Figure 1.
Organisation of the ram locus in Klebsiella pneumoniae / Enterobacter spp.
In K. pneumoniae romA and ramA are usually co-transcribed from the TSS depicted with *, under the control of the PI promoter. RamR can bind the inverted repeat (atgagtgn6cactcat), which in turn represses the transcription of both romA and ramA. Our analysis shows that the small regulatory RNA, sRamA5 and romA in Klebsiella pneumoniae share the TSS (depicted with *).
Figure 2.
A: Northern blot analysis of sRamA5.
15 μg of total RNA extracted from Ecl8/pGEMTpI+romA were loaded into wells. The blots were either incubated with the 32P-end labelled sRamA5 specific DNA probe or romA specific DNA probe. The bands pointed as 1 and 2 are primary transcripts for both the RNA codes for sRamA5 and romA. The band referred to as sRamA5 is specifically detected by the sRamA5 DNA probe, sized at around 60 nucleotides; the band referred to as romA was specifically detected by the romA DNA probe. B: EMSA of RamR-sRamA5 or RamR-PI interaction in the presence of sRamA5. (i). RamR-sRamA5 interaction. The concentrations of sRamA5 and RamR were 40 nM and 1 μM respectively. (ii). RamR-pI interaction in the absence/presence of sRamA5. Radioactive labeled pI was 2 nM from lane 1 to 4. RamR’s concentrations from lane 1 to 4 were: 0, 2, 0, 2 μM. Cold sRamA5’s concentrations from lane 1 to 4 were: 0, 0, 1, 1 μM. FR = free RNA, CR = RNA-protein complex, FD = free DNA, C = RNA-DNA-protein complex. C: qPCR for the level of romA and sRamA5’s transcriptions in Ecl8ΔramR. qPCR using LNA probe for determining the levels of sRamA5 transcription in Ecl8 and Ecl8ΔramR. Despite sharing the same TSS, the transcript levels of sRamA5 are not linked to romA levels, thereby reducing the likelihood of sRamA5 being a 5’ untranslated region of romA. The log2 fold changes in Ecl8ΔramR displayed in the bar chart are relative to their transcript levels in Ecl8. One-way ANOVA analyses (P<0.001) were performed to demonstrate statistical significance. D: qPCR assay for the level of ramR, romA and sRamA5 in Ecl8ΔramR pACYCramR and Ecl8ΔramR pACYC177. The log2 fold changes in the two strains displayed in the bar chart are relative to their transcript levels in Ecl8ΔramR. sRamA5 levels are elevated in the presence of ramR, implying that RamR could stabilise the sRamA5 transcript. One-way ANOVA analyses (P<0.001) were performed to demonstrate statistical significance.
Figure 3.
Venn diagram representing the RNA sequencing results.
Ecl8ΔramA or Ecl8ΔramR were used as calibrators in the pairwise comparisons. The arrows ⬇ indicates a lower than 0.5 fold decrease in transcription compared to calibrator; ⬆ indicates a higher than 2 fold transcription compared to calibrator. The numbers beneath A, B and C indicate the number of transcripts showing higher or lower transcription (based on statistical cut-off) compared to calibrator. The genes under the different categories A, B and C represent pairwise comparisons between Ecl8/Ecl8ΔramA, Ecl8ΔramR/Ecl8ΔramA and Ecl8/Ecl8ΔramR comparison respectively; the genes in Area AB were found to be differentially transcribed in both the Ecl8/Ecl8ΔramA and the Ecl8ΔramR/Ecl8ΔramA comparisons; the genes in Area AC were found to be differentially transcribed in both the Ecl8/Ecl8ΔramA and the Ecl8/Ecl8ΔramR comparisons; the genes found in the area BC were found to be differentially transcribed in both the Ecl8/Ecl8ΔramR and Ecl8ΔramR/Ecl8ΔramA comparisons. The romA gene in Area ABC was found to be differentially transcribed in all the three comparisons.
Figure 4.
A: Quantitative real-time RT-PCR validation of differentially expressed genes in Ecl8∆ramR.
All qPCR experiments were performed as outlined in materials and methods. Expression levels were normalized to 16S levels, and fold change values were generated by calibrating against Ecl8∆ramA. Genes designated BN373_36191, BN373_39031, BN373_03291 encode a putative membrane protein, oxidoreductase family and conserved hypothetical protein respectively. All data is a mean of 3 experiments. B: Electrophoretic Mobility Shift Assay (EMSA) using purified RamA protein. Following PCR amplification, each promoter region was end-labelled with 32P-γ ATP. Purified RamA (200 nM) and the different labelled DNA probes (2 nM) were incubated on ice. All reactions were performed on ice prior to electrophoresis on 7.5% native gel. Lane 1 of each panel indicates the labelled DNA probe only, Lane 2 is the BSA control and Lane 3 contains RamA+DNA. C: Transcription in vitro assay of different promoters using the purified RamA protein. The test DNA (2 nM- yrbF, ybhT, yhbW, acrA, nfnB, lpxO and lpxC) with the control template (gnd) were incubated for transcription in vitro [32P]α-UTP with (+) or without (-) 200 nM purified RamA. Samples were fractionated by polyacrylamide/urea gel electrophoresis prior to drying and exposure to the phosphorimager. Relative fold increase was determined using densitometric analysis as described previously [55], by first normalizing all test transcription levels to the control promoter (gnd) prior to comparison to the no protein control. Statistics was done using One way ANOVA (P value < 0.05) where transcription levels were found to be statistically significant in the presence of purified RamA compared to the no protein control.
Figure 5.
A. Gene Expression analyses of lpx genes.
All qPCR experiments were performed as outlined in materials and methods. Expression levels were normalized to 16S levels, and fold change values were generated by calibrating against Ecl8∆ramA. Two-way ANOVA analyses (P<0.05) were performed to demonstrate statistical significance. B. Regulation of lpx genes. (i) EMSA using lpxC and lpxK promoter regions. Purified RamA (200nM) and the different labelled DNA probes (2 nM) were incubated on ice. All reactions were performed on ice prior to electrophoresis on 7.5% native gel. (ii) Transcription in vitro of lpxC promoter region. Relative fold increase was determined using densitometric analysis as described previously [55]. Fold increases in the presence of RamA were determined by first normalising to the control promoter (gnd) prior to comparison to the no protein control. C: Lipid A analysis from K. pneumoniae Ecl8 (WT), Ecl8ΔramA, Ecl8ΔramR and Ecl8ΔramRA. Lipid A analysis was undertaken as described before [31]. Negative ion MALDI-TOF mass spectrometry of lipid A isolated from K. pneumoniae Ecl8 and its derivatives. Peaks in bold correspond to LpxO dependent 2’ secondary chain modifications.
Figure 6.
Survival assay of K. pneumoniae (Ecl8, Ecl8∆ramA, Ecl8∆ramR, Ecl8∆ramRA) to polymyxin B, colistin and the antimicrobial peptide LL-37.
The relative survival of bacteria (expressed as a percentage of the number of colonies obtained from the unexposed control of the same strain) in the presence of different concentrations of polymyxin B (A), colistin (B) and LL-37 (C) are shown. Asterisks indicate that results obtained for the ramA expresser, Ecl8ΔramR is significantly different (P < 0.05 by Two-way ANOVA) compared to Ecl8, Ecl8ΔramA and Ecl8∆ramRA.
Figure 7.
A: Attachment of K. pneumoniae Ecl8, Ecl8ΔramA, Ecl8ΔramR or Ecl8∆ramRA to murine macrophage RAW 264.7 cell line.
One-way ANOVA analyses were performed to demonstrate statistical significance. B: Microscopy to assess attachment to RAW 264.7 cell line. (i) Infection of the RAW264.7 cell line was carried out with K. pneumoniae Ecl8 (WT), Ecl8ΔramA, Ecl8∆ramR or Ecl8ΔramRA transformed with plasmid pRSMgfp. MOI was 1:100 and infections were carried out for 2 hrs. The actin cytoskeleton was stained with Acti stain 555 phalloidin (red) and host cell nuclei were stained with DAPI (blue). Images are representative of 80 fields. (ii) Graph representating mean values are derived from 3 independent experiments. One-way ANOVA analyses (P<0.001) were performed to demonstrate statistical significance. C: Internalisation of K. pneumoniae Ecl8, Ecl8ΔramA, Ecl8ΔramR or Ecl8∆ramRA by RAW 264.7 cells. Bacterial internalisation was assessed by the gentamicin protection assay. One-way ANOVA analyses were performed to demonstrate statistical significance. D: Enumeration of the extracellular non-phagocytosed K. pneumoniae Ecl8, Ecl8ΔramA, Ecl8ΔramR or Ecl8∆ramRA. One-way ANOVA analyses were performed to demonstrate statistical significance.
Figure 8.
Effect of RamA on bacterial recovery using an intranasal infection model.
Bacterial recovery (cfu/ml) was determined from lung (A) and spleen (B) homogenates following a 24h infection of 5–7 week old, C57BL6, female mice (n = 5) using the previously described intranasal infection model. Unpaired t-test analyses were performed to demonstrate statistical significance.
Table 1.
Strains used in this study.