Poly(A) polymerase is required for RyhB sRNA stability and function in Escherichia coli

Small regulatory RNAs (sRNAs) are an important class of bacterial post-transcriptional regulators that control numerous physiological processes, including stress responses. In Gram-negative bacteria including Escherichia coli, the RNA chaperone Hfq binds many sRNAs and facilitates pairing to target transcripts, resulting in changes in mRNA transcription, translation, or stability. Here, we report that poly(A) polymerase (PAP I), which promotes RNA degradation by exoribonucleases through the addition of poly(A) tails, has a crucial role in the regulation of gene expression by Hfq-dependent sRNAs. Specifically, we show that deletion of pcnB, encoding PAP I, paradoxically resulted in an increased turnover of certain Hfq-dependent sRNAs, including RyhB. RyhB instability in the pcnB deletion strain was suppressed by mutations in hfq or ryhB that disrupt pairing of RyhB with target RNAs, by mutations in the 3′ external transcribed spacer of the glyW-cysT-leuZ transcript (3′ETSLeuZ) involved in pairing with RyhB, or an internal deletion in rne, which encodes the endoribonuclease RNase E. Finally, the reduced stability of RyhB in the pcnB deletion strain resulted in impaired regulation of some of its target mRNAs, specifically sodB and sdhCDAB. Altogether our data support a model where PAP I plays a critical role in ensuring the efficient decay of the 3′ETSLeuZ. In the absence of PAP I, the 3′ETSLeuZ transcripts accumulate, bind Hfq, and pair with RyhB, resulting in its depletion via RNase E-mediated decay. This ultimately leads to a defect in RyhB function in a PAP I deficient strain.

. Strains and plasmid used in this study.
Amp r , Cm r ; oriR6Kɣ; cat cassette flanked by FRT sites (Datsenko and Wanner 2000) pKD4 Amp r , Kan r ; oriR6Kɣ; kan cassette (Datsenko and Wanner 2000) flanked by FRT sites   Table S2. Figure 4 F. Northern blot analysis was used to determine RyhB and SsrA (loading control) expression at indicated time points following rifampicin addition in the indicated strain backgrounds under exponential growth conditions. (B) RyhB decay curves corresponding illustrating RyhB stability in the wild-type strain (NRD1138; WT (fur + )) and its derived mutants (NRD1198; ∆pcnB, DS060; hfqR17A, and DS073; hfqR17A ∆pcnB). sRNA decay curves were generated as described in Figure 4 legend and corresponding half-life measurements are listed in Table 1. Points and error bars in the curves represent the means and the standard errors (SEM) of at least three independent experiments. FIGURE S6. Determination of the relative abundance of sodB transcript in wild-type (fur + ; WT) and its derived isogenic fur and pcnB mutants. Northern blot analysis performed to determine the transcript steady state levels of RyhB target sodB under RyhB inducing and noninducing conditions as described in the legend of Figure 5B in a wild-type (WT (fur + ); NRD1138) and its derived isogenic mutants (∆pcnB, NRD1198; ryhBmut, LM11; ryhBmut ∆pcnB; ∆fur, DS024; ∆fur ∆pcnB, DS025). Samples for RNA extraction from DS024 and DS025 were collected as described in the legend of Figure 3A. Representative blots showing relative steadystate levels of sodB, RyhB, and SsrA in the indicated strain backgrounds are presented. SsrA is used as the loading control.

FIGURE S8. Northern blot analysis to determine transcript steady-state levels of LeuZ, 3'ETS LeuZ and RyhB in wild-type and derived isogenic pcnB and leuZ mutants under RyhB inducing and non-inducing conditions.
Representative northern blots corresponding to Figure 6C. Strains and growth conditions used are described in the legend of Figure 6C. Experiment was performed in triplicate and northern blots representing transcript steady-state levels of LeuZ, 3'ETS LeuZ and RyhB are shown. SsrA was used as the loading control. LeuZ term probe (Table S2) was used to determine LeuZ and 3'ETS LeuZ levels.

FIGURE S9
. Analysis of 3′-end polyadenylation state of RyhB. 3′ RACE was used to determine the sequence at the 3' ends of RyhB from exponential cultures of the ∆fur mutant and the wild type (WT; control) strain. Full-length RyhB (90 nt.) sequence was detected in 7 out of 47 clones sequenced from the ∆fur mutant and 5 out of 7 clones consisted of a template independent terminal adenine at the 3′ end (A). RyhB sequence as annotated in E. coli MG1655 is highlighted in red (A). The remaining 40 clones from the ∆fur mutant and all clones sequenced from the WT control yielded RyhB degradation products of varying lengths (C) and most RyhB degradation products terminated at nucleotide positions 40, 61 and 64 (B).

Strain construction
Strains generated by P1 vir transduction are indicated in Supplementary Table S1 with the donor strain indicated in brackets, using the protocol described by Miller (Miller 1992) and appropriate antibiotic selection.

Succinate growth assay
Strains were initially grown on M9-glucose agar plates and single colonies for each strain was inoculated in 2 mL of M9-glucose broth and grown overnight at 37 o C aerobically.
Each overnight culture was subcultured separately into 5 mL of fresh M9-glucose broth and M9succinate broth to a starting OD 600 of 0.01. Cultures were grown aerobically at 37 o C and growth was determined by measuring OD 600 after 24 h and 48 h. Final growth yield in succinate at the end of each time point was expressed as a ratio of the OD 600 obtained for growth in M9succinate to that in M9-glucose corresponding to each strain culture.

Northern blot analysis of ompA
Northern blot analysis of ompA was carried out as described previously (De Lay and Gottesman 2009). Briefly, 8 µg of each RNA sample was loaded on a 1.2% agarose gel that was pre-run at 12V/cm for at least 5 min and subsequently run at 5V/cm for 2 h in 1X MOPS (morpholinepropanesulfonic acid) buffer. Next, the RNA samples were transferred to a Zeta-Probe GT membrane (Bio-Rad) via capillary transfer. Transferred RNA was UV crosslinked and hybridized overnight with 100 ng/mL of 5′ biotinylated OmpA probe (Supplemental Table S2) as described in the Materials and Methods. Signal intensity corresponding to ompA was normalized to that of 16S rRNA, which served as internal loading control.