The mRNA derived MalH sRNA contributes to alternative carbon source utilization by tuning maltoporin expression in E. coli.

Previous high-throughput studies in Gram-negative bacteria identified a large number of 3'UTR fragments that potentially function as sRNAs. Here we extensively characterize the MalH sRNA. We show that MalH is a stable degradation intermediate derived from the 3' end of malG, which is part of the maltose uptake operon transcript malEFG. Unlike the majority of bacterial sRNAs, MalH is transiently expressed during the transition from the exponential to the stationary growth phase, suggesting that it contributes to adaptation to changes in nutrient availability. Over-expression of MalH reduces expression of general outer membrane porins and MicA, a repressor of the high-affinity maltose/maltodextrin transporter LamB. Disrupting MalH production and function significantly reduces lamB accumulation when maltose is the only available carbon source, presumably due to the accumulation of the MicA repressor. We propose that MalH is part of a regulatory network that, during the transition phase, directly or indirectly promotes accumulation of high-affinity maltose transporters in the outer membrane by dampening competing pathways.

Introduction

Efficient adaptation to changing environmental conditions is fundamental to bacterial survival and success. The environments that bacteria naturally encounter typically lack optimal concentrations of nutrients, and available substrates can be found in complex mixtures. Thus, free-living bacteria often have to switch from catabolising one nutrient to another and co-utilize multiple substrates. Moreover, nutrient availability and bacterial metabolic strategies are closely linked to virulence [1-4]. Understanding bacterial physiology from a systems point of view requires a thorough understanding of how adaptation to changes in nutrient composition in the environment is controlled.

The most common experimental approach used to study changes in nutrient adaptation is using LB (Lysogeny broth) as a growth medium. Although not fully replicating natural environments, LB batch cultures have been widely used to model bacterial growth and provide a suitable system for studying bacterial growth under nutrient-limited conditions. LB is particularly well suited for investigating glucose-limited growth because glucose is only found in trace amounts in this medium [5]. Initially, Escherichia coli utilizes carbon sources from LB in a sequential manner [6,7]. First, it consumes its preferred nutrient, glucose. When the concentration of glucose drops below levels that allow sustained population growth, E.coli can transition to a scavenging mode [8], utilizing multiple alternative carbon sources from LB’s carbohydrate mix, including maltodextrins, D-mannose, melibiose, D-galactose, and ribose [5]. These carbon sources are not highly abundant in LB; thus by the time growth reaches the stationary phase, the population relies on the highly abundant amino acids and peptides [7,9]. The growth on sub-saturating nutrient levels and the changes that occur during the transition from exponential to stationary growth phase requires the expression of the appropriate transport systems for nutrient import, metabolic enzymes, as well as additional changes in gene expression for maintaining cellular homoeostasis. Given their general role in nutrient uptake, outer membrane proteins (OMPs) expression is coordinated with changes in nutrient availability [8,10]. Moreover, cells must tightly balance the expression of general and specialized transporters in the outer membrane to maintain membrane integrity [11].

While transcription factors are considered to have the main contribution in coordinating these changes in gene expression, post-transcriptional regulators, including Hfq and its associated base-pairing sRNAs, provide a key additional layer of regulation by controlling target mRNA transcription termination, degradation and/or translation [12]. Some sRNAs have direct roles in controlling nutrient uptake metabolism and transporter expression. For example, SdhX is involved in acetate metabolism [13,14], SgrS downregulates the expression of a major glucose transporter [15], Spot42 regulates catabolite repression [16], GcvB regulates peptide and amino acid transporters [17-19] and MicA represses the expression of maltoporin (LamB), the maltose-specific porin [20,21]. Additionally, MicA and RybB downregulate the expression of several OMPs [21-24].

A rapidly emerging class of sRNAs are encoded within mRNA genomic loci rather than at independent sites in the genome. Indeed, the 3ʹUTR fragments of many mRNAs co-immunoprecipitate with Hfq [25,26] and a subset have been shown to be functional Hfq-dependent sRNAs. While some mRNA-derived sRNAs are transcribed from regions overlapping a 3ʹUTR (e.g. DapZ [26]), others originate by the cleavage of a parental mRNA by RNases [13,14,27-29] or are independently transcribed before they are processed by RNases [30]. RNase E has been shown to be involved in the biogenesis of certain mRNA-cleaved sRNAs, but for others, the RNase responsible for processing has yet to be identified [30]. Moreover, 3ʹUTR-derived sRNAs can also associate with the RNA chaperone ProQ (e.g. RaiZ [28,31,32]). The recovery of hybrid fragments involving 3ʹUTRs of mRNAs in Hfq CLASH and RIL-seq data expanded the interactomes of known 3ʹUTR-derived sRNAs, but also uncovered many new 3ʹUTR-derived sRNA candidates, highlighting the prevalence of this class of riboregulators [33,34]. A few mRNA-derived sRNAs have been characterized in detail and have been shown to regulate critical aspects of bacterial physiology, ranging from the regulation of membrane assembly and composition [27,30,35,36], to controlling nitrate [29], acetate [13] and amino acid metabolism [37]. Their pervasiveness and key roles underscore the importance of 3ʹUTRs as reservoirs of sRNAs in enterobacteria.

Despite playing a fundamental role in rewiring gene expression, the functions of most Hfq-associated sRNAs during the transition from exponential to stationary phase have remained unexplored. To fill this gap, we recently employed CLASH to map the sRNA interactomes mediated by Hfq in E.coli during the exponential, transition and stationary phases of growth [33]. Our interaction maps have expanded the regulatory networks [34] with hundreds of novel sRNA–mRNA interactions, many of them specific to the transition stage. Here, we focussed our analyses on interactions that were specifically recovered during this phase. We identified and characterized a 3ʹUTR-derived sRNA, which we refer to as MalH. Unlike the majority of bacterial sRNAs, MalH is transiently expressed during the transition phase. We demonstrate that MalH is a degradation intermediate of the 3ʹ end of the malG transcripts, the last transcript of the malEFG polycistron encoding components of the maltose transport system [38]. We show that over-expression of MalH downregulates several mRNAs encoding major OMPs and directly or indirectly reduces the levels of MicA, an sRNA that downregulates the high-affinity maltose transporter LamB [20,21]. In addition, mutations in the seed sequence of the endogenous MalH reduce the accumulation of the lamB mRNA, presumptively caused by MicA de-repression.

We propose that the MalH sRNA is part of a regulatory network that, during the transition phase, promotes accumulation of high-affinity maltose transporters in the outer membrane by dampening competing pathways.

Materials and methods

Bacterial strains and culture conditions

An overview of the bacterial strains used in this study is provided in Supplementary Table 2. The E. coli MG1655 and TOP10 strains served as parental strains. Cells were grown in Lysogeny Broth (LB) or minimal medium with supplements (1xM9 salts, 2 mM MgSO4,0.1 mM CaCl2, 0.03 mM thiamine, 0.2% carbon source) at 37°C under aerobic conditions with shaking at 200 rpm. The media were supplemented with antibiotics where required at the following concentrations: ampicillin (Sigma, UK, A9518) – 100 µg/ml, chloramphenicol (Corning, – S, C239RI) – 25 µg/ml, kanamycin (Gibco, US–11,815-024) – 50 µg/ml. Where indicated, 0.2% glucose or maltose were used. For induction of sRNA expression from plasmids 0.2% L-arabinose (Sigma, A3256) was used. An overview of plasmids used in this study is provided in Supplementary Table 3.

Construction of sRNA expression plasmids

For the sRNA pulse over-expression constructs (Supplementary Table 3), the sRNA gene of interest was cloned at the transcriptional +1 site under PlacOara control by amplifying pBAD+1 plasmid [25] by inverse PCR using Q5 DNA Polymerase (NEB). The pBAD+1 template is derived from pBADmycHisA. The sRNA genes and seed mutants (SM) were synthesized as ultramers (IDT, Belgium), which served as the forward primers (Supplementary Table 4). The reverse primer (oligo pBAD+1_5P_rev) bears a monophosphorylated 5ʹ-end. The template plasmid was digested with 10 U DpnI (NEB) for 1 h at 37°C and the PCR product purified by ethanol precipitation. The pBAD-sRNA linear PCR product was circularized by self-ligation and transformed in DH5α competent cells. Positive transformants were screened by Sanger sequencing (Edinburgh Genomics, Edinburgh, UK). The control plasmid pBAD+1 was constructed similarly by self-ligation of the PCR product generated from oligonucleotides pBAD+1_XbaI_fwd and pBAD+1_5P_rev.

Construction of mRNA-superfolder GFP fusions

To generate constitutively expressed mRNA-sfGFP fusions for the fluorescence reporter studies, the 5ʹUTR, start codon and first ~5 codons of target genes were cloned under the control of PLtetO-1 promoter in a pXG10-SF backbone as previously described [19,39]. Derivatives of the target–GFP fusion plasmids harbouring seed mutations (SM) were generated using synthetic-mutated gene-fragments (Supplementary Table 4) (IDT, Belgium). To prepare the inserts, the target region of the mRNA of interest was either amplified by PCR from E. coli genomic DNA or synthesized as g-blocks (IDT, Belgium) and cloned using NheI and NsiI restriction sites and transformed in TOP10 cells. Transformants were screened by restriction digest analysis and verified by Sanger sequencing (Edinburgh Genomics, Edinburgh, UK).

GFP reporter system to quantify sRNA effect on target expression

A two-plasmid system was used to express each sRNA, and mRNA-sfGFP fusions [19,39] with modifications. The sRNA and sfGFP-fusion plasmids were co-transformed in E. coli TOP10 cells by electroporation and cells were maintained on dual selection with ampicillin and chloramphenicol. In TOP10 cells, the mRNA-sfGFP constructs are constitutively expressed, whereas the sRNA expression requires L-arabinose induction. TOP10 cells were chosen for this experiment as they are deficient in RecA recombinase, which reduces intermolecular recombination between plasmids. The expression of sfGFP-fused targets in the presence or absence of sRNAs was quantified at the protein level, by plate reader experiments and at the RNA level, by RT-qPCR (see below).

For the plate reader experiments, a single colony of the bacterial strain harbouring an sRNA-target-sfGFP combination was inoculated in a 96-well Flat Bottom Transparent Polystyrene plate with lid (Thermo Scientific, 152,038) and cultured overnight at 37°C in 100 μl LB supplemented with antibiotics and L-arabinose (Sigma, A3256) to induce expression of sRNAs. Next day, each overnight inoculum was diluted 1:100 by serial dilution, in triplicate, in LB with freshly prepared L-arabinose to a final volume of 100 μl. Cultures were grown in a 96-well plate in an Infinite 200 Pro plate reader (Tecan) controlled by i-control software (Tecan) for 192 cycles at 37°C with 1 min orbital shaking (4 mm amplitude) every fifth minute. To monitor the optical density over time, the following parameters were used: wavelength 600 nm, bandwidth 9 nm. Fluorescence was monitored with excitation wavelength 480 nm, bandwidth 9 nm and emission wavelength 520 nm, bandwidth 20 nm. Measurements were recorded at 5-min intervals, by top reading. Raw data were processed following guidance from previous reports [19]. First, the range of linearity of increase of fluorescence with OD600 was identified for all individual triplicates. Only the linearity range common to all triplicates was considered for further analysis. For each set of triplicates, the mean fluorescence was calculated at each OD600. To correct for background and cell autofluorescence, the mean fluorescence of a strain with plasmid pXG-0 was subtracted from all strains with GFP plasmids at the equivalent OD600. Ultimately, a curve was generated for each sample, plotting the background-corrected fluorescence (GFP) versus OD600. The experiments were performed for three biological replicates, and mean values and standard error of the means calculated for each strain.

RT-qPCR

For the superfolder GFP expression experiments, total RNA (12.5 µg) was treated with 2 U of Turbo DNase (Thermo Scientific, AM2238) for 1 hour at 37°C in a 10 μl reaction in the presence of 2 U SuperaseIn RNase inhibitor (Thermo Scientific, AM2694). The DNase was inactivated by 10 minutes incubation at 75°C. Reverse transcription (RT) was performed in a single reaction for all target genes of interest using a mix of gene-specific RT primers at 3.5 μM concentration each. After the addition of 2.5 μl RT primer mix, the RNA and primers were denatured at 70°C for 3 min, then snap chilled and incubated on ice for 5 min. RT was performed for 1 h at 55°C with SuperScript III (Thermo Scientific, 18,080,051) using 5 μl of RNA-RT primers mix in 10 μl final volume (100 U Superscript III, 2.5 mM DTT, 1xFS Buffer, 0.75 mM dNTPs) in the presence of 1 U RNasin (Promega, N2115). RT was followed by treatment with 5 U RNase H for 30 min at 37°C to remove the RNA from the RNA-cDNA duplexes. The cDNA was diluted 10-fold with DEPC water. Quantitative PCR was performed on 50 ng of DNAse I-treated total RNA using the Brilliant III UltraFast SYBR Green QPCR Master Mix (Agilent, #600,883). For sRNA pulse-over-expression and sRNA deletion experiments, the Luna Universal One-Step RT-qPCR Kit (NEB, E3005E) was used according to the manufacturer’s instructions. The qPCRs were run on a LightCycler 480 (Roche), and the specificity of the product was assessed by generating melting curves, as follows: 65°C-60s, 95°C (0.11 ramp rate with 5 acquisitions per °C, continuous). The data analyses were performed with the lightcycler software, at default settings: Absolute Quantification/Fit Points for Cp determination and Melt Curve Genotyping. The qPCR for all samples was performed in technical triplicate. Outliers from the samples with technical triplicate standard deviations of Cp >0.3 were discarded from the analyses. To calculate the fold-change relative to the control, the 2−ddCp method was employed, using recA or 5S rRNA (rrfD) as the reference genes where indicated. Experiments were performed for minimum of two biological replicates, and the mean fold-change and standard error of the mean were computed. Unless otherwise stated, the significance of the fold-change difference compared to the reference sample control (for which fold-change = 1 by definition) was tested with a one-sample t-test. Oligonucleotides used for RT-qPCR are provided in Supplementary Table 4.

MalH over-expression studies

For pulse over-expression studies overnight MG1655 cultures containing pBAD::sRNA and empty pBAD+1 control plasmids were inoculated in fresh LB-ampicillin medium at a starting OD600 of 0.05 and grown aerobically at 37°C to OD600 0.4. Pre-induction (0 min) and post-induction samples were harvested. For induction, cultures were supplemented with L-arabinose (Sigma, A3256) and rapidly collected by filtration and flash-frozen in liquid nitrogen at the indicated time-points. RNA was extracted from samples harvested 15 minutes post-induction (OD600 between 0.6 and 0.7) three biological replicate time-series, followed by RNA-seq library preparation, next-generation sequencing and DESeq2 [40] analysis of differentially expressed genes (Supplementary Table 1).

RNA extraction

RNA extraction was performed using the guanidium thiocyanate (GTC) method as previously described [41,42]. Briefly, 5 ml of cell culture was harvested at exponential phase by centrifugation. Cell pellets were resuspended in 100 µl GTC-Phenol (pH 7; 1:1 ratio) and lysed by vortexing the cells for 5 min with 100 µl of Zirconia beads (0.1 mm; Biospec products 11079101z). Subsequently, 550 µl of GTC was added and the mixture was vortexed for several minutes and incubated at 65ºC for 10 minutes and then 10 minutes on ice. Phases were separated by adding 300 µl chloroform isoamylalcohol (24:1) and 1/10th of a volume of 3 M NaAc (pH 5.2) followed by vigorous vortexing. After 5 minutes of centrifugation in an Eppendorf centrifuge (14 k rpm), the RNA was purified from the upper phase via two additional rounds of phenol-chloroform extraction and precipitated with 3 volumes of 96% cold ethanol.

RNA-seq

For the over-expression analysis of MalH, we generated RNA-seq libraries using an in-house protocol. Genomic DNA was removed by incubating 10 μg of total RNA with 2 U Turbo DNase (Thermo Scientific, AM2238) in a 50 μl final volume for 30 minutes at 37°C in the presence of 10 U SuperaseIn RNase Inhibitor (Thermo Scientific, AM2694). RNA was subsequently phenol-chloroform extracted and purified by ethanol-precipitation. Ribosomal RNA was removed with the Ribo-Zero rRNA Removal Kit (Gram-Negative Bacteria; Illumina, MRZGN126) according to the manufacturer’s instructions. Successful rRNA depletion was verified on the Agilent 2100 Bioanalyzer. The RNA was fragmented for 5 min at 95°C in the presence of Superscript III buffer (Invitrogen) followed by a five-minute incubation on ice. Reverse-transcription (RT) was performed with Superscript III (Thermo Scientific, 18,080,044) in 20 μl reactions according to manufacturer’s procedures using 250 ng of ribosomal RNA depleted RNA and 2.5 μM random hexamers (PE_solexa_hexamer, oligo 73, Supplementary Table 4). The RNA and free primers were degraded using 20 U of Exonuclease I (NEB, M0293l) and 50 U RNaseIf (NEB, M0243S) and the cDNA was purified with the DNA Clean & Concentrator 5 kit (Zymo Research). Ligation of the 5ʹ adapter (P5_phospho_adapter, oligo 39) to the cDNA was performed using CircLigase II (Lucigen, CL9021K) for 6 hours at 60°C, followed by a 10-min inactivation at 80°C. The cDNA was purified with the DNA Clean & Concentrator 5 kit. Half of the cDNA library was PCR amplified using Pfu polymerase (Promega, M7745) using the P5 forward PCR oligonucleotide and barcoded BC reverse oligonucleotides (200 nM; Supplementary Table 4; 95°C for 2 min, 95°C for 20s, 52°C for 30 s and 72°C for 1 min, and a final extension of 72°C for 5 min. 20 cycles of amplification). The PCR products were treated with Exonuclease 1 (NEB, M0293L) for 1 h at 37°C and purified by ethanol precipitation. Libraries were resolved on a 2% MetaPhor agarose gel 200–500 bp fragments were gel-extracted using the MinElute kit. All libraries were quantified on a 2100 Bioanalyzer using the High-Sensitivity DNA assay (Agilent, 5067–4627). Individual libraries were pooled in equimolar amounts. Paired-end sequencing (75 bp) was performed by Edinburgh Genomics on the Illumina HiSeq 4000 platform.

Seed-mutant studies

Wild-type MG1655 and seed-mutant strains were grown overnight in minimal medium with glucose. Next day, each starter culture was split and inoculated at OD600 0.05 in fresh M9 medium with glucose or maltose as the sole carbon source. Growth was monitored and cells were harvested at OD600 0.5. Total RNA was extracted, and gene expression was quantified by RT-qPCR or Northern Blot.

Polysome profiling analyses

Wild-type E. coli MG1655 containing empty pBAD plasmid and an isogenic strain containing pBAD: MalH was grown in LB until OD600 0.4, then treated for 15 minutes with L-arabinose (Sigma, A3256) to induce over-expression of MalH, and cycloheximide (Sigma, C4859-1 ML) at a final concentration of 100 µg/ml for 3 minutes. 200 ml of cells (OD600 0.6–0.7) were harvested by rapid filtration and flash frozen. The cells were washed in ice-cold PBS supplemented with 100 µg/ml cycloheximide.

Polysomal profiling was performed according to previously described protocols [43,44] with minor changes in the lysis buffer (10 mM NaCl, 10 mM MgCl2, 10 mM Tris-HCl (pH 7.5), 100 µg/ml cycloheximide, 1% (w/v) Na-Deoxycholate, 1 U RQ DNAse I (Promega, M6101), 0.6 U/mL RiboLock (Thermo Scientific, EO0381) in DEPC water). Lysates were kept on ice for 30 min, centrifuged 3X at 15,000 g for 10 min. The supernatants were loaded on a linear 10%–30% [w/v] sucrose gradient and centrifuged for 4 hours using an SW41 rotor at 40,000 rpm in a Beckman Optima XPN-100 Ultracentrifuge. Fractions of 1 mL in volume were collected monitoring the absorbance at 254 nm with the UA-6 UV/VIS detector (Teledyne Isco). Fractions from the entire gradient (total RNA) and from the fractions corresponding to ribosomes (70S) or polysomes (polysomal RNA) were pooled and RNA was purified by acid phenol–chloroform extraction as previously described [45].

Terminator™ 5′-PhosphateDependent Exonuclease treatment

Ten micrograms of total RNA extracted from cell-samples at OD600 1.2 and 1.8 were treated with 5ʹ-Terminator Dependent Exonuclease (Lucigen, TER51020) as per manufacturer instructions using Buffer A. The reaction was terminated by phenol extraction and ethanol precipitation, and the RNA was loaded on 8% polyacrylamide-urea gels and transferred to nylon membranes that were probed for MalH, CpxQ, RybB and 5S rRNA (Supplementary Table 4).

Primer extension analysis

One microgram total RNA was reverse-transcribed using SuperScript III reverse transcriptase (Thermo Scientific, 18,080,051) using 32P -radiolabelled oligonucleotides as primers (Supplementary Table 4). Primers were added to the RNA and annealing was performed by heating the samples at 85°C for 3 min and then snap chilling them on ice. The RT was performed for 1 h at 45°C, followed by Exonuclease I and RNaseIf (NEB M0293L and M0243S) (0.5 μl each) treatment for 30 minutes at 37°C. Reactions were stopped by mixing with an equal volume of 2XRNA loading dye (NEB, B0363S), 2 minute incubation at 95°C and snap chilled. The sequencing ladders were prepared with Sequenase v2.0 (Thermo Scientific, 70775Y200UN) according to specified instructions. Samples were resolved on 6% PAA/8 M TBE-urea gels and visualized using the FujiFilm FLA5100 scanner.

Construction of the MalH seed-mutant strain

To mutate the chromosomal copy of MalH, we used the λRed system [46]. We amplified the integration cassette from plasmid pKD4 with ultramers 895 and 896, containing homology regions to the coding sequence of malG, the desired MalH sequence and to the region immediately downstream of the Rho-independent terminator, respectively. With this design, the scar after removal of the Kanr cassette was expected at a site outside the MalH/malG sequence. The PCR product was electroporated in E. coli MG1566 strains carrying the pKD46 plasmid from which λRed recombinase was induced with 10 mM L-arabinose. Correct replacement of the MalH seed sequence was screened by colony PCR using primer pairs: 725 & 909 and 726 & 910 (Supplementary Table 4). The antibiotic resistance cassette was removed from substitution mutants by FLP-recombinase expressed constitutively from pE-FLP [47]. Successful allele replacement was confirmed by Sanger sequencing.

Northern blot analysis

Total RNA was extracted from cell lysates by GTC-Phenol extraction. For large RNA fragments, 10 μg of total RNA was resolved on a 1.25% BPTE-gel (pH 7) and transferred to a nylon membrane (HyBond N+, GEHealthcare, RPN1210B) by capillarity. For short RNA fragments, 10 μg total RNA was separated on an 8% polyacrylamide TBE-Urea gel and transferred to a nitrocellulose membrane by electroblotting for 4 h at 50 V. Membranes were pre-hybridized in 10 ml of UltraHyb Oligo Hyb (Thermo Scientific, AM8663) for 1 h and probed with 32P-labelled DNA oligo at 42°C for 12–18 hours in a hybridization oven. The sequences of the probes used for Northern blot detection are detailed in Supplementary Table 4. Membranes were washed twice with 2xSSC +0.5% SDS solution for 10 minutes and visualized using a Phosphor imaging screen and FujiFilm FLA-5100 Scanner (IP-S mode). For detection of highly abundant species (5S rRNA), autoradiography was used for exposure.

Computational analysis

Pre-processing of the raw sequencing data

Raw sequencing reads in fastq files were processed using a pipeline developed by Sander Granneman, which uses tools from the pyCRAC package [48]. The entire pipeline is available at https://bitbucket.org/sgrann/. The CRAC_pipeline_PE.py pipeline first demultiplexes the data using pyBarcodeFilter.py and the in-read barcode sequences found in the L5 5ʹ adapters. Flexbar then trims the reads to remove 3ʹ-adapter sequences and poor-quality nucleotides (Phred score <23). Using the random nucleotide information present in the L5 5ʹ-adaptor sequences, the reads are then collapsed to remove potential PCR duplicates. The reads were then mapped to the E. coli MG1655 genome using Novoalign (www.novocraft.com). To determine to which genes the reads mapped to, we generated an annotation file in the Gene Transfer Format (GTF). This file contains the start and end positions of each gene on the chromosome as well as what genomic features (i.e. sRNA, protein-coding, tRNA) it belongs to. To generate this file, we used the Rockhopper software [49] on E. coli rRNA-depleted total RNA-seq data (generated by Christel Sirocchi), a minimal GTF file obtained from ENSEMBL (without UTR information). The resulting GTF file contained information not only on the coding sequences but also complete 5ʹ and 3ʹ UTR coordinates. We then used pyReadCounters.py with Novoalign output files as input and the GTF annotation file to count the total number of unique cDNAs that mapped to each gene.

Differential expression analyses.

DESeq2 [40] was used on raw read counts generated by pyReadCounters from the pyCRAC package [48] to detect differentially expressed genes. Three MalH pulse-over-expression datasets were compared to three pBAD Control over-expression datasets. Only differentially expressed genes that had an adjusted p-value of 0.05 or lower were considered significant.

Normalization steps

To normalize the read count data generated with pyReadCounters.py and to correct for differences in library depth between time-points, we calculated Transcripts Per Million reads (TPM) for each gene. Briefly, for each time-point the raw counts for each gene were first divided by the gene length and then divided by the sum of all the values for the genes in that time-point to normalize for differences in library depth. The TPM values for each OD600 studied were then log2-normalized.

Multiple sequence alignments and conservation analyses

The homologous sequences of MalH in other enterobacteria were retrieved by BLAST. JalView was used for the multiple sequence alignments, using the MAFFT algorithm [50].

Data and Code availability

The MalH over-expression RNA sequencing data have been deposited on the NCBI Gene Expression Omnibus (GEO) with accession number GSE149059. The RNA-seq data generated from cells harvested at the various growth stages are available from NCBI GEO with accession number GSE123048. The python pyCRAC [48], kinetic-CRAC and GenomeBrowser software packages used for analysing the data are available from https://bitbucket.org/sgrann (pyCRAC up to version 1.4.3), https://git.ecdf.ed.ac.uk/sgrannem/, https://pypi.org/project/pyCRAC and https://pypi.org/project/GenomeBrowser.

Results

To characterize MalH, we analysed published E. coli RNA–RNA interaction data [33,34]. This revealed that MalH could potentially base-pair with 5ʹUTRs of mRNAs encoding several major porins, including the highly abundant ompC and ompA mRNAs (Fig. 1A-C). An interaction between MalH and another outer membrane porin (ompW) was also detected in the RIL-seq data [34], implying that MalH could potentially regulate many outer membrane porins. The most abundant interactions of MalH with mRNAs in the published datasets (ompC and ompA) utilized mostly one region in the malG 3ʹUTR for base-pairing (Fig. 1C, Stem 1). This suggests that the corresponding site on the predicted sRNA is likely the main seed. The predicted interaction between MalH and ompC is unusually long and consists of two stems interrupted by a bulge (Fig. 1C), suggesting that these two RNAs form a stable complex. Conservation analyses and in silico target predictions (CopraRNA [51,52]) indicated that the seed sequence part of stem 1 is relatively well conserved (Supplementary Fig. 1A) and could be utilized for the regulation of multiple mRNA targets (Supplementary Fig. 1B). The sequences in MalH and ompC that form stem 2 (Fig. 1C) are less well conserved (Supplementary Fig. 1A and 2A).

Figure 1.

MalH is a 3ʹUTR-derived sRNA that base-pairs with major porin mRNAs

To obtain the full-length sequence of MalH, we performed primer extension analyses to map the 5ʹ end of the sRNA (Fig. 2A). This revealed that MalH is a 104 nt sRNA that contains part of the malG coding sequence, the stop codon and the Rho-independent terminator (Fig. 2B). Next, we analysed the expression of this operon using our previously published RNA-seq data that contained gene expression information at various different growth stages in LB medium [33] (see Fig. 1B of this study for LB growth curves). The expression of malEFG, similarly to all other transporters and enzymes of the maltose system, is under the control of the MalT transcription factor that is bound to the maltotriose inducer and dependent on CRP-cAMP (Fig. 3A). Thus, these genes are expressed in the presence of maltose or maltodextrins and repressed in the presence of glucose due to catabolite repression. In LB, the steady-state levels of malG recapitulate a similar transient expression profile peaking between OD600 1.2 and OD600 2.4 (Fig. 3B).

Figure 2.

MalH is a 104nt sRNA

Figure 3.

MalH is a degradation product that emerges at the transition between exponential and stationary phases of growth

MalH sRNA levels generally peaked around OD600 1.8–2.4 (Fig. 2A and Fig. 3C). We suspect that this variability in MalH expression is likely due to experimental noise as a result of slight differences in sugar content of the various batches of LB used during the project. Nevertheless, we consistently observed a transient induction of MalH.

The malEFG operon is known to be subjected to RNase E-mediated processing [54], which led us to hypothesize that MalH is produced through endonucleolytic cleavage of malG. To test this, Northern blot analyses were performed with a malG 3ʹUTR probe to detect not only MalH, but also any processed malEFG RNA fragments containing this region. For this experiment, we used RNA extracted from cells harvested at different optical densities. We detected many shorter malG 3ʹUTR-containing RNA fragments including a very short and abundant RNA species that peaks at an optical density of 1.8, which we speculated could be MalH (Fig. 3C). RNase E cleavage was detected 10 nucleotides downstream of the site identified in MalH in the 3ʹUTR of malG in the Salmonella TIER-seq data [55] by primer extension (Supplementary Figure 1A). If MalH is indeed produced by RNase E cleavage, then the expectation would be that it has a 5ʹ-monophosphate and therefore be susceptible to 5ʹ-monophosphate dependent exonuclease (TEX) degradation. Consistent with this idea, data from published 5ʹ end mapping analyses that used TEX to identify processed RNAs, supported the existence of a short RNA in the same region that has a 5ʹ-monophosphate [56]. In this analysis, the 5ʹ end of MalH was quite heterogeneous; however, a potential RNAse E cleavage site was located 4 nucleotides downstream of the site mapped by primer extension (Supplementary Figure 1A). To confirm that MalH indeed has a 5ʹ monophosphate, we performed Northern blot analyses on RNA samples that were incubated with TEX (Fig. 3D). MalH could not be detected in our TEX treated RNA samples (Fig. 3D, lanes 2 and 4), confirming that it bears a 5ʹ monophosphate. The positive control CpxQ [27] was also degraded in the presence of TEX (Fig. 3D). In contrast, the independently transcribed RybB sRNA was a poor substrate for the exonuclease (Fig. 3D). Note that although the 5S rRNA has a 5ʹ monophosphate it is not degraded by TEX. TEX treatment of total RNA also reduced the levels of the longer intermediate species. These data support a previously described mechanism by which the full-length polycistronic RNA undergoes decay that is initiated at a site in the upstream malEFG region. The distal gene malE is clipped off by the degradosome and selectively stabilized, allowing its expression at higher levels than other members of the operon [54]. The malG 3ʹ end, however, would presumably be less susceptible to further degradation because it is stabilized by Hfq binding.

MalH directly regulates the expression of major outer membrane

To verify the MalH CLASH data we pulse over-expressed MalH under the control of an arabinose inducible promoter followed by RNA sequencing (Fig. 4A-B). Northern blot analyses of an arabinose time-series experiments revealed that MalH was maximally expressed already after 15 minutes of induction. Hence, to minimize secondary changes in gene expression, cells were harvested at this time-point and the resulting RNA was analysed by RNA sequencing (Fig. 4A, lane labelled with red rectangle). Note that the induction was performed at OD600 = 0.4, when endogenous levels of MalH (the band visible by Northern blot in the no sRNA control samples; Fig. 4A) are very low. Differential gene expression analysis identified ~20 transcripts that were significantly reduced upon MalH over-expression (Fig. 4B; Supplementary Table 1), implying these transcripts are regulated by MalH in vivo. This set of transcripts included those encoded by the sigma factor RpoE (σE) and the anti-σE protein RseA. Both are members of the same operon and play an important role in controlling gene expression during stress responses, including the envelope stress response [57-60].

Figure 4.

MalH destabilizes ompC and ompA mRNAs and downregulates key members the σE regulon

Intriguingly, MalH over-expression also reduced the levels of sRNAs RyeA and MicA, the latter depending on RpoE (σE) for its expression [61]. In Salmonella, MicA downregulates LamB, a high-affinity maltose/maltodextrin transporter [20]. Fragments of three mRNAs (ompC, ompA and ptsH) that were found in MalH chimeric reads were also differentially expressed, providing strong evidence that these are direct MalH targets.

To substantiate these analyses, we repeated the MalH over-expression experiments and performed RT-qPCR to verify some of the results from the differential expression analyses (Fig. 4C). For this experiment, we also included a MalH mutant in which the seed sequence (stem 1 sequence in Fig. 1C and Supplementary Fig. 2C) was changed into its complementary sequence (referred to as MalH SM mutant). As control, we also included the RybB sRNA, which regulates rpoE and ompC expression [21,22,24]. These results suggest that MalH, but not the seed mutant, can significantly reduce the levels of ompC, MicA and rseA (Fig. 4C). Although the qPCR data were generally in agreement with the RNA-seq data, we could not reproducibly detect significant changes in rpoE mRNA levels upon MalH over-expression. To our surprise, we did not observe significant changes in rpoE mRNA levels upon RybB over-expression, while levels of another target (ompC) were dramatically reduced. Notably, over-expression of the MalH seed mutant also resulted in a significant reduction in RyeA levels, implying that the observed changes are not a result of direct interactions between MalH and RyeA or that the regulation involves regions of MalH located outside of the main seed region.

To demonstrate the direct target regulation in vivo, we employed a well-established reporter system where an sRNA is co-expressed with a construct containing the mRNA target region fused to the coding sequence of superfolder green fluorescent protein (sfGFP) [19,39] (Fig. 5A). Fusions for OmpC, OmpA and σE were constructed, but only the OmpC and OmpA-sfGFP reporters produced stable fusions that could be analysed. We also included a MalH sRNA seed sequence mutant (MalH SM) and an ompC mutant containing compensatory mutations (OmpC SM; Fig. 5B). Importantly, the seed mutations in MalH did not affect the stability of the sRNA expressed from pBAD (Supplementary Fig. 3A and B). On the contrary, our OmpA-GFP reporter construct containing compensatory mutations in the target region did not generate a stable fusion. Therefore, we were unable to use this reporter system to fully verify the MalH-ompA interaction. As positive controls, we used sRNAs MicC and RybB sRNAs as they both regulate E. coli ompC expression [21,62]. Fluorescence measurements confirmed that levels of OmpC-sfGFP and OmpA-sfGFP fusions were significantly lower in cells expressing MalH (Fig. 5C). Importantly, MalH over-expression did not change the expression of the GFP reporter itself (Fig. 5C). Mutations in the MalH seed region largely restored OmpA- and OmpC-sfGFP reporter levels to the levels of the no sRNA negative control. The MalH SM mutant partially restored the expression of the OmpC SM-sfGFP mutant, suggesting that the base-pairing interaction between these two mutants is less stable compared to the wild-type (Fig. 5C). As expected, MicC was able to suppress OmpC SM-sfGFP levels as it base-pairs with the Stem 2 sequence in ompC (Supplementary Fig. 2C). The wild-type MalH was also able to partially suppress the expression of the OmpC SM mutant. We suggest that the predicted base-pairing interactions between MalH and ompC mRNA in the second stem (Supplementary Fig. 2C; ‘stem 2’) might be sufficient to at least partially suppress ompC expression. Regardless, the data strongly imply that MalH directly regulates ompC expression.

Figure 5.

Validation of MalH-ompC interaction using GFP reporters

To determine whether the changes in fluorescence were associated with changes in reporter mRNA levels, we measured the expression of the GFP reporters by RT-qPCR. The results were essentially identical to the GFP fluorescence measurements (Fig. 5D). Over-expression of the wild-type MalH, but not of the seed mutant, reduced ompC-sfGFP mRNA levels. The ompC seed mutation (SM) did not fully disrupt the regulation by wild-type MalH. However, the MalH mutant containing compensatory mutations (SM mutant) was able to suppress the ompC SM mRNA levels, consistent with the idea that base-pairing was partially restored.

Next, we performed polysome profiling experiments to assess the level of ompC recruitment on polysomes upon over-expression of MalH. Over-expression of MalH did not noticeably affect 70S and polysome levels (Fig. 6A). We observed a significant (~37%) reduction of ompC mRNA in the polysomal fractions, relative to the upper fractions (Fig. 6B). We hypothesize that MalH regulates ompC expression at the post-transcriptional and translational level.

Figure 6.

MalH regulates ompC mRNA translation in E. coli.

MalH tunes maltoporin expression during maltose fermentation

Considering that MalH is produced from the malEFG operon, we reasoned that the novel sRNA may be involved in adaptation to growth in maltose containing medium. Because our previous analyses demonstrated that mutating the stem 1 sequence in MalH did not completely disrupt the regulation of ompC (Fig. 5), we generated a strain in which the predicted base-pairing interactions of MalH with its targets (both stem 1 and 2; Supplementary Fig. 2C and Fig. 5B; MalH dSM (double seed mutant)) was disrupted. Subsequently, overnight cultures grown in glucose were split and (re)inoculated in fresh medium containing either glucose or maltose as the sole carbon sources, and expression of several mal regulon genes and MalH targets were quantified. We show that MalH and its parental transcript malG are almost undetectable during growth in glucose, and relatively highly expressed during growth in maltose (~35-fold increase, Fig. 7A and 7B). This is consistent with catabolite repression of the mal regulon by glucose, and its induction by maltose [38]. Intriguingly, we observed that ompC mRNA levels are overall significantly lower during growth in maltose, compared to glucose (Fig. 7A). This suggests that porin expression is also regulated by nutrient source in E. coli. Similarly, levels of MicA, a repressor of LamB synthesis [20,21], were reduced by ~30% in maltose compared to glucose (Fig. 7A). Notably, Northern blot (Fig. 7B) and qPCR (Fig. 7C) analyses revealed that when cells are grown in maltose, the MalH dSM mutant less abundant than the wild-type but longer processed fragments that contained upstream malG regions could still be detected (Fig. 7B). Although the main RNase E cleavage site was not mutated in the MalH dSM mutant, it appears that the mutations partially blocked MalH processing, which could impact the half-life of the malG mRNA as well. Unfortunately, because the probe designed to detect both MalH WT and MalH dSM detected only shorter MalH degradation intermediates but not the longer malEFG degradation intermediates (Supplementary Fig. 4; Supplementary Table 4; MalG_3ʹUTR), we were unable to accurately quantify to what extent processing of the MalH was impaired. As a compromise, we measured malG and malH levels by RT-qPCR (Fig. 7C). From these results, we estimated that in the MalH dSM strain malG levels increased roughly 1.6-fold and MalH levels decreased ~40% in maltose. To measure the impact of these MalH seed mutations, we first compared the expression levels of ompC, MicA, and lamB between the wild type and MalH dSM strains. In maltose, MicA levels increased by roughly 30% in the MalH dSM seed-mutant strain (Fig. 7D). Concomitantly, lamB levels were reduced by ~20% in this medium. OmpC expression did not significantly change in the mutant in either maltose or glucose (Fig. 7D). Because MalH over-expression slightly (but not significantly) reduced σE mRNA levels (rpoE; Fig. 4B-C) and σE controls the synthesis of MicA, we also analysed this mRNA by RT-qPCR in these strains. In the mutant strain, rpoE levels also did not significantly change in maltose, corroborating that MalH does not directly control its transcript levels (Fig. 7D). Based on these results, we hypothesize that when cells need to use maltose as a main carbon source, subsequent MalH expression enhances the uptake of maltose by (directly or indirectly) reducing MicA levels, independently of rpoE. This in turn could enhance the production of the LamB maltoporin (Figs. 7D and 8A). However, because the seed mutations also affected the synthesis of MalH, we cannot completely rule out the possibility that the observed changes in lamB and MicA levels were, at least in part, the result of defective RNase E cleavage of malG.

Figure 7.

MalH modulates expression of key factors involved in maltose intake

Discussion

Our previous Hfq CLASH study [33] identified over 100 distinct interactions between 3ʹUTRs and other mRNA regions, which implicate direct mRNA-mRNA communication. We proposed that many of these represent novel 3ʹUTR-derived sRNAs that target 5ʹUTRs of mRNAs. MalH is a new sRNA that was uncovered by Hfq CLASH and RIL-seq and consists entirely of the 3ʹUTR of the malG transcripts. This sRNA was of particular interest as it was mainly detected during the transition from late exponential to early stationary phase. Here we propose that MalH plays a role in improving maltose uptake by enhancing expression of the high-affinity maltose transporter LamB (Fig. 8A).

Figure 8.

An sRNA-target interaction networks that contributes to adaptation to nutrient availability

For our studies, we used Lysogeny broth (LB). LB is a well-defined medium that is rich in peptides and has a low content of utilizable sugars [9]. Although catabolisable amino acids are the main carbon source that support growth in LB [9], LB also contains sugar mixtures in short supply, which made it an ideal medium for our studies on nutrient limitation. These sugars are consumed one after another or co-utilized in E. coli cultures depending on their hierarchy and concentration [5,10,63]. Glucose, the preferred carbon source, is consumed first, followed by other fermentable sugars such as maltodextrins, nucleotides, and free sugars. During growth in LB, glucose is utilized during the exponential phase when the mal regulon genes are not expressed due to catabolite repression [38]. Glucose depletion is followed by simultaneous activation of systems for the utilization of non-glucose carbon sources. Accordingly, proteomics studies have shown that the entry into stationary phase in LB is characterized by increased expression of proteins required for scavenging nutrients and the uptake of alternative carbon sources: proteins of the phosphotransferase (PTS) system, maltose, amino acid and peptide transporters [64]. Studies in LB have shown that maltodextrins are the first substrates to support growth after glucose limitation [5]. Maltose/maltodextrins need to be transported into cells before their use in catabolic and anabolic processes. Efficient uptake of these carbon sources involves the maltose-specific transport system, which includes inner membrane, periplasmic and outer membrane (OMP) components encoded by malEFG and lamBmalK operons, respectively (Fig. 8A). The MalT transcription factor bound by maltotriose, induces transcription of genes encoding maltose transporters and mal enzymes. Although malEFG genes are transcribed as a single unit, RNase E processing of the polycistronic transcript can generate a differential expression of the monocistronic mRNAs [65]. Here we show that the MalH sRNA is a product of malEFG processing, which we hypothesize is stabilized by Hfq binding.

The induction of the Mal system and MalH during the transition phase in LB is a cumulative effect of glucose limitation, increased levels of cAMP and CRP, exogenous induction and synthesis of maltotriose endoinducer. By monitoring the expression of the targets of MalH during growth on maltose as the sole carbon source, we were able to specifically ascribe a hypothetical role to MalH in the optimization of exogenous maltose uptake. MalH is unique in a sense that it is not only an mRNA-derived sRNA targeting multiple pathways, but it is also because it is part of a ‘mixed coherent feed forward loop’ (FFL [66]) that we predict promotes maltose uptake via LamB. A major advantage of a coherent FFL is that it can accelerate or delay responses to stimuli, such as changes in carbon source availability. In this FFL, the key activator is the MalT transcription factor, which induces the expression of both malEFG and lamB when cells start to consume maltose (Fig. 8B). We propose that malEFG indirectly promotes LamB accumulation via the MalH sRNA by (directly or indirectly) reducing the expression of MicA. This downregulation is important as MicA interferes with translation of the lamB mRNA through base-pairing interactions [20]. Consistent with this model, in the MalH seed-mutant MicA was expressed at ~30% higher levels. Although a 15 min over-expression of MalH was not sufficient to significantly impact lamB levels (Supplementary Table 1), mutating the MalH seed sequences in the genome did reduce lamB levels by ~20% (Fig. 7D). While these changes appear modest, it is important to take into consideration the very high abundance of the lamB mRNAs. MicA expression is controlled by the sigma factor rpoE (σE) and we see a modest reduction of rpoE and rseA mRNA levels from the σE operon upon over-expression of MalH (Fig. 4C). This raises the possibility that MalH indirectly suppresses MicA by downregulating rpoE. However, the strain carrying mutations in the MalH seed sequence did not show significant changes in rpoE mRNA levels. Direct downregulation of MicA by MalH, however, is less likely as no evidence for base-pairing between these sRNAs was found in either our Hfq CLASH [33] data or by computational prediction. Furthermore, besides ompA and ompC, other mRNAs directly regulated by MicA (tsx, ygiM) were downregulated upon MalH over-expression (Fig. 4B, Supplementary Table 1) which would be expected to have gone up if MalH acted directly on MicA. Some of these genes are also regulated by other sRNAs, such as MicC (ompC), RybB (tsx, ompC, ompA), or RseX (ompA, ompC) [62,67,68] (Supplementary Fig. 2B-C), while the RNA binding protein CsrA, a carbon storage regulator, represses rpoE expression [69]. Therefore, it is possible that MalH acts in concert with these other post-transcriptional regulators to control the expression of these differentially expressed genes.

Our data imply that MalH base-pairs with the ompC mRNA and, when over-expressed, MalH can effectively reduce its steady-state levels. However, in medium with maltose as sole carbon source, endogenous MalH expression was not sufficient to significantly affect ompC mRNA levels. However, given the very high abundance and stability of ompC (minutes-long half-life [70]), we predict that even a mild reduction in ompC levels can profoundly relieve the pressure on the translation machinery and OMP assembly pathways [30]. The potential dual regulation by MalH (repression of OmpC and upregulation of LamB) could help balance levels of OMPs in response to changes in the available nutrients (Fig. 8A). Although this model is intriguing, the biological significance of MalH remains unclear. Interestingly, maltose has been shown to be a key nutrient for efficient colonization of mouse intestines by pathogenic and commensal strains of E. coli [71,72]. It would therefore be interesting to see if the MalH seed mutant has a competitive disadvantage in intestinal colonization.

MalH is reminiscent of the Vibrio cholerae MicX sRNA that also maps to the 3ʹ region of malG and regulates levels of outer membrane proteins [73]. However, the MicX targets (uncharacterized OMP and a peptide ABC transporter) are different from the MalH ones. Furthermore, unlike MalH, MicX is produced from an independent promoter. Conditions that change the expression of the upstream Mal operon in V. cholerae do not affect accumulation of MicX [73], whereas the levels of MalH strongly correlate with malEFG mRNA levels. Nevertheless, both studies demonstrate that an sRNA resides in the 3ʹUTR of the malG transcript and pinpoints a mechanism where the expression of multiple porins is connected by sRNAs encoded within transporter mRNAs.

Analysis of the Hfq CLASH and RIL-seq data revealed that MalH could also potentially control sRNA levels or activity. For example, we frequently recovered chimaeras that contained MalH fragments fused to fragments of a cis-encoded sRNA, OhsC. OhsC is an sRNA that is part of a toxin-antitoxin system that suppresses the production a short hydrophobic and toxic protein (ShoB) of which the expression needs to be tightly controlled [74,75]. It is tempting to speculate that MalH indirectly controls ShoB expression by sponging OhsC. Vice versa, OhsC could also control MalH activity/stability.

We also found interactions between MalH and transcripts that are part of the pts operon (ptsI and ptsH [76]) in both RIL-seq and Hfq CLASH data. This operon codes for three components of the phosphoenolpyruvate-dependent phosphotransferase system (PTS) that regulates the uptake, phosphorylation and metabolism of a number of energetically preferred sugars (PTS sugars) [77], such as glucose. Uptake of non-PTS sugars, such as maltose and maltodextrins, is regulated by the EIIAglc enzyme. In the presence of an excess of glucose, this enzyme accumulates in an unphosphorylated state, which blocks maltose/maltodextrin uptake by binding to the MalK homodimer (Fig. 8A) [78]. PtsI and PtsH are non-sugar specific components of the PTS system that can phosphorylate EIIAglc. Thus, we hypothesize that MalH post-transcriptionally enhances the production of PtsI and PtsH (presumably by stimulating translation) to maintain sufficiently high levels of phosphorylated EIIAglc.

In conclusion, our work provides an intriguing example of how an mRNA-derived sRNA can enhance the output of its regulon by not only dampening pathways that inhibit the accumulation of the encoded maltose transporters but possibly also by regulating the enzymes that control the activity of the maltose transporters.

Click here for additional data file.