Autoinducer-2 Quorum Sensing Contributes to Regulation of Microcin PDI in Escherichia coli

Abstract

The Escherichia coli quorum sensing (QS) signal molecule, autoinducer-2 (AI-2), reaches its maximum concentration during mid-to-late growth phase after which it quickly degrades during stationary phase. This pattern of AI-2 concentration coincides with the up- then down-regulation of a recently described microcin PDI (mccPDI) effector protein (McpM). To determine if there is a functional relationship between these systems, a prototypical mccPDI-expressing strain of E. coli 25 was used to generate ΔluxS, ΔlsrACDBFG (Δlsr), and ΔlsrR mutant strains that are deficient in AI-2 production, transportation, and AI-2 transport regulation, respectively. Trans-complementation, RT-qPCR, and western blot assays were used to detect changes of microcin expression and synthesis under co-culture and monoculture conditions. Compared to the wild-type strain, the AI-2-deficient strain (ΔluxS) and -uptake negative strain (Δlsr) were >1,000-fold less inhibitory to susceptible bacteria (P < 0.05). With in trans complementation of luxS, the AI-2 deficient mutant reduced the susceptible E. coli population by 4-log, which was within 1-log of the wild-type phenotype. RT-qPCR and western blot results for the AI-2 deficient E. coli 25 showed a 5-fold reduction in mcpM transcription with an average 2-h delay in McpM synthesis. Furthermore, overexpression of sRNA micC and micF (both involved in porin protein regulation) was correlated with mcpM regulation, consistent with a possible link between QS and mcpM regulation. This is the direct first evidence that microcin regulation can be linked to quorum sensing in a Gram-negative bacterium.

Keywords: autoinducer-2; bacteriocin; lsr; luxS; mccPDI; microcin; quorum sensing.

Figure 1

Delayed mccPDI inhibition when luxS is deleted. Competition assays between mccPDI-positive E. coli strain (25 or 25 ΔluxS) and target E. coli strain (BW25113 or BW25113 ΔluxS) in M9 media for 4, 8, 12, and 24 h. Results are expressed as the difference of mean log CFU during co-culture and mono-culture of the target strain (n = 3 independent replicates; error bar = SEM). ^*P < 0.05 compared to wild-type co-culture (black bars) based on two-way ANOVA.

Figure 2

Complementation restores of luxS mccPDI phenotype. CFU counts for E. coli BW25113 following competition with microcin-PDI producer E. coli 25, E. coli 25 ΔluxS, E. coli 25 ΔluxS/pBAD18-Cm, and E. coli 25 ΔluxS/pBAD18-Cm::luxS. Non-induced (black bar) and induced with 0.02% _L-arabinose (white bar). Results are expressed as the difference in CFU counts of BW25113 grown in co-culture and monoculture (n = 3 independent replicates; error bar = SEM). ^*P < 0.05 compared to wild-type co-culture based on one-way ANOVA.

Figure 3

Deletion of AI-2 ABC cassette, lsrACDBFG limits the mccPDI phenotype. Competition assay between isogenic mccPDI-producing E. coli strains (25, 25 ΔluxS, and 25 Δlsr) and target E. coli strains (BW25113 or BW25113 ΔluxS) in M9 media for 8 h. Results are expressed as the difference of mean log CFU during co-culture and mono-culture (n = 3 independent replicates; error bar = SEM). ^*P < 0.05 compared to wild-type co-culture based on one-way ANOVA.

Figure 4

Transcription of mcpM is significantly down regulated in AI-2 QS deficient E. coli 25 strains. Transcriptional analysis of the mccPDI effector mcpM for E. coli 25 ΔluxS, E. coli 25 ΔlsrR mutant and isogenic wild-type strain in M9 media over time by qPCR. Fold change is expressed relative to mcpM expression in M9 at 24 h (error bars = SEM; three independent replicates). ^*P < 0.05 based on two-way ANOVA.

Figure 5

E. coli 25 ΔluxS mutant causes delay in McpM production. (A) Western blot of McpM. Whole-cell lysate samples from E. coli 25 ΔmcpM/pCR2.1::P_{mic−10/−210}mcpM and E. coli 25 ΔluxS/pCR2.1::P_{mic−10/−210}mcpM complemented strains were collected for every 2 h from 2 to 12 h, and 24 h. Endogenous DnaK served as a loading control. (B) Western blot densitometry analysis of McpM. Whole-cell lysate samples from E. coli 25 ΔmcpM/pCR2.1::P_{mic−10/−210}mcpM (black circle) and E. coli 25 ΔluxS/pCR2.1::P_{mic−10/−210}mcpM (white circle) complemented strains were collected for every 2 h from 2 to 12 h, and 24 h. Endogenous DnaK served as a loading control. Normalization of the McpM against DnaK are represented by arbitrary unit (AU) over 24 h. Error bars = SEM; three independent experiments.

Figure 6

Overexpression of sRNA micC and micF in PDI-producer strain. The co-culture competition assay of over expression micC and micF in wild-type E. coli 25 (E. coli 25/pGEM-2-micC, E. coli 25/pGEM-2-micF, E. coli 25/pGEM-2) induced with 0.5 mM IPTG against target E. coli BW25113 in M9 media for 8 h. Results are expressed as the difference of mean log CFU during co-culture and mono-culture (n = 3 independent replicates; error bar = SEM). ^*P < 0.05 based on one-way ANOVA.

Figure 7

Microcin-PDI regulation model. The proposed regulatory mechanism of mcpM through the AI-2 uptake pathway (modified from Li et al., 2007). The AI-2 molecule produced by LuxS is actively transported into the cell by LsrACBD where it is phosphorylated by LsrK. The phosphorylated AI-2 interacts with LsrR and the signal is transduced via LsrR through (1) an unknown mechanism (?) that influences the two component system, EnvZ/OmpR (modified from Delihas and Forst, ; Blain et al., 2010) and induces expression of sRNA micC and/or micF that subsequently bind mcpM mRNA to inhibit translation, or (2) via an alternative pathway (?) that regulates transcription of micF and/or micC.