Small RNA-mediated activation of sugar phosphatase mRNA regulates glucose homeostasis

Abstract

Glucose homeostasis is strictly controlled in all domains of life. Bacteria that are unable to balance intracellular sugar levels and deal with potentially toxic phosphosugars cease growth and risk being outcompeted. Here, we identify the conserved haloacid dehalogenase (HAD)-like enzyme YigL as the previously hypothesized phosphatase for detoxification of phosphosugars and reveal that its synthesis is activated by an Hfq-dependent small RNA in Salmonella typhimurium. We show that the glucose-6-P-responsive small RNA SgrS activates YigL synthesis in a translation-independent fashion by the selective stabilization of a decay intermediate of the dicistronic pldB-yigL messenger RNA (mRNA). Intriguingly, the major endoribonuclease RNase E, previously known to function together with small RNAs to degrade mRNA targets, is also essential for this process of mRNA activation. The exploitation of and targeted interference with regular RNA turnover described here may constitute a general route for small RNAs to rapidly activate both coding and noncoding genes.

Figure 1. YigL is induced by SgrS and required to counteract phosphosugar stress

(A) YigL and PtsG proteins or SgrS RNA were probed on Western or Northern blots, respectively, prior to (0 min) and following α-MG treatment (10–80 min) of Salmonella enterica wild-type (WT) and ΔsgrS in exponential growth phase. (B) Overnight growth of selected E. coli gene disruption strains (as indicated) on sugar-supplemented minimal media plates show that loss of YigL, unlike other HAD phosphatases (OtsB, YniC, YidA, YfbT, YbiV), phenocopies the growth defect of ΔsgrS in the presence of α-MG or 2-deoxyglucose. (C) Salmonella wild-type or ΔsgrS strains carrying inducible pBAD expression plasmids were stressed with α-MG in logarithmic phase. Subsequent ectopic expression (induced with L-arabinose) of YigL fully rescues ΔsgrS bacteria; active site mutant YigL^K191A fails to rescue. Data are represented as mean +/− SEM.

Figure 2. SgrS promotes discoordinate expression of the pldB-yigL operon

(A) The pldB-yigL-yigM locus of Salmonella. Gene synteny is observed in other species including E. coli (Fig. S2C). Wavy lines represent the predicted mRNAs from this region (Kröger et al., 2012). (B) qRT-PCR analysis of pldB, yigL and yigM transcript levels in response to SgrS overproduction (plasmid pPL-SgrS). RNA levels in cells with control vector were set at a value of 1.0, to calculate fold expression changes. (C) Primer extension analysis detects SgrS-induced processing of the pldB-yigL mRNA, after sRNA induction for 15 and 30 min from a pBAD plasmid in Salmonella ΔsgrS cells (lanes 1–6). The pldB-yigL specific extension product at U+841 (cf. panel A) was visualized using a radiolabeled primer. Lanes 7–10 prove the expected lack of signal in ΔpldB-yigL. (D) RNase E is required for SgrS-mediated yigL activation. Primer extension reveals no processing at U+841 in an RNase E-temperature sensitive strain (rne-TS, lanes 5–8) at 44 °C when RNase E is inactive. (E) Western or Northern blots show YigL or SgrS expression in the same context of RNase E activity as above.

Figure 3. SgrS binds in the pldB coding sequence

(A) Western blot of YigL protein in Salmonella ΔsgrS carrying the indicated expression plasmids. YigL is equally up-regulated by Salmonella and E. coli SgrS (SgrS^Eco), independent of a functional sgrT ORF (premature stop codon in SgrSUGA). However, a G→C mutation in the SgrS seed region (SgrS*, see also below) abrogates up-regulation, suggesting activation by base pairing. (B) Overview of the pldB-yigL::gfp reporters employed to identify the SgrS-pldB interaction site. Scissors indicate the RNase E-dependent processing at U+841. (C) Regulation of the different pldB-yigL::gfp reporters by SgrS. (D) The SgrS-pldB RNA interaction. Numbering for pldB is relative to A of start codon AUG, and for SgrS counting from +1 site. Vertical arrows denote compensatory nucleotide changes. Note that the yigL* mutation lies in pldB ORF. (E) Western blot showing activation of gfp reporters for yigL or yigL* mutant by SgrS or compensatory SgrS* RNA, respectively, validates base pairing in vivo.

Figure 4. Post-transcriptional activation of yigL is independent of translation initiation

(A) RNA half-life measurements show that base-pairing by SgrS stabilizes the yigL mRNA. Decay of yigL mRNA was followed after stop of transcription in Salmonella ΔsgrS expressing SgrS or pairing-deficient SgrS*, induced from pBAD plasmids 10 min before the half-life assay. SgrS RNA was probed in the pBAD-SgrS sample. Data are represented as mean +/− SEM. (B) The pldB-ompX::gfp reporter and mini-gene constructs, and location of RNase E cleavage (scissors) and SgrS binding sites. The mini-gene (3′ end of pldB) produces a noncoding RNA containing the SgrS site. Ribosome occupancy of the SgrS site was restored in “pldB mini+21aa ORF” with a short artificial reading frame. (C) Successful regulation of pldB-ompX::gfp reporters by SgrS proves that activation is determined by sequences in pldB (compare to Figure 3C). Data are represented as mean +/− SEM. (D) qRT-PCR analysis of pldB mini-gene activation, with or without an artificial 21aa ORF, in a Salmonella ΔsgrS background. Fold-activation was determined 5, 10 and 15 min after induction of pBAD-SgrS, relative to empty pBAD vector. Data are represented as mean +/− SEM.

Figure 5. SgrS antagonizes pldB cleavage by RNase E

(A) Northern blot of non-coding pldB mini-gene RNA in Salmonella ΔsgrS and ΔsgrS/rneTS (RNase E inactive at 44°C) strains. Cells were grown at 28°C to OD₆₀₀ of 1.0, split (time-point 0 min) and cultivated further at the indicated temperatures. (B) In vitro synthesized, radiolabeled pldB RNA (see Fig. S5C) was incubated with RNase E for 60 min, in the absence or presence of non-labeled SgrS RNA (ratios as indicated). The autoradiograph shows the pldB degradation products after denaturing electrophoresis. Size markers to the left. An arrow indicates the 948–955 cleavage in the pldB RNA (see Fig. S5C). (C) Analogous to (B), but radiolabeled SgrS RNA incubated with RNase E and unlabeled pldB RNA. The 177–179 cleavage product (see Fig. S5E) is indicated. (D) Quantification of signals from panels B and C, based on triplicate experiments. Data are represented as mean +/− SEM.

Figure 6. Post-transcriptional activation of YigL is required for α-MG efflux

(A) Northern blot of SgrS RNA to follow phosphosugar stress recovery of E. coli cells mutated for SgrS-pldB base-pairing. Samples were taken from untreated (−) cells and cells treated with α-MG for 15 min; and treated cells at 10, 20 and 40 min after wash and re-inoculation into fresh media. In strain pldB*, a chromosomal point mutation in the SgrS site in pldB prevents base pairing. The pldB strain serves as isogenic control to pldB*. (B) Export of α-MG is retarded when yigL is missing or SgrS cannot activate it (pldB* mutant). The above strains were treated with radiolabeled [14C] α-MG and then diluted (time-point 0 min) to follow α-MG efflux for 50 min.

Figure 7. Model of YigL activation within the phosphosugar stress response

During membrane translocation glucose is phosphorylated by importing PTS (green barrel), a process which impairs efflux and activates the sugar for downstream metabolic tasks. Accumulation of phosphosugars is toxic and activates SgrS RNA (through the SgrR transcription factor). Aided by Hfq protein, SgrS targets the pldB coding sequence and blocks sustained 5′ to 3′ endonucleolytic turnover of the pldB-yigL discistronic transcript by RNase E. Initial mRNA decay is essential for SgrS-mediated activation. Stabilization of the ‘pldB-yigL mRNA fragment increases the levels of HAD phosphatase YigL, which dephosphorylates the accumulated sugars to facilitate their export via undefined efflux systems (gray barrel).