1. Introduction
Small non-coding RNAs (sRNAs) are bacterial post-transcriptional regulators that control gene expression through varied molecular mechanisms affecting target mRNA/s availability, stability, and/or translation [
1]. sRNAs are classically divided into two main categories:
cis-acting and
trans-acting, depending on whether the sRNA is encoded within the same locus as its target mRNA(s) or a different locus, respectively.
Cis-acting sRNAs are either regulatory segments included in untranslated regions (UTRs) or encoded on the opposite strand of target mRNAs (e.g., antisense RNAs and excludons) and regulate exclusively that particular single gene. By contrast,
trans-acting sRNAs can regulate multiple targets through imperfect binding to the mRNAs [
2]. However, cumulative evidence in the last few years revealed a somehow overlooked mechanistic diversity in sRNA-based regulation, which suggests that this basic classification might be limited [
3].
sRNA-based regulation has been associated with virulence in most model pathogenic bacteria [
4,
5]. Hints linking sRNA function with virulence have most often relied on two facts. First, the sRNA locus belongs to a genetic segment carrying virulence-associated genes, which is absent in non-pathogenic strains (i.e., a pathogenicity island or PAI) [
6,
7], and second, the sRNA is induced during the course of the infection of eukaryotic cells [
8,
9,
10,
11]. Subsequently, the definite proof was that null mutants for the sRNA showed reduced virulence in in vitro and/or in vivo models of infection. However, the number of sRNAs whose deletion leads to an unequivocal attenuation phenotype is still exceedingly low compared to the number of sRNAs located in PAIs and/or induced in infection conditions. For instance, comparative genome-wide analyses showed that around one-fifth of all identified sRNAs (29 out of 150) in the foodborne pathogen
Listeria monocytogenes (
L. monocytogenes) are specifically expressed in intracellular bacteria infecting macrophages. On the other hand, a similar number of sRNA loci were exclusively encoded in pathogenic
L. monocytogenes strains and were absent in related non-pathogenic
Listeria species [
9]. However, to our knowledge, a solid functional association of
L. monocytogenes sRNAs with virulence has been established only for a handful of them [
8,
9,
12]. Thus, at present, the precise regulatory role of the majority of sRNAs induced in infection conditions and/or located in PAIs has not been described yet.
L. monocytogenes is a Gram-positive bacterial pathogen that causes listeriosis, one of the most serious foodborne diseases in humans and animals [
13]. The most relevant virulence factors are clustered in the genome making up PAIs, from which the
Listeria pathogenicity island 1 (LIPI-1) is present in most isolates of pathogenic species
L. monocytogenes and
L. ivanovii [
13]. LIPI-1 is highly induced in intracellular bacteria and comprises seven well-characterized protein-coding genes that are critical for intracellular survival and bacterial dissemination:
prfA,
plcA,
hly,
mpl,
actA,
plcB, and
orfX [
14]. In addition, there are two annotated sRNAs in LIPI-1. Rli74 and Rli51 are located in the 5′-UTR of the virulence factor genes
actA and
mpl, respectively [
7,
8,
9]. A single promoter drives the intracellular expression of actA, which encodes a cell surface protein essential for intracellular motility and the invasion of neighboring cells. This promoter generates an mRNA with a long UTR that includes Rli74. Deletion of
rli74 halted ActA-mediated actin recruitment and cell-to-cell spread, indicating that Rli74 is a
cis-acting sRNA essential to achieve a sufficient ActA level to drive bacterial motility within cell host cytoplasm [
15]. Similarly, the transcription start site of
mpl, which codes for a metalloprotease that activates other important virulence factors, such as the phospholipase PlcB, and is essential for cell-to-cell spread, is located 150 bps upstream of the start codon, suggesting that Rli51 represents a long 5′UTR of the
mpl transcript [
16,
17]. Intriguingly, there are no further molecular or functional analyses investigating the role of Rli51 as a
cis-encoded sRNA. Thus, currently, it is unknown whether Rli51 regulates the expression of the metalloprotease Mpl and the relevance that such
cis regulation would have for the virulence of
L. monocytogenes.
Evidence retrieved from different transcriptomic analyses of
L. monocytogenes challenges the view of Rli51 as a 5′-UTR of
mpl transcript. Specifically, tiling arrays suggested that Rli51 could be a short 121-nucleotide-long RNA induced in blood and the intestine of infected mice [
8]. Consistently, Term-seq, a genome-wide approach aimed to identify the 3′ end of expressed transcripts, identified a transcription termination for Rli51 upstream of the
mpl start codon [
18]. In fact, the occurrence of Rli51 as a short transcript has been recently verified by northern blot in
L. monocytogenes grown in PrfA-inducing conditions [
19]. Collectively, these observations suggest that, at least in some conditions, the transcription of Rli51 might not be coupled to
mpl mRNA. Therefore, it is not yet clear if Rli51 sRNA is embedded within the 5′-UTR of
mpl, is an independent transcript, or undergoes a condition-specific premature transcriptional termination (PTT). Furthermore, we do not know the molecular mechanisms that regulate this process and the functional role of Rli51.
In this work, we aimed to investigate the function of Rli51, an sRNA encoded in the LIPI-1 of L. monocytogenes. Analysis of the transcriptional landscape of the rli51-mpl locus under nutrient-rich laboratory conditions and in intracellular bacteria indicated the occurrence of a conditional regulation of Rli51 transcriptional read-through into mpl. Then, using translational fusions with egfp, we identified the sequence elements and potential trans-acting factors relevant to Rli51 transcription termination regulation. Finally, we explored the functional role of Rli51 in infection-relevant conditions through a transcriptomic analysis of rli51 deletion mutants collected from infected epithelial JEG-3 cells. Our work shows that Rli51 does not participate in mpl upregulation during the infection of eukaryotic cells. Instead, we found that the LIPI-1-encoded sRNA Rli51 functions as a trans-acting sRNA in intracellular bacteria.
3. Discussion
Rli51 is an sRNA specific from pathogenic
Listeria species induced under infection-relevant conditions [
7,
8,
9,
23,
32,
33]. However, both the mechanism of action and the physiological role of Rli51 have remained largely unexplored. Here, we have found that although Rli51 belongs to the 5′UTR of
mpl mRNA, it does not play any role in the upregulation of
mpl under infection conditions. Instead, Rli51 mediates a premature termination of transcription (PTT) in bacteria actively growing in a rich medium at 37 °C, which results in the production of a short 121-nucleotide-long sRNA (
Figure 5). Under these conditions, the transcriptional activity of the promoter is very low, which could allow Rli51 to fold as a transcriptional terminator with the assistance of
trans-acting sRNAs, such as Rli21/RliI (
Figure 5). By contrast, PrfA-dependent induction of
rli51 in intracellular bacteria leads to an increased abundance of short Rli51 while, at the same time, enables a transcriptional read-through into
mpl (
Figure 5). Consequently, under infection-relevant conditions, Rli51 occurs in two similarly abundant RNA species: a 121-nucleotide-long sRNA and the 5′UTR of
mpl mRNA. The accumulation of short Rli51 sRNA allows it to regulate target transcripts in intracellular bacteria, such as Lmo2186 and Lmo0937 (
Figure 5). Thus, Rli51 can regulate gene expression through two different mechanisms: (i) as a
cis-acting sRNA, where Rli51 would contribute to preventing leaky
mpl transcription under non-infection conditions by mediating a conditional PTT, and (ii) as a
trans-acting sRNA, where increased Rli51 levels in intracellular bacteria would contribute to fine tuning a pathogen’s response to the environment encountered within eukaryotic cell compartments.
RNA-mediated post-transcriptional control of genes involved in virulence is a common regulatory mechanism in bacterial pathogens [
4,
34], and in
L. monocytogenes in particular [
35]. Essential virulence factors in
L. monocytogenes are post-transcriptionally regulated through their 5′UTRs [
17]. For instance, the 5′UTR of
hly,
actA, and
inlA is necessary to achieve maximal expression levels under infection-relevant conditions [
15,
20,
36]. By contrast, in the case of the 5′UTR of
mpl, i.e., Rli51, it functions as a transcriptional attenuator and does not have a relevant role in the induction of
mpl in intracellular bacteria. The 5′UTR of
prfA is also a negative regulator that acts as a temperature-sensing regulator, which allows for PrfA translation at 37 °C (the temperature of the host) while preventing it at lower temperatures [
37]. In addition, the S-adenosylmethionine riboswitches in
L. monocytogenes can release the sRNAs SreA and SreB, which, in turn, regulate
prfA by binding to its 5’UTR, suggesting a link between nutrient availability and virulence [
26]. Similarly, it would be interesting to identify the physiological signals that ultimately control Rli51 function and, thus, Mpl.
Transcriptional attenuation refers to a blockade imposed by a specific conformation of an RNA segment (referred to as terminator) that functions conditionally. In a particular condition, binding of
trans-acting factors (such as the ribosome, a small molecule, an RNA-binding protein, or an sRNA) can stabilize the formation of a terminator hairpin that impedes transcriptional read-through into the downstream gene [
38]. In addition, in Gram-negative bacteria, the Rho factor serves as a global attenuator that terminates transcription of specific genes conditionally; sRNA binding within 5′ UTR of those genes disrupts Rho-mediated termination, thus leading to transcriptional read-through [
39]. Therefore, sRNA-mediated antitermination is a common mechanism of bacterial gene regulation. We show that Rli51 is predicted to fold as a transcriptional terminator and as an antiterminator; the latter being the most stable conformation. We experimentally verified that the sRNA Rli21 (RliI) binds Rli51, and the predicted binding site lies in loop L2, a single-stranded RNA region that is critical for Rli51 to mediate PTT. This suggests that Rli21 binding stabilizes the formation of the terminator fold in Rli51. Interestingly, Rli21/RliI expression does not change under infection-relevant conditions compared to growth in nutrient-rich broth [
8]. Therefore, in conditions of low transcriptional activity of the Rli51 promoter, Rli21/RliI might saturate Rli51 molecules, while under Rli51-inducing conditions, the excess free Rli51 would fold as an antiterminator (
Figure 5). Such a regulatory scheme exerted by Rli21, and/or additional
trans-acting factors, could explain the co-existence of both Rli51 and Rli51-
mpl in intracellular bacteria.
Rli51 mediates condition-dependent PTT, producing a 121 nt long sRNA that acts as a
trans-acting regulator for other genes. This conclusion is supported by three observations: (i) the absence of Rli51 in intracellular Δ
rli51 bacteria altered the expression of a set of genes; (ii) we predicted binding of some of these candidate target transcripts to Rli51 within ssRNA loops L1 and L2; and (iii) we experimentally confirmed they bind Rli51 in vivo. Specifically, we verified that Rli51 binds and positively regulates the expression of Lmo2186.
lmo2186 is transcribed as a bicistronic mRNA together with
lmo2185. lmo2185 mRNA levels (
Supplementary Table S2) are decreased in intracellular Δ
rli51 compared to EGD-e, suggesting that
lmo2185 and
lmo2186 are co-regulated by Rli51.
Lmo2185 and Lmo2186 were functionally annotated with the enriched GO term “response to iron” in our comparative transcriptomics (
Supplementary Table S3).
lmo2185 and
lmo2186 encode the fur-regulated hemin-binding proteins Hbp1 and Hbp2, respectively, which are homologs of IsdA and IsdC in
Staphylococcus aureus and function as a hemophore that scavenges ferric heme from the environment [
40]. The IsdG family heme monooxygenase Lmo0484, which breaks down heme to release free iron [
41], was also downregulated in intracellular Δ
rli51. Additionally,
lmo0541 (
fhud) and
lmo1960 (
fhuc), encoding two subunits of the iron siderophore ABC transporter FhuBCDG, were also downregulated in Δ
rli51 (
Supplementary Table S2). These findings suggest that Rli51 plays a role in upregulating iron scavenging systems under iron-limiting conditions, such as those found in intracellular compartments during infection. Interestingly, envelope stress induced by excess heme causes the LhrC4 sRNA to bind to
lmo2185,
lmo2186, and
lmo0484 transcripts, and, at least for
lmo0484, this resulted in target transcript downregulation [
42]. These results suggest that genes involved in heme uptake and degradation are post-transcriptionally regulated in response to both the excess and limitation of heme.
Our comparative transcriptomics revealed a consistent upregulation of genes belonging to LIPI-1:
hly, mpl, actA, plcB,
orfX, and
orfZ (see
lmo0202 to
lmo0207 in
Supplementary Table S2).
mpl, actA, and
plcB belong to the lecithinase operon and can be transcribed from the
rli51 promoter as a single mRNA [
14]. Consistently, the deletion of
rli51 resulted in upregulated endogenous
mpl or
mpl::gfp translational fusions in intracellular bacteria, as determined by targeted transcript quantification by RT-qPCR (
Figure 1A,C). Together, these results also suggest a
cis-acting mechanism of regulation by Rli51 in intracellular bacteria. The transcriptional activity of the promoter is very high, so it seems that the impact of this negative regulation mechanism is limited during intracellular infection. On the other hand,
hly,
orfX, and
orfZ are transcribed as monocistrons, which suggests that Rli51 could also regulate the expression of other virulence factors through a
trans-acting mechanism. In fact,
orfZ was the target candidate with the highest predicted interaction energy (
Figure 4A and
Supplementary Table S5). Future studies will help to ascertain if Rli51 has any role as a general repressor of the virulence program of
L. monocytogenes.
Our study sheds light on the function and mechanism of regulation mediated by an sRNA from LIPI-1 but entails limitations. For example, although we have identified Rli21/RliI as an sRNA that binds Rli51 at loop L2, our results do not preclude the involvement of additional sRNAs or other
trans-acting factors (such as a small molecule) in the regulation of Rli51-mediated PTT. On the other hand, the deletion of a single sRNA typically results in hard-to-detect subtle phenotypes with modest expression changes in DEGs [
43]. Our intracellular comparative transcriptomics was not meant to analyze comprehensively Rli51 regulon in intracellular bacteria though, for which other approaches might have shown clearer expression changes in putative target transcripts [
5,
44]. The aim of [
5,
44] was to identify physiologically relevant Rli51 targets in intracellular bacteria so as to prove a
trans-acting mechanism. Further studies will help to determine a more detailed view of the function of Rli51 in virulence.
Our work provides compelling evidence of the dual regulatory mechanism of Rli51, acting both in
cis and in
trans. Similarly, other sRNAs exhibit
cis- and
trans-acting regulatory mechanisms, influencing virulence [
26,
45], metabolism [
46,
47,
48], and nutrient uptake [
49], which suggests that this dual mechanism might be more common than previously anticipated [
50]. In fact, global mapping of 3′ ends in different bacterial species revealed that pervasive conditional PTT populates the transcriptome with 5′UTR-derived sRNAs that might have regulatory functions [
18,
50,
51]. As conditional PTT most often relies on riboswitches, these findings open a yet unexplored level of regulation that links the concentration of particular intracellular metabolites with the release of sRNA regulators [
26], suggesting an exciting intracellular sensing and regulatory crosstalk to control bacterial homeostasis.
4. Materials and Methods
4.1. Bacterial Strains and Growth Conditions
The
Listeria monocytogenes (
L. monocytogenes) strain used here was isogenic to wild-type strain EGD-e (serotype 1/2a) [
52]. The
rli51-null mutant was generated by allelic replacement. Homology regions flanking the
rli51 locus were generated by PCR and cloned one after another into the thermo-sensitive suicide integrative vector pMAD between the BamHI and KpnI restriction sites. Primer pairs used for the amplifications were pMAD-BamHI-AB-Fw with AB-StuI-CD-Rv and AB-StuI-CD-Fw with CD-KpnI-pMAD-Rv. Integration, excision, and screening steps were performed as described [
53]
. rli51 deletion was confirmed by Sanger sequencing.
Unless otherwise stated,
L. monocytogenes strains were grown at 37 °C in brain heart infusion (BHI) broth without shaking. To prepare the cultures in the exponential phase of growth in gene expression analysis experiments, bacteria were allowed to reach an OD
600 = 0.2 and then diluted two consecutive times. Specifically, a single colony was inoculated in 10 mL of BHI and grown at 37 °C in standing conditions. When it reached an OD
600 = 0.2, bacteria were diluted with fresh pre-warmed BHI and grown for at least 10 generations such that an OD
600 = 0.2 is not exceeded. For inducible expression experiments, strains carrying constructs cloned in the tetracycline-inducible expression vector pRMC2 [
54] were grown to the exponential phase (OD
600 = 0.2) (see above) and then exposed to anhydrotetracycline (final concentration: 0.05 µM or 0.1 µM) for 40 min. To transform plasmids into
L. monocytogenes, 1 µg of plasmidic DNA was electroporated at 2.4 kV, 200 ohms, and 25 μF in electrocompetent EGD-e and selected on plates supplemented with 10 µg/mL chloramphenicol (pRMC2 derivatives) or 5 µg/mL erythromycin (pMAD derivatives).
4.2. Cloning Procedures
Cloning details for each specific construct made for this work are included in
Supplementary Table S6. Cloning procedures were performed in the DH5α
E. coli strain. Bacteria were routinely grown in Luria Bertani (LB) broth at 37 °C with shaking at 180 rpms, and strains bearing pRMC2 or pMAD derivatives were selected with 100 µg/mL ampicillin. Ligation reactions and pure plasmids were transformed into chemically compentent bacteria [
53]. Q5 DNA polymerase (M0491S), T4 DNA ligase (M0202S), and restriction enzymes (KpnI: R3142S, SalI: R3138S, EcoRI: R3101S, BglII: R0144S, EcoRV: R0195S, DpnI: R0176S) were supplied by New England Biolabs. Kits for extracting plasmidic DNA (NucleoSpin Plasmid: 740588.50) and DNA from agarose gels (NucleoSpin Gel and PCR Clean-up: 740609.50) were supplied by Macherey Nagel. Rli51and MS2::Rli51 constructs and GFP translational fusions were cloned into pRMC2 plasmid [
54]. The GFP coding sequence was obtained from pAD1-cGFP [
55]. All constructs were assessed by Sanger sequencing.
4.3. Determination of L. monocytogenes Intracellular Proliferation (Gentamicin-Resistance Assay)
Human placental choriocarcinoma JEG-3 epithelial cells (ATCC HTB-36) and NRK-49F-9F (ATCC CRL-1570) rat fibroblasts were propagated in MEM and DMEM, respectively, supplemented with 10% FBS, 4 mM L-glutamine, and nonessential amino acids. Infections were carried on cells at an 80% confluence seeded on 24-well culture dishes. Prior to infection, L. monocytogenes strains were grown in BHI medium for 14 h at 37 °C in standing conditions. Bacteria were collected by centrifugation at 800× g for 1 min at room temperature and then resuspended in the corresponding cell culture medium. Bacterial suspension was added to the eukaryotic cells medium at a multiplicity of infection (MOI) of 10:1 (bacteria/eukaryotic cell). After a 60 min long co-incubation, non-internalized bacteria were removed by three consecutive washing steps with pre-warmed Hank’s Balanced Salt Solution (HBSS) (Gibco). To kill the remaining extracellular bacteria, a fresh medium containing 10 µg/mL gentamicin was added to infected cells. After incubation for 2 or 6 h in these conditions, the eukaryotic cells were lysed with 100 µL of a solution containing 1%TritonX-100–0.1%SDS for 5 min at room temperature. Serial dilutions of this suspension were spread onto BHI agar plates, and the colony-forming units (CFUs) were counted after overnight incubation at 37 °C. The intracellular proliferation rate was calculated by dividing the number of CFUs at 6 h post-infection by that at 2 h post-infection.
4.4. RNA Isolation from Cultured and Intracellular Bacteria
To prepare the samples for RNA extraction from bacteria grown in BHI broth (supplemented with 10 µg/mL chloramphenicol in pRMC2 plasmid-bearing strains), 0.2 volumes of an ice-cold solution containing 5% phenol and 95% ethanol were added to the culture and then kept for 30 min in ice to quench RNA metabolism. The suspension was scrapped off the plate and the bacteria collected by centrifugation at 9300× g at 4 °C for 10 min. Afterwards, the bacterial pellet was washed with 1 mL of an ice-cold solution containing 1% phenol and 19% ethanol diluted in 0.1% DEPC-treated water, transferred into a 1.5 mL tube, re-centrifuged at 9600× g for 5 min, and stored at −80 °C until used.
To obtain intracellular bacteria, human epithelial JEG-3 cells (ATCC HTB-36) were cultured with MEM supplemented with 10% FBS, 4 mM L-glutamine, and nonessential amino acids on 500- cm2-BioDish-XL plates (351040, BD Biosciences, Franklin Lakes, NJ, USA) to an ∼80% confluence (∼5.6 × 107 cells per plate). A single colony was inoculated in 1.5 mL of BHI (supplemented with 10 µg/mL chloramphenicol in pRMC2 plasmid-bearing strains) in a 10 mL glass tube with a vented cap and grown overnight at 37 °C without shaking. This culture was spun down for 1 min at 800× g at room temperature and resuspended in a serum-free culture medium at the same cell density. For each intracellular bacterial sample, two 500 cm2 plates were infected for 30 min at an MOI of 10:1 (bacteria/eukaryotic cell). Non-internalized bacteria were washed away with pre-warmed HBSS and killed by adding a fresh culture medium containing 100 µg/mL gentamicin. After a 2-h incubation, the infected cell culture was washed twice with pre-warmed HBSS, and a fresh medium supplemented with 10 µg/mL gentamycin was added. At 6 h after infection, the eukaryotic cell culture was washed five times with HBSS and lysed in a chilled solution containing 0.1% SDS, 1% acidic phenol, and 19% ethanol diluted in 0.1% DEPC-treated water for 30 min at 4 °C. The eukaryotic cell lysate was scrapped off the plate, and the intracellular bacteria were collected by centrifugation (27,500× g, 4 °C, 30 min). Bacteria were washed with the same lysis solution and re-pelleted (15,500× g, 4 °C, 10 min), and the pellet was stored at −80 °C.
Frozen pellets from bacteria cultured in BHI (total amount 3 ODs: 15 mL at an OD
600 = 0.2) and from intracellular bacteria (obtained from two 500 cm
2 plates of infected JEG-3 cells) were thawed, resuspended in a freshly prepared solution containing 10% glucose, 12.5 mM Tris-HCl pH 7.6, and 65 mM EDTA. Resuspended cells were transferred to a tube containing 1 volume of acidic phenol. This cell suspension was beaten with acid-washed glass beads (0.4 g ≈ ca. 0.2 mL) twice for 45-s-long cycles at 6 m/s in a FastPrep 24 (MP Biomedicals). As a centrifugation step to pellet unsoluble material and separate the phenolic phase (14,000×
g, 4 °C, 10 min), RNA was extracted from the aqueous phase with 1 mL of NZYzol (MB18501, Nzytech, Lisboa, Portugal) followed by two consecutive extraction steps with 0.5 mL of chloroform and then precipitated with isopropanol for 30 min at room temperature. RNA yield was quantified by spectrophotometry with a nanodrop ND2000 (Thermo Fisher Sci., Waltham, MA, USA), and 5 µg of RNA was treated with 3 U of DNase I (Turbo DNA-free kit: AM1907, Thermo Fisher Sci.) for 60 min at 37 °C. RNA integrity was assessed by running a 200 ng sample of the extracted RNA through agarose TAE electrophoresis. Genomic DNA carryover contamination-free RNA was verified by a negative result in a 30-cycle PCR using rpoB-F and rpoB-R primers (
Supplementary Table S7) and 20 ng of each RNA sample.
4.5. Reverse Transcription and Real-Time Quantitative PCR (qPCR)
cDNA samples were prepared from 1 µg of total DNA-free RNA using the Applied Biosystems High Capacity cDNA Reverse Transcription Kit (4368814, Thermo Fisher Sci.) including a mix of random hexamers. The RT-PCR protocol was 10 min at 25 °C, 120 min at 37 °C, and 5 min at 85 °C. qPCR was performed in a 10 µL final reaction volume, including 1 ng of cDNA as a template, 500 nM of each target gene-specific primer (
Supplementary Table S7), and 5 µL of the Applied Biosystems Power SYBR Green PCR Master Mix (4367659, Thermo Fisher Sci.). qPCR runs were performed in an Applied Biosystems 7500 Real-Time PCR System instrument (Thermo Fisher Sci.) using a standard amplification protocol (10 min at 95 °C; 45 cycles of 15 s at 95 °C and 1 min at 60 °C; dissociation curve of 15 s at 95 °C, 1 min at 60 °C, and a progressive temperature increase until 95 °C). Target-specific primers (listed in
Supplementary Table S7) were designed using Applied Biosystems Primer Express 3.0 (Thermo Fisher Sci.). As a quality control of each primer pair, the amplification efficiency of the corresponding qPCR assay was determined by running a standard curve comprising 12.5 ng of a pool of cDNA samples and five consecutive 1/5 serial dilution points. For each sample, a reverse transcription-minus control including 80 ng of RNA was analyzed with the same primer set to assess the extent of genomic DNA contamination.
For absolute RNA quantification, every plate included a standard curve made up of 5 points represented by serial dilutions of an EGD-e genomic DNA sample (12.5, 2.5, 0.5, 0.1, and 0.02 ng of gDNA). To calculate the number of molecules of target genes in each point of the standard curve, we considered a molecular weight of EGD-e gDNA of 1.82 × 109 g/mol (gDNA length 2.94 × 106 bps, average MW per nucleotide in dsDNA 619.96 g/mol) and the Avogadro No. 6.023 × 1023 molecules/mol. The fluorescence threshold was set such that it crossed the amplification curves (cycle vs. ΔRn) of all concentration points in the standard curve within the linear region of the amplification plot (exponential phase). Each experimental cDNA sample was run in triplicate, and the corresponding Ct was calculated as the mean of the three determinations. Individual technical replicates that deviated more than 0.75 cycles from the other two ones were not considered to calculate the mean Ct value. The No. of molecules of target mRNA transcripts in each experimental sample was calculated by interpolation of the corresponding mean-Ct values in the standard curve plot (Log10 [No. of molecules] vs. Ct). Absolute quantification of each target mRNA transcript was expressed as the No. of molecules per nanogram of total extracted RNA in each sample. The relative quantification of target mRNAs was calculated by dividing its absolute quantification by that of the rpoB transcript. Enrichment ratios in RNA pull-down experiments were calculated by dividing the relative quantification of each specific RNA in MS2-Rli51 eluate by that in the eluates of Rli51 control pull downs. rpoB was used as a housekeeping gene. Every qPCR assay included a non-template control (NTC), which included the same volume of H2O instead of the cDNA template.
4.6. Production of Recombinant His-MBP-CP
A 500 mL culture of a BL21
E. coli strain with pMAL-C2::His-MBP-CP [
56] was grown to OD
600 = 0.6 in LB supplemented with 10 mM glucose and 100 µg/mL ampicillin (final volume = 500 mL). His-MBP-CP ectopic expression was induced for 2 h after the addition of IPTG (final concentration: 0.3 mM). Afterwards, bacteria were collected by centrifugation (4000×
g, 4 °C, 20 min), and the pellet was frozen for 24 h at −20 °C. Bacteria were thawed, resuspended in 10 mL of a lysis buffer (50 mM Na
2HPO
4-NaH
2PO
4 pH = 8, 300 mM NaCl, 1% triton X-100, 10 mM β-mercaptoethanol, 10% glycerol, and 1× Roche EDTA-free complete protease inhibitor cocktail [04693116001, Merck, Rahway, NJ, USA]) per gram of bacterial wet weight and run through four consecutive freeze and thaw cycles in a dry ice/ethanol bath. The suspension was then sonicated (15 pulses of 15 s with 50% amplitude with 1-min-long breaks in ice in between) in a Q500 sonicator with a CL334 converter (QSonica, Newtown, CT, USA), and the lysate was clarified by centrifugation (9000×
g, 4 °C, 30 min). To purify His-MBP-CP, 3 mL (bed volume) of TALON
® Metal Affinity Resin (635503, Takara, San Jose, CA, USA) were equilibrated by incubating the resin in batch with 6 volumes of washing buffer (lysis buffer without glycerol and protease inhibitors) for 3 times. The equilibrated resin was incubated with the clarified lysate at a volume ratio of 1/10 (resin/lysate) on a wheel for 2 h at 4 °C, washed 6 times in batches with 6 volumes of washing buffer, and then eluted with washing buffer supplemented with 220 mM imidazole (12 fractions, 250 µL each). Protein concentration was determined with the Pierce BCA reagent (23225, Thermo Fisher Sci.), and the protein-containing fractions were pooled (ca. total volume: 1 mL) and dialyzed against 600 mL of RLN buffer (50 mM Na
2HPO
4-NaH
2PO
4 pH = 7.4, 140 mM NaCl, 1.5 mM MgCl
2, 1 mM DTT) at 4 °C for 16 h.
4.7. MS2-Affinity Purification of Rli51-Binding RNAs (RNA Pull Down)
To prepare bacterial lysates, L. monocytogenes EGD-e strains Δrli51:pRMC2-rli51-MS2 and Δrli51:pRMC2-Rli51 were grown to exponential phases (final volume of 600 mL, OD600 = 0.12) in BHI supplemented with 10 µg/mL chloramphenicol. Then, Rli51 and MS2-Rli51 expression was induced by adding anhydrotetracycline (ATc; final concentration: 0.2 µM). When the cultures reached an OD600 = 0.2 (ca. after a 40-min-long incubation with ATc), bacteria were collected by centrifugation (8500× g, 4 °C, 10 min), washed with column buffer (20 mM Tris-HCl pH 7.4, 200 mM NaCl, 1 mM EDTA, 1 mM DTT), and frozen at −80 °C for 24 h. The pellet was resuspended in 1 mL of column buffer, subjected to three consecutive freeze and thaw cycles in a dry ice/ethanol bath, and beaten with 0.4 g (ca. 0.2 mL) of acid-washed glass beads for three 20-s-long cycles at 2 m/s in a FastPrep 24 (MP Biomedicals, Auckland, New Zealand). After removing the beads, 4 µL of the Ambion™ RNase Inhibitor (cloned, 40 U/µL; AM2684, Thermo Fisher Sci.) were added to the lysate and then clarified by centrifugation (17,500× g, 4 °C, 30 min).
To prepare the affinity resin for MS2-tagged RNA pull-down experiments, the recombinant purified His-MBP-CP protein was first bound to amylose resin. A total of 500 µL (bed volume) of amylose resin (E8021S, New England Biolabs, Ipswich, MA, USA) was equilibrated in column buffer (6 resin volumes, 5 min, 3 times). This equilibrated resin sample was incubated in a batch with 10 volumes of column buffer containing 0.75 mg/mL of His-MBP-CP on the wheel (4 °C, 2 h). The resin was washed 4 times with 6 resin volumes of column buffer and stored at 4 °C until used (≤5 days).
To conduct the RNA pull down, a 100 µL aliquot (bed volume) of the amylose resin bound to His-MBP-CP (ca. 0.715 mg of His-MBP-CP) was mixed with 1 mL of the bacterial lysate and let to incubate for 2 h on the wheel at 4 °C. The resin was washed 4 times with 6 volumes of column buffer and subsequently transferred to a disposable empty Poly-Prep® chromatography column (7311550, Bio-Rad, Hercules, CA, USA). Elution was performed with 3 volumes of the column buffer supplemented with 10 mM maltose. The fractions were pooled (total volume: 300 µL), and the RNA was extracted with NZYol (see above).
4.8. PASIFIC Structural Analysis of Rli51
In silico analysis to predict the occurrence of alternative terminator/antiterminator RNA conformation compatible with a riboswitch or an attenuator was carried out using the PASIFIC webserver (
https://www.weizmann.ac.il/molgen/Sorek/PASIFIC/; accessed on 16 March 2022) [
25]. A 128-nucleotide-long sequence starting from the transcriptional start site of rli51 was selected as the input. This included the 121 nucleotides annotated as the Rli51 transcript plus a U-rich 7-nucleotide-long downstream sequence (AGUUUUA). Default parameters optimized to detect attenuators were used.
4.9. sRNA-sRNA Interaction Prediction
In silico prediction of sRNA-Rli51 pairwise interactions was performed by conducting multiple one-to-one interaction analyses of each sRNA with Rli51 using the command-line version of IntaRNA 2.0 [
28,
57]. A batch analysis was run using a manually created
L. monocytogenes sRNAs list file in FASTA format as the query and the Rli51 sequence as the target. sRNA sequences were obtained from the Listeriomics database [
29].
4.10. RNA Sequencing: Library Construction and Next-Generation DNA Sequencing
Library construction and DNA sequencing were performed by Novogene. Prokaryotic RNA samples were depleted of ribosomal RNA and purified by ethanol precipitation. After fragmentation, the first strand of cDNA was synthesized using random hexamer primers. During the second strand cDNA synthesis, dUTPs were replaced with dTTPs in the reaction buffer. The directional library was ready after end repair, A-tailing, adapter ligation, size selection, USER enzyme digestion, amplification, and purification. The library was checked with Qubit and real-time PCR for quantification and a bioanalyzer for size distribution detection. Quantified libraries will be pooled and sequenced on the Illumina platform Novaseq6000, according to effective library concentration and data amount required.
4.11. RNA Sequencing Data Analysis
For data analysis, raw reads were curated by trimming the adapters and removing poly-N sequences and low-quality reads. Only high-quality data were used for downstream analysis. Bowtie2 software [
58] was used for indexing the reference genome (NC_003210.1) and mapping the reads. Reads mapping to each gene were counted by FeatureCounts and used to calculate the FPKM for estimating the expression levels of each gene [
59]. Principal component analysis (PCA) using these expression levels was conducted using the FactoMineR R package to determine the overall reproducibility of our RNAseq biological replicates. Differential expression analysis between EDG-e (control) and ∆
rli51 sample groups was performed by the DeSEQ2 R package based on the negative binomial distribution, and
p-values were adjusted by the Benjamini–Hochberg correction to control the false discovery rate. Only those features with a
p-value < 0.05 were considered significant. A heatmap of the normalized expression levels of each significant gene was visualized using Morpheus from the Broad Institute (
https://software.broadinstitute.org/morpheus; accessed 20 November 2023). For normalization, each gene FPKM was normalized by the average of the FPKM of that gene in all samples and then transformed by log2. The ClusterProfiler R package was used to perform statistical enrichment of differential expression genes based on GO terms and known regulons of
L. monocytogenes available in curated databases: PRODORIC (
http://www.prodoric.de/; accessed 5 February 2021), RegPrecise (
https://regprecise.lbl.gov/index.jsp; accessed 5 February 2021), and CollectTF (
http://www.collectf.org/browse/home/; accessed 5 February 2021). Protein–protein interaction (ppi) networks were built by searching the differential expression genes in the STRING database (
http://string-db.org/; accessed 20 November 2023).