phase was dependent upon RpoS, with increasing signal only observed in the wt and not in the absence of RpoS (wt = 2.29-fold increase, ∆rpoS = 1.2-fold decrease) (Fig 2C). The results showed only a minor difference in arfA promoter activity between wt and ∆rpoS (Fig 2D). The arfA promoter activity was also similar during exponential and stationary phases in both the wt and ∆rpoS strains. Taken together, these results show that arfA promoter activity is between 4 and 10 times greater, and in contrast to the mscL promoter is neither significantly affected by growth phase nor by RpoS.

Next, promoter activity was assessed in arfA and mscL gene–deleted E. coli strains, carrying the promoter reporter plasmids, grown in minimal media with low and high osmolality (215 and 764 mOsm) (Fig 3A and B, upper panels, respectively). Consistent with the previous report (15), mscL promoter activity in wt strain was up-regulated at higher osmolality (∼3.5-fold) (Fig 3A upper panel). The pattern was similar in ΔarfA and ΔmscL strains, indicating that mscL promoter activity is independent of the presence of both genes (Fig 3A upper panel). In contrast to mscL, the arfA promoter activity showed little response to the different osmolalities across all strains (Fig 3B upper panel). These results indicate that arfA is not transcriptionally regulated in response to different external osmolalities and that its regulatory function upon mscL expression does not occur at the transcriptional level.

Analysis of the mscL transcript abundance from the corresponding strains showed an increase in the higher osmolality environment (∼1.8-fold, P < 0.0092) (Fig 3A lower panel) consistent to the increase in promoter activity observed under the same conditions (Fig 3A upper panel). Although both positively regulated in response to osmolality increase, comparison of the mscL promoter (3.5-fold) and transcript (1.8-fold) changes suggest that mscL is potentially negatively regulated at the post-transcriptional level at high osmolality. In absence of arfA, mscL transcript abundance also increases independent of the external osmolality (∼2.5-fold, P < 0.0004) (Fig 3A lower panel) with no effect observed on the mscL promoter activity (Fig 3A upper panel). In contrast, analysis of the arfA transcript abundance showed no significant change in wt under different osmolalities (Fig 3B lower panel). However, in absence of mscL, the arfA transcript abundance increased only at high osmolality (2.6-fold, P < 0.0004), whereas no effect upon the arfA promoter activity was detected (Fig 3B). Collectively, these results suggest that arfA and mscL mediate down-regulation of each other at the post-transcriptional level. More specifically, the negative regulation of arfA upon mscL is observed at both high and low osmolalities, whereas the negative regulation of mscL upon arfA is only observed at high osmolality.

arfA regulates MscL-dependent excretory activity

MscL excretion activity is dependent upon arfA in E. coli cells that encounter both osmotic and translation stress conditions (17); therefore, we sought to investigate if the MscL excretion activity is regulated by arfA. First, cells were grown under high-cell density conditions, where a change in external osmolality is observed between early- and late-stage exponential growth phases, resulting in an osmolality drop of ∼70 mOsm (Fig S3A) (17). Consistent with the earlier observation (Fig 3A and B), mscL transcript levels were highest in the absence of arfA and before the drop in osmolality (Fig S3B and C). This result indicates a synergetic role of arfA and osmotic drop in mediating the reduction in mscL transcript abundance. Second, to study the impact of arfA up-regulation upon mscL, we generated a SmpB-deleted strain (ΔsmpB) in which the primary ribosome rescue system is impaired (11). arfA transcript abundance is significantly increased in the ΔsmpB strain under all conditions, whereas no change in mscL transcript abundance is observed pre-osmolality drop. The greatest reduction in mscL transcript abundance was observed in the ΔsmpB strain post-osmolality drop, providing further evidence of mscL down-regulation via a combination of arfA and osmolality. Finally, we investigated the phenotypic role (MscL production and excretion) of the arfA-mediated down-regulation of mscL in cells undergoing both growth-induced osmotic drop and arfA-mediated translational stress. To quantify MscL production, we generated wt and ΔarfA expression strains (BL21(DE3)), encoding a tagged MscL (mscL::mscLHis) and bearing an inducible (IPTG) expression plasmid (pET44-sfGFP). The His-tag modification showed no significant impact on both mscL expression and MscL excretion activity (Fig S4A and B); and the increase of arfA transcript abundance in the presence of IPTG induction confirmed that gfp overexpression results in an arfA-mediated response to translational stress as previously reported (Fig S4C) (17). Consistent with our previous finding (17), excretion of cytoplasmic GFP (%ECP) was significantly reduced (≥twofold) in absence of arfA, following the drop in media osmolality (Fig 4A upper panel). mscL transcript abundance in absence of arfA increases (≥twofold) and reflects an increase in MscL protein level (≥twofold) (Fig 4A lower panel). Here, we demonstrate that increased mscL transcript level results in increased MscL abundance but counter-intuitively with decreased MscL excretion activity in the absence of arfA. Yet, it was previously shown that clustering of MscL, which occurs with increased MscL concentration, promotes channel closure and decreases jettisons activity (32).

mscL transcript is a target of arfA asRNA

The E. coli genomic arrangement of mscL and arfA genes results in an overlapping sequence between the mscL 3′ UTR and the arfA 3′ CDS (Fig 2A). Previous studies (10, 11) identified a regulatory stem-loop structure within the 3′ CDS of arfA transcript of E. coli, from the nucleotide 146 to 206 (Fig 4B). This stem-loop structure contains two sites recognised by RNaseIII that once cleaved produce a truncated transcript without a stop codon and an sRNA 54 nucleotides in length, which can be further processed into sRNAs 24 and 30 nucleotides long (10, 11). To date, the target and function of the arfA sRNA was unknown; we therefore investigated if the arfA sRNA is involved in the post-transcriptional regulation of mscL expression functioning as an antisense RNA. arfA gene variants were generated to examine the effect of arfA variation upon mscL expression. These included: arfA full-length (FL); a truncated version missing the regulatory stem-loop (arfA_Δ(154-216 nt)) (11); an inverted loop using synonymous codons to maintain correct amino acid coding but disrupt the 3′ CDS stem-loop structure (arfA_IL) (11); a mutant encoding a catalytically inactive version of ArfA (arfA_A18T) (33); and the hairpin loop (from nucleotide 160 to 203) (arfA_sRNA) (Fig 4C). For context, both arfA mutants Δ(154–216 nt) and IL are known to still produce functional ArfA proteins (10, 11). The A18T mutant carries amino acid substitution that has been shown previously to diminish ArfA protein function to rescue stalled ribosomes (33). First, we observed that arfA deletion leads to an increase in mscL transcript abundance (2.13-fold, P = 0.0004) (Fig 4D), consistent with our earlier results (Figs 2B and 3A). Next, we analysed the transcript level of mscL in ΔarfA strains. In the strains expressing arfA_FL, arfA_A18T, and arfA_sRNA, mscL transcript abundance similar to the wt strain was observed. In contrast, in the strains expressing arfA_Δ(154–216 nt) and arfA_IL mscL transcript abundance was ≅ 2.0-fold higher than wt, and similar to ΔarfA (Fig 4D). These results show that only those arfA gene variants containing the minimal complementary sequence region (nucleotide 160–203) in their 3′ CDS can rescue mscL transcript levels, and therefore that this region of arfA, and/or the sRNA derived from it, plays a direct role in the negative regulation of mscL.

To investigate if the mscL regulation mediated by the sRNA of arfA has a phenotypic role, we monitored the MscL excretion activity in an E. coli arfA–deleted strain expressing sfGFP in combination with arfA_FL or arfA_sRNA undergoing osmotic and translation stress (Fig 4E). Consistent with previous analysis (Fig 4A), arfA deletion leads to decreased excretion (twofold, P = 0.014) relative to wt (Fig 4E upper panel). Intriguingly, mscL expression levels observed in the wt strain produce the greater ECP activity, whereas the higher mscL expression observed in the ΔarfA strain leads to reduced ECP (Fig 4E). Upon expression of the arfA variants (∆arfA_pFL and ∆arfA_psRNA), the excretion phenotype activity was rescued and compared with the wt strain. Steady-state analysis of mscL transcript levels (Fig 4E lower panel) confirms that the function of arfA sRNA is to negatively regulate mscL expression at the post-transcriptional level, and in turn to increase MscL excretion activity. Implicitly, this indicates that arfA sRNA functions as an antisense RNA (asRNA) and targets the complementary 3′ UTR region of mscL mRNA.

arfA asRNA regulates mscL transcript stability

Having established that arfA asRNA mediates the negative regulation of mscL expression, we sought to investigate if it regulates mscL transcript stability. We therefore measured the stability of the native mscL transcript in the presence (wt) and absence of arfA (∆arfA). An ∼twofold increase in the mscL transcript half-life was observed in the absence of arfA (Fig 5A). These results support a hypothesis that the observed down-regulation of mscL is caused by formation of a heteroduplex between arfA sRNA and mscL mRNA which results in a decreased stability of mscL transcript. Indeed, antisense mRNA heteroduplexes are known to be cleaved by either endo- (i.e., RNaseIII and RNase E) or exoribonucleases, resulting in the destabilization of the target RNA (34).

RNaseIII activity plays a role in the post-transcriptional regulation of mscL

arfA sRNAs are known to be released from the arfA transcript via RNaseIII cleavage (10, 11); we therefore next investigated whether RNaseIII affects the mscL transcript abundance via arfA sRNA. The strain missing a functional RNaseIII protein (RNaseIII(38)) (35) (Δrnc) showed a modest but significant decrease (1.5-fold, P = 0.0005) in mscL transcript level compared with the wt strain (Fig 5B). To further investigate the role of RNaseIII upon mscL expression in the absence of arfA, we sought to compare mscL transcript abundance in absence of both arfA and functional RNaseIII (ΔrncΔarfA) and of solely arfA (ΔarfA). A decrease in mscL transcript level (3.7-fold, P < 0.0001) was observed in ΔrncΔarfA strain (Fig 5B), suggesting that RNaseIII positively regulates mscL expression in an arfA independent manner. Comparison of mscL transcript abundance between both Δrnc genetic backgrounds (Δrnc and ΔrncΔarfA) show similar levels indicating that unprocessed arfA, present in Δrnc strain, is unable to down-regulate mscL expression. To validate the latter, we measured the steady-state level of mscL transcript in the ΔrncΔarfA genetic background bearing the arfA-sRNA or arfA-FL plasmid (Fig 5C). We observed a reduction in mscL transcript of about threefold (P = 0.0007) when arfA asRNA was expressed, consistent with the earlier results, confirming the role of arfA asRNA in mscL regulation (Fig 4E). In contrast, when arfA_FL was expressed, no reduction was observed, in fact, mscL abundance increased by around twofold (P = 0.0003) (Fig 5C), which could be due to the formation of a heteroduplex between mscL–arfA_FL. However, arfA transcript is believed to predominately exist in the truncated form in vivo (36), so it is unclear if this heteroduplex is natively found or is an artefact driven by the high levels of arfA expression from the rescue plasmid (Fig S5). Overall, these results indicate that RNaseIII has an arfA-dependent negative effect and an arfA-independent positive effect upon mscL expression.

mscL transcript is processed by RNaseIII in vitro

Our results suggest that RNaseIII may also regulate mscL expression in an arfA independent manner (Fig 5B); intriguingly, Kawano and co-workers identified a 49 nt small asRNA derived from the 3′ UTR of mscL (37). We therefore sought to investigate if the stem-loop within the mscL 3′ UTR (nucleotide +1 to +63) (Fig S6) is cleaved in vitro by RNaseIII similarly to arfA transcript. We synthesised two mscL transcript species, one bearing the mscL CDS and 3′ UTR stem loop (nucleotide +1 to +70) (mscL_FL) and one terminating at the stop codon (mscL_TAA). We also synthesised arfA transcript as a positive control. As previously reported, we observed that arfA transcript is processed by RNaseIII to afford two sRNAs ∼24 and ∼30 nts, which are derived from the 3′ CDS of arfA (nucleotide 166–189 and 190 to stop codon, respectively) (Fig 5D upper panel) (10, 11). In addition, we observed that mscL_FL is also a substrate of RNaseIII in vitro affording ∼25 nt sRNA as a major product which is not released from mscL_TAA transcript, following the same treatment (Fig 5D upper panel and Fig S7). Furthermore, Northern blot analysis performed with a probe complementary to mscL 3′ UTR, spanning between nucleotide +13 and +38, detected a band of ∼25 nt further confirming that the RNaseIII cleavage of mscL occurs within the 3′ UTR (Fig 5D lower panel). In summary, these results identified mscL as an RNaseIII substrate in vitro, which once cleaved releases ∼25 nt sRNA from its 3′ UTR, and therefore a truncated mscL transcript with a reduced/removed target site for the arfA asRNA.

Gene content analysis

3,822 prokaryotic proteomes were recovered from NCBI on 15/3/22. These translated CDS proteomes were recovered based on the criteria that they were first from prokaryota and second were listed as coming from either a reference or representative genome that was fully assembled to the “complete genome” level. These proteomes were filtered to remove any proteins tagged “pseudo” to remove proteins marked as problematic by the NCBI Prokaryotic Genome Annotation Process. HMMs corresponding to the ArfA protein (PF03889.16) and MscL protein (9PF01741.21) were recovered from Pfam (29) on 28/3/22 (ArfA) and 31/3/22 (MscL), respectively. HMMER (version 3.1b2) (http://hmmer.org/) was used to search for matches to the two HMMs within the collected proteomes. Hmmsearch was run using the gathering bit scores (25.5 for ArfA and 27.4 for MscL) provided by the HMMs to set model-specific score thresholds and the default values for all other settings (Supplemental Data 1). A Bonferroni corrected P-value was calculated by dividing the HMMER calculated P-value by two.

Intergenic distance analysis

Python (56) and several packages (NumPy (57), pandas (58), Biopython (59), and SciPy (60)) were used to calculate the distance between instances of the two genes found in the same genome. Distances were calculated from the last nucleotide of one CDS to the last nucleotide of the other CDS taking the minimum possible intergenic distance assuming genome circularity. Genomes were binned according to which genes they had (arfA only, mscL only, both genes, and neither of the genes) and then sub-binned within the “both genes” bin according to distance between the genes. Genomes in which the genes overlapped with one another were labelled “overlap,” those which were non-overlapping and closer than or at a distance of 110 nt were labelled “proximal,” and those which were also non-overlapping, but further away than 110 nt were labelled “distal.” In one case, there was a genome with one gene on a chromosome and the other on a plasmid. This genome was labelled as “distal” (Supplemental Data 1). Metadata and taxonomic information were retrieved for each of the genomes from NCBI on 12/4/22 (metadata) and 23/6/22 (taxonomic information) (Supplemental Data 1). Data were graphed using R, tidyverse (61), and Polychrome (62) packages. A cladogram of the genomes that we surveyed was produced from the NCBI taxonomy database using phyloT (https://phylot.biobyte.de). This was then visualised using iTOL (63).

Bacterial strains, plasmids, and media

E. coli strains and plasmids used in this study are listed in Table S3. To generate target gene deletions in the strain backgrounds MG1655 and BL21(DE3), lambda Red recombineering strategy was performed as described before (64), using pRL128 (65) and pSIM18 (66) plasmids. The primers used to generate FRT-flanked kanamycin selection cassettes and to construct the rescue plasmids are listed in Tables S4 and S3. For cloning purposes and general overnight starter culture, cells were grown in the LB medium (0.5% yeast extract, 0.5% NaCl, 1.0% bactotryptone). For gene expression study in minimal media, cells were grown in citrate–phosphate-defined medium, pH 7, per litre contains 8.58 g Na₂HPO₄, 0.87 g K₂HPO₄, 1.34 g citric acid, 1.0 g (NH4)₂SO₄, 0.001 g thiamine, 0.1 g MgSO₄.7H₂O, and 0.002 g (NH₄)₂SO₄.FeSO₄.6H₂O, with 0.04 (starter culture) or 0.2% of glucose (experimental culture), and supplemented with 0 or 0.3 M NaCl (media osmolality of ∼215 and 764 mOsm, respectively). For gene expression studies in high cell density culture, cells were grown in TB medium (2.7% yeast extract, 4.5% glycerol, 1.3% bactotryptone, 0.2% glucose). For gene expression study in cells missing a functional RNaseIII protein (Δrnc), cells were grown in LB medium (10 g/liter NaCl, 5 g/l yeast extract, 10 g/liter tryptone).

Construction of promoter–reporter plasmids

P_mscL-sfGFP and P_arfA-sfGFP fragments were synthesised as gBlocks (Integrated DNA Technologies). P_mscL-sfGFP contained a 321-bp fragment of the mscL promoter, including the first 11 codons of mscL (15). P_arfA-sfGFP contained arfA promoter referred to the sequence from Chadani et al (11). The sfGFP CDS is located after the promoter. The DNA fragments were cloned into the low-copy p131B plasmid using NEBBuilder HiFi DNA Assembly following the manufacture’s protocol, and the resultant plasmids were used for promoter–reporter activity study.

Construction of rescue plasmids

All arfA gene variant fragments were synthesised as gBlocks (Integrated DNA Technologies) and contained upstream the T5 promoter sequence and downstream the rrnB T1, T7, and lambda T0 terminator sequences. The variant arfA_Δ(154-216 nt) contained a truncated version of full-length gene missing nucleotide 154–216 nt. The variant arfA_IL contained inverted repeats: nucleotide position at 147 (G), 156 (G), and 159 (C) were substituted with A; nucleotide position at 162 (T) was substituted with C; and nucleotide position at 171 (A) was substituted with G. The variant arfA_A18T contained the substitution of the alanine amino acid codon (AGC) in position 18 with threonine (ACC). The variant arfA_sRNA contained only the sequence from nucleotide 160 to 203. The DNA fragments were cloned into the pCA24N plasmid using NEBBuilder HiFi DNA Assembly following the manufacture’s protocol, and the resultant plasmids were used for experiments in ΔarfA-rescued strains.

Growth conditions in minimal media

Cells were first inoculated in 10 ml of citrate–phosphate medium (50 ml falcon tube) supplemented with 0.04% glucose, 50 μg/ml carbenicillin (plasmid maintenance), and when required 50 μg/ml kanamycin, then grown overnight 12–16 h at 37°C 200 rpm. The following morning, cultures were supplemented with 0.2% glucose. After one doubling, cultures were diluted 20-fold into fresh citrate–phosphate media supplemented with 0 and 0.3 M NaCl (215 and 764 mOsm, respectively), 0.2% glucose, and 50 μg/ml carbenicillin, and continued to grow at 37°C at 200 rpm (New Brunswick Scientific I26 shaker) to reach exponential growth phase (4 h) or stationary growth phase (O/N). Samples were collected at indicated time points for further analysis.

Growth conditions for high-density culture

High-density cultures were obtained by growing the cells in TB medium with an osmolality of ∼350 mOsm. After an overnight pre-inoculum in TB, cultures were diluted to OD₆₀₀ 0.02 in 30 ml of fresh TB medium supplemented with 0.2% glucose (UY flask), and then continued to grow at 37°C at 200 rpm (New Brunswick Scientific I26 shaker) until the desired OD₆₀₀ was reached. When required, induction of recombinant protein (sfGFP) was performed by adding 250 μM IPTG at OD₆₀₀ ∼1. Under these conditions, the cultures can grow to an OD₆₀₀ up to ∼25 and undergo a growth-induced drop in media osmolality.

Media osmolality measurement

Cultures were centrifuged (17,000 g, 1 min or 5,000 g, 10 min) to separate media fraction from the cell pellet. The osmolality of the media fraction was measured with a cryoscopic osmometer (Osmomat 030; Gonotec) following the manufacture’s protocol.

GFP expression and ECP analysis

For promoter–reporter activity assay, GFP expressions from the intact cell were measured. Cultures were harvested (5,000g, 10 min) and washed twice in an equal volume of 1 × PBS. Relative fluorescence units (RFUs) and OD₆₀₀ were measured, and GFP expressions were presented as RFU normalised to OD (RFU/OD). For ECP analysis, GFP expressions from the media and cell fractions (periplasm and spheroplast) obtained as previously described (17) were measured directly as RFU. The extracellular localisation of GFP (ECP) was expressed as %ECP (RFU in the medium/RFU intracellular + RFU media). A BMG CLARIOstar microplate reader was used to measure the colorimetric fluorescence and cell density (OD₆₀₀) of intact cells.

RNA analysis by qRT-PCR

Samples for RNA analysis were collected from ∼10⁹ cells (OD₆₀₀ of 1) grown in minimal or rich media. Cellular RNA was stabilised using RNAlater (Invitrogen) following the manufacture’s protocol. RNA extraction was performed following the RNeasy mini protocol (QIAGEN). For quality control of extracted RNA, RNA integrity (RIN) was assessed using the Agilent 2100 Bioanalyser (Agilent Technologies) following the preparation protocol from RNA 6000 Nano kit (Agilent Technologies). Only RNA with a RIN > 7 (67) were used to generate the cDNA by the reverse transcriptase PCR. The RNA samples were snap-frozen in liquid nitrogen and then stored at −80°C. cDNA synthesis was performed following the SuperScript IV VILO protocol (Invitrogen), using ∼0.7 μg total extracted cellular RNA as a template. No RT enzyme control was also included to assess the presence of contaminating genomic DNA. The cDNA was stored for the short term at −20°C and long term at −80°C. The qRT-PCR reaction was performed in a volume of 10 μl in 96-well clear plates sealed with adhesive (MicroAmp Applied Biosystems) following the SYBR Green Master Mix manufacturer’s protocol (Thermo Fisher Scientific). In brief, each 10 μl reaction, including the no RT control for genomic DNA contamination, comprises 2 μl of 5× diluted cDNA, a final concentration of 1 × Syber mix (Thermo Fisher Scientific), and 300 nM forward and reverse primers (Table S5). Optimal gene–specific primers (Table S4) were designed using the Roche Online Assay Design Centre and submitted to the Basic Local Alignment Search Tool (BLAST) to check for the nonspecific binding. To compensate for error between samples (variations in cell number, RNA isolation, reverse transcription, etc), two endogenous “house-keeping” transcripts were chosen using the GeNorm algorithm (68). Initially, four reference genes used in previous analysis (recA, idnT, ffh, and cysG) (69, 70) were assayed under the experimental conditions used for this study; cysG and recA, whose expression did not change across the assayed conditions, were selected as the two most stable transcript. In the case of analysis in the nonfunctional RNaseIII strain, cysG was unstable and therefore recA was used for normalisation. Samples were amplified on a QuantStudio3 real-time PCR machine, and the temperature programme was set at 50°C for 2 min and then 95°C for 2 min followed by 40 cycles of 95°C for 15 s and 57°C for 15 s. Two to three technical and at least three biological replicates were performed for each sample. Transcript abundance of target genes expressed in a sample were normalised to the geometric mean of the two endogenous control genes. This is given by ∆C_t, where ∆C_t is determined by subtracting the geometric mean of endogenous genes C_t value from the average of target gene C_t value (C_t GOI–C_t Ref). Relative abundance (RA) was calculated by the following formula, RA = 100 × (E^-∆C_t), where E stands for amplification efficiency, as determined from the slope of the standard curve of each pair primers (E = 10^(−1/slope)). Only pair primers with a % E value (= −1 + 10(−1/slope) × 100) ranging between 90% and 110% were used in this study.

RNA decay analysis

To determine the stability of the mscL transcripts, cells were grown in LB medium at 37°C at 200 rpm (New Brunswick Scientific I26 shaker) until OD₆₀₀ ∼0.7, at which rifampicin (1 mg/ml) was added to block transcription. Samples were then collected 0,1,2,3, and 4 min after rifampicin addition for analysis of RNA decay by qRT-PCR as described above. The stability of mscL RNA was determined by plotting the relative amount of mscL versus the time, with the amount of mscL RNA at time 0 set to 100%. The relative mRNA abundance was calculated from C_t values of detected mscL normalised to the transcript level of the stable M1 gene (RNA component of RNaseP) (71). Half-life in minutes was determined by one-phase decay nonlinear fit in GraphPad Prism (version 9).

Western blot

Samples for protein analysis by Western blot were collected from cell cultures normalised to a volumetric OD₆₀₀ of 10. Cultures were centrifuged (5,000 g, 10 min), and pellet cells were then resuspended in 400 μl of lysis buffer (PBS 1× supplemented with Roche protease inhibitor and benzonase). Cells were lysed by sonication (30% Amp, 7 s “ON” 7 s “OFF,” cold condition). Lysates were mixed with 2 × SDS–PAGE loading buffer (1:1), and then were boiled for 10 min. Equal amounts of samples were separated by SDS–PAGE and then transferred onto PVDF 0.2 μM membrane using Trans-Blot Turbo apparatus (Bio-Rad). The membrane was blocked with 5% (wt/vol) skim milk in PBS 1 × (30 min, 50 rpm, room temperature), and then probed with primary antibody (mouse monoclonal anti-His from Pierce, 1:3,000; rabbit monoclonal anti-β-Pol from Abcam, 1:2,000) overnight, 90 rpm at 4°C. After washes with 1× PBS (three times), the membrane was then incubated with secondary antibody (anti-mouse and anti-rabbit 1:30,000), 30 min, 90 rpm, at room temperature. The protein signals were detected with a LI-COR Odyssey scanner and quantified based on densitometry analysis using LI-COR Image Studio 5.0. All data were measured in biological triplicate.

In vitro transcription

DNA templates containing mscL sequence spanning from the transcription starting site to the 3′UTR (70 nts long) (mscL_FL) or to the stop codon (mscL_TAA) and arfA sequence spanning from the transcription starting size to the stop codon, both bearing the T7 promoter sequence upstream, were synthesized as gBlocks (Integrated DNA Technologies). Transcription of RNAs in vitro by T7 RNA polymerase from the DNA templates was then carried out using the MEGAscript T7 kit (Ambion) according to the manufacturer’s instructions. RNAs were then checked for quality by electrophoresis on denaturing urea gel (10% TBU, Novex).

In vitro RNaseIII cleavage assay

In vitro transcribed RNAs were subjected to RNaseIII cleavage assay as follow: 10 pmol of RNA were used in 20 μl reaction volume containing 1× enzyme buffer (AM2290; Ambion). Reactions were pre-incubated at 37°C for 5 min, followed by the addition of 0.5 U of E. coli RNaseIII (AM2290; Ambion) and further incubation at 37°C for up to 15 min. Reactions were stopped by the addition 5 mM EDTA and purified with ZipTip C18 (10 μl bed; Millipore). Cleavage products were separated on denaturing urea gel (15% TBU; Novex) which was then stained with 1× SYBRGold and exposed to a PhosphorImager screen.

Northern blot

The Northern blot was performed according the “Northern blot analysis using bitoin PCR–labelled probes” protocol from LI-COR Odyssey website with few modifications detailed below. The mscL_UTR biotin–labelled probe, containing nucleotide +13 to +38 after the mscL stop codon, was synthesized as 5′/52-Bio/modified DNA oligo (5′/52-Bio/CACTTTTTTACCACTGGTCTTCTGCT) (IDT). The transfer of RNA from denaturing urea gel onto nylon membrane was performed at 1.0 A constant for 15′ using a Trans-Blot Turbo apparatus (Bio-Rad). EDC (E7750; Sigma-Aldrich) was used to cross-link the RNA to the nylon membrane as described previously (72).