Tigecycline challenge triggers sRNA production in Salmonella enterica serovar Typhimurium

Abstract

Background: Bacteria employ complex transcriptional networks involving multiple genes in response to stress, which is not limited to gene and protein networks but now includes small RNAs (sRNAs). These regulatory RNA molecules are increasingly shown to be able to initiate regulatory cascades and modulate the expression of multiple genes that are involved in or required for survival under environmental challenge. Despite mounting evidence for the importance of sRNAs in stress response, their role upon antibiotic exposure remains unknown. In this study, we sought to determine firstly, whether differential expression of sRNAs occurs upon antibiotic exposure and secondly, whether these sRNAs could be attributed to microbial tolerance to antibiotics.

Results: A small scale sRNA cloning strategy of Salmonella enterica serovar Typhimurium SL1344 challenged with half the minimal inhibitory concentration of tigecycline identified four sRNAs (sYJ5, sYJ20, sYJ75 and sYJ118) which were reproducibly upregulated in the presence of either tigecycline or tetracycline. The coding sequences of the four sRNAs were found to be conserved across a number of species. Genome analysis found that sYJ5 and sYJ118 mapped between the 16S and 23S rRNA encoding genes. sYJ20 (also known as SroA) is encoded upstream of the tbpAyabKyabJ operon and is classed as a riboswitch, whilst its role in antibiotic stress-response appears independent of its riboswitch function. sYJ75 is encoded between genes that are involved in enterobactin transport and metabolism. Additionally we find that the genetic deletion of sYJ20 rendered a reduced viability phenotype in the presence of tigecycline, which was recovered when complemented. The upregulation of some of these sRNAs were also observed when S. Typhimurium was challenged by ampicillin (sYJ5, 75 and 118); or when Klebsiella pneumoniae was challenged by tigecycline (sYJ20 and 118).

Conclusions: Small RNAs are overexpressed as a result of antibiotic exposure in S. Typhimurium where the same molecules are upregulated in a related species or after exposure to different antibiotics. sYJ20, a riboswitch, appears to possess a trans-regulatory sRNA role in antibiotic tolerance. These findings imply that the sRNA mediated response is a component of the bacterial response to antibiotic challenge.

Figure 1

A pie chart representation illustrating the cDNA sequences distribution pattern on the SL1344 chromosome. The results showed that 50% of the sequences are encoded within IGRs, 90% of which are situated between 16S and 23S rRNA (shown on the right), 31% are tRNA sequences, 6% are part of rRNA sequences, 9% completely overlap with ORFs, and 4% partially overlap with ORFs.

Figure 2

(A) Northern blot analysis for the four sRNAs (sYJ5, sYJ20 (SroA), sYJ75 and sYJ118) that were upregulated in the presence of tigecycline, and (B) bar chart illustration of the overexpressed sRNAs and (C) chromosomal locations and the directions of transcription of sYJ5, sYJ20, sYJ75 and sYJ118.A) Northern blot analysis for sYJ5, 20, 75 and 118. Image on top: all lanes marked by - were loaded with SL1344 total RNA extracted from cells grown under normal conditions (RDM, shaking, 37°C); all lanes marked by + were loaded with SL1344 total RNA extracted from cells challenged with half the MIC of tigecycline (0.125 μg/ml). Image below: representative image of the internal reference of 5S RNA levels in the same RNA samples. B) Densitometric analysis of the data from northern blot experiments of challenged / unchallenged cells with half the MIC of tigecycline. After normalisation to the 5S RNA levels, relative fold increases for sYJ5, 20, 75 and 118 were found to be 8, 2, 2, and 8 fold, respectively compared to unchallenged cells. Error bars are generated based on three independent experiments. C) The three coding sequences of sYJ5 are located in (1) SL1344_rRNA0001-rRNA0002, (2) SL1344_rRNA0014-rRNA0015 and (3) SL1344_rRNA0017-rRNA0018. The two identical copies of sYJ118 are encoded in (1) SL1344_rRNA0010-rRNA0009 and (2) SL1344_rRNA0011-rRNA0012, and the other five paralogs are found in (1) SL1344_rRNA0001-rRNA0002, (2) SL1344_rRNA0006-rRNA0005, (3) SL1344_rRNA0014-rRNA0015, (4) SL1344_rRNA0017-rRNA0018 and (5) SL1344_rRNA0020-rRNA0021.

Figure 3

Northern blots for sYJ5, sYJ20 (SroA), sYJ75 and sYJ118 A) in SL1344 challenged with half the MIC of tetracycline, B) ciprofloxacin or ampicillin, and the four sRNAs level in E. coli and K. pneumoniae challenged with half the MIC of tigecycline.A) Lanes with - were loaded with control samples; lanes with + were loaded with total RNA extracted from cells challenged with half the MIC of tetracycline. This image is composite from different experiments. B) Lanes marked by - were loaded with control total RNA extracted from S. Typhimurium. Lanes marked as C were loaded with the total RNA extracted from S. Typhimurium that was challenged with half the MIC of ampicillin (1 µg/ml). Lanes marked by K- were loaded with the control total RNA extracted from K. pneumoniae. Lanes marked as K + were loaded with the total RNA extracted from K. pneumoniae that was challenged with half the MIC of tigecycline. Lanes marked as E- were loaded with the control total RNA extracted from E. coli. Lanes marked as E + were loaded with the total RNA extracted from E. coli that was challenged with half the MIC of tigecycline. Probe sequences were checked for 100% identity match in K. pneumoniae and E. coli prior to use.

Figure 4

Northern blots for A) the 5S RNA level in SL1344 and B) sYJ20 level in SL1344 and the Δhfqstrain (JVS-0255) in the presence of ciprofloxacin.A) Lane 1 and 3 (also labelled as -) were loaded with SL1344 total RNA extracted from cells grown under normal conditions (RDM, shaking, 37°C); lane 2 was loaded with SL1344 total RNA extracted from cells challenged with half the MIC of tigecycline (0.125 μg/ml); lane 4 was loaded with SL1344 total RNFA extracted from cells challenged with half the MIC of tetracycline (1 μg/ml). All lanes were loaded with 125 ng of total RNA. The experiment was repeated 4 times. Densitometric analysis of the results showed little or no difference in 5S RNA expression level in the three growing conditions (5S_tigecycline: 5S_control = 0.88, 5S_tetracycline : 5S_control = 1.15, average of 4 different experiments). B) Both strains (SL1344 and the hfq deletion strain (JVS-0255, Table 2)) were challenged with sub-inhibitory concentration of ciprofloxacin (0.0078 μg/ml) before the total RNA was extracted and probed for sYJ20 by northern blot. As shown above, the Δhfq strain (right lane) produced less sYJ20 compared to SL1344 (left lane). 5S RNA was used as a loading control.

Figure 5

The chromosomal location of the sYJ20 (SroA) encoding region and its encoding sequence. sYJ20 is encoded upstream of the tbpA-yabK-yabJ operon, and the shared TSS of sYJ20 and tbpA as determined by 5’ RACE analysis is represented by the dark-black arrow. The DNA sequence of sYJ20 (SroA) is shown in bold letters, which is also the region that was deleted in YJ104 and used for TargetRNA prediction (Table 1). The THI-box sequence is underlined. The start codon of tbpA is displayed at larger size as GTG, where the first G is considered +1 in the numbering system.

Figure 6

qPCR on sYJ20, tbpA and stress responsive genes (dinF and ycfR) on SL1344 control (no challenge with antibiotics), SL1344 challenged with half the MIC of tigecycline (0.125 μg/ml), and SL1344 challenged with half the MIC of tetracycline (1 μg/ml). QPCR was performed as described in Materials and Methods. All the fold changes are calculated relative to the value of the control (SL1344, unchallenged). Error bars are generated from at least 4 experiments.

Figure 7

Survival rate assays of SL1344, YJ104, YJ107 and YJ110 when cells were challenged with MIC of tigecycline. Fresh overnight culture was spread on RDM plates either supplemented with MIC of tigecycline (0.25 μg/ml) or nothing (as a control). Colony number was determined after overnight incubation at 37°C. Survival rate was calculated as follows: cfu/ml on the tigecycline plate divided by cfu/ml on the control plate. P values were also calculated from at least three biological replicates. We found that statistical comparisons of SL1344 versus YJ104 (ΔsYJ20) and YJ107 (YJ104/pACYC177·sYJ20) versus YJ110 (YJ104/pACYC177) are significant (P < 0.05)